article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
this paper presents an importance sampling based approximate optimal planning and control algorithm .
optimal motion planning in deterministic and continuous systems is computationally np - complete @xcite except for linear time invariant systems . for nonlinear systems
, there is a vast literature on approximate solutions and algorithms . in optimal planning ,
the common approximation scheme is discretization - based . by discretizing the state and input spaces ,
optimal planning is performed by solving the shortest path problem in the discrete transition systems obtained from abstracting the continuous dynamics , using heuristic - based search or dynamic programming . comparing to discretization - based methods , _ sampling - based graph search _ ,
includes probabilistic roadmap ( prm ) @xcite , rrt @xcite , rrt * @xcite , are more applicable for high - dimensional systems .
while rrt has no guarantee on the optimality of the path @xcite , rrt * compute an optimal path asymptotically provided the cost functional is lipschitz continuous .
however , such lipschitz conditions may not be satisfied for some cost functions under specific performance consideration .
the key idea in the proposed sampling - based planning method builds on a unification of importance sampling and approximate optimal control @xcite .
in approximate optimal control , the objective is to approximate both the value function , i.e. , optimal cost - to - go , and the optimal feedback policy function by weighted sums of _ known _ basis functions . as a consequence
, the search space is changed from infinite trajectory space or policy space to a continuous space of weight vectors , given that each weight vector corresponds to a unique feedback controller . instead of solving the approximate optimal control through training actor and critic neural networks ( nns ) using trajectory data @xcite
, we propose a sampling - based method for sampling the weight vectors for a policy function approximation and searching for the optimal one .
this method employs @xcite , a probabilistic complete global optimization algorithm , for searching the optimal weight vector that parametrizes the approximate optimal feedback policy .
the fundamental idea is to treat the weight vector as a random variable over a parameterized distribution and the optimal weight vector corresponds to a dirac s delta function which is the target distribution .
the algorithm iteratively estimates the parameter that possesses the minimum kullback - leibler divergence with respect to an intermediate reference model , which assigns a higher probability mass on a set of weights of controllers with improved performance over the previous iteration . at the meantime ,
a set of sampled weight vectors are generated using the parameterized distribution and the performance of their corresponding policies are evaluated via simulation - based policy evaluation . under mild conditions ,
the parameterized distribution converges , with probability one , to the target distribution that concentrates on the optimal weight vector with respect to given basis functions . resembles another adaptive search algorithm called cross - entropy(ce ) method and provides faster and stronger convergence guarantee for being less sensitive to input parameters @xcite
previously , ce algorithm has been introduced for motion planning @xcite based on sampling in the trajectory space . the center idea is to construct a probability distribution over the set of feasible paths and to perform the search for an optimal trajectory using ce .
the parameters to be estimated is either a sequence of motion primitives or a set of via - points for interpolation - based trajectory planning .
differ to these methods , ours is the first to integrate importance sampling to estimate parameterization of the optimal policy function approximation for continuous nonlinear systems .
since the algorithm performs direct policy search , we are able to enforce robustness and stability conditions to ensure the computed policy is both robust and approximate optimal , provided these conditions can be evaluated efficiently . to conclude
, the contributions of this paper are the following : first , we introduce a planning algorithm by a novel integration of model reference adaptive search and approximate optimal control .
second , based on contraction theory , we introduce a modification to the planning method to directly generate stabilizing and robust feedback controllers in the presence of bounded disturbances . last but not the least , through illustrative examples , we demonstrate the effectiveness and efficiency of the proposed methods and share our view on interesting future research along this direction .
notation : the inner product between two vectors @xmath0 is denoted @xmath1 or @xmath2 . given a positive semi - definite matrix @xmath3 , the @xmath3-norm of a vector is denoted @xmath4 .
we denote @xmath5 for @xmath3 being the identity matrix .
@xmath6 is the indicator function , i.e. , @xmath7 if event @xmath8 holds , otherwise @xmath9 . for a real @xmath10
, @xmath11 is the smallest integer that is greater than @xmath12 .
we consider continuous - time nonlinear systems of the form @xmath13 where @xmath14 is the state , @xmath15 is the control input , @xmath16 is the initial state , and @xmath17 is a vector field .
we assume that @xmath18 and @xmath19 are compact .
a feedback controller @xmath20 takes the current state and outputs a control input .
the objective is to find a feedback controller @xmath21 that minimizes a finite - horizon cost function for a nonlinear system @xmath22 where @xmath23 is the stopping time , @xmath24 defines the running cost when the state trajectory traverses through @xmath25 and the control input @xmath26 is applied and @xmath27 defines the terminal cost . as an example
, a running cost function can be a quadratic cost @xmath28 for some positive semi - definite matrices @xmath29 and @xmath30 , and a terminal cost can be @xmath31 where @xmath32 is a goal state .
we denote the set of feedback policies to be @xmath33 . for infinite horizon optimal control ,
the optimal policy is independent of time and a feedback controller suffices to be a minimizing argument of ( see ref .
@xcite ) . for finite - horizon optimal control ,
the optimal policy is time - dependent .
however , for simplicity , in this paper , we only consider time - invariant feedback policies and assume the time horizon @xmath23 is of sufficient length to ignore the time constraints . algorithm , introduced in @xcite , aims to solve the following problem : @xmath34 where @xmath35 is the solution space and @xmath36 is a deterministic function that is bounded from below .
it is assumed that the optimization problem has a unique solution , i.e. , @xmath37 and for all @xmath38 , @xmath39 .
the following regularity conditions need to be met for the applicability of .
[ assume1 ] for any given constant @xmath40 , the set @xmath41 has a strictly positive lebesgue or discrete measure .
this condition ensures that any neighborhood of the optimal solution @xmath42 will have a positive probability to be sampled .
[ assume2 ] for any constant @xmath43 , @xmath44 , where @xmath45 , and we define the supremum over the empty set to be @xmath46 . *
selecte a sequence of reference distributions @xmath47 with desired convergence properties .
specifically , the sequence @xmath47 will converge to a distribution that concentrates only on the optimal solution .
* selecte a parametrized family of distribution @xmath48 over @xmath18 with parameter @xmath49 .
* optimize the parameters @xmath50 iteratively by minimizing the following kl distance between @xmath51 and @xmath52 .
@xmath53 where @xmath54 is the lebesgue measure defined over @xmath55 .
the sample distributions @xmath56 can be viewed as compact approximations of the reference distributions and will converge to an approximate optimal solution as @xmath57 converges provided certain properties of @xmath47 is retained in @xmath58 .
note that the reference distribution @xmath47 is unknown beforehand as the optimal solution is unknown .
thus , the algorithm employs the estimation of distribution algorithms @xcite to estimate a reference distribution that guides the search . to make the paper
self - contained , we will cover details of in the development of the planning algorithm .
in this section , we present an algorithm that uses in a distinguished way for approximate optimal feedback motion planning . the _ policy function approximation _ @xmath59 is a weighted sum of basis functions , @xmath60 where @xmath61 are basis functions , and the coefficients @xmath62 are the weight parameters , @xmath63 .
an example of basis function can be polynomial basis @xmath64 $ ] for one - dimensional system .
a commonly used class of basis functions is .
it can be constructed by determining a set of centers @xmath65 , and then constructing basis functions @xmath66 , for each center @xmath67 , where @xmath68 is a pre - defined parameter . in vector form
, a policy function approximation is represented by @xmath69 where vector @xmath70^\intercal$ ] and @xmath71^\intercal$ ] .
we let the domain of weight vector be @xmath72 and denote it by @xmath73 the set of all policies that can be generated by linear combinations of pre - defined basis functions . in the following context ,
unless specifically mentioned , the vector of basis functions is @xmath74 .
clearly , for any weight vector @xmath75 , @xmath76 .
thus , we aim to solve @xmath77 so as to minimize the error in the optimal cost introduced by policy function approximation .
[ def : optweight ] given a basis vector @xmath74 , a weight vector @xmath78 with respect to @xmath74 is _ optimal _ if and only if @xmath79 and for all @xmath80 such that @xmath81 , @xmath82 the approximate optimal feedback policy is @xmath83 . by requiring @xmath84
, it can be shown that the optimal weight vector @xmath78 minimizes the difference between the optimal cost achievable with policies in @xmath85 and the cost under the global optimal policy .
for clarity in notation , we denote @xmath86 by @xmath87 as @xmath74 is a fixed basis vector throughout the development of the proposed method .
clearly , if the actual optimal policy @xmath88 can be represented by a linear combination of selected basis functions , then we obtain the optimal policy by computing the optimal weight vector , i.e. , @xmath89 .
[ [ remark ] ] remark : + + + + + + + here , we assume a feedback policy can be represented by @xmath90 for some weight vector @xmath91 . in cases
when the basis functions are continuous , a feedback policy must be a continuous function of the state .
however , this requirement is hard to satisfy for many physical systems due to , for example , input saturation . in cases when a feasible controller is discontinuous
, we can still use a continuous function to approximate , and then project the continuous function to the set @xmath33 of applicable controllers .
using function approximation , we aim to solve the optimal feedback planning problem in approximately by finding the optimal weight vector with respect to a pre - defined basis vector .
the main algorithm is presented next . in this section
, we present an adaptive search - based algorithm to compute the approximate optimal feedback policy .
the algorithm is `` near '' anytime , meaning that it returns a feasible solution after a small number of samples .
if more time is permitted , it will quickly converge to the globally optimal solution that corresponds to the approximate optimal feedback policy .
the algorithm is probabilistic complete under regularity conditions of .
we start by viewing the weight vector as a random variable @xmath92 governed by a multivariate gaussian distribution with a compact support @xmath72 .
the distribution is parameterized by parameter @xmath93 , where @xmath94 is a @xmath95-dimensional mean vector and @xmath96 is the @xmath95 by @xmath95 covariance matrix .
recall @xmath95 is the number of basis functions .
the optimal weight vector @xmath78 can be represented as a _ target distribution _ @xmath97 as a dirac s delta , i.e. , @xmath98 and @xmath99 for @xmath100 .
dirac s delta is a special case of multivariate gaussian distribution with zero in the limit case of vanishing covariance .
thus , it is ensured that the target distribution can be arbitrarily closely approximated by multivariate gaussian distribution by a realization of parameter @xmath101 .
recall that the probability density of a multivariate gaussian distribution is defined by @xmath102 where @xmath95 is the dimension of weight vector @xmath103 and @xmath104 is the determinant of @xmath96 .
now , we are ready to represent the main algorithm , called , which includes the following steps . 1 .
* initialization * : the initial distribution is selected to be @xmath105 , for @xmath106 which can generate a set of sample to achieve a good coverage of the sample space @xmath72 .
for example , @xmath107 and @xmath108 which is an identity matrix .
the following parameters are used in this algorithm : @xmath109 $ ] for specifying the quantile , the _ improvement parameter _
@xmath110 , a _ sample increment percentage _ @xmath12 , an initial sample size @xmath111 , a _ smoothing coefficient _ @xmath112 $ ] .
let @xmath113 .
* sampling - based policy evaluation * : at each iteration @xmath114 , generate a set of @xmath115 samples @xmath116 from the current distribution @xmath117 . for each @xmath118 , using simulation we evaluate the cost @xmath119 from the initial state @xmath120 and the feedback policy @xmath121 with system model in . the cost @xmath87 is determined because the system is deterministic and has a unique solution .
policy improvement with elite samples * : next , the set @xmath122 is ordered from largest ( worst ) to smallest ( best ) among given samples : @xmath123 we denote @xmath124 to be the estimated @xmath125-quantile of cost @xmath126 , i.e. , @xmath127 .
the following cases are distinguished . *
if @xmath113 , we introduce a threshold @xmath128 . *
if @xmath129 , the following cases are further distinguished : * * @xmath130 , i.e. , the estimated @xmath125-quantile of cost has been reduced by the amount of @xmath131 from the last iteration , then let @xmath132 .
let @xmath133 and continue to step 4 ) .
* * otherwise @xmath134 , we find the largest @xmath135 , if it exists , such that the estimated @xmath136-quantile of cost @xmath137 satisfies @xmath138
. then let @xmath139 and also let @xmath140 .
continue to step 4 ) .
however , if no such @xmath141 exists , then there is no update in the threshold @xmath142 but the sample size is increased to @xmath143 .
let @xmath144 , @xmath145 , and continue to step 2 ) .
* * parameter(policy ) update * : we update parameters @xmath146 for iteration @xmath147 .
first , we define a set @xmath148 of _ elite samples_. note that the parameter update in @xmath101 is to ensure a higher probability for elite samples . to achieve that , for each elite sample @xmath149 , we associated a weight such that a higher weight is associated with a weight vector with a lower cost and a lower probability in the current distribution .
the next parameter @xmath146 is selected to maximize the weighted sum of probabilities of elite samples . to this end
, we update the parameter as follows .
@xmath150\end{gathered}\ ] ] where @xmath151 is the expected value of a random variable @xmath152 given distribution @xmath153 , @xmath154 is a strictly decreasing and positive function . or @xmath155 if @xmath25 is strictly positive . ]
@xmath156 is the weight for parameter @xmath75 .
[ assume3 ] the optimal parameter @xmath157 is the interior point of @xmath158 for all @xmath114 .
[ lma ] assuming [ assume1],[assume2 ] , and [ assume3 ] and the compactness of @xmath72 , with probability one , @xmath159 where @xmath78 is the optimal weight vector and @xmath160 is an @xmath95-by-@xmath95 zero matrix .
note that since @xmath161 converges in the limit a zero matrix , the stopping criterion is justified .
building on the convergence result of , the proposed sampling - based planner ensures a convergence to a dirac delta function concentrating on the optimum . in practice
, the parameter update is performed using the expectation maximization ( em ) algorithm .
* em - based parameter update / policy improvement * since our choice of probability distribution is the multivariate gaussian , the parameter @xmath162 is computed as follows @xmath163i_{w\in e } w } { { \mathbb{e}}_{\theta_k } [ s(j(x_0,w))^k / p(w,\theta_k)]i_{w\in e } } \\
\approx \frac{\sum_{w\in w_k } [ s(j(x_0,w))^k / p(w,\theta_k)]i_{w\in e } w } { \sum_{w\in w_k } [ s(j(x_0,w))^k / p(w,\theta_k)]i_{w\in e}},\end{gathered}\ ] ] and @xmath164i_{w\in e } ( w - { \mathbf{\mu}})(w - { \mathbf{\mu}})^\intercal}{{\mathbb{e}}_{\theta_k } [ s(j(x_0,w))^k / p(w,\theta_k)]i_{w\in e}}\\ \approx \frac { \sum_{w\in w_k } [ s(j(x_0,w))^k / p(w,\theta_k)]i_{w\in e } ( w - { \mathbf{\mu}})(w - { \mathbf{\mu}})^\intercal } { \sum_{w\in w_k}[s(j(x_0,w))^k / p(w,\theta_k)]i_{w\in e}},\end{gathered}\ ] ] where we approximate @xmath165 with its estimate @xmath166 for @xmath167 and the fraction @xmath168 was canceled as the term is shared by the numerator and the denominator . * smoothing * : due to limited sample size , a greedy maximization for parameter update can be premature if too few samples are used .
to ensure the convergence to the _ global optimal solution _ , a _ smoothing _ update is needed . to this end
, we select the parameter for the next iteration to be @xmath169 where @xmath170 is the smoothing parameter .
let @xmath145 .
we check if the iteration can be terminated based on a given stopping criterion .
if the stopping criterion is met , then we output the latest @xmath171 .
otherwise , we continue to update of @xmath101 by moving to step 2 ) . *
stopping criterion * given the probability distribution will converge to a degenerated one that concentrates on the optimal weight vector .
we stop the iteration if the covariance matrix @xmath161 becomes near - singular given the convergence condition in lemma [ lma ] . to conclude , the proposed algorithm using is probabilistic complete and converges to the global optimal solution .
if the assumptions are not met , the algorithm converges to a local optimum .
being able to directly search within continuous control policy space , one major advantage is that one can enforce stability condition such that the search is restricted to stable and robust policy space . in this subsection
, we consider contraction theory to compute conditions that need to be satisfied by weight vectors to ensure stability and robustness under bounded disturbances .
@xcite given the system equation for the closed - loop system @xmath172 , a region of the state space is called a _ contraction region _ if the jacobian @xmath173 is uniformly negative definite in that region , that is , @xmath174 where @xmath175 is a positive definite matrix for all @xmath176 .
@xcite [ thm : contraction ] given the system model @xmath177 , any trajectory , which starts in a ball of constant radius with respect to the matrix @xmath178 , centered about a given trajectory and contained at all times in a contraction region with respect to the matrix @xmath178 , remains in that ball and converges exponentially to this trajectory .
furthermore , global exponential convergence to the given trajectory is guaranteed if the whole state space is a contraction region .
theorem [ thm : contraction ] provides a necessary and sufficient condition for exponential convergence of an autonomous system . under
bounded disturbances , the key idea is to incorporate a contraction analysis in the planning algorithm such that it searches for a weight vector @xmath75 that is not only optimal in the nominal system but also ensures that the closed - loop actual system under the controller @xmath179 has contraction dynamics within a tube around the nominal trajectory . using a similar proof in @xcite
, we can show that for systems with contracting dynamics , the actual trajectory under bounded disturbances will be ultimately uniformly bounded along the nominal trajectory .
consider a closed - loop system @xmath180 where @xmath181 is a disturbance with @xmath182 , let a state trajectory @xmath183 be in the contraction region @xmath184 at all time @xmath185 , then for any time @xmath185 , the deviation between @xmath183 and the nominal trajectory @xmath186 , whose dynamic model is given by @xmath187 , satisfies @xmath188 in other words , the error is uniformly ultimately bounded with the ultimate bound @xmath189 .
let s pick the lyapunov function @xmath190 whose time derivative is @xmath191 where the following property is used : @xmath192 for some @xmath193 $ ] if @xmath194 or @xmath195 $ ] otherwise .
since the trajectories stays within the contraction region , the following condition holds @xmath196 , and we have @xmath197 meanwhile , @xmath198 as the trajectory @xmath183 stays within the region of contraction , and also @xmath199 we conclude that as @xmath200 , @xmath201 since @xmath202 and under the condition that @xmath203 , we obtain @xmath204 , and therefore @xmath205 thus , to search for the optimal and robust policies , we modify the algorithm by introducing the following step .
* contraction verification step : * suppose the closed - loop system is subject to bounded disturbances , the objective is to ensure the trajectory is contracting within the time - varying tube @xmath206 , for all @xmath207 , where @xmath208 is the nominal state trajectory .
the following condition translates the contraction condition into verifiable condition for a closed - loop system : choose positive constants @xmath209 , a positive definite symmetric and constant matrix @xmath210_{i=1,\ldots , n , j=1,\ldots , n}$ ] , and verify whether , at each time step along the nominal trajectory @xmath211 in the closed - loop system under control @xmath212 , the following condition holds .
@xmath213 where @xmath214 is the @xmath215th component in the matrix @xmath216 .
we verify this condition numerically at discrete time steps instead of continous time .
further , if the function @xmath217 is semi - continuous , according to the extreme value theorem , this condition can be verified by evaluating @xmath217 at all critical points where @xmath218 and the boundary of the set @xmath219 .
the modification to the planning algorithm is made in step 3 ) , if a controller @xmath220 of elite sample @xmath75 does not meet the condition , then @xmath75 is rejected from the set of elite samples . alternatively , one can do so implicitly by associating @xmath75 with a very large cost . however , since the condition is sufficient but not necessary as we have the matrix @xmath178 , constant @xmath221 and @xmath222 pre - fixed and @xmath178 is chosen to be a constant matrix , the obtained robust controller may not necessary be optimal among all robust controllers in @xmath85 . a topic for future work is to extend joint planning and control policies with respect to adaptive bound @xmath221 , @xmath222 , and a uniformly positive definite and time - varying matrix @xmath223 .
in this section , we use two examples to illustrate the correctness and efficiency of the proposed method .
the simulation experiments are implemented in matlab on a desktop with intel xeon e5 cpu and 16 gb of ram . to illustrate the correctness and sampling efficiency in the planning algorithm
, we consider an optimal control of systems with non - quadratic cost . for this class of optimal control problems ,
since there is no admissible heuristic , one can not use any planning algorithm facilitated by the usage of a heuristic function .
moreover , the optimal controller is nonlinear given the non - quadratic cost .
consider a system @xmath224 where @xmath225 and @xmath226 with @xmath227 and @xmath228 .
the initial state is @xmath229 $ ] .
the cost functional is @xmath230 for a non - quadratic cost functional , the optimal controller is no longer linear and can not be computed by lqr unless the running cost can be written in the sum - of - square form .
thus , we consider an approximate feedback controller with basis vector @xmath231^\intercal$ ] .
suppose the magnitude of external disturbance is bounded by @xmath232 .
the following parameters are used in stability verification : @xmath233 , at any time @xmath207 , for all @xmath25 such that @xmath234 , the controller ensures @xmath235 because @xmath236 . with this choice for stability analysis , the constraint @xmath237 in this case ,
if we select @xmath238 nonpositive , @xmath239 and @xmath240 , then closed - loop system , which is a nonlinear polynomial system , will become globally contracting .
0.32 under feedback controller computed with .,title="fig : " ] 0.33 under feedback controller computed with .,title="fig : " ] 0.33 figures [ fig : cost ] and [ fig : mean ] show the convergence result with in one simulation in terms of cost and the mean of the multivariant gaussian over iterations .
the following parameters are used : initial sample size @xmath241 , improvement parameter @xmath242 , quantile percentage @xmath243 , smoothing parameter @xmath244 , sample increment parameter @xmath245 .
the algorithm converges after @xmath246 iterations with @xmath247 samples to the mean @xmath248^\intercal $ ] and the covariance matrix with a norm @xmath249 .
each iteration took less than 10 seconds .
the approximate optimal cost under feedback controller @xmath250 is @xmath251 .
figure [ fig : statelinnoquad ] shows the state trajectory for the closed - loop system with bounded disturbances . with 25 independent runs of , the mean of @xmath252 is @xmath253 and the standard deviation is @xmath254 , @xmath255 of the approximate optimal cost .
note , if we only use linear feedback @xmath256 , the optimal cost is @xmath257 , which is about three times the optimal cost that can be achieved with a nonlinear controller .
consider a dubins car dynamics @xmath258 where @xmath259 being the state ( coordinates and turning angle with respect to @xmath25-axis ) and @xmath26 and @xmath260 are control variables including linear and angular velocities .
the system is kinematically constrained by its positive minimum turning radius @xmath261 which implies the following bound @xmath262 . without loss of generality
, we assume @xmath263 and @xmath264 are the input constraints . the control objective is to reach the goal @xmath265 while avoiding static obstacles . the cost function @xmath266 where @xmath267 , the running cost is @xmath268 , and the terminal cost is @xmath269 .
the initial state is @xmath270 . in simulation
, we consider the robot reaches the target if @xmath271 for @xmath272 $ ] . in simulation , @xmath273 .
we select as basis functions and define @xmath274^\intercal$ ] for @xmath95 center points . in the experiment ,
the center points are includes uniform grids in @xmath275 coordinates with step sizes @xmath276 , @xmath277 ; and vertices of the obstacle .
we also include linear basis functions @xmath278 $ ] .
the basis vector is @xmath279^\intercal$ ] .
we consider a bounded domain @xmath280 and @xmath281 and @xmath282 $ ] and thus the total number of basis functions is @xmath283 .
the control input @xmath284^\intercal$ ] where @xmath285 and @xmath286 .
the total number of weight parameters is twice the number of bases and in this case @xmath287 . 0.45 computed using the mean of multivariate gaussian over iterations ( from the lightest to the darkest ) .
( b ) the convergence of the covariance matrix .
( c ) the total cost evaluated at the mean of the multivariate gaussian over iterations .
, title="fig : " ] 0.45 the following parameters are used : initial sample size @xmath288 , improvement parameter @xmath289 , smoothing parameter @xmath244 , sample increment percentage @xmath290 , and @xmath291 . in fig .
[ fig : dubinstraj ] we show the trajectory computed using the estimated mean of multivariate gaussian distribution over iterations , from the lightest ( @xmath292-th iteration ) to the darkest ( the last iteration when stopping criterion is met ) .
the optimal trajectory is the darkest line . in fig .
[ fig : dubinscost ] we show the cost computed using the mean of multivariate gaussian over iterations .
converges after 22 iterations with @xmath293 samples and the optimal cost is @xmath294 .
each iteration took about @xmath295 to @xmath296 seconds .
however , it generates a collision - free path only after @xmath297 iterations . due to input saturation
, the algorithm is only ensured to converge to a local optimum .
however , in 24 independent runs , all runs converges to a local optimum closer to the global one , as shown in the histogram in fig .
[ fig : dubins_histo ] .
our current work is to implement trajectory - based contraction analysis using time - varying matrices @xmath223 and adaptive bound @xmath221 , which are needed for nonlinear dubins car dynamics .
in this paper , an importance sampling - based approximate optimal planning and control method is developed . in the control - theoretic formulation of optimal motion planning ,
the planning algorithm performs direct policy computation using simulation - based adaptive search for an optimal weight vector corresponding to an approximate optimal feedback policy .
each iteration of the algorithm runs time linear in the number of samples and in the time horizon for simulated runs .
however , it is hard to quantify the number of iterations required for to converge .
one future work is to consider incorporate multiple - distribution importance sampling to achieve faster and better convergence results .
based on contraction analysis of the closed - loop system , we show that by modifying the sampling - based policy evaluation step in the algorithm , the proposed planning algorithm can be used for joint planning and robust control for a class of nonlinear systems under bounded disturbances . in future extension of this work ,
we are interested in extending this algorithm for stochastic optimal control .
l. e. kavraki , p. svestka , j .- c .
latombe , and m. h. overmars , `` probabilistic roadmaps for path planning in high - dimensional configuration spaces , '' _ ieee transactions on robotics and automation _ , vol . 12 , no . 4 , pp . 566580 , 1996 .
s. c. livingston , e. m. wolff , and r. m. murray , `` cross - entropy temporal logic motion planning , '' in _ proceedings of the 18th international conference on hybrid systems : computation and control_.1em plus 0.5em minus 0.4emacm , 2015 , pp . | in this paper , we propose a sampling - based planning and optimal control method of nonlinear systems under non - differentiable constraints . motivated by developing scalable planning algorithms
, we consider the optimal motion plan to be a feedback controller that can be approximated by a weighted sum of given bases . given this approximate optimal control formulation ,
our main contribution is to introduce importance sampling , specifically , model - reference adaptive search algorithm , to iteratively compute the optimal weight parameters , i.e. , the weights corresponding to the optimal policy function approximation given chosen bases .
the key idea is to perform the search by iteratively estimating a parametrized distribution which converges to a dirac s delta that infinitely peaks on the global optimal weights .
then , using this direct policy search , we incorporated trajectory - based verification to ensure that , for a class of nonlinear systems , the obtained policy is not only optimal but robust to bounded disturbances .
the correctness and efficiency of the methods are demonstrated through numerical experiments including linear systems with a nonlinear cost function and motion planning for a dubins car . |
it was realized long ago on theoretical grounds that most melting transitions should be first order,@xcite although exceptions include the possibility of defect - mediated melting in two - dimensional ( 2d ) crystals@xcite and the mean - field transition between the superconducting vortex lattice and the normal state in type - ii superconductors.@xcite as with most first - order transitions , there is no good simple theory of melting to compare with our theoretical understanding of many continuous phase transitions : a first - order transition occurs when the free energy of one phase as a function of temperature crosses the free energy of another phase ( see fig . [ fig:1 ] ) .
these two phases are qualitatively different and generally demand theoretical treatments with distinct approximations , such that a detailed comparison of free energies is not possible .
the solid - liquid transition is a prime example of this problem .
we can treat the solid phase well within elasticity theory , but this is not a useful description of the liquid .
a common approach to this problem is to concentrate on the solid phase and calculate its stability limit@xcite ( a complementary method is to treat the liquid , e.g. within density functional theory,@xcite and find the lowest temperature for the liquid phase to exist ) .
a plausible mechanism for the crystal instability comes from anharmonic effects that may soften the lattice when thermal fluctuations are large . to this end
one can develop a self - consistent harmonic approximation ( scha),@xcite where an effective harmonic theory is used that better approximates the true anharmonic system for finite thermal fluctuations .
we can then identify the limit to a self - consistent solution with the instability temperature of the solid . while it is unclear whether or not the instability point @xmath0 is close to the melting temperature @xmath1 , it does give an upper bound ( see section [ sec : scha ] ) , and represents the first step to understanding the possible mechanisms of melting for a given system .
= 6.5 cm an example of this hunt for the anharmonic instability point occurred several years ago in the context of 2d melting .
an early scha calculation by platzman and fukuyama@xcite for a 2d electron system found a dramatic stiffening of the wigner crystal as a function of temperature , and the anharmonic instability point was reached well above the dislocation - mediated melting temperature .
later , fisher considered anharmonic corrections from a perturbation expansion around the harmonic theory , and could show that the lowest - order temperature correction to the shear modulus was downwards.@xcite the opposite result of the scha calculation comes from the fact that it misses contributions from odd anharmonicities .
a series of papers by lozovik and coworkers@xcite emphasized this fact , and developed an alternative self - consistent procedure from the dyson equation for the two lowest order diagrams of the expansion .
this alternative method was named the ( 3 + 4)-self consistent approximation ( sca ) .
( the sca is distinguished from the scha in that self - consistency comes from the dyson resummation of a set of skeletal diagrams in the perturbation theory , rather than by using an effective harmonic theory to take averages over the true hamiltonian . )
an important result of lozovik s work was that for long - range interactions in 2d , the anharmonic instability point lies below the dislocation mediated melting temperature , implying a first - order rather than continuous transition .
our own interest in the melting problem originates in studies of the vortex system in high-@xmath2 superconductors ( see ref . for a review .
) there is excellent experimental evidence for a first - order melting transition of the vortex lattice in clean crystals of bi@xmath3sr@xmath3cacu@xmath3o@xmath4 ( bscco)@xcite and yba@xmath3cu@xmath5o@xmath6 ( ybco),@xcite with a latent heat and magnetization step consistent with our theoretical models of the vortex system.@xcite however , the detailed understanding of where and why this transition takes place has so far relied on numerical simulations.@xcite with this in mind , we have tried to apply the above analytic extensions to elasticity theory within the london model@xcite to derive the instability point of the vortex crystal . by comparing our results to the numerical simulations of ref .
we have found , in agreement with lozovik , that the scha seriously underestimates the thermal softening of the vortex lattice ( see section [ sec : vl ] ) .
however , application of the diagrammatic ( 3 + 4)-sca did not lead to a good quantitative comparison , and so we have developed an improved self - consistent method that includes all diagrams in the ( 3 + 4)-sca and the scha . because the improved method uses an effective harmonic theory , but is equivalent to the dyson equation for all two - vertex diagrams , we call it the two - vertex - scha . the successful comparison of this new method to numerical results for the vortex lattice forms the main result of this paper . the article is structured as follows .
we first define the harmonic approximation for a generalized crystal . then in section [
sec : scha ] we describe the self - consistent harmonic approximation and its justification within a variational approach . in section [ sec : pert ] we review the perturbation expansion about the harmonic approximation , look at the diagrams corresponding to the scha and see why this approximation can give such misleading results .
the alternative ( 3 + 4)-sca of lozovik _ et al .
_ is also introduced , and we discuss the advantages and disadvantages over the scha . after learning the diagrammatic details of these self - consistent approximations , in section [ sec : new ] we introduce our improved self - consistent method , the two - vertex - scha , that includes all diagrams in both the scha , ( 3 + 4)-sca and more , while keeping some of the simplicity in treatment of the scha . finally in section [ sec : real ] we apply this improved method to some physical problems .
first we consider the thermal melting in two dimensions of the `` one - component plasma '' ( with @xmath7 interactions ) and the wigner crystal ( with @xmath8 interactions ) .
in contrast to the ( 3 + 4)-sca , our results from the two - vertex - scha leave open the possibility of dislocation - mediated melting .
we also calculate the quantum melting of 2d bosons , equivalent to thermal melting of vortex lines .
we find a significant thermal softening of the vortex lattice , which could be responsible for the observed peak effect@xcite in critical current close to the melting transition in ybco .
we consider a classical @xmath9-dimensional crystal at finite temperature , where the particle positions make small fluctuations @xmath10 about their equilibrium sites @xmath11 .
all thermodynamic properties are controlled by the partition function , @xmath12/t},\ ] ] where the hamiltonian @xmath13 $ ] gives the energy increase from the ground state for a given configuration of displacements.@xcite we work in fourier space ( which diagonalizes a harmonic hamiltonian ) , and define , @xmath14 where @xmath15 is the particle density .
the inverse transform is an integral over the brillouin zone , @xmath16 in the harmonic approximation ( ha ) we expand the hamiltonian to second order in displacements , @xmath17=h_2[{\bf u}_\mu ] + { \cal o}(u^3),\ ] ] with @xmath18 & = & \frac{1}{2n^2}\sum_\mu\sum_\nu \phi_0^{\alpha\beta}({\bf r}_\mu-{\bf r}_\nu ) u_\mu^\alpha u_\nu^\beta,\\ & = & \frac{1}{2 } \int_{\rm bz}\frac{d^dk}{(2\pi)^d } \phi_0^{\alpha\beta}({\bf k})u^\alpha_{\bf k}u^\beta_{-{\bf k}}. \end{aligned}\ ] ] the elastic matrix is defined by @xmath19 here @xmath20 is the total volume , which determines the discretization of @xmath21-space . we define the harmonic green s function as the inverse of the elastic matrix , @xmath22 . the ha , being a quadratic theory , is completely solvable , with the size of fluctuations given by the equipartition theorem .
the harmonic propagator is the thermal average , @xmath23/t}\nonumber\\ & = & t\ , \delta^d({\bf k}_1+{\bf k}_2)g_0^{\alpha\beta}({\bf k}_1),\end{aligned}\ ] ] where @xmath24 is the partition function for the harmonic hamiltonian @xmath25 . for dimensions
@xmath26 a useful measure of the fluctuations is the mean - square displacement of each particle ( in 2d we need to be careful about divergences at long wavelengths , see appendix [ ap : nn ] ) .
this is just the integral over @xmath27 of the propagator , @xmath28
from eq . ( [ eq : equi ] ) we see how the mean displacements increase with temperature . at some stage the anharmonic corrections to eq .
( [ eq : h2 ] ) become important .
one way to treat these anharmonic effects is to take the quadratic form as a trial hamiltonian , @xmath29 = \frac{1}{2 } \int_{\rm bz}\frac{d^dk}{(2\pi)^d } \phi_t^{\alpha\beta}({\bf k})u^\alpha_{\bf k}u^\beta_{-{\bf k}},\ ] ] with the elastic matrix as a set of variational parameters
. we can then use the general inequality@xcite for the free energy @xmath30 , @xmath31 so that for a given trial hamiltonian we get the best approximation to the free energy by minimizing the right hand side .
it is straightforward to show that for the quadratic form this minimization is satisfied by the effective elastic matrix , @xmath32 where the average is over the distribution defined by the trial hamiltonian ( [ eq : trial ] ) at the given temperature .
this must be solved self - consistently , which defines the scha .
let us consider the general case of a system of particles with pairwise interactions , @xmath33 where @xmath34 are the ideal crystal positions with nearest neighbor separation @xmath35 .
( [ eq : eff ] ) then becomes , @xmath36 this average over the second derivative of the potential is conveniently written in terms of the displacement fluctuations by fourier transforming the interaction potential , @xmath37 because the fields in a harmonic potential will have gaussian fluctuations , we have used the property @xmath38 for a gaussian distributed variable @xmath39 . although not necessary for calculations , the formalism simplifies greatly if we can make the approximation that @xmath40 , which means ignoring correlations between the fluctuations of different particle positions . in this case
we find , @xmath41 where @xmath42 are the reciprocal lattice vectors , and @xmath43 .
the fluctuations @xmath44 are calculated from the equipartition theorem ( [ eq : equi ] ) but with the green s function given as the inverse of the effective elastic matrix .
thus within this homogeneous approximation , we have turned the problem of finding @xmath45 for all @xmath27 to a self - consistent equation for a single parameter , @xmath44 . from ( [ eq : effsimple ] ) it is clear that at large @xmath44 the elastic matrix will be suppressed from the zero temperature limit , which in turn increases @xmath44 above the harmonic expectation .
it is this feature which drives an anharmonic instability of the crystal phase at a certain temperature .
the scha in the `` independent fluctuations '' approximation has an interesting feature for the case of coulomb interactions ( @xmath46 in three dimensions , @xmath47 in two dimensions etc . ) . by definition
the laplacian of a coulomb interaction is zero , and one can show that the renormalized interaction , and therefore the effective elastic matrix , is a non - perturbative function of @xmath44 .
this is shown explicitly for the case of logarithmic interactions in two dimensions in appendix [ ap : log ] .
for the 2d case , the single particle fluctuations diverge ( see appendix [ ap : nn ] ) even at low temperatures where the shear modulus remains non - zero.@xcite however , the shear modulus @xmath48 $ ] is dominated by the nearest - neighbor fluctuations @xmath49 ; the formula for @xmath50 is given in section [ sec:2d ] and shown in the inset of fig .
[ fig:2d_c66ofu2 ] .
this allows us to illustrate the scha in its simplest form . for this 2d coulomb system
we can approximate the equipartition result as @xmath51 ( i.e. we ignore compression modes , and assume no dispersion ) .
the two equations relating @xmath52 and @xmath53 can then be solved graphically for any temperature as in fig .
[ fig:2d_c66ofu2 ] , and the maximum value of @xmath52 for a stable self - consistent solution is easily found . in this case
we find the large value of @xmath54 at @xmath0 .
= 8 cm the following argument indicates that this instability point in the scha is a rigorous upper bound on the true stability limit of the crystal phase .
if there is no solution to the scha then by definition there is no stationary value of the rhs of eq .
( [ eq : ineq ] ) , which implies that this rhs is an unbounded function of trial parameters .
because of the inequality , there can be no finite free energy @xmath55 , and no equilibrium phase corresponding to the exact hamiltonian @xmath56 $ ] .
therefore the only possible equilibrium phase is one without a well - defined mapping from particle positions to reference lattice points . in this case
the fourier transforms @xmath57 are not the relevant variables of the system , and the liquid phase is still allowed .
unfortunately , it is known that the scha is not always a reliable approximation , and we referred in the introduction to an example where it drastically underestimates the thermal softening of the lattice.@xcite to understand why this is so we must consider which diagrams the scha represents in a perturbation expansion around the harmonic limit , and this is done in section [ sec : diagsscha ] .
we now consider the standard perturbation expansion about the ha.@xcite the exact partition function is , @xmath58 + h'[{\bf u}_{\bf k}]\right\}/t},\ ] ] where @xmath59=h[{\bf u}_{\bf k } ] - h_2[{\bf u}_{\bf k}]$ ] . as usual we expand the exponential in @xmath60 , @xmath61}{t}\right)^n e^{- h_2[{\bf u}_{\bf k}]/t}.\nonumber\end{aligned}\ ] ] with the same expansion , the propagator is , @xmath62 + h'[{\bf u}_{{\bf k}'}]\right\}/t}\nonumber\\ & = & \langle u^\alpha_{{\bf k}}u^\beta_{-{\bf k}}\rangle_0\\ & & + \sum_{n=1}^\infty\frac{(-1)^n}{n ! } \left\langle\left\|
u^\alpha_{{\bf k}}u^\beta_{-{\bf k}}\ , \left(\frac{h'[{\bf u}_{{\bf k}'}]}{t}\right)^n\right\|\right\rangle_{c , g_0}. \nonumber\end{aligned}\ ] ] the brackets @xmath63 denote a cumulant average in the ha , that is , using wick s theorem the @xmath64 fields must be paired up to give harmonic green s functions , @xmath65 , and all terms with disconnected pairings are ignored , as these cancel exactly with the expansion of the partition function @xmath66 in the denominator . to establish the terms in this series at each order in @xmath67
, we need to separate each order in @xmath64 of the hamiltonian , @xmath68=h_3[{\bf u}_{\bf k } ] + h_4[{\bf u}_{\bf k}]+ h_5[{\bf u}_{\bf k}]\,+\,\ldots\ ] ] where @xmath69 $ ] contains all terms of order @xmath70 , and will lead to a vertex in the diagrammatic expansion with @xmath71 legs .
explicitly , @xmath72 the crystal symmetry of the ground state lattice implies that the interaction tensor conserves pseudomomentum , @xmath73 where @xmath74 .
the important point is that in the ha the expectation of the propagator @xmath75 is proportional to @xmath67 , so that the average value of @xmath76 is of order @xmath77 .
we can write the perturbation expansion as , @xmath78 and then take all possible pairings using wick s theorem . we can represent this series in diagrams and the terms to order @xmath79 are shown in fig .
[ fig : diagrams ] .
the order of each diagram ( in @xmath67 ) is counted by adding up the number of lines , and subtracting the number of vertices .
notice that to order @xmath80 , only @xmath81 and @xmath82 contribute , while to order @xmath79 we also have diagrams including @xmath83 and @xmath84 .
this is a general rule : the terms @xmath85 and @xmath86 first appear to order @xmath87 .
= 8.5 cm we now make a resummation that allows us to express the full propagator @xmath88 in terms of the harmonic propagator @xmath65 and a self energy @xmath89 .
note that , because we do not include disconnected diagrams , each @xmath90 in ( [ eq : pert ] ) must join with a @xmath91 in @xmath60 when using wick s theorem .
this means we can rewrite ( [ eq : pert ] ) in the form , @xmath92 inspection of fig .
[ fig : diagrams ] shows that many possible terms will be trivial `` copies '' of lower order terms .
these are the reducible diagrams , for instance group ( b ) in fig .
[ fig : diagrams ] , which have the property that when we group all of the reducible diagrams corresponding to a given irreducible diagram , they form a resummable series .
we can write this in the compact form @xmath93 with the self energy defined as , @xmath94 the subscript @xmath95 in the average tells us to only include irreducible contributions , that is , any term which can be separated by cutting just one leg should be ignored , as it is already included in the full green s function on the rhs of ( [ eq : greens ] ) .
a further resummation is possible by evaluating the rhs of ( [ eq : selfenergy ] ) with respect to the renormalized , rather than the harmonic , green s function .
the functional form of the self energy will then depend self - consistently on @xmath88 , @xmath96 where the subscript @xmath97 tells us only to include `` skeletal '' diagrams in the average .
we define a skeletal diagram as one that has not already been included in the resummations of lower order diagrams by using the full green s functions . for instance
the diagrams of group ( c ) in fig .
[ fig : diagrams ] must be ignored , as they are included when we evaluate the self - energy contribution from group ( a ) with the renormalized propagator .
after these resummations we have a dyson - type equation , which must be solved self - consistently for @xmath88 , @xmath98 note that this inverse propagator is exactly the response function @xmath99 of the fluctuating crystal for an external force .
this is demonstrated in appendix [ ap : greens ] . due to the difficulties of calculating all possible diagrams in the perturbation expansion
, one looks for approximate resummations .
a simple example is to only take the lowest order skeletal diagrams ( group ( a ) in fig .
[ fig : diagrams ] ) in the self energy , which is known as the ( 3 + 4)-sca.@xcite this is shown in fig .
[ fig : diags_34 ] , and it corresponds to solving the dyson equation with the self energy , @xmath100 note that this method gives the exact propagator to order @xmath80 , although it only includes six of the thirteen diagrams at order @xmath79 . the two skeletal diagrams are typically of opposite sign and , importantly , for all the cases of 2d power law interactions looked at by lozovik _
the magnitude of the positive @xmath101 term is slightly greater than the negative @xmath82 term .
this means the lattice first softens with temperature , but any approximation that does not include the @xmath101 term will incorrectly find a hardening of the lattice with increasing temperature .
more details of the ( 3 + 4)-sca and calculations for the example of logarithmic interactions in 2d are given in appendix [ ap:34 ] .
= 7.5 cm it is insightful to consider which diagrams the scha ( see section [ sec : scha ] ) corresponds to . writing @xmath102 we see that eq .
( [ eq : eff ] ) becomes @xmath103 comparing this to ( [ eq : dyson ] ) and equating @xmath104 to a green s function @xmath105 we see that the scha corresponds to solving the dyson equation with the self energy given by @xmath106 which is equivalent to the @xmath107 part of the sum in ( [ eq : selfenergyr ] ) ( for @xmath107 there are no non - skeletal diagrams , so that the averages are equivalent ) .
therefore the scha is formally identical to the diagrammatic equation in fig .
[ fig : diags_scha ] .
note that only even order vertices are present , as the average of any odd combination over the elastic hamiltonian is zero .
this is an important feature of the scha .
( an identical diagrammatic series appears for a form of scha in the context of the sine - gordon model@xcite . in this case
, the cosine potential only has terms even in the field , and so the scha has a better chance of being an accurate approximation . ) for our problem the scha only includes one of the two diagrams that contribute to the propagator at order @xmath80 , and only three of the thirteen diagrams at the next order ( compare to fig .
[ fig : diagrams ] ) .
the usefulness of the scha is that it is straightforward to calculate averages of the true hamiltonian over a gaussian distribution .
this avoids the difficult task of actually evaluating high order vertices in a diagrammatic expansion .
( e.g. for a pairwise interaction one must evaluate a sum over @xmath71 derivatives of the potential to find the @xmath71-th vertex .
this is done for @xmath108 and @xmath109 in appendix [ ap:34 ] . )
we want to develop a method that combines the simplicity and elegance of the scha with the accuracy at lowest order of the ( 3 + 4)-sca .
we note that the scha is equivalent to evaluating just the @xmath107 terms in the self energy ( [ eq : selfenergy ] ) , while it is the @xmath110 part that includes the `` flying - saucer '' diagram that is so important in the ( 3 + 4)-sca .
the contribution to the self energy from @xmath110 is , @xmath111 to write this as an average over the full hamiltonian with a given distribution of displacements , we must subtract the diagrams that were included in resummations of the @xmath107 diagrams . allowing for this , and separating @xmath112 leads to the form ,
@xmath113 note that terms explicitly including @xmath25 have cancelled with each other .
this corresponds to calculating , within an scha type method , the effective elastic matrix , @xmath114.\nonumber\end{aligned}\ ] ] we call this equation the `` two - vertex - scha '' .
the extra terms look rather like a fluctuation contribution to the elasticity , but we should remember that it is simply the two - vertex part of the dyson series , and there will in principle be further corrections with higher numbers of vertices . the diagrammatic representation of the two - vertex - scha is shown in fig .
[ fig : diags_n2 ] .
this scheme now incorporates the diagrams from the scha and the ( 3 + 4)-sca along with more terms not included in these last two approaches .
for instance , it includes nine of the thirteen diagrams for the propagator at order @xmath79 , see fig .
[ fig : diagrams ] .
it also has the feature of the scha that one only needs to calculate averages of derivatives of the full hamiltonian , and not any diagrams explicitly .
the 2d coulomb system , also called the one - component plasma , consists of particles interacting with a pairwise , repulsive logarithmic potential , @xmath115 it has been studied in the past partly because of some special mathematical properties@xcite ( the system is exactly soluble at one particular temperature @xmath116 in the liquid phase ) .
numerical simulations@xcite have found evidence for a weak first order transition at @xmath117 .
physically , this system is of interest as it is applicable to a thin superconducting film with field - induced vortices which interact logarithmically@xcite up to a screening length @xmath118 where @xmath9 is the thickness of the film and @xmath119 is the bulk penetration depth , typically of order 1000 .
such films were predicted to undergo a 2d melting transition of the vortex lattice well below the zero - field transition temperature.@xcite ( a similar scenario takes place in layered superconductors at high fields , see ref . . )
the ground state of the 2d coulomb model is a triangular lattice , with nearest - neighbor spacing @xmath120 .
this lattice has a shear modulus of @xmath121= \frac{nv}{8}\ ] ] ( this result was known over thirty years ago@xcite ) and a diverging compression modulus at large wavelengths , @xmath122 in refs . and
it was presumed that the melting transition would be of the continuous kosterlitz - thouless - halperin - nelson - young ( kthny ) type@xcite with a dislocation - unbinding mechanism , which is predicted to occur at @xmath123 with @xmath124 the renormalization constant for the shear modulus at finite temperature , @xmath125 . however ,
this is not consistent with the first - order result of simulations.@xcite one possible reason for this discrepancy is that regular fluctuations in an anharmonic potential , such as we considered in the last section , can destabilize the lattice before singular fluctuations such as free dislocations are generated . therefore in this section we consider the anharmonic softening and instability of the 2d coulomb lattice , and compare to the predictions of the kthny theory .
we calculate @xmath126 within the scha .
we also find the first order correction in temperature to the shear modulus using perturbation theory , by calculating the diagrams of group ( a ) in fig .
[ fig : diagrams ] .
this shows that the scha greatly underestimates the thermal softening .
we then present our results for @xmath126 from the ( 3 + 4)-sca and the two - vertex scha and compare to the kthny universal value of @xmath127 .
the scha for the 2d coulomb system was briefly looked at in ref . .
in fact , the scha with `` independent fluctuations '' has some nice features for this system due to the laplacian of the interaction being zero . in appendix [ ap : log ] we derive the non - perturbative form of the elastic moduli as a function of @xmath128 , which leads to the result for the shear modulus , @xmath129 this formula is shown in the inset of fig .
[ fig:2d_c66ofu2 ] ( the nearest - neighbor approximation here is extremely accurate ) .
the non - perturbative feature ( the extreme flatness at small @xmath52 ) will be lost when we include the correlations in fluctuations between different particles . in fact , we have found that the scha predicts an initial stiffening of lattice with small fluctuations , which we understand to be an incorrect result due to the neglect of diagrams with odd vertices .
inspection of eq .
( [ eq : dyson ] ) shows the correct temperature expansion of the change in shear modulus to be , @xmath130 we can find the first correction to the shear modulus to order @xmath67 by evaluating the bare self - energy contributions in the diagrams of fig . [
fig : diags_34 ] .
formulas for the ( 3 + 4 ) self - energies for a system with pairwise interactions are given in appendix [ ap:34 ] .
we have numerically evaluated these formulas for the 2d coulomb system , and our results for the transverse @xmath131 and longitudinal @xmath132 parts taking @xmath133 are shown in fig .
[ fig : s3s4_2d ] .
the large wavelength limits are found to go as , @xmath134 @xmath135 = 8 cm from eqs .
( [ eq : deltamu ] ) and ( [ eq : selftrans ] ) the shear modulus at low temperatures will be given by @xmath136 so that within first - order perturbation theory , @xmath126 is about @xmath137 of its zero temperature value at the numerically observed transition at @xmath117 .
the large wavelength limit of the longitudinal part of the self energy in ( [ eq : selflong ] ) is a constant , so that the compressional modes are also softened to first order in temperature , @xmath138 this thermal softening of the long - wavelength compression energy ( a special feature of the 2d coulomb system ) might seem to contradict the physically intuitive expectation that the compressibility is fixed by the bare charge of the system . a full understanding of this result will be discussed in a future paper , but it is due to the non - linear relation between changes in the density and the longitudinal displacements , @xmath139 .
we can go beyond the lowest order in @xmath67 using the ( 3 + 4 ) self energy by iterating the equation of fig .
[ fig : diags_34 ] until a stable solution is found within this ( 3 + 4)-sca .
we have also found numerical solutions within the two - vertex - scha of section [ sec : new ] .
details for calculating the effective elastic moduli of eq .
( [ eq : neweff ] ) for pairwise interactions are given in appendix [ ap : n=2 ] . as we have shown in section [ sec : new ] , this approximation gives identical results to lowest order in @xmath67 as the ( 3 + 4)-sca .
however , we find important deviations at larger temperatures . in fig .
[ fig : c66oft_2d ] the results for @xmath126 as a function of temperature from both methods are shown .
the shear modulus was recently measured in the same system with langevin - dynamics simulations,@xcite and we find good agreement between the measured temperature dependence ( the circles in fig .
[ fig : c66oft_2d ] ) and the two - vertex - scha calculation .
notice that the instability point from the ( 3 + 4)-sca is below the numerically measured melting temperature of @xmath140 , while the upper limit to solutions of the two - vertex - scha is slightly above this temperature . in fig .
[ fig : c66oft_2d ] we also compare to the predictions of the kthny theory of 2d melting .
as described above , this theory predicts a universal value of the shear modulus at the dislocation - unbinding transition , @xmath141 , shown as the dot - dashed line in fig .
[ fig : c66oft_2d ] .
the fact that the two - vertex - scha result crosses this value tells us that the results at temperatures above this value are unphysical , as the crystal is unstable to free dislocations ( if the ( 3 + 4)-sca were correct , it would practically rule out the kthny mechanism for this system ) .
therefore the crossing of the universal jump and the two - vertex - scha result is an upper bound to a melting transition and we find @xmath142 ( the effect of thermally generated dislocation pairs will push @xmath126 , and the melting temperature , down further ) . the fact that at melting the shear modulus is so close to the kthny universal value tells us that dislocation - pairs ( not included in our analysis ) as well as anharmonicities both play an important role in the physics of the melting transition . whether the weak first - order nature seen in simulations is correct , or a numerical artifact , remains unclear .
= 8 cm it is experimentally feasible to confine electrons to a flat surface in various systems , such that they are free to move in only two dimensions . in particular , for electrons trapped on the surface of liquid helium by an applied electric potential , a classical melting transition from a 2d `` wigner crystal '' to a liquid state has been observed.@xcite the elastic properties of the 2d wigner crystal are well studied,@xcite and platzman and fukuyama@xcite considered the effect of thermal fluctuations using the scha , reporting a dramatic stiffening of the lattice with temperature . using the perturbation expansion of section [ sec : pert ] , fisher@xcite showed this to be incorrect , as the first order term in @xmath67 gives a downward correction to @xmath126 .
et al . _ have applied their ( 3 + 4)-sca to this problem,@xcite and stated that an anharmonic instability is reached below the kthny temperature . however , the discrepancy in section [ sec:2d ] for logarithmic interactions between the ( 3 + 4)-sca and our two - vertex - scha motivates us to reconsider the 2d wigner crystal .
electrons interact with a repulsive pairwise potential , @xmath143 where @xmath144 is the electron s charge .
the ground state is a triangular lattice , spacing @xmath35 , and at finite temperatures all properties depend on the dimensionless ratio @xmath145 . in the literature
the parameter @xmath146 is more often used .
when calculating the elastic matrix , care must be taken to split the interaction to long- and short - range parts , and treat them separately as in the ewald technique.@xcite then the zero - temperature shear modulus is found to be , @xmath147 and the compression modulus diverges at small @xmath21 ( as for the coulomb crystal but with a different power ) as @xmath148 repeating the perturbation theory of section [ sec : pert ] we find the first - order correction to the shear modulus to be , @xmath149 in agreement with ref . .
this correction again comes from a negative contribution from @xmath150 and a positive , but smaller in magnitude , contribution from @xmath151 .
we have also confirmed that the scha gives an incorrect stiffening of @xmath126 with increasing temperature .
= 8 cm in fig .
[ fig : c66oft_wig ] we plot our results for the shear modulus @xmath152 from the ( 3 + 4)-sca and the two - vertex - scha .
the resulting picture is very similar to that for logarithmically interacting particles in fig .
[ fig : c66oft_2d ] .
our instability point for the ( 3 + 4)-sca of @xmath153 is close to the quoted value from lozovik _ et al .
_ of @xmath154 .
the two - vertex - scha is again more stable , and together with the kthny condition on the minimum allowed @xmath152 gives @xmath155 .
this is close to the experimental melting temperature of @xmath156,@xcite valid over a range of experimental densities .
also , numerical simulations have found a melting transition in the range from @xmath157 ( ref . ) to @xmath158 ( ref . and ) to @xmath159 ( ref . ) .
( it is not clear why there is such disagreement between the numerical results .
the rather high melting temperature from the earliest simulation@xcite is probably an out of equilibrium artifact ; the system was started in the crystal state and heated up . ) in the previous sections we have ignored the effect of quantum fluctuations .
the uncertainty principle can prevent the formation of a lattice phase if the mass of the particles is sufficiently low . in this section
we will consider the effect of quantum fluctuations on a 2d coulomb lattice of bosons at zero temperature . in the path - integral representation , the action of the boson system in imaginary time is , @xmath160 with @xmath161 given in eq .
( [ eq : logint ] ) .
the properties of the system are determined by the statistical sum , or partition function , @xmath162 $ ] . at zero temperature
all properties depend only on the unitless de boer parameter , @xmath163 , which measures the ratio of zero - point kinetic energy to the interaction energy .
this model maps directly to a 3d classical system at finite temperature , where the world lines of the bosons correspond to fluctuating elastic strings .
in fact , the coulomb interaction between the strings provides a model of the interactions between vortex lines in a type - ii superconductor ( this fruitful analogy was discovered by nelson@xcite ) .
thus the path - integral representation of 2d bosons has been studied with the motivation of describing vortex - lattice melting@xcite which occurs at finite magnetic fields in high-@xmath2 superconductors.@xcite to be specific , the 3d vortex system is obtained by replacing @xmath164 with @xmath67 , the time @xmath165 with the distance along the field direction @xmath166 and the mass @xmath71 with the line elasticity @xmath167 .
( in fact , this model only approximates the 3d interactions of the vortex system . a more correct mapping would be from a boson system with interactions non - local in time . )
the strength of the interaction per unit length is determined by the penetration depth @xmath119 of the superconductor via the relation @xmath168 with @xmath169 where the quantum of flux is @xmath170 . in these units
the de boer parameter is @xmath171 in the numerical simulations@xcite a melting transition occurs at @xmath172 , above which the system is in a line - liquid phase .
although it has proved difficult to measure the equilibrium values of the elastic moduli in the same simulations , the value of @xmath44 was measured , and shows deviations from a linear dependence on temperature before melting occurs.@xcite in this section we use our self - consistent treatment to find the effect of anharmonicity and the instability point in this system and compare our results to the measured @xmath44 in the simulations .
we also calculate the quantitative effect of fluctuations on the shear modulus of the vortex crystal . in the following we use the notation for the classical vortex system , but
all results are easily transferred to the 2d quantum system at zero temperature .
the generalizations to fluctuating lines of the results of sections 25 are straightforward , where the displacement variables are 2d vectors in 3d @xmath27-space , @xmath173 the elastic matrix and the self energy are now functions of @xmath174 as well as the 2d vector @xmath175 .
brief details of the ( 3 + 4)-sca and the two - vertex - scha are given in appendix [ ap : string ] . at large wavelengths
the transverse elastic matrix is given by @xmath176 and the longitudinal modes by @xmath177 where @xmath178 , @xmath179 and @xmath180 are the voigt notation for the compression , shear and tilt moduli respectively .
we can again find the first correction to the shear modulus by evaluating the self energy to lowest order in @xmath67 .
we find that @xmath181 and @xmath182 , where @xmath183 .
therefore to lowest order in the de boer parameter , @xmath184 where the shear modulus without fluctuations is @xmath185 .
the compression modulus at large wavelengths is not renormalized .
note that the `` hartree '' self - energy @xmath186 is independent of the @xmath166-component of the wave - vector ( see appendix [ ap : string ] ) , but the contribution @xmath187 does depend on @xmath174 .
for example , for @xmath175 at the brillouin zone we find , @xmath188 , which gives a fluctuation induced stiffening of the tilt modulus , @xmath189 we have made detailed calculations for the two - vertex - scha for this model , and the results for the shear modulus as a function of @xmath190 are shown in fig .
[ fig : c66oft_3d ] .
it turns out that the effective elastic matrix ( see appendix [ ap : string ] ) depends on the full correlations of displacement fluctuations of a given vortex line in the @xmath166-direction , and we need to numerically find a self - consistent solution for the function , @xmath191 in practise we discretize along the @xmath166-direction , making sure that @xmath192 is smoothly varying .
= 8 cm we again see that the stability limit from the two - vertex - scha is higher than that from the ( 3 + 4)-sca .
surprisingly , the crystal maintains a self - consistent solution for a shear modulus which is only 20% of the zero - temperature value .
this contrasts with our results for a 2d crystal , and is mainly due to the stiffness of the elastic lines : the average displacements within the equipartition theorem only grow with the inverse square - root of the shear modulus,@xcite @xmath193 .
the result is qualitatively similar to the measured shear modulus in simulations of coupled vortex pancakes in a layered superconductor in ref . .
in their somewhat different model of a vortex lattice , a reduction in the shear modulus at melting of about 30% of the zero temperature value was found . also , we find that the tilt modulus stiffens slightly with the fluctuations , allowing for a still softer shear modulus before instability is reached .
= 8 cm in fig .
[ fig : u2oft_3d ] we compare the mean fluctuation per particle @xmath194 with the values measured in simulations ( see fig . 8 of ref . ) .
the simulation results come from fitting a debye - waller factor to the measured widths of the crystal structure factor.@xcite the main point is that the instability point predicted by the two - vertex - scha , @xmath195 is very close the numerically measured melting point @xmath196 .
the fact that we find @xmath197 slightly below melting means that the two - vertex - scha still slightly overestimates the level of softening , and correspondingly there is an increasing discrepancy with the numerical results for @xmath44 close to the melting temperature .
we should of course remember that the two - vertex - scha is only a limited resummation of the perturbation theory .
however , our new method clearly gives much superior results compared to the ( 3 + 4)-sca , also shown in the figures , and to the scha , which we found to give very little change to the elastic moduli below the numerically found melting and an instability bound over twice the value of @xmath198 .
the thermal softening of the shear modulus of the vortex lattice is rarely explicitly calculated in the literature .
the measurements from simulations in ref . agree with fig .
[ fig : c66oft_3d ] in that this softening is really significant .
it may help to explain the peak effect in the critical current seen close to the vortex - lattice melting transition in ybco.@xcite the weak collective pinning theory predicts that , for a non - dispersive tilt modulus , the critical current depends on the elastic moduli as @xmath199 where @xmath200 is a pinning strength parameter.@xcite however , the effect of finite temperature is not only to reduce the shear modulus , but also to smooth the effective pinning landscape over @xmath201 , thus reducing the pinning force . to see if our results lead to a critical - current peak as a function of temperature requires a careful analysis of the @xmath27 dependent elastic moduli ( the critical pinning volume depends on long - wavelength distortions but @xmath201 comes from integrating over all modes ) , and we leave this for a future project .
in this paper we have introduced a new method , the two - vertex - scha , to treat the anharmonicity of simple lattices in the presence of thermal , or quantum , fluctuations .
the method is based on the diagrammatic expansion about the ha , but takes the form of an extension to the scha calculation of effective elastic moduli . by considering two long - range interacting systems at finite temperature , and a system of zero - temperature 2d bosons , we have seen that the two - vertex - scha is a significant improvement over previously used self - consistent techniques , the scha and the ( 3 + 4)-sca , which respectively under- and overestimated the fluctuation - induced softening . it is the comparison to simulation results that gives us the most confidence in the reliability of the two - vertex - scha as a tool for calculating the softening and instability of such simple crystals due to thermal or quantum fluctuations .
one should therefore not trust the results of previous treatments of the anharmonic softening of 2d lattices,@xcite but use instead the two - vertex - scha .
in addition to establishing this new method , we have found important results for the vortex - lattice melting problem , using the bose model for the finite - temperature vortex system .
we have found a much larger than expected thermal softening of vortex lattice with the shear modulus reducing to 20% of its zero temperature value at the melting transition , which may have important consequences for the pinning properties of the vortex lattice .
it should be noted , however , that the real vortex system does not only have interactions within planes of equal height , as in the bose model . the full london model of vortices corresponds to a 2d bose system with time - nonlocal interactions.@xcite an investigation to extend the results of the bose model to the full time - nonlocal case is underway.@xcite another facet of the vortex system in high-@xmath2 superconductors is the layered nature of the cuprates .
for very weak inter - layer coupling , the system can be described by magnetically coupled `` pancake '' vortices . in a recent paper
the two - vertex - scha has been used to find the instability line of such a pancake - vortex lattice.@xcite it lies just above the melting line as calculated with a free energy comparison using results from 2d simulations for the pancake - vortex liquid phase .
finally , it would also be interesting to investigate the stability of lattices in the presence of quenched disorder .
this is also relevant for the vortex lattice , where pinning disorder exists , and a disorder - induced melting transition occurs as the vortex density , or magnetic field , is increased.@xcite we thank anne van otterlo , henrik nordborg , alex koshelev , boris ivlev and valerii vinokur for useful discussions , and the swiss national foundation for financial support .
mjwd is supported by an epsrc advanced fellowship af/99/0725 .
in the scha defined by eq .
( [ eq : effpair ] ) , we have to take averages over the second derivative of the potential between two particles as they fluctuate within the crystal ( similar averages are needed in the two - vertex - scha , see appendix [ ap : n=2 ] ) . within the gaussian approximation these averages only depend on the mean square , @xmath128 , which in two dimensions is given by @xmath202 \approx \frac{t}{\varepsilon_v}\ln \left(r_\mu / a\right),\ ] ] ( @xmath203 is an energy - density scale set by the interaction @xmath204 ) .
note that the fluctuations of a single particle , @xmath205 diverges in 2d .
we show here that the logarithmic divergence at large distances is not important for the elastic moduli which have dominant contributions from fluctuations at small distances .
consider the difference between the renormalized and the zero temperature elastic matrix @xmath206.\ ] ] now the fourier transform only has weight at values of @xmath207 , so as long as @xmath208 we can write @xmath209 and , @xmath210 now the @xmath211 dependence of @xmath212 is logarithmic and weak compared to the short - range variation in @xmath213 , so short distances will dominate and we are justified to approximate fluctuations by the nearest - neighbor result @xmath214 .
similar arguments apply for the sums that appear in the two - vertex scha .
( for logarithmic interactions , @xmath215 , and the perturbation theory is not helpful .
however , for this case the non - perturbative result ( [ eq : c66ofu2 ] ) derived in appendix [ ap : log ] is clearly dominated by the nearest neighbors . )
in this appendix we show how to calculate the elastic matrix of a 2d lattice with logarithmic interactions within the scha as given by ( [ eq : effpair ] ) and demonstrate the non - perturbative dependence on the size of fluctuations @xmath44 .
we need the thermal average of the second derivatives of the interaction potential between each pair of particles . assuming the displacement fluctuations to be gaussian distributed , this is , @xmath216 where the derivatives on the rhs do not act on the distribution function , @xmath217 the width of the fluctuations is given by the matrix @xmath218 [ u^\beta({\bf r}_\mu)-u^\beta({\bf 0})]\rangle,\ ] ] and @xmath219^{\frac{1}{2}}$ ] .
we take the isotropic fluctuation approximation , @xmath220 , which we find to be good from numerical evaluation of ( [ eq : equiofk ] ) in the ha . in this case , because the potential is circular symmetric , the average will be independent of the angle @xmath221 of the ground state position @xmath34 , so @xmath222 where we have made use of the divergence theorem .
we are interested in the case of logarithmic interactions , , and we use the fact that @xmath223=2\pi\delta^2({\bf r})$ ] to give @xmath224 = -v\ln(r_\mu/\xi)- v \int^\infty_{r_\mu}\frac{dr}{r } e^{-r^2/2\sigma_{\mu 0 } } + \hbox{const},\ ] ] where the integration constant depends on @xmath225 and @xmath226 but not on @xmath211 . notice that the correction term to the zero - temperature potential is non - perturbative in the fluctuation @xmath225 .
the physics of this result is clear : because the 2d - laplacian of the interaction potential is zero , gauss theorem applies and the circular symmetric potential gives the same result as if all of the `` mass '' was at a point in the center of the circle ( the ground state position @xmath34 ) .
the non - perturbative correction appears from the exponentially small tail of the distribution which is outside of the radius @xmath211 .
this `` shell '' does not contribute to the average interaction , so it s weight must be subtracted from the bare interaction . we can easily differentiate the correction term with respect to the lower limit of the integral , to find @xmath227 e^{-\frac{{r_\mu}^2}{2\sigma_{\mu 0}}}.\ ] ] this expression should be inserted into ( [ eq : effpair ] ) to give the renormalized elastic matrix .
it is then straightforward to derive the softening of the shear modulus within the scha as given in eq .
( [ eq : c66ofu2 ] ) and fig . [ fig:2d_c66ofu2 ] .
we demonstrate in this appendix that the elastic response to an external force is given by the inverse propagator ( or green s function ) . in the presence of a force @xmath228 acting on each particle , the hamiltonian is changed to , @xmath229=h[{\bf u}_\mu ]
-\sum_\mu { \bf f}_\mu\cdot { \bf u}_\mu.\ ] ] the definition of elastic response within the thermally fluctuating region is that the force is given by hooke s law , @xmath230 which is valid to lowest order in the mean displacements @xmath231 .
the displacements are calculated with straightforward statistical mechanics , @xmath232-\sum_{\lambda}{\bf f}_{\lambda}\cdot { \bf u}_{\lambda}\right ) /t},\\ & = & \frac{1}{t}\sum_\mu f_\mu^\alpha \langle u_\nu^\beta u_\mu^\alpha\rangle_{f=0 } + { \cal o}(f^2).\end{aligned}\ ] ] if we have lattice - translational symmetry , the fourier transform of this reads as @xmath233 , where the green s function is defined by @xmath234 , so that the elastic matrix in @xmath27-space is exactly the inverse green s function , @xmath235 note that this example of the classical fluctuation - dissipation theorem goes beyond the equipartition result , which is only valid for the harmonic approximation .
the true elastic response of the fluctuating system is always given by the inverse of the propagator . if we construct the free energy corresponding to @xmath236 , we will find the result , @xmath237 this is for a system with forces specified .
however , one is often interested in the case where a certain displacement is enforced ( e.g. when boundary conditions are changed in a numerical simulation ) .
we must then make a legendre transformation to the relevant thermodynamic potential that depends on a fixed displacement , @xmath238)&=&f_f + \sum_\mu f_\mu^\alpha u_\mu^\alpha\\ & = & f_{f=0}+\frac{1}{2}\int_{\rm bz } \frac{d^dk}{(2\pi)^d } ( g^{-1})^{\alpha\beta}({\bf k } ) u^\alpha_{\bf k } u^\beta_{-{\bf k}}. + { \cal o}(u^3).\end{aligned}\ ] ] with this formulation , we see that the elastic matrix may be found as the second derivative of the thermodynamic potential with fixed displacements , @xmath239 . for example
, the @xmath240 shear modulus may be measured as the response to the change in the angle @xmath241 determining the boundaries of a finite system as in ref . .
as the hamiltonian will depend on the `` displacement '' @xmath241 , we can differentiate the relevant free energy , @xmath242 , to give , @xmath243\ ] ] the important point is that this shear modulus is exactly the same as the shear modulus found from the @xmath240 limit of the inverse of the transverse propagator @xmath244 $ ] .
in this appendix we show how to explicitly calculate the ( 3 + 4 ) diagrammatic expansion of fig .
[ fig : diags_34 ] , for the case of pairwise interactions . for a hamiltonian of the form ( [ eq : pairwise ] ) , the third and fourth rank anharmonic tensors defined by eqs .
( [ eq : hm ] ) and ( [ eq : atophi ] ) will be given by , @xmath245 and @xmath246 with @xmath247 .
the above poisson resummations to reciprocal - lattice sums allow for more convenient evaluation , especially in the case of long - ranged interactions .
the contribution to the self energy from the `` flying - saucer '' diagram is , @xmath248 while the contribution from the first , `` hartree''-type diagram is @xmath249
we now describe the details of the two - vertex - scha of section [ sec : new ] for a system with pairwise interactions . splitting the effective elastic matrix of eq .
( [ eq : neweff ] ) term by term we have , @xmath250 with @xmath251 given by the scha formula ( [ eq : effpair ] ) , @xmath252 and , @xmath253 where the averages are taken with respect to the effective elastic hamiltonian . in the independent fluctuation approximation , we can write the entire elastic matrix as a function of @xmath44 and @xmath67 ( note that in 2d , although @xmath44 diverges , we can take the nearest - neighbor value @xmath52 as the sum in ( [ eq : phi2real ] ) is dominated by short distances ) , @xmath254 + \frac{n^2}{2t}\ ! \int \!\!\ !
\frac{d^dq'}{(2\pi)^d}\ ! \sum_{\bf q } \left [ 2 h^{\alpha\beta}_{{\bf q}-{\bf q}',{\bf q } ' } - h^{\alpha\beta}_{{\bf q}-{\bf q}'-{\bf k},{\bf q } ' } - h^{\alpha\beta}_{{\bf q}-{\bf q}'+{\bf k},{\bf q } ' } \right],\ ] ] with @xmath255,\ ] ] and @xmath256 ^ 2.\ ] ] finally the correction term is given by , @xmath257 to complete the two - vertex - scha one must find a self - consistent solution numerically for @xmath44 by inserting the effective elastic matrix in the equipartition result ( [ eq : equi ] ) .
we now outline the generalization of the self - consistent methods to the problem of elastic strings directed along @xmath166 and interacting within 2d planes of equal @xmath166 .
first , the harmonic energy is given by , @xmath258 u_{\bf k}^\alpha u_{-{\bf k}}^\beta.\ ] ] the generalization of eq .
( [ eq : flyingsaucer ] ) is , @xmath259 and of eq .
( [ eq : hartree ] ) is , @xmath260 the matrices with @xmath261 subscripts are to be calculated for the interactions within a given 2d plane .
for the two - vertex - scha we now have to be careful to include the @xmath166-correlations in the fluctuation terms , @xmath262 .
we can again split the effective elastic matrix as in ( [ eq : split ] ) , with @xmath263 being the 2d scha elastic modulus plus the single string elasticity , and @xmath264\\ & & + \frac{n^2}{2t}\int \frac{d^2q'}{(2\pi)^2}\sum_{\bf q } \int^\infty_{-\infty}dz \,\,e^{-ik_zz } \left [ 2 h^{\alpha\beta}_{{\bf q}-{\bf q}',{\bf q}'}(z ) - h^{\alpha\beta}_{{\bf q}-{\bf
q}'-{\bf k},{\bf q}'}(z ) - h^{\alpha\beta}_{{\bf q}-{\bf q}'+{\bf k},{\bf q}'}(z ) \right],\nonumber\end{aligned}\ ] ] with @xmath265,\ ] ] and @xmath266 ^ 2,\ ] ] where @xmath192 is defined in eq . ( [ eq : u2ofz ] ) . the last ,
correction term is given by , @xmath267 w. k. kwok , j. a. fendrich , c. j. van der beek , g. w. crabtree , phys .
lett . * 73 * , 2614 ( 1994 ) ; k. deligiannis , p. a. j. de groot , m. oussena , s. pinfold , r. langan , r. gagnon , and l. taillefer , phys .
79 * , 2121 ( 1997 ) ; t. nishizaki , t. natio , and n. kobayashi , phys .
b * 58 * , 11169 ( 1998 ) .
we ignore the cohesion energy of the ground state as it does not influence the fluctuations considered in this paper .
however , if we wanted to calculate the full free energy , e.g. to compare to the liquid state , we would need to include this ground state energy .
d. ertas and d. r. nelson , physica c * 272 * , 79 ( 1996 ) ; t. giamarchi and p. le doussal , phys .
b * 55 * , 6577 ( 1997 ) ; a. van otterlo , r. t. scalettar and g. t. zimanyi , phys . rev . lett . * 81 * , 1497 ( 1998 ) . | we consider the thermal softening of crystals due to anharmonicity .
self - consistent methods find a maximum temperature for a stable crystal , which gives an upper bound to the melting temperature .
previous workers have shown that the self - consistent harmonic approximation ( scha ) gives misleading results for the thermal stability of crystals .
the reason is that the most important diagrams in the perturbation expansion around harmonic theory are not included in the scha .
an alternative approach is to solve a self - consistent dyson equation ( sca ) for a selection of diagrams , the simplest being a ( 3 + 4)-sca .
however , this gives an unsatisfactory comparison to numerical results on the thermal and quantum melting of two - dimensional ( 2d ) coulomb - interacting particles ( equivalent to vortex - lattice melting in two and three dimensions ) .
we derive an improved self - consistent method , the two - vertex - scha , which gives much better agreement to the simulations .
our method allows for accurate calculation of the thermal softening of the shear modulus for 2d crystals and for the lattice of vortex - lines in type - ii superconductors . 2 |
the far ultraviolet ( fuv ) and extreme ultraviolet ( euv ) continua of active galactic nuclei ( agn ) are thought to form in the black hole accretion disk ( krolik 1999 ; koratkar & blaes 1999 ) , but their ionizing photons can influence physical conditions in the broad emission - line region of the agn as well as the surrounding interstellar and intergalactic gas .
the metagalactic background from galaxies and agn is also an important parameter in cosmological simulations , as a dominant source of ionizing radiation , critical for interpreting broad emission - line spectra of agn , intergalactic metal - line absorbers , and fluctuations in the ratio of the ly@xmath1 absorbers of and . since the deployment of the first space - borne ultraviolet ( uv ) spectrographs
, astronomers have combined spectral observations of agn into composite spectra .
these composites constrain the intensity and shape of the agn component of the ionizing photon background .
the most direct probe of the fuv and euv continua in the agn rest frame comes from observations taken by instruments such as the _ international ultraviolet explorer _ ( iue ) and a series of uv spectrographs aboard the _ hubble space telescope _ ( _ hst _ ) : the goddard high resolution spectrograph ( ghrs ) , the faint object spectrograph ( fos ) , the space telescope imaging spectrograph ( stis ) , and the cosmic origins spectrograph ( cos ) .
ultraviolet spectra were also obtained by the _ far ultraviolet spectroscopic explorer _ ( _ fuse _ ) .
for agn at modest redshifts , all of these instruments provide access to the rest - frame lyman continuum ( lyc , @xmath10 ) , and at @xmath11 they avoid strong contamination from the ly@xmath1-forest absorbers in the spectra of high - redshift agn .
thus , obtaining access to high - s / n , moderate - resolution uv spectra is crucial for finding a reliable underlying continuum .
+ this is our second paper , following a agn composite spectrum presented in paper i ( shull et al .
2012 ) based on initial results from cos spectra of 22 agn at redshifts @xmath12 .
we analyzed their rest - frame fuv and euv spectra , taken with the g130 m and g160 m gratings , whose resolving power @xmath13 ( 17 km s@xmath14 velocity resolution ) allows us to distinguish line blanketing from narrow interstellar and intergalactic absorption lines . here , in paper ii , we enlarge our composite spectrum from 22 to 159 agn , confirm the validity of our early results , and explore possible variations of the euv spectral index with agn type and luminosity .
both studies were enabled by high - quality , moderate - resolution spectra taken with the cosmic origins spectrograph installed on _ hst _ during the may 2009 servicing mission .
the cos instrument ( green et al . 2012 ) was designed explicitly for point - source spectroscopy of faint targets , particularly quasars and other agn used as background sources for absorption - line studies of the intergalactic medium ( igm ) , circumgalactic medium ( cgm ) , and interstellar medium ( ism ) .
our survey is based on high - quality spectra of the numerous agn used in these projects .
+ our expanded survey of 159 agn finds a composite spectral energy distribution ( sed ) with frequency index @xmath9 in the rest - frame euv .
this confirms the results of paper i , where we found @xmath15 .
we adopt the convention in which rest - frame flux distributions are fitted to power laws in wavelength , @xmath16 , and converted to @xmath5 in frequency with @xmath17 .
we caution that these spectral indices are _ local _ measures of the slope over a small range in wavelength , @xmath18 . because of
curvature of the agn spectral distributions , local slopes can be misleading , when compared to different wavelength bands and to objects at different redshift .
we return to this issue in section 3.3 where we discuss possible correlations of indices @xmath19 with agn type , redshift , and luminosity and compare indices measured by both _ hst_/cos and _ fuse_. + the cos composite spectrum is somewhat harder than that in earlier _ hst_/fos and stis observations ( telfer et al . 2002 ) who fitted the continuum ( 5001200 ) with @xmath20 for 184 qsos at @xmath21 .
their sample of 39 radio - quiet agn had @xmath22 .
our fit differs considerably from the _ fuse _ survey of 85 agn at @xmath23 ( scott et al .
2004 ) who found a harder composite spectrum with @xmath24 .
the different indices could arise in part from the small numbers of targets observed in the rest - frame euv . even in the current sample ,
only 10 or fewer agn observations cover the spectral range @xmath25 . another important difference in methodology is our placement of the euv continuum relative to strong emission lines such as ( 770 , 780 ) , ( 570 ) , ( 834 ) , ( 833 , 702 ) , ( 788 , 554 ) , ( 630 ) , and ( 1032 , 1038 ) . identifying and fitting these emission lines requires high s / n spectra .
a complete list of lines appears in table 4 of paper i. we also use the higher spectral resolution of the cos ( g130 m and g160 m ) gratings to distinguish the line - blanketing by narrow absorption lines from the ly@xmath1 forest .
increasingly important at higher redshifts , we need to identify and correct for absorption from lyman - limit systems ( lls ) with n@xmath26 @xmath3 and partial lyman - limit systems ( plls ) with n@xmath27 @xmath3 .
the historical boundary at @xmath28 @xmath3 occurs where the photoelectric optical depth @xmath29 at the 912 lyman edge .
+ in paper i , our 22 agn ranged in redshift from @xmath30 to @xmath31 , but included only four targets at sufficient redshift to probe the rest - frame continuum below 550 .
our new survey contains 159 agn out to @xmath32 with 16 targets at @xmath33 , sufficient to probe below 600 by observing with g130 m down to 1135 . in all agn spectra
, we identify the prominent broad emission lines and line - free portions of the spectrum and fit the underlying continua , excluding interstellar and intergalactic absorption lines . in section 2
we describe the cos data reduction and our new techniques for restoring the continua with a fitting method that corrects for the effects of absorption from the igm and ism . in section 3
we describe our results on the fuv and euv spectral indices in both individual and composite spectra .
section 4 presents our conclusions and their implications .
table 1 lists the relevant cos observational parameters of our 159 agn targets , which include agn type and redshift ( @xmath34 ) , from ned , continuum index ( @xmath19 ) , fitted rest - frame flux normalization at 1100 , observed flux @xmath35 at 1300 , and s / n ratios .
we also provide power - law fits to their _
hst_/cos spectra ( see section 2.2 ) and their monochromatic luminosities , @xmath36 , at 1100 , given by @xmath37 ^ 2 \left [ \frac { f_{\lambda } } { 10^{-14 } } \right ] \left [ \frac { \lambda } { 1100~{\rm \aa } } \right ] \;.\ ] ] we converted flux , @xmath35 ( in erg @xmath3 s@xmath14 @xmath14 ) to monochromatic luminosity , @xmath38 , using the luminosity distance , @xmath39 , computed for a flat @xmath40cdm universe with @xmath41 km s@xmath14 mpc@xmath14 and density parameters @xmath42 and @xmath43 ( komatsu et al .
+ the redshift distribution of the sample is shown in figure 1 , and the target distribution in accessible rest - frame wavelength in figure 2 .
our sample consists only of those agn observed with both the g130 m ( 11351460 ) and g160 m ( 13901795 ) cos gratings , providing the broad wavelength coverage at 17 km s@xmath14 resolution needed for our study of the continuum , emission lines , and absorption line blanketing .
the cos instrument and data acquisition are described by osterman et al .
( 2011 ) and green et al .
we retrieved all cos agn spectra publicly available as of 2013 april 25 , but excluded spectra with low signal - to - noise per pixel ( s / n @xmath44 ) and all bl lac objects , which are over - represented in the cos archives .
we also excluded three targets with abnormal spectra that would complicate the analysis : sdssj004222.29 - 103743.8 , which exhibits broad absorption lines ; sdssj135726.27 + 043541.4 , which features a plls longward of the cos waveband ; and ugca 166 , which gil de paz & madore ( 2005 ) classify as a blue compact dwarf galaxy .
this leaves 159 agn for analysis .
we follow the same procedure as in paper i for obtaining , reducing , and processing the data .
below , we briefly summarize the procedure and explicitly note improvements or deviations from earlier methods .
many of these techniques of coaddition and continuum fitting were also discussed in our igm survey ( danforth et al .
of particular interest in this paper are new techniques for identifying lls and plls absorbers , measuring their column densities , and using that information to correct the continuum .
our analysis proceeds through the following steps : 1 . *
retrieve exposures . *
the calcos calibrated exposures were downloaded from the mikulski archive for space telescopes ( mast ) and then aligned and co - added using idl procedures developed by the cos gto team .
typical wavelength shifts were a resolution element ( @xmath450.1 ) or less , and the co - added flux in each pixel was calculated as the exposure - weighed mean of the flux in aligned exposures
fit spline functions to spectra . *
the raw data contain narrow absorption features that should be excluded from the agn composite spectrum .
identifying and masking each of these features by hand in all of our spectra would be extremely tedious .
for this reason , we utilize a semi - automated routine that removes narrow absorption features and fits the remaining data with a piecewise - continuous function composed of spline functions and legendre polynomial functions .
this spline - fitting process involves first splitting the spectra into 5 - 10 segments and calculating the average s / n per pixel ( flux / error ) in each segment .
pixels with s / n less than 1.5 @xmath46 below their segment s / n are rejected from the fitting process to exclude absorption features and regions of increased noise .
this process is repeated iteratively until there is little change between iterations .
the median flux values in the segments are then fitted with a spline function .
we manually inspect the fits and adjust the identification of rejected regions as necessary .
smoothly varying data are well described by this spline - only method . near broad emission and other cusp - like features , short segments of piecewise - continuous legendre polynomials
are preferred .
more details on the process are given in our igm survey paper ( danforth et al .
* deredden spectra . *
we correct the fine - grained data and their corresponding spline fits for galactic reddening , using the empirical mean extinction curve of fitzpatrick ( 1999 ) with a ratio of total - to - selective extinction @xmath47 and color excesses @xmath48 from ned . in this paper
we use values of @xmath48 based on dust mapping by schlegel et al .
( 1998 ) with a 14@xmath49 recalibration by schlafly & finkbeiner ( 2011 ) .
we do not correct for reddening intrinsic to the agn , although we do not think this could be a substantial effect .
we can probably rule out a large amount of dust ( see discussion in section 3.2 ) .
4 . * identify plls and lls absorption . * in paper
i , we identified plls absorption by inspecting the spectra for flux decrements or lyman breaks . for this paper
we employ a custom computer script that scans each spectrum for correlated down - pixels at the locations of higher - order lyman lines of plls and lls absorbers .
first , the script divides the spectra by their respective spline fits , normalizing the flux unaffected by igm absorption to unity .
we determine the median flux for 15 pixels that have the same relative spacing as the first 15 lyman lines of a plls with a redshift equal to the source agn .
if there is a plls , the median will be much less than unity .
we then step one pixel to the left , recalculate the relative spacing of the first 15 lyman lines at this redshift and the median flux for this group of 15 pixels .
we repeat this process until we reach the end of the spectrum or a plls redshift of zero .
the script returns a list of redshifts of system candidates to be inspected . when a system is confirmed , we measure the equivalent widths of up to the first 12 lyman lines and fit them to a curve of growth ( cog ) to determine the column density and doppler parameter of the system . in paper
i , we found a total of 17 lls and plls systems in 8 of the 22 sight lines , and we were sensitive to systems with column density @xmath50 . in this paper , using our new identification method , we confirm the 17 previously identified systems plus 13 unidentified systems above the sensitivity limit @xmath51 in the same 22 sight lines from paper i. figure 3 shows examples of plls identification and continuum restoration .
the lowest column density measurement derived from cog fitting in this paper is @xmath52 .
we detect 221 systems ( 7 lls and 214 plls ) in 71 of the 159 agn sight lines , with absorber redshifts @xmath53 .
these absorbers are listed in table 2 together with our measurements of their redshifts , column densities , and doppler parameters . of the 221 systems , 167 have column densities @xmath54 whose distribution in column density is shown in figure 4 .
we only correct for these 167 systems in our analysis .
systems with @xmath55 and @xmath56 have opacity that depress the flux immediately blueward of the lyman limit by less than 2@xmath49 and 0.7@xmath49 respectively . owing to the multiple correlated lyman lines used in this identification technique ,
our sensitivity is better than the local s / n over most of the spectral coverage . correcting for the opacity of weaker systems ( @xmath57 ) would have a negligible effect on our analysis of agn continuum , changing the euv slope of our composite spectrum by only 0.006 .
restore flux depressed by plls and mask unrecoverable flux .
* we account for lyman continuum absorption by measuring the equivalent widths of the first 12 lyman lines and fitting them with a cog to estimate the column density and doppler parameter .
we use these measurements to correct for the @xmath58 opacity shortward of each lyman edge .
we correct only the flux below the lyman limit .
when a spectrum has plls absorption with column density @xmath59 , we mask the flux between the lyman limit ( 911.753 ) and lyman-9 ( 916.429 ) or @xmath454.7 redward in the plls rest - frame . for a plls with @xmath60
, we mask from the lyman limit to lyman-13 ( 920.963 ) or @xmath459.2 redward . when data blueward of the lyman limit of lls or plls had @xmath61 or did not appear continuous with two or more regions of continuum redward of the lyman limit , we masked the data .
we also mask regions of the spectra affected by broad absorption from damped ly@xmath1 systems and h@xmath62 lyman bands after qualitative visual inspection .
the amount of masking varies for individual cases .
some spectra have one or two gaps of @xmath6310 in the data from observations that were not planned with contiguous wavelength coverage over the entire cos - fuv spectral range .
these gaps are masked prior to our continuum analysis .
* shift to rest - frame . *
we shift each spectrum to the rest - frame of the agn by dividing the wavelength array by @xmath64 .
* mask non - agn features . * in every spectrum we exclude galactic ly@xmath1 absorption ( 1215.67 ) by masking 14 on both sides of the line center in the observed frame .
we exclude geocoronal emission lines of @xmath65 and @xmath66 by masking 2 on both sides of and 5 on both sides of . in five spectra we masked the absorption due to the ly@xmath1 line of damped ly@xmath1 systems
* resample the spectra . * as in paper i and telfer et al .
( 2002 ) , we resample the spectra to uniform 0.1 bins .
after resampling , each flux pixel corresponds to a new wavelength bin and is equal to the mean of the flux in the old pixels that overlap the new bin , weighted by the extent of overlap .
the error arrays associated with the resampled spectra are determined using a weighting method similar to the flux rebinning .
see equations ( 2 ) and ( 3 ) in telfer et al .
( 2002 ) for the rebinning formulae .
to construct the overall composite spectrum we start by following the bootstrap method of telfer et al .
( 2002 ) and then adapt the construction technique for our unique dataset . to summarize the bootstrap technique ,
we start near the central region of the output composite spectrum , between 1050 and 1150 , and normalize the spectra that include the entire central region to have an average flux value of 1 within the central region , which we refer to as the central continuum window . "
we then include spectra in sorted order toward shorter wavelengths .
lastly , we include the spectra in sorted order toward longer wavelengths .
when a spectrum does not cover the central continuum window , we normalize it to the partially formed composite by finding the weighted - mean normalization constant within multiple emission - line - free continuum windows , calculated using equation ( 4 ) of telfer et al .
we form two independent composite spectra simultaneously : one of the fine - grained spectra showing the line - blanketing by the ly@xmath1 forest and interstellar absorption lines , and another of the spline fits to the individual spectra .
the spline fits pass over the narrow absorption lines . with our unique dataset and spline
fits , we adjust the composite construction method in five ways .
first , with the identification in paper i of broad emission lines from highly ionized species below 800 , we were able to restrict the normalization of the spectra at the highest redshifts to two narrow regions of continuum - like windows at 660 - 670 and 715 - 735 .
this is in contrast to using all of the flux , including that from emission lines below 800 in the calculation of the normalization constant , as was done in our initial method .
our second adjustment also limits the normalization calculation to regions of continuum , which is our primary interest .
we refine and narrow the continuum windows above 800 to wavelengths 855 - 880 , 1090 - 1105 , 1140 - 1155 , 1280 - 1290 , 1315 - 1325 , and 1440 - 1465 . for our third adjustment
we choose the region between 855 - 880 as the central continuum window , because it is the largest of the narrowed euv continuum windows with a large number of contributing spectra .
fourth , we note that the bootstrapping technique can be sensitive to the ordering in which one includes the spectra , especially at the beginning of the process when only a handful of spectra determine the shape of the composite
. therefore , we increase the number of spectra normalized at only the central continuum window from 40 to 70 by decreasing the required overlap with the central continuum window from 100% to 50% .
lastly , because we are interested in characterizing the shape of the underlying continuum as a power law , we follow the approach of vanden berk et al .
( 2001 ) and combine the spectra as a geometric mean , which preserves power - law slopes .
we also provide a median - combined composite , which preserves the relative shape of the emission features . below 700 ,
the number of agn spectra contributing to each 0.1 bin in figure 2 declines steadily .
several agn listed in table 1 do not appear in this figure because their short - wavelength spectra are masked out , owing to lls absorption , airglow , and plls edges .
the final overall sample composite spectra ( both geometric - mean and median ) are presented in two panels of figure 5 , covering rest - frame wavelengths from 475 - 1785 . in each panel , we show both the fine - grained data with absorption lines included and the spline - fit continuum composites . in the geometric - mean spectrum , which we regard as the better characterization of the agn composite ,
the effects of line - blanketing by the ly@xmath1 forest can be seen in the difference between the spline - fit composite and the real data composite .
figure 6 shows the optical depth , @xmath67 , arising from line - blanketing of the continuum by the ly@xmath1 forest at @xmath68 .
we derive optical depths from the difference in fluxes ( red and black ) in the geometric mean composite , shown in the top panel of figure 5 . to characterize the continuum slope of the composite spectrum , we follow the simple approach of vanden berk et al .
we calculate the slope between continuum regions of maximum separation on either side of the spectral break , which is clearly present in the composite around 1000 . because the flux distribution , @xmath35 , flattens at shorter wavelengths , the two power - law fits pass under the observed spectrum and match at the break .
to satisfy this condition , we first calculate the slope of a line connecting the minima of the two best continuum regions in log - log space .
we then divide the entire spectrum by this line , find the wavelengths of the minima , and calculate the slope of the line that connects the second pair of minima .
this results in a line that does not cross the composite .
we perform this calculation twice , once in the euv and again in the fuv .
we find a mean euv spectral index @xmath69 between line - free windows centered at 724.5 and 859 , and mean fuv index @xmath70 between 1101 and 1449 .
these wavelength indices correspond to frequency indices @xmath71 of @xmath72 ( euv ) and @xmath73 ( fuv ) .
we now discuss sources of random and systematic uncertainty in the composite spectral indices . as in paper
i , we fit two power laws to the spline composite spectrum , with indices @xmath74 and @xmath75 , and match them at a break wavelength , which we find to be @xmath7 , consistent with paper i and accurate to @xmath76 .
although this gradual break is apparent in the composite , its presence is less clear in the individual spectra , owing to the limited spectral range of the cos observations and to strong emission lines of @xmath77 and ly@xmath78 @xmath79 near the break wavelength .
because the sample of agn in paper i had no targets between @xmath80 , most rest - frame spectra lay either blueward or redward of the 1000 break . in our new sample , 55 agn have redshifts in that range , but we do not distinguish a clear break in individual spectra . with the limited wavelength coverage of cos ( g130 m , g160 m ) , any single agn spectrum does not have access to the four continuum windows needed to measure two distinct power laws that straddle the break . to quantify the uncertainty in the fitting of the composite spectrum
, we explore the sources of uncertainty described by scott et al .
( 2004 ) , including the effects of intrinsic variations in the shape of the seds , galactic extinction , and formal statistical fitting errors . as in paper
i , we do not include the effects of intrinsic absorbers or interstellar lines , as these absorption lines are easily removed with our moderate - resolution cos spectra .
however , we do consider the effects from the strongest systems in the ly@xmath1 forest .
the largest source of uncertainty comes from the natural variations in the slope of the contributing spectra .
we estimate this uncertainty by selecting 1000 bootstrap samples with replacement from our sample of 159 agn spectra .
the resulting distributions of spectral index in frequency lead to mean values : @xmath81 and @xmath82 in the euv and fuv .
figure 7 shows a montage of spectra for individual agn , illustrating the wide range of spectral slopes and emission - line strengths .
we also investigated the range of uncertainties arising from uv extinction corrections from two quantities : @xmath48 and @xmath83 .
we alter the measured @xmath48 by @xmath84% ( @xmath85 ) as reported by schlegel et al .
we deredden the individual spectra with @xmath48 multiplied by 1.16 or 0.84 , compile the spectra into a composite , and fit the continua . over these ranges , we find that the index @xmath75 changes by ( @xmath86 while @xmath74 changes by @xmath87 . next , we estimate the sensitivity to deviations from the canonical value @xmath88 , which clayton , cardelli , & mathis ( 1988 ) found to vary from @xmath89 to @xmath90 .
we follow scott et al .
( 2004 ) and deredden individual spectra with @xmath91 and @xmath92 and compiling the spectra into composites .
we find that @xmath75 changes by @xmath93 and @xmath74 by @xmath94 .
we estimate the uncertainties arising from correcting plls absorption of strength @xmath54 by altering the measured column densities by @xmath95 as reported in table 2 .
we find that @xmath75 changes by @xmath96 and @xmath74 by @xmath97 .
the formal statistical errors for the spectral indices are negligible ( @xmath98 ) , owing to the high s / n ratio of our composite spectra .
thus , we do not include them in the final quoted uncertainties .
we add the random uncertainties of cosmic variance with the systematic effects of correcting for extinction in quadrature and estimate the total uncertainties to be @xmath99 for @xmath75 and @xmath100 for @xmath74 .
as noted earlier , we do not correct for agn dust , but small amounts could be present , as long as they do not produce a strong turnover in the far - uv fluxes .
we can rule out a large amount of intrinsic extinction , if it obeys a selective extinction law , @xmath101 , that rises steeply at short ( uv ) wavelengths .
otherwise , we would see steep curvature in the rest - frame euv , rather than a power law .
ultraviolet spectra of agn have been surveyed by many previous telescopes , including the _ international ultraviolet explorer _
( obrien et al . 1988 ) and the _ hst _ faint object spectrograph ( zheng et al .
more recent agn composite spectra were constructed from data taken with _
( telfer et al . 2002 ) , _ fuse _ ( scott et al . 2004 ) , and _ hst_/cos ( shull et al .
2012 and this paper ) .
figure 8 compares the _
hst_/cos composites with previous studies with _ hst_/fos+stis and _ fuse_. our current cos survey finds essentially the same euv spectral index , @xmath9 , as found in paper i , @xmath102 , but with better statistics and coverage to shorter wavelengths ( below 500 ) .
this consistency is reassuring , as our current composite includes 159 agn spectra , compared with 22 agn in the initial cos study ( paper i ) .
the _ hst_/cos euv index , @xmath103 is slightly harder than the _ hst_/fos+stis value , @xmath22 , found by telfer et al .
( 2002 ) for 39 radio - quiet agn .
however , both _
hst _ surveys found indices steeper than the _ fuse _ slope , @xmath24 ( scott et al .
2004 ) , a puzzling discrepancy that we now investigate . the differences between continuum slopes found in _ fuse _ and cos composite spectra are likely to arise from four general factors : ( 1 ) continuum placement beneath prominent euv emission lines ; ( 2 ) line blanketing by the ly@xmath1 forest and stronger ( plls ) absorbers ; ( 3 ) continuum windows that span an intrinsically curved agn spectrum ; and ( 4 ) possible correlation of slope and agn luminosity .
high - s / n spectra at the moderate resolution of cos ( g130m / g160 m ) are critical for identifying the underlying continuum near the strong euv emission lines of , , and and the doublet ( @xmath104 ) .
the cos spectral resolution also allows us to fit over the narrow absorbers in the ly@xmath1 forest ( factor 2 ) and restore the continuum absorbed by the stronger systems ( lls and plls ) .
factor 3 is a more subtle effect , but it may be the most important .
the cos wavelength coverage ( 11351795 ) is broader than that of _ fuse _ ( 9121189 ) , and it provides line - free continuum windows above and below 1100 , spanning an intrinsically curved sed .
this allows us to construct a two - component spectrum with indices @xmath9 in the euv ( 500 - 1000 ) and @xmath8 in the fuv ( 1200 - 2000 ) with a break at @xmath105 .
many of the cos sight lines observe higher - redshift agn that sample different regions of the sed than those of _
shortward of 912 , we place the continuum below a number of prominent emission lines , using nearly line - free continuum windows at @xmath106 , @xmath107 , and @xmath108 .
factor 4 refers to possible selection effects of agn luminosity with redshift .
previous samples used targets at a variety of redshifts and luminosities , observed with different spectral resolution , fuv throughput , and instruments .
all of the uv composite spectra ( _ hst _ and _ fuse _ ) are based on the available uv - bright targets ( type-1 seyferts and qsos ) studied with _
iue _ and the _ galaxy evolution explorer _ ( _ galex _ ) .
most of these agn were chosen as background sources for studies of igm , cgm , and galactic halo gas .
although these targets are not a complete , flux - limited sample of the agn luminosity function ( e.g. , barger & cowie 2010 ) , they probably are representative of uv - bright qsos , at least at redshifts @xmath109 . figure 9 compares the average agn redshift per wavelength bin for the cos and _ fuse _ surveys , overlaid on the line - free continuum windows .
evidently , the cos targets are at systematically higher redshift , and their wavelength coverage is broader than that of _ fuse_. the average agn luminosity also differs , longward and shortward of the break . at @xmath110 , the cos and _ fuse _ composites are both probing similar luminosities . as shown in the top panel of figure 9 , the two spectral slopes are fairly similar between 6501000 , and the only difference comes from the sudden decline in _ fuse _ fluxes between 10901140 . lacking the longer - wavelength continuum windows , the _ fuse _
spectra were unable to fit the break in spectral slope at longer wavelengths .
figure 10 shows the distributions of spectral index @xmath19 and the effects of the available continuum windows falling longward or shortward of the 1000 break .
the two - power - law fits possible with cos data allow us to measure the spectral curvature and distinguish between fuv and euv slopes .
this was not done with the _ fuse _ composite fits . in summary
, we believe the _
composite spectra are superior owing to their higher spectral resolution ( g130 m and g160 m gratings ) allowing us to resolve and mask out the ly@xmath1 forest and restore the continuum from stronger ( lls and plls ) absorbers .
the higher s / n of the cos spectra allow us to identify and resolve prominent uv / euv emission lines and fit a more accurate underlying continuum . as shown in figures 2 and 7 , the cos composite still contains fewer than 10 agn at @xmath111 that probe the rest - frame continuum at @xmath112 .
these numbers are larger than in the earlier surveys , but the small sample means that the composite spectrum remains uncertain at the shortest wavelengths . as in paper
i , we explore trends within the _
hst_/cos agn sample by constructing composite spectra based on various parameters and subsamples .
figure 11 shows the distributions of index @xmath19 in redshift , agn activity type , galactic foreground reddening , and monochromatic ( 1100 ) luminosity . in each panel , two horizontal lines denote the sample - mean values : @xmath113 for the rest - frame euv ( 500 - 1000 ) band and @xmath114 for the rest - frame fuv ( 1200 - 2000 ) band .
the spectral indices extend over a wide range of agn luminosities and activity types , with no obvious trend or correlation .
galactic reddening does not appear to produce any difference in the index .
there may be subtle trends in the distribution of @xmath19 with redshift , because we are observing the rest - frame flux from an intrinsically curved spectral energy distribution ( sed ) . at low redshift ( @xmath115 )
there are many agn with steep slopes , @xmath116 , indicating hard uv spectra .
however , only seven agn have spectra with @xmath116 . at higher
redshift ( @xmath117 ) there are few agn with slopes @xmath118 , and the survey contains few agn at the most extreme redshifts ( @xmath111 ) .
many more have soft spectra with slopes @xmath119 .
accretion disk ( ad ) model spectra have recently been investigated by a number of groups ( davis et al . 2007 ; laor & davis 2011 ; done et al . 2012 ;
slone & netzer 2012 ) with a goal of comparing to uv and euv spectra .
the observed far - uv spectral turnover at @xmath120 limits the maximal disk temperature to @xmath121 k. model atmospheres computed with the _ tlusty _ code ( hubeny et al .
2001 ) and including winds driven from inner regions of the disk predict a spectral break near 1000 , arising from the lyman edge ( 912 ) and wind - truncation of the hot inner part of the disk ( laor & davis 2014 ) . in a standard multi - temperature accretion - disk models with blackbody spectra in annular rings ( pringle 1981 ) the radial temperature distribution scales as @xmath122 . in their models of wind - ejecting disks ,
slone & netzer ( 2012 ) suggest that the spectral shape is governed by the radial profile of @xmath123 , and the radius @xmath124 where half the disk mass has been ejected .
their observational predictions are based on the removal of hot accreting gas from the inner regions of the ad and accompanying removal of energy from the uv - emitting portions of the sed . a large mass accretion rate throughout the ad produces higher luminosities and shifts the sed to shorter ( uv ) wavelengths .
the closer @xmath124 comes to the innermost stable circular orbit , @xmath125 , the more fuv and euv radiation will be emitted .
slone & netzer ( 2012 ) used the sensitivity of the euv spectral index , @xmath126 between 456912 , to constrain accretion properties outside @xmath127 , the radius of the innermost stable circular orbit around a black hole ( see their figure 7 ) .
their model is governed by mass accretion rates , @xmath128 and @xmath129 , at the inner and outer disk radii of the disk , relative to the eddington accretion rate @xmath130 and luminosity @xmath131 relative to the eddington luminosity , @xmath132 . from the observed euv spectral index , @xmath133 ,
we constrain the mean agn accretion rate and luminosity to values @xmath134 and @xmath135 .
we caution that these inferences are subject to the validity of accretion - disk model atmospheres , including effects of external irradiation , uncertainty in where energy is being deposited , and the role of magnetic field energy dissipation .
in addition , disk photospheres may differ from those of hot stars , with spatially variable @xmath136 surfaces .
laor & davis ( 2014 ) explore similar disk - truncation models , solving for the radial structure of a disk with mass loss .
they find that the wind mass loss rate , @xmath137 , becomes comparable to the total accretion rate @xmath138 at radii a few tens of gravitational radii , @xmath139 .
line - driven winds set a cap of @xmath140 k on their disks , which in most cases are truncated well outside the isco radius .
these models are consistent with the observed sed turnover at @xmath120 that is weakly dependent on luminosity @xmath141 and black - hole mass @xmath142 .
their models of line - driven winds also cap ad effective temperatures , @xmath143 k. the uv spectral turnover is produced by both an lyman edge and the limit on disk temperature .
standard models of accretion disk atmospheres are predicted to exhibit and continuum edges at 912 and 504 , respectively .
this issue and the euv ( soft x - ray ) spectra of accretion disks have been discussed by many authors ( e.g. , kolykhalov & sunyaev 1984 ; koratkar & blaes 1999 ; done et al .
the absence of any continuum absorption at 912 in the composite spectrum was noted in paper i , where we set an optical depth limit of @xmath144 . from the 159-agn composite ( see figures 5 and 8) our limit is now @xmath145 derived from the flux around 914.5 and 910.5 .
the limit for the edge at 504 is less certain because of the difficulty in fitting the local continuum under neighboring broad euv emission lines .
however , from the general continuum shape between 480 - 520 , we can limit the continuum optical depth to @xmath146 .
additional cos / g140l data now being acquired toward 11 agn at redshifts @xmath147 probe the rest - frame c ontinua at @xmath148 with good spectral coverage at the 504 edge .
we continue to see no continuum edge .
we now summarize the results and implications of our _ hst_/cos survey of agn spectral distributions in the agn rest - frame fuv and euv . using spectra of 159 agn taken with _
hst_/cos g130 m and g160 m gratings , we constructed a 2-component composite spectrum n the euv ( 500 - 1000 ) and fuv ( 1200 - 2000 ) .
these two spectral fits match at a break wavelength @xmath7 , below which the sed steepens to @xmath149 .
the euv index is the same as found in paper i , but with smaller error bars .
it is slightly harder than the index , @xmath150 , found from the hst / fos+stis survey ( telfer et al .
2002 ) for radio - quiet agn , but much softer than the index , @xmath24 , from the _ fuse _ survey ( scott et al .
these composite spectra are based on small numbers of agn with redshifts ( @xmath151 ) sufficient to probe below 600 .
however , the _ hst_/cos survey provides a superior measure of the true underlying continuum .
our g130m / g160 m data have sufficient spectral resolution and signal to noise to mask out narrow lines from the ly@xmath1 forest and restore the continuum from stronger ( lls and plls ) absorbers .
we also fit the continuum below the prominent broad euv emission lines using nearly line - free continuum windows at @xmath106 , @xmath107 , and @xmath108 . 1
. the hst / cos composite spectrum follows a flux distribution with @xmath152 for agn rest - frame wavelengths 1200 - 2000 and @xmath153 for 500 - 1000 .
this euv spectral index is slightly harder than that used in recent simulations ( haardt & madau 2012 ) of igm photoionization and photoelectric heating .
2 . individual spectra of the 159 agn surveyed exhibit a wide range of spectral indices in the euv , with typical values between @xmath6 .
these indices are _ local _ slopes and not characteristic of the spectral energy distribution over the full uv / euv band , 3 .
the composite sed exhibits a turnover at @xmath120 , characteristic of accretion disk models in which the maximum temperature @xmath140 k and the inner disk is truncated by line - driven winds .
we see no continuum edges of ( 912 ) or ( 504 ) , with optical depth limits @xmath145 and @xmath146 .
the absence of these edges suggests that accretion disk atmospheres differ from those of hot stars , because of external irradiation or inverted temperature structures arising from magnetic energy dissipation .
we find no obvious correlations of the euv spectral index with interstellar reddening , agn type , redshift , or luminosity ( @xmath36 at 1100 ) . such trends are difficult to pick out , because the observable _ hst_/cos ( g130m / g160 m ) wavelength band ( 1135 - 1795 ) covers different portions of the sed over the agn redshifts @xmath154 ) in our sample .
the quoted indices , @xmath19 are _ local _ slopes that fall either in the fuv or euv depending on agn redshift .
the mean euv slopes , compared to models of wind - truncated thin accretion disks , constrain the mean accretion rate in the inner disk and the agn luminosity to values @xmath134 and @xmath135 relative to their eddington rates .
the order - of - magnitude improvement in sensitivity offered by cos over previous spectrographs has greatly increased the number of targets available for moderate - resolution uv spectroscopy .
some of these spectra have s / n below the threshold chosen for this survey , and many are low - resolution ( g140l ) rather than g130m
/ g160 m used here .
nevertheless , some of these archival spectra will provide euv coverage down to 500 ( with @xmath4540 agn ) and to 912 ( with @xmath45100 agn ) .
additional data at agn rest wavelengths 400 - 500 would be helpful in fitting the sed deeper into the euv , where fewer than ten agn sight lines have been probed to date with cos / g130 m .
currently , our composite spectrum includes 10 agn that contribute at @xmath155 but only two agn at @xmath156 .
as noted in paper i , one can explore even shorter rest - frame wavelengths ( 304500 ) using the sample of
quasars " ( worseck et al . 2011 ; syphers et al .
2011 ; shull et al .
2010 ) that probe the epoch of reionization at @xmath157 . in hubble cycle 21
, we are observing 11 new agn targets at @xmath158 to @xmath159 , using the lower - resolution ( g140l ) grating .
the first ten of these spectra have now been acquired . after reduction
, they should improve the accuracy of the composite spectrum down to 400 and provide more sight lines that cover the 504 continuum edge .
our intent is to create a composite agn spectrum between 350 and 1800 , using _
archival spectra of agn with a variety of types and luminosities .
we thank the cos / gto team for help on the calibration and verification of cos data .
we acknowledge helpful discussions with shane davis , ari laor , and jim pringle on accretion disk models and thank the referee for helpful comments that encouraged us to explore the differences between cos and _ fuse _ composite spectra .
this research was supported by nasa grants nnx08-ac14 g and nas5 - 98043 and the astrophysical theory program ( nnx07-ag77 g from nasa ) at the university of colorado boulder .
jms thanks the institute of astronomy , cambridge university , for their stimulating scientific atmosphere and support through the sackler visitor program .
, resampled to 0.1 , plotted in 1 bins , normalized to unit flux at 1100 , and showing broad fuv and euv emission lines atop a power - law continuum .
eight continuum windows are shown as small red boxes along bottom .
composite data are shown in black ; red curve is composite of individual spline fits .
( top ) geometric - mean - combined _ hst_/cos spectrum with frequency distribution , @xmath5 , with break at @xmath7 and spectral indices @xmath9 ( euv , @xmath120 ) and @xmath8 ( fuv , @xmath160 ) .
( bottom ) median - combined composite , with break at @xmath161 and spectral indices @xmath162 ( euv ) and @xmath163 ( fuv ) .
, title="fig : " ] , resampled to 0.1 , plotted in 1 bins , normalized to unit flux at 1100 , and showing broad fuv and euv emission lines atop a power - law continuum .
eight continuum windows are shown as small red boxes along bottom .
composite data are shown in black ; red curve is composite of individual spline fits .
( top ) geometric - mean - combined _ hst_/cos spectrum with frequency distribution , @xmath5 , with break at @xmath7 and spectral indices @xmath9 ( euv , @xmath120 ) and @xmath8 ( fuv , @xmath160 ) .
( bottom ) median - combined composite , with break at @xmath161 and spectral indices @xmath162 ( euv ) and @xmath163 ( fuv ) .
, title="fig : " ] , all normalized at the 860 line - free continuum window . dashed lines show underlying power - law continua .
the cos ( 2012 , 2014 ) composite spectra , @xmath5 , have essentially the same euv spectral index , @xmath164 .
fos / stis spectrum has @xmath150 for 39 radio - quiet qsos .
fuse spectrum is harder with @xmath24 based on lower - redshift agn . _
( bottom . ) _ spectra offset vertically for clarity .
the different slopes arise from fitting the continuum beneath prominent broad euv emission lines ( , , , ) and hst / cos wavelength coverage spanning an intrinsically curved agn spectrum .
access to line - free continuum windows above and below 1100 ( red bars along bottom ) allows us to fit different fuv and euv continuum slopes with a break at @xmath105 .
, title="fig : " ] , all normalized at the 860 line - free continuum window .
dashed lines show underlying power - law continua .
the cos ( 2012 , 2014 ) composite spectra , @xmath5 , have essentially the same euv spectral index , @xmath164 .
fos / stis spectrum has @xmath150 for 39 radio - quiet qsos .
fuse spectrum is harder with @xmath24 based on lower - redshift agn . _
( bottom . ) _ spectra offset vertically for clarity .
the different slopes arise from fitting the continuum beneath prominent broad euv emission lines ( , , , ) and hst / cos wavelength coverage spanning an intrinsically curved agn spectrum .
access to line - free continuum windows above and below 1100 ( red bars along bottom ) allows us to fit different fuv and euv continuum slopes with a break at @xmath105 .
, title="fig : " ] ) but the _ fuse _ fluxes show a sudden drop - off longward of 1100 in continuum windows at 1100 and 1145 . _ ( middle ) .
_ average agn luminosity per wavelength bin .
we plot the geometric mean luminosity , @xmath36 , at 1000 . _
( bottom ) . _ average redshift of agn per wavelength bin .
the _ fuse _ survey samples agn at lower redshift than cos , with agn luminosities comparable at @xmath165 , but much lower at @xmath166 . ] ngc4395 & sy 1.8 & 0.001064 & @xmath167 & @xmath168 & @xmath169 & 39.70 & 4 , 4 + ngc4051 & nlsy 1 & 0.002336 & @xmath170 & @xmath171 & @xmath172 & 41.02 & 16 , 13 + ngc3516 & sy 1.5 & 0.008836 & @xmath173 & @xmath174 & @xmath175 & 42.46 & 14 , 12 + ngc3783 & sy 1.5 & 0.00973 & @xmath176 & @xmath177 & @xmath178 & 43.40 & 21 , 17 + ngc7469 & sy 1.5 & 0.016317 & @xmath179 & @xmath180 & @xmath181 & 43.88 & 28 , 18 + mrk1044 & nlsy 1 & 0.016451 & @xmath182 & @xmath183 & @xmath184 & 43.41 & 19 , 13 + ngc5548 & sy 1.5 & 0.017175 & @xmath185 & @xmath186 & @xmath187 & 43.83 & 28 , 19 + akn564 & liner & 0.024684 & @xmath188 & @xmath189 & @xmath190 & 43.25 & 9 , 7 + mrk335 & sy 1 & 0.025785 & @xmath191 & @xmath192 & @xmath193 & 43.98 & 24 , 18 + eso031g008 & sy 1.2 & 0.027619 & @xmath194 & @xmath195 & @xmath196 & 43.25 & 9 , 7 + mrk290 & sy 1.5 & 0.029577 & @xmath197 & @xmath198 & @xmath199 & 43.59 & 20 , 16 + mrk279 & sy 1 & 0.030451 & @xmath200 & @xmath201 & @xmath202 & 42.98 & 8 , 7 + mrk817 & sy 1.5 & 0.031455 & @xmath203 & @xmath204 & @xmath205 & 44.34 & 38 , 24 + mrk509 & sy 1.5 & 0.034397 & @xmath206 & @xmath207 & @xmath208 & 44.72 & 63 , 55 + pg1011 - 040 & sy 1.2 & 0.058314 & @xmath209 & @xmath210 & @xmath211 & 44.36 & 26 , 13 + mrk1513 & sy 1.5 & 0.062977 & @xmath212 & @xmath213 & @xmath214 & 44.42 & 27 , 15 + mr2251 - 178 & sy 1.5 & 0.06398 & @xmath215 & @xmath216 & @xmath217 & 44.65 & 33 , 25 + rxj0503.1 - 6634 & sy 1 & 0.064 & @xmath218 & @xmath219 & @xmath220 & 44.14 & 16 , 8 + sdssj145108.76 + 270926.9 & nlsy 1 & 0.0645 & @xmath221 & @xmath222 & @xmath223 & 43.77 & 10 , 8 + rbs563 & sy 1.5 & 0.069 & @xmath224 & @xmath225 & @xmath226 & 43.85 & 13 , 7 + sdssj031027.82 - 004950.7 & sy 1 & 0.080139 & @xmath227 & @xmath228 & @xmath229 & 44.15 & 10 , 8 + pg0804 + 761 & sy 1 & 0.101 & @xmath230 & @xmath177 & @xmath231 & 45.48 & 51 , 33 + iras - f22456 - 5125 & sy 1.5 & 0.101 & @xmath232 & @xmath233 & @xmath234 & 44.77 & 44 , 24 + uks-0242 - 724 & sy 1.2 & 0.1018 & @xmath235 & @xmath236 & @xmath175 & 44.58 & 13 , 9 + iras - f04250 - 5718 & sy 1.5 & 0.104 & @xmath237 & @xmath238 & @xmath239 & 45.05 & 63 , 34 + tons210 & sy 1 & 0.116 & @xmath240 & @xmath241 & @xmath239 & 45.19 & 36 , 21 + q1230 + 0115 & nlsy 1 & 0.117 & @xmath242 & @xmath243 & @xmath244 & 45.10 & 51 , 31 + hs0033 + 4300 & sy 1 & 0.12 & @xmath245 & @xmath246 & @xmath247 & 43.99 & 6 , 7 + mrk106 & sy 1 & 0.122951 & @xmath248 & @xmath249 & @xmath175 & 44.79 & 23 , 14 + sdssj152139.66 + 033729.2 & sy 1 & 0.126354 & @xmath250 & @xmath251 & @xmath252 & 43.94 & 5 , 4 + mrk876 & sy 1 & 0.129 & @xmath253 & @xmath254 & @xmath255 & 45.33 & 56 , 31 + irasl06229 - 6434 & sy1 & 0.129 & @xmath256 & @xmath257 & @xmath258 & 44.68 & 16 , 9 + pg0838 + 770 & sy 1 & 0.131 & @xmath259 & @xmath260 & @xmath172 & 44.71 & 24 , 11 + pg1626 + 554 & sy 1 & 0.133 & @xmath261 & @xmath262 & @xmath220 & 45.08 & 24 , 14 + qso0045 + 3926 & sy 1 & 0.134 & @xmath263 & @xmath264 & @xmath265 & 44.79 & 29 , 20 + pks0558 - 504 & nlsy 1 & 0.1372 & @xmath266 & @xmath267 & @xmath268 & 45.33 & 15 , 7 + sdssj094733.21 + 100508.7 & sy 1.5 & 0.139297 & @xmath269 & @xmath270 & @xmath271 & 44.68 & 11 , 6 + sdssj135712.61 + 170444.1 & qso & 0.1505 & @xmath272 & @xmath273 & @xmath274 & 44.58 & 12 , 8 + sdssj112114.22 + 032546.7 & nlsy 1 & 0.152033 & @xmath275 & @xmath276 & @xmath277 & 44.01 & 4 , 3 + pg1115 + 407 & sy 1 & 0.154567 & @xmath278 & @xmath279 & @xmath229 & 44.86 & 19 , 11 + sdssj095915.65 + 050355.1 & qso & 0.162296 & @xmath280 & @xmath281 & @xmath282 & 44.72 & 11 , 7 + sdssj135625.55 + 251523.7 & sy1 & 0.164009 & @xmath283 & @xmath284 & @xmath285 & 44.57 & 8 , 5 + sdssj015530.02 - 085704.0 & sy 1 & 0.164427 & @xmath286 & @xmath287 & @xmath288 & 44.51 & 9 , 6 + pg1202 + 281 & sy 1.2 & 0.1653 & @xmath289 & @xmath290 & @xmath169 & 44.18 & 7 , 5 + pg1048 + 342 & sy 1 & 0.167132 & @xmath291 & @xmath292 & @xmath293 & 44.74 & 18 , 12 + sdssj121114.56 + 365739.5 & sy 1 & 0.170796 & @xmath294 & @xmath295 & @xmath274 & 44.64 & 10 , 7 + sdssj134231.22 + 382903.4 & sy 1 & 0.171869 & @xmath296 & @xmath297 & @xmath202 & 44.63 & 10 , 6 + 1saxj1032.3 + 5051 & qso & 0.173128 & @xmath298 & @xmath299 & @xmath300 & 43.93 & 9 , 5 + sdssj021218.32 - 073719.8 & sy 1 & 0.17392 & @xmath301 & @xmath302 & @xmath303 & 44.36 & 8 , 5 + pg1116 + 215 & sy 1 & 0.1763 & @xmath304 & @xmath305 & @xmath306 & 45.67 & 34 , 22 + 2masx - j01013113 + 4229356 & sy 1 & 0.19 & @xmath307 & @xmath308 & @xmath309 & 44.89 & 8 , 6 + phl1811 & nlsy 1 & 0.192 & @xmath310 & @xmath311 & @xmath312 & 45.95 & 34 , 18 + sdssj123604.02 + 264135.9 & qso & 0.208995 & @xmath313 & @xmath314 & @xmath315 & 44.35 & 7 , 5 + pg1121 + 422 & sy 1 & 0.225025 & @xmath316 & @xmath317 & @xmath318 & 45.13 & 16 , 10 + sdssj001224.01 - 102226.5 & qso & 0.228191 & @xmath319 & @xmath320 & @xmath303 & 44.67 & 7 , 5 + pg0953 + 414 & sy 1 & 0.2341 & @xmath321 & @xmath322 & @xmath323 & 45.83 & 33 , 20 + sdssj092909.79 + 464424.0 & qso & 0.239959 & @xmath324 & @xmath325 & @xmath326 & 45.37 & 14 , 9 + sdssj133053.27 + 311930.5 & sy 1.5 & 0.242204 & @xmath327 & @xmath328 & @xmath274 & 44.77 & 10 , 7 + rxj0439.6 - 5311 & sy 1 & 0.243 & @xmath329 & @xmath330 & @xmath331 & 44.81 & 14 , 8 + fbqsj1010 + 3003 & qso & 0.255778 & @xmath332 & @xmath333 & @xmath334 & 44.68 & 15 , 9 + sdssj115758.72 - 002220.8 & sy 1 & 0.260247 & @xmath335 & @xmath336 & @xmath337 & 44.79 & 8 , 5 + sdssj134206.56 + 050523.8 & sy 1.2 & 0.266015 & @xmath338 & @xmath339 & @xmath340 & 45.02 & 10 , 7 + pks1302 - 102 & sy 1.2 & 0.2784 & @xmath341 & @xmath342 & @xmath343 & 45.67 & 23 , 16 + ton580 & sy 1 & 0.290237 & @xmath344 & @xmath345 & @xmath346 & 45.45 & 18 , 12 + sdssj092837.98 + 602521.0 & sy 1 & 0.29545 & @xmath347 & @xmath348 & @xmath303 & 44.74 & 5 , 5 + h1821 + 643 & sy 1.2 & 0.2968 & @xmath349 & @xmath350 & @xmath351 & 46.10 & 52 , 7 + sdssj091235.42 + 295725.4 & qso & 0.305331 & @xmath352 & @xmath353 & @xmath354 & 44.50 & 6 , 5 + sdssj082633.51 + 074248.3 & qso & 0.310643 & @xmath355 & @xmath356 & @xmath357 & 45.22 & 9 , 6 + sdssj120720.99 + 262429.1 & qso & 0.323529 & @xmath358 & @xmath359 & @xmath360 & 45.16 & 8 , 7 + sdssj134251.60 - 005345.3 & sy 1 & 0.325 & @xmath361 & @xmath362 & @xmath196 & 45.30 & 10 , 7 + sdssj092554.43 + 453544.4 & qso & 0.329478 & @xmath363 & @xmath364 & @xmath365 & 45.22 & 13 , 11 + pg1001 + 291 & sy 1 & 0.3297 & @xmath366 & @xmath367 & @xmath175 & 45.59 & 18 , 14 + pg0832 + 251 & qso & 0.329773 & @xmath368 & @xmath369 & @xmath370 & 45.04 & 12 , 10 + sdssj132704.13 + 443505.0 & qso & 0.330709 & @xmath371 & @xmath372 & @xmath373 & 44.29 & 4 , 4 + pg1216 + 069 & sy 1 & 0.3313 & @xmath374 & @xmath375 & @xmath376 & 45.63 & 20 , 15 + rxj2154.1 - 4414 & sy 1 & 0.344 & @xmath377 & @xmath378 & @xmath229 & 45.54 & 24 , 16 + b0117 - 2837 & sy 1 & 0.348858 & @xmath379 & @xmath279 & @xmath376 & 45.66 & 21 , 17 + pg1049 - 005 & sy 1.5 & 0.3599 & @xmath380 & @xmath257 & @xmath381 & 45.69 & 11 , 10 + sdssj094952.91 + 390203.9 & qso & 0.365562 & @xmath382 & @xmath383 & @xmath384 & 45.53 & 11 , 9 + sdssj132222.68 + 464535.2 & qso & 0.374861 & @xmath385 & @xmath386 & @xmath387 & 45.00 & 8 , 6 + sdssj122035.10 + 385316.4 & sy 1 & 0.375767 & @xmath388 & @xmath389 & @xmath390 & 45.18 & 6 , 6 + sdssj024250.85 - 075914.2 & qso & 0.377651 & @xmath391 & @xmath392 & @xmath393 & 45.15 & 7 , 6 + sdssj123335.07 + 475800.4 & qso & 0.38223 & @xmath394 & @xmath395 & @xmath370 & 45.22 & 9 , 8 + sdssj134246.89 + 184443.6 & qso & 0.382 & @xmath396 & @xmath328 & @xmath360 & 45.23 & 8 , 7 + sdssj121037.56 + 315706.0 & sy 1.2 & 0.389041 & @xmath397 & @xmath398 & @xmath399 & 45.20 & 7 , 8 + hb89 - 0202 - 765 & sy 1 & 0.38939 & @xmath400 & @xmath401 & @xmath402 & 44.60 & 9 , 8 + sdssj110312.93 + 414154.9 & qso & 0.401023 & @xmath403 & @xmath290 & @xmath303 & 45.06 & 9 , 7 + sdssj133045.15 + 281321.4 & qso & 0.416754 & @xmath404 & @xmath405 & @xmath303 & 45.01 & 10 , 7 + sdssj111754.31 + 263416.6 & qso & 0.420466 & @xmath406 & @xmath407 & @xmath334 & 45.32 & 9 , 7 + sdssj143511.53 + 360437.2 & qso & 0.428593 & @xmath408 & @xmath409 & @xmath410 & 45.22 & 7 , 6 + he0435 - 5304 & qso & 0.427 & @xmath411 & @xmath412 & @xmath410 & 45.12 & 12 , 7 + sdssj110406.94 + 314111.4 & qso & 0.434356 & @xmath413 & @xmath414 & @xmath357 & 45.51 & 10 , 8 + b0120 - 28 & qso & 0.436018 & @xmath415 & @xmath416 & @xmath417 & 45.62 & 15 , 12 + sdssj161649.42 + 415416.3 & qso & 0.440417 & @xmath418 & @xmath419 & @xmath402 & 45.26 & 8 , 6 + sdssj080359.23 + 433258.4 & sy 1 & 0.448706 & @xmath420 & @xmath421 & @xmath354 & 45.45 & 7 , 5 + ton236 & sy 1.2 & 0.4473 & @xmath422 & @xmath423 & @xmath293 & 45.65 & 15 , 12 + pg0003 + 158 & sy 1.2 & 0.4509 & @xmath424 & @xmath425 & @xmath258 & 45.80 & 20 , 15 + he0153 - 4520 & qso & 0.451 & @xmath426 & @xmath427 & @xmath428 & 46.07 & 22 , 15 + sdssj100902.06 + 071343.8 & qso & 0.455631 & @xmath429 & @xmath430 & @xmath169 & 45.21 & 6 , 5 + sdssj091029.75 + 101413.6 & qso & 0.463194 & @xmath431 & @xmath432 & @xmath373 & 45.02 & 4 , 5 + sdssj082024.21 + 233450.4 & qso & 0.470212 & @xmath433 & @xmath434 & @xmath303 & 45.27 & 9 , 5 + sdssj161916.54 + 334238.4 & qso & 0.470946 & @xmath435 & @xmath436 & @xmath360 & 45.67 & 12 , 11 + sdssj092554.70 + 400414.1 & qso & 0.471139 & @xmath437 & @xmath438 & @xmath439 & 45.08 & 6 , 4 + sdssj123304.05 - 003134.1 & qso & 0.471167 & @xmath440 & @xmath441 & @xmath442 & 45.26 & 9 , 6 + pg1259 + 593 & sy 1 & 0.4778 & @xmath443 & @xmath264 & @xmath428 & 46.06 & 29 , 19 + he0226 - 4110 & sy 1 & 0.493368 & @xmath444 & @xmath445 & @xmath446 & 46.22 & 29 , 19 + sdssj155048.29 + 400144.9 & qso & 0.496843 & @xmath447 & @xmath448 & @xmath315 & 45.44 & 7 , 6 + hs1102 + 3441 & qso & 0.508847 & @xmath449 & @xmath450 & @xmath370 & 45.49 & 15 , 10 + sdssj113327.78 + 032719.1 & qso & 0.525073 & @xmath451 & @xmath452 & @xmath247 & 45.14 & 7 , 5 + sdssj025937.46 + 003736.3 & qso & 0.534178 & @xmath453 & @xmath454 & @xmath455 & 45.78 & 7 , 7 + sdssj094331.61 + 053131.4 & qso & 0.564336 & @xmath456 & @xmath457 & @xmath252 & 45.29 & 6 , 4 + sdssj040148.98 - 054056.5 & qso & 0.570076 & @xmath458 & @xmath459 & @xmath169 & 45.70 & 8 , 6 + pks0405 - 123 & sy 1.2 & 0.574 & @xmath460 & @xmath461 & @xmath462 & 46.73 & 63 , 29 + sdssj124154.02 + 572107.3 & qso & 0.583237 & @xmath463 & @xmath464 & @xmath442 & 45.34 & 9 , 6 + sdssj095000.73 + 483129.3 & sy 1 & 0.588734 & @xmath465 & @xmath466 & @xmath467 & 45.62 & 8 , 5 + sdssj225738.20 + 134045.4 & qso & 0.593937 & @xmath468 & @xmath469 & @xmath315 & 45.59 & 7 , 6 + sdssj022614.46 + 001529.7 & qso & 0.6151 & @xmath470 & @xmath471 & @xmath472 & 45.60 & 10 , 6 + sdssj105945.23 + 144142.9 & qso & 0.630543 & @xmath473 & @xmath474 & @xmath475 & 45.78 & 10 , 8 + he0238 - 1904 & qso & 0.631 & @xmath476 & @xmath477 & @xmath326 & 46.34 & 23 , 18 + sdssj111239.11 + 353928.2 & qso & 0.635784 & @xmath478 & @xmath479 & @xmath480 & 45.32 & 6 , 4 + 3c263 & sy 1.2 & 0.646 & @xmath481 & @xmath482 & @xmath483 & 46.33 & 30 , 20 + sdssj093518.19 + 020415.5 & qso & 0.649117 & @xmath484 & @xmath485 & @xmath252 & 45.49 & 6 , 5 + pks0637 - 752 & sy 1.5 & 0.653 & @xmath486 & @xmath487 & @xmath258 & 46.43 & 20 , 15 + sdssj080908.13 + 461925.6 & qso & 0.656338 & @xmath488 & @xmath489 & @xmath490 & 46.11 & 12 , 10 + sdssj105958.82 + 251708.8 & qso & 0.661907 & @xmath491 & @xmath492 & @xmath393 & 45.61 & 9 , 7 + sdssj154553.48 + 093620.5 & qso & 0.665 & @xmath493 & @xmath494 & @xmath495 & 45.81 & 1 , 6 + 3c57 & sy 1.2 & 0.670527 & @xmath496 & @xmath497 & @xmath265 & 46.03 & 21 , 12 + pks0552 - 640 & sy 1 & 0.68 & @xmath498 & @xmath499 & @xmath229 & 46.53 & 24 , 18 + sdssj151428.64 + 361957.9 & qso & 0.694596 & @xmath500 & @xmath501 & @xmath439 & 45.33 & 4 , 4 + sdssj144511.28 + 342825.4 & qso & 0.696951 & @xmath502 & @xmath503 & @xmath480 & 45.21 & 7 , 5 + sdssj113457.62 + 255527.9 & qso & 0.710078 & @xmath504 & @xmath505 & @xmath410 & 45.81 & 8 , 7 + sdssj155504.39 + 362848.0 & qso & 0.713654 & @xmath506 & @xmath507 & @xmath169 & 45.35 & 7 , 4 + sdssj124511.25 + 335610.1 & qso & 0.717 & @xmath508 & @xmath509 & @xmath402 & 45.51 & 7 , 8 + sdssj155304.92 + 354828.6 & qso & 0.721814 & @xmath510 & @xmath511 & @xmath512 & 46.02 & 8 , 6 + sdssj091440.38 + 282330.6 & qso & 0.735345 & @xmath513 & @xmath457 & @xmath402 & 45.57 & 8 , 6 + sdssj100102.55 + 594414.3 & qso & 0.746236 & @xmath514 & @xmath515 & @xmath516 & 46.18 & 11 , 9 + sbs1108 + 560 & qso & 0.766619 & @xmath517 & @xmath201 & @xmath518 & 46.10 & 4 , 10 + sdssj143726.14 + 504555.8 & sy 1 & 0.783319 & @xmath519 & @xmath520 & @xmath521 & 45.16 & 4 , 4 + sdssj102218.99 + 013218.8 & qso & 0.789304 & @xmath522 & @xmath523 & @xmath472 & 46.15 & 8 , 6 + sdssj234500.43 - 005936.0 & qso & 0.789429 & @xmath524 & @xmath168 & @xmath247 & 45.78 & 7 , 5 + sdssj101622.60 + 470643.3 & qso & 0.821527 & @xmath525 & @xmath526 & @xmath360 & 45.92 & 8 , 5 + sbs1122 + 594 & qso & 0.852 & @xmath527 & @xmath528 & @xmath399 & 46.08 & 12 , 9 + sdssj141910.20 + 420746.9 & qso & 0.873501 & @xmath529 & @xmath530 & @xmath303 & 45.73 & 7 , 4 + sdssj112244.89 + 575543.0 & qso & 0.905906 & @xmath531 & @xmath532 & @xmath533 & 45.87 & 7 , 5 + fbqsj0751 + 2919 & qso & 0.915 & @xmath534 & @xmath535 & @xmath536 & 46.48 & 26 , 19 + pg1407 + 265 & qso & 0.946 & @xmath537 & @xmath538 & @xmath175 & 46.56 & 33 , 20 + hb89 - 0107 - 025-ned05 & qso & 0.956 & @xmath539 & @xmath540 & @xmath247 & 45.92 & 12 , 8 + lbqs-0107 - 0235 & qso & 0.957039 & @xmath541 & @xmath542 & @xmath373 & 45.63 & 11 , 10 + pg1148 + 549 & qso & 0.975 & @xmath543 & @xmath421 & @xmath544 & 46.28 & 25 , 15 + sdssj084349.49 + 411741.6 & qso & 0.989986 & @xmath545 & @xmath546 & @xmath300 & 45.77 & 5 , 5 + he0439 - 5254 & qso & 1.053 & @xmath547 & @xmath246 & @xmath334 & 46.20 & 14 , 9 + sdss - j100535.24 + 013445.7 & qso & 1.0809 & @xmath548 & @xmath549 & @xmath550 & 46.00 & 11 , 11 + first - j020930.7 - 043826 & qso & 1.131 & @xmath551 & @xmath552 & @xmath387 & 45.82 & 13 , 13 + pg1206 + 459 & qso & 1.16254 & @xmath553 & @xmath554 & @xmath360 & 46.39 & 21 , 16 + pg1338 + 416 & qso & 1.21422 & @xmath555 & @xmath556 & @xmath557 & 45.77 & 16 , 11 + lbqs-1435 - 0134 & qso & 1.30791 & @xmath558 & @xmath559 & @xmath293 & 46.67 & 26 , 20 + pg1522 + 101 & qso & 1.32785 & @xmath560 & @xmath561 & @xmath562 & 46.64 & 19 , 14 + q0232 - 042 & qso & 1.43737 & @xmath563 & @xmath485 & @xmath442 & 46.35 & 15 , 12 + pg1630 + 377 & qso & 1.47607 & @xmath564 & @xmath565 & @xmath258 & 46.43 & 25 , 10 lllc sdssj115758.72 - 002220.8 & 0.25661 & @xmath566 & 25 + sdssj092554.43 + 453544.4 & 0.25057 & @xmath567 & 25 + & 0.30959 & @xmath568 & 25 + pg1216 + 069 & 0.28231 & @xmath569 & 34 + b0117 - 2837 & 0.34833 & @xmath570 & 25 + & 0.34866 & @xmath571 & 25 + sdssj122035.10 + 385316.4 & 0.27332 & @xmath572 & 25 + sdssj123335.07 + 475800.4 & 0.28495 & @xmath573 & 25 + hb89 - 0202 - 765 & 0.30657 & @xmath574 & 25 + sdssj110312.93 + 414154.9 & 0.27116 & @xmath575 & 25 + sdssj133045.15 + 281321.4 & 0.27553 & @xmath576 & 25 + sdssj111754.31 + 263416.6 & 0.35193 & @xmath577 & 25 + sdssj143511.53 + 360437.2 & 0.26246 & @xmath578 & 25 + & 0.37292 & @xmath579 & 25 + & 0.3876 & @xmath580 & 25 + pg0003 + 158 & 0.30573 & @xmath581 & 25 + & 0.31215 & @xmath582 & 25 + & 0.34787 & @xmath583 & 17 + & 0.36619 & @xmath584 & 25 + & 0.37034 & @xmath585 & 25 + & 0.38612 & @xmath586 & 25 + & 0.40137 & @xmath587 & 25 + & 0.42184 & @xmath588 & 25 + he0153 - 4520 & 0.40051 & @xmath589 & 25 + sdssj100902.06 + 071343.8 & 0.35586 & @xmath590 & 25 + & 0.3745 & @xmath591 & 25 + & 0.37554 & @xmath592 & 25 + & 0.37624 & @xmath593 & 25 + & 0.41401 & @xmath594 & 25 + sdssj091029.75 + 101413.6 & 0.2634 & @xmath595 & 25 + & 0.26375 & @xmath596 & 25 + & 0.26432 & @xmath597 & 25 + & 0.41924 & @xmath598 & 29 + sdssj082024.21 + 233450.4 & 0.45424 & @xmath599 & 25 + sdssj161916.54 + 334238.4 & 0.2694 & @xmath600 & 25 + & 0.26988 & @xmath601 & 25 + & 0.27086 & @xmath602 & 25 + & 0.2716 & @xmath603 & 25 + & 0.42676 & @xmath604 & 25 + & 0.44231 & @xmath605 & 25 + & 0.47091 & @xmath606 & 25 + & 0.47179 & @xmath607 & 25 + sdssj092554.70 + 400414.1 & 0.2477 & @xmath608 & 59 + & 0.25283 & @xmath609 & 25 + sdssj123304.05 - 003134.1 & 0.31875 & @xmath610 & 25 + & 0.43061 & @xmath611 & 25 + pg1259 + 593 & 0.2924 & @xmath612 & 25 + he0226 - 4110 & 0.24525 & @xmath613 & 25 + & 0.49252 & @xmath614 & 25 + sdssj155048.29 + 400144.9 & 0.31257 & @xmath615 & 41 + & 0.42739 & @xmath616 & 25 + & 0.4919 & @xmath617 & 25 + & 0.49255 & @xmath618 & 25 + hs1102 + 3441 & 0.26164 & @xmath574 & 25 + & 0.28916 & @xmath619 & 25 + & 0.28986 & @xmath620 & 25 + & 0.31039 & @xmath621 & 25 + & 0.33246 & @xmath622 & 25 + & 0.50607 & @xmath623 & 25 + sdssj113327.78 + 032719.1 & 0.24663 & @xmath624 & 25 + & 0.30216 & @xmath625 & 25 + & 0.45225 & @xmath626 & 25 + sdssj094331.61 + 053131.4 & 0.35464 & @xmath627 & 91 + sdssj040148.98 - 054056.5 & 0.32381 & @xmath628 & 25 + & 0.36547 & @xmath629 & 25 + pks0405 - 123 & 0.36077 & @xmath587 & 25 + & 0.4057 & @xmath630 & 25 + sdssj095000.73 + 483129.3 & 0.48502 & @xmath631 & 25 + sdssj225738.20 + 134045.4 & 0.37712 & @xmath623 & 25 + & 0.3787 & @xmath632 & 25 + & 0.49905 & @xmath633 & 25 + sdssj022614.46 + 001529.7 & 0.4161 & @xmath588 & 25 + sdssj105945.23 + 144142.9 & 0.34074 & @xmath634 & 25 + & 0.46567 & @xmath635 & 25 + & 0.57638 & @xmath636 & 25 + & 0.61727 & @xmath637 & 25 + he0238 - 1904 & 0.3441 & @xmath638 & 25 + & 0.35534 & @xmath639 & 25 + & 0.40102 & @xmath640 & 25 + & 0.42424 & @xmath641 & 25 + sdssj111239.11 + 353928.2 & 0.24679 & @xmath642 & 25 + 3c263 & 0.32545 & @xmath643 & 25 + & 0.52796 & @xmath644 & 25 + sdssj093518.19 + 020415.5 & 0.35457 & @xmath645 & 25 + & 0.42852 & @xmath646 & 25 + pks0637 - 752 & 0.24326 & @xmath647 & 38 + & 0.41755 & @xmath648 & 25 + & 0.4528 & @xmath649 & 25 + & 0.46847 & @xmath650 & 25 + sdssj080908.13 + 461925.6 & 0.61917 & @xmath651 & 25 + sdssj154553.48 + 093620.5 & 0.47379 & @xmath652 & 25 + & 0.47623 & @xmath653 & 25 + 3c57 & 0.24988 & @xmath654 & 25 + & 0.29224 & @xmath655 & 25 + & 0.32332 & @xmath569 & 76 + & 0.32827 & @xmath656 & 25 + & 0.38329 & @xmath657 & 25 + & 0.5332 & @xmath658 & 25 + pks0552 - 640 & 0.34517 & @xmath659 & 25 + & 0.34592 & @xmath660 & 25 + & 0.446 & @xmath661 & 30 + & 0.63017 & @xmath662 & 25 + sdssj151428.64 + 361957.9 & 0.41065 & @xmath663 & 42 + sdssj144511.28 + 342825.4 & 0.60722 & @xmath664 & 25 + sdssj113457.62 + 255527.9 & 0.2469 & @xmath665 & 25 + & 0.43233 & @xmath600 & 53 + & 0.50265 & @xmath666 & 25 + & 0.66824 & @xmath667 & 25 + sdssj155504.39 + 362848.0 & 0.30689 & @xmath668 & 25 + & 0.36504 & @xmath669 & 25 + & 0.57611 & @xmath670 & 25 + & 0.60275 & @xmath671 & 25 + sdssj124511.25 + 335610.1 & 0.31802 & @xmath672 & 25 + & 0.44947 & @xmath673 & 25 + & 0.55682 & @xmath674 & 25 + & 0.58762 & @xmath567 & 25 + & 0.63215 & @xmath675 & 65 + & 0.64496 & @xmath676 & 25 + & 0.64862 & @xmath677 & 25 + & 0.68918 & @xmath678 & 25 + & 0.71297 & @xmath679 & 35 + sdssj155304.92 + 354828.6 & 0.4756 & @xmath680 & 25 + & 0.52027 & @xmath681 & 25 + sdssj091440.38 + 282330.6 & 0.24426 & @xmath682 & 25 + & 0.59969 & @xmath634 & 25 + sdssj100102.55 + 594414.3 & 0.30355 & @xmath683 & 25 + & 0.4159 & @xmath684 & 25 + sbs1108 + 560 & 0.28646 & @xmath685 & 50 + & 0.46334 & @xmath686 & 35 + & 0.61765 & @xmath631 & 25 + & 0.68267 & @xmath671 & 25 + sdssj143726.14 + 504555.8 & 0.25065 & @xmath687 & 25 + & 0.56945 & @xmath688 & 25 + & 0.76901 & @xmath689 & 25 + & 0.77109 & @xmath690 & 25 + & 0.77248 & @xmath691 & 36 + sdssj102218.99 + 013218.8 & 0.39907 & @xmath692 & 25 + & 0.7425 & @xmath693 & 25 + sdssj234500.43 - 005936.0 & 0.2539 & @xmath650 & 25 + & 0.54818 & @xmath694 & 34 + sdssj101622.60 + 470643.3 & 0.4321 & @xmath695 & 25 + & 0.66475 & @xmath696 & 25 + & 0.72766 & @xmath697 & 14 + & 0.74627 & @xmath698 & 25 + sbs1122 + 594 & 0.31236 & @xmath699 & 25 + & 0.31784 & @xmath700 & 25 + & 0.35115 & @xmath701 & 25 + & 0.3919 & @xmath702 & 25 + & 0.55744 & @xmath703 & 25 + & 0.55817 & @xmath704 & 25 + & 0.5698 & @xmath705 & 25 + & 0.67835 & @xmath706 & 19 + sdssj141910.20 + 420746.9 & 0.289 & @xmath707 & 25 + & 0.42561 & @xmath571 & 25 + & 0.52221 & @xmath708 & 25 + & 0.53461 & @xmath709 & 25 + & 0.60842 & @xmath633 & 25 + & 0.80463 & @xmath710 & 25 + & 0.84523 & @xmath711 & 25 + sdssj112244.89 + 575543.0 & 0.39798 & @xmath712 & 25 + fbqsj0751 + 2919 & 0.43187 & @xmath713 & 25 + & 0.49455 & @xmath714 & 25 + & 0.82902 & @xmath715 & 25 + pg1407 + 265 & 0.29717 & @xmath716 & 25 + & 0.32569 & @xmath717 & 25 + & 0.57488 & @xmath718 & 25 + & 0.59964 & @xmath719 & 26 + & 0.68278 & @xmath720 & 33 + & 0.81699 & @xmath653 & 25 + hb89 - 0107 - 025-ned05 & 0.39909 & @xmath721 & 30 + & 0.53546 & @xmath722 & 25 + & 0.7178 & @xmath628 & 25 + & 0.8093 & @xmath723 & 25 + & 0.87569 & @xmath724 & 25 + lbqs-0107 - 0235 & 0.53635 & @xmath725 & 86 + & 0.71892 & @xmath726 & 25 + & 0.87636 & @xmath727 & 25 + pg1148 + 549 & 0.25242 & @xmath728 & 25 + & 0.57785 & @xmath729 & 25 + & 0.68864 & @xmath670 & 25 + & 0.90485 & @xmath730 & 25 + sdssj084349.49 + 411741.6 & 0.53258 & @xmath731 & 25 + & 0.53346 & @xmath732 & 25 + & 0.54106 & @xmath733 & 25 + & 0.54353 & @xmath734 & 25 + he0439 - 5254 & 0.328 & @xmath735 & 25 + & 0.61508 & @xmath736 & 52 + & 0.86515 & @xmath737 & 25 + sdss - j100535.24 + 013445.7 & 0.41753 & @xmath738 & 25 + & 0.41853 & @xmath739 & 25 + & 0.41963 & @xmath740 & 25 + & 0.83711 & @xmath741 & 25 + & 0.83938 & @xmath742 & 25 + first - j020930.7 - 043826 & 0.39035 & @xmath743 & 49 + & 0.8268 & @xmath744 & 25 + pg1206 + 459 & 0.40852 & @xmath745 & 25 + & 0.41412 & @xmath746 & 25 + & 0.92772 & @xmath747 & 46 + pg1338 + 416 & 0.34886 & @xmath748 & 41 + & 0.46369 & @xmath749 & 25 + & 0.62136 & @xmath750 & 58 + & 0.68617 & @xmath751 & 32 + lbqs-1435 - 0134 & 0.26228 & @xmath752 & 25 + & 0.29907 & @xmath733 & 25 + & 0.39214 & @xmath641 & 25 + & 0.43834 & @xmath753 & 25 + & 0.61283 & @xmath733 & 25 + & 0.68124 & @xmath726 & 25 + pg-1522 + 101 & 0.51841 & @xmath754 & 16 + & 0.57179 & @xmath755 & 17 + & 0.67518 & @xmath756 & 25 + & 0.72879 & @xmath757 & 26 + q0232 - 042 & 0.32239 & @xmath758 & 30 + & 0.73888 & @xmath759 & 35 + & 0.80773 & @xmath760 & 32 + pg1630 + 377 & 0.27395 & @xmath761 & 44 + & 0.27821 & @xmath762 & 25 + & 0.41774 & @xmath633 & 25 + & 0.41856 & @xmath763 & 25 + & 0.8111 & @xmath670 & 25 + & 0.91449 & @xmath764 & 25 | the ionizing fluxes from quasars and other active galactic nuclei ( agn ) are critical for interpreting their emission - line spectra and for photoionizing and heating the intergalactic medium ( igm ) . using far - ultraviolet spectra from the cosmic origins spectrograph ( cos ) on the _ hubble space telescope _ ( _ hst _ ) , we directly measure the rest - frame ionizing continua and emission lines for 159 agn at redshifts @xmath0 and construct a composite spectrum from 4751875 .
we identify the underlying agn continuum and strong euv emission lines from ions of oxygen , neon , and nitrogen after masking out absorption lines from the ly@xmath1 forest , 7 lyman - limit systems ( @xmath2 @xmath3 ) and 214 partial lyman - limit systems ( @xmath4 ) .
the 159 agn exhibit a wide range of fuv / euv spectral shapes , @xmath5 typically with @xmath6 and no discernible continuum edges at 912 ( ) or 504 ( ) . the composite rest - frame continuum shows a gradual break at @xmath7 , with mean spectral index @xmath8 in the fuv ( 12002000 ) steepening to @xmath9 in the euv ( 5001000 ) .
we discuss the implications of the uv flux turnovers and lack of continuum edges for the structure of accretion disks , agn mass inflow rates , and luminosities relative to eddington values . |
a wide variety of astrophysical objects are thought to be powered by accretion . in the case of accretion onto a black hole
, the properties of these flows mainly depend on the mass and spin of the black hole and on the mass accretion rate ( @xmath4 ) .
variations in @xmath4 are probably responsible for the spectral and variability states of black hole x - ray binaries ( bhxbs ; @xcite ) .
transient bhxbs spend long periods at very low x - ray luminosities , referred to as ` quiescence ' , but during occasional outbursts the luminosity increases by as much as 7 to 8 orders of magnitude , typically reaching values of tens of per cent of the eddington luminosity ( @xmath5 ) .
has shown that below around a few percent of @xmath5 bhxbs typically return to the so called hard state , in which the spectrum is dominated by a power law component with an index @xmath61.5 ( e.g. @xcite ) .
bhxbs are still in that state when they have decayed down to @xmath7@xmath5 . with _
chandra _ and _ xmm - newton _ it is possible to explore the spectral properties of bhxbs all the way to quiescence , for which log(@xmath8 ) is typically @xmath630.533.0 ( e.g.@xcite ; ) .
this range corresponds to @xmath63@xmath99@xmath10@xmath5for a 8.5@xmath11 black hole ( which is the average black - hole mass from @xcite ) .
about 15 bhxbs have been observed in quiescence with _
chandra _ and _ xmm - newton _ , but only in a handful of cases could the spectral properties be well constrained . nevertheless , these few cases suggest an interesting property of quiescent spectra : when fitted with a power - law the indices tend to be considerably softer than the index of @xmath12 found in the hard state ( e.g. @xcite ) .
although the errors for individual sources are still large , the overall trend seems to indicate that , at least spectrally , the quiescent state is different from the hard state that is observed three to five decades higher in @xmath8 .
corbel et al .
( 2003 ) and @xcite have demonstrated the existence of a relation between observed x - ray and radio emission from hard state black holes , of the form @xmath13 ; this relation is nearly the same for v404 cyg and gx 3394 .
however , , @xcite , @xcite , @xcite and @xcite found that the relation might not be as universal as previously thought ( see the compilation in @xcite ) . on the other hand @xcite and
@xcite found further relations between the x ray and near infrared fluxes in black hole transients : @xmath14 .
both relations extend over more than three decades in luminosity down to @xmath15@xmath5 and suggest that the x ray , near infrared and radio emission in the hard state are intimately connected .
correlations between optical and x ray light in the black hole source gx 3394 point in the same direction ( @xcite ) .
_ chandra _ and vla observations of a 062000 in quiescence by @xcite showed that the radio and x ray flux lie on the extension of the @xmath16 correlation , suggesting that it holds all the way down to quiescence .
some models predict that the relation should break down ( i.e. significantly steepen ) around @xmath17@xmath5(@xcite ) .
the result of gallo et al .
( 2006 ) seems to rule this out , but given the contradicting results on the decay rate and normalisation of the radio x ray correlations , this needs to be investigated further .
adaf and jet models both predict that the x ray luminosity scales as @xmath18 below @xmath7@xmath5 ( narayan et al .
1997 ; @xcite ) .
observational evidence for this scaling was deduced from x - ray / radio correlations (;
@xcite ) . in adaf models a gradual softening of the power
law photon index is expected until an index of @xmath62.2 in quiescence ( @xcite ; @xcite ) . however
, in adaf models the disk should recede in the hard state , something that is recently debated by @xcite ( but see @xcite ; @xcite ) . was discovered in august , 1977 by the ariel5 all sky monitor .
the position was more accurately determined by heao
i ruling out an association with 4u 1755388 ( @xcite ; @xcite ) .
the source has shown repeated outbursts since 2003 after being rediscovered by the international gamma
ray astrophysics laboratory ( integral ; igr j17464 - 3213 @xcite ) .
@xcite suggested that the integral found source is the same as .
recently , the source showed several outbursts ; one in 2007/2008 ( the data presented in this paper is from this outburst ; @xcite ) , one in late 2008 ( ) and one in 2009 ( @xcite ) .
@xcite list as a strong black hole candidate .
phase resolved optical or near infrared observations to determine the mass function of the black hole have not yet been reported , even though the source was found in outburst in near infrared and optical bands ( @xcite ; @xcite ) . near
infrared observations of the source in quiescence show an unrelated star within 1 and found the counterpart @xmath19band magnitude to be 17.1 ( @xcite ) .
is one of the few sources where x ray jets have been imaged ( @xcite ) .
the other sources where an x ray jet has been found are the black hole sources 4u 175533 and xte j1550564 ( @xcite ; @xcite ) , the neutron star source cir x1 ( @xcite ; @xcite ) and the peculiar source ss 433 ( @xcite ) .
in addition to x ray jet outflows @xcite found a strongly variable disc wind from . the distance to this source
is so far not well constrained , however , @xcite find from the fact that the jet speed is limited to maximally the speed of light and the observed x ray jet proper motion that the upper limit to the distance is [email protected] kpc , consistent with the often assumed galactic center distance of 7.5 or 8 kpc .
this assumption is based on the galactic coordinates of the source ( @xmath20357.255 and @xmath21 - 1.83 ) ; it lies towards the galactic bulge . in this paper
we assume a distance of 7.5 kpc . here , we report on contemporary _ rxte _ , and x ray and very large array ( vla ) radio observations of aimed at following the x ray and radio light curves and establishing the x ray radio correlation while the source decays to quiescence .
the observations have been obtained in the last stages of the outburst ending early 2008 . for this outburst
@xcite reported that the source returned to the low
hard state between jan , 24 , 2008 and feb .
, 1 , 2008 . to determine the quiescent spectral properties we further analysed three observations of obtained when it was in quiescence ( @xcite ) .
we observed with the satellite using the back illuminated s3 ccd chip of the advanced ccd imaging spectrometer ( acis ) detector ( @xcite ) on several occasions during the final parts of the decay to quiescence . during the first two observations in 2008 ( obs ids 8985 and 8986 ) we employed the high energy transmission grating ( hetg ) to mitigate effects of photon pile up .
in the subsequent observations we windowed the acis s ccd , providing a frame time of 0.4104 s. we have reprocessed and analysed the data using the _ ciao 4.0.2 _ software developed by the chandra x ray center . in our analysis
we have selected events only if their energy falls in the 0.37 kev range .
all data has been used , as background flaring is very weak or absent in all data . [
cols="^,^,^,^,^ " , ] @xmath22 vla 8.46 ghz observation reported in @xcite .
+ @xmath23 three @xmath24 rms limit . on several nights near
simultaneous radio data has been obtained at different frequencies .
we use this data to assess the radio spectral index .
initially , the spectral index is negative , implying optically thin radio emission .
e.g. the 1.4 , 4.86 to 8.46 ghz spectral index on mjd 54486 is [email protected] .
after the transition to the low hard state the spectral index is consistent with being 0 . to obtain a more accurate measurement of the radio spectral index ,
we have averaged three short 4.86 ghz and 5 8.46 ghz observations close in time ( in the range mjd 5449954505 ) where the radio flux was consistent with being constant .
the 4.868.46 ghz radio spectral index for these averages is [email protected] .
this is consistent with optically thick emission as has been observed in low
hard radio spectra .
however , the radio spectral index does not stay close to 0 during the decay to quiescence .
we have averaged the 1.4 ghz and 8.46 ghz data in the mjd range 5451654526 .
this data coincides with the radio flare observable at 8.46 ghz .
combining the three 1.4 ghz observations does provide a detection at [email protected] mjy . at 8.46 ghz
the average flux is [email protected] mjy .
the 1.48.46 ghz radio spectral index is [email protected] .
this implies optically thin radio emission late in the low hard state .
in figure [ rxcorr ] we plot the observed correlation between the x ray and 8.46 ghz radio fluxes for using the x ray observations closest in time to the radio observations ( see footnotes to the journals of the x ray observations and table [ vla - log ] ) . the best fitting power law with index [email protected] ( 1 @xmath24 ) is overplotted .
the index is less steep than the index of @xmath250.7 as found for several sources before ( s@xmath26@xcite ; ) .
there are several effects that could distort the picture .
first , the radio and x ray data are not strictly simultaneous ( see table [ vla - log ] ) .
there is typically less than 1 day between the start of the radio and x ray observations , however , in one case the difference in start times is as much as 1.5 days .
this together with the flares apparent in the radio light curve as well as in the _ rxte _ and x ray observations could result in a higher flux in either radio or x ray which might confuse the apparent correlation .
short duration radio flares have been found in e.g. v404 cyg ( @xcite ) .
last , since the spatial resolution of our vla observations is several arcseconds , the observed radio flux might include emission from a ballistic jet
ejection event earlier in the outburst .
although the atca non detection shows that the initial radio emission probably related to jet ejection events occurring before the transition to the low
hard state faded away , shocks further down the flow can lead to rebrightenings in radio ( and x ray ) at a later stage .
the fact that the radio spectral index during the epoch related to the radio flare at mjd 54520 is showing that the radio emission is optically thin strongly argues in favour of this scenario .
given the low spatial resolution of our radio observations this emission can significantly contribute to the radio flux density during the subsequent measurements .
xte j1550564 dramatically demonstrates this , with radio ( and x ray ) emission from the state transition ejecta still brightening years after the outburst ( @xcite ) .
the part of the radio decay of that our observations sample with detections lasts about 33 days . in the most extreme case of a jet / flow speed of c , in 33 days a distance of 0.028 pc
can be travelled . at a source distance of 7.5 kpc
this corresponds to 078 for a jet in the plane of the sky .
such an angular extent would still be unresolved by the vla in the cnb and certainly c configuration .
this indicates that similar to the case of xte j1550564 , shocks in the jet flow could indeed be responsible for the enhanced radio emission .
since such an angular extend could potentially be detected by we averaged the four observations obtained after march 20 , 2008 .
we registered the four observations to a common astrometric frame using the position of the brightest x ray source ( besides ) detected in the observations . after merging the registered frames
, we determined the position of the x ray source associated with in the same way as described above .
we find that the position of in the average of the last 4 observations in 2008 is consistent with the best known radio position found by @xcite , therefore , the shocked radio emission is either not responsible for x ray emission or it has travelled significantly less than the 078 calculated above . in order for radio emission from shocks
downstream to be the cause of the flattening of the index of the radio x ray correlation , the radio emission should not be accompanied by a similar amount of x ray emission as during the processes responsible for the index of @xmath00.7 during the normal low
hard state decay . in order to compare the normalisation of this correlation
with that found in other sources we need to know the distance of ( see @xcite ) . as mentioned in the introduction we use a distance of 7.5 kpc for . comparing the normalisation of the radio
x ray correlation for a distance of 7.5 kpc with that of other sources we find that lies above the area traced out by gx 3394 and v 404 cyg ( @xcite ) . even lowering the distance by a factor 2 one still finds the source above gx 3394 and v 404 cyg .
finally , our fluxes and luminosities are given in the 0.5 - 10 kev band whereas the correlation was plotted in the 211 kev band by @xcite .
the influence of the different energy bands on the slope of the radio x ray correlation is minimal since the x ray spectrum does not change significantly during the decay .
the normalisation is influenced by the difference in energy bands , though .
for comparison we also computed the 211 kev x
ray fluxes of our x ray observations and we find that the 211 kev values are 80@xmath25 per cent of the values at 0.510 kev .
thus , our 0.510 kev luminosities are slightly higher than the 211 kev luminosities bringing artificially closer to the area traced by gx 3394 and v 404 cyg .
the difference in normalisation would be larger if we had used the exact same energy range .
using , and _ rxte _ x ray observations with ( near)simultaneous radio vla observations we have observed the decay towards quiescence of the black hole candidate x ray binary while the source was in the low hard state
. the overall shape of the x ray decay light curve can be described with an exponential with an e folding timescale of @xmath04 days that levels off towards quiescence .
ray flares are superposed on the exponential decay light curve . during several of the observations flares
were observed as well . the unabsorbed 0.510 kev quiescent flux is 4.3@xmath27 . for an assumed source distance of 7.5 kpc
this implies a quiescent 0.5 - 10 kev x
ray luminosity of @xmath28 .
the source luminosity in quiescence after the outburst ending early 2008 is consistent with being the same as that after the 2004 outburst ( @xcite ) .
the quiescent x ray luminosity of @xmath29 of implies an orbital period longward of @xmath010 hours assuming follows the general trend between orbital period and quiescent x ray luminosity reported in @xcite .
such an orbital period suggests that the mass donor star in has evolved off the main sequence .
alternatively , the mass of the black hole in could be substantially larger than that in the sources defining the relation between orbital period and quiescent x ray luminosity . the low
hard state radio light curve shows evidence for radio flares superposed on an exponential decay with an e folding timescale of @xmath30 days .
when plotted against the x ray light curve the radio x ray correlation index is lower ( @xmath31 ) than found before ( @xmath32 ; ; @xcite ) .
the radio flares could be responsible for extended jet emission that due to the relatively large vla beamsize in our observations using c and cnb configurations was not resolved .
the 1.48.46 ghz radio spectral index during the mjd 54520 flare is [email protected] indicating that the radio emission is probably optically thin .
relativistic arcsecond scale jets have not been found before for sources in the low hard state ( cf .
@xcite ; @xcite ) .
therefore , it seems more likely that jet ejections earlier in the outburst of caused shocks when travelling along the jet .
we note that similar events could be behind the apparent scatter and variability in the observed radio x ray correlation in other sources ( cf .
the resulting index then depends on the amount and energy in the shocks caused by earlier ejection events .
these will change from source to source and from outburst to outburst .
for instance , plotting the radio and x ray detections in the low hard state during the decay @xcite we find that two of the three 2003 data points fall approximately a factor 2 to 3 below the curve in figure [ rxcorr ] .
increases in the radio flux density in the low hard state by factors of 310 have been observed in the short timescale radio flares in the quiescent state of v 404 cyg ( @xcite ) .
given that the radio and x ray data that we presented here are not exactly simultaneous , any hour long radio flare would have no corresponding flare in the x ray band .
furthermore , even if the observations had been strictly simultaneous , the short integration times of our initial radio observations implies that an x ray flare could have been missed since such an x ray flare would take tens of minutes or longer to propagate downstream to the point in the compact jet where the optical depth is @xmath01 at 8.46 ghz .
nevertheless , in order to produce a slope of @xmath33 the radio observations must have been close to the peak of flares , which seems unlikely .
we conclude that optically thin radio emission from shocks caused by jet ejections earlier in the outburst influence the late time low hard state radio emission .
previous studies of black hole ( candidate ) sources decaying via the hard state to quiescence suggests that the power law index of the x ray spectrum in the quiescent state is different from that in the low hard state .
@xcite observed gs 112468 with _
ginga _ and found a power - law index that remained essentially constant at @xmath61.6 during the time the source was decaying in the hard state , except at the lowest observed luminosity of @xmath34@xmath5 ( for @xmath35=7@xmath11 ) when it was found to be [email protected] .
have reported a power - law spectral index of @xmath62.2 around @xmath36@xmath5 for gx 3394 ( for @xmath35=5.8@xmath11 ) .
@xcite followed xte j1650500 down to @xmath37@xmath5 ( for @xmath35=8.5@xmath11 ) with _ chandra _ , and also found evidence for spectral softening .
finally , @xcite show that the power law spectral index is softer in v404 cyg in quiescence compared to that in the brighter low hard state . in figure [ plindex ]
we have plotted the power law index from the and spectral modelling on during the decay towards quiescence and in quiescence .
the data point at the lowest x ray flux is determined by combining the quiescent data obtained right after the outburst of early 2008 as well as the quiescence observations used by @xcite .
there is no evidence for a softening of the power
law index towards and in quiescence , although for this source the relatively high neutral hydrogen column density limits the accuracy of the power
law determination .
a fit of a constant power
law spectral index gives a best fitting index of [email protected] with a @xmath38 for 9 degrees of freedom .
we have compared the x ray flux decay rate of with that observed during the last part of the outburst decay from the black hole candidate xte j1908 + 094 and we find that the fit function describing the decay of is a good approximation to the last phase of approximately three weeks presented in jonker et al .
( 2004 ) for xte j1908 + 094 .
note however , that due to a reduced sampling that source was only observed twice during those three weeks . nevertheless , it is interesting to see that the sources have a similar decay rate . if other sources also follow the same decay rate in x
rays this could provide constraints on the accretion disc model , such as the accretion disc to adaf evaporation .
pgj acknowledges support from a vidi grant from the netherlands organisation for scientific research .
mapt acknowledges support from nasa grant go8 - 9042a .
ds acknowledges an stfc advanced fellowshipthe national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
e. , 2007 , in american institute of physics conference series , vol .
924 , the multicolored landscape of compact objects and their explosive origins , di salvo t. , israel g. l. , piersant l. , burderi l. , matt g. , tornambe a. , menna m. t. , eds . , pp .
715722 h. a. , barthelmy s. d. , baumgartner w. , cummings j. , fenimore e. , gehrels n. , markwardt c. b. , palmer d. , parsons a. , sakamoto t. , skinner g. , tueller j. , ukwatta t. , 2009 , the astronomer s telegram , 2058 , 1 t. , romano p. , mangano v. , moretti a. , cusumano g. , la parola v. , troja e. , campana s. , chincarini g. , tagliaferri g. , capalbi m. , perri m. , giommi p. , burrows d. , 2006 , nuovo cimento b serie , 121 , 1521 | in this paper we report on radio ( vla and atca ) and x ray ( _ rxte _ , and ) observations of the outburst decay of the transient black hole candidate in early 2008 .
we find that the x ray light curve followed an exponential decay , levelling off towards its quiescent level .
the exponential decay timescale is @xmath04 days and the quiescent flux corresponds to a luminosity of 3@xmath1 .
this together with the relation between quiescent x ray luminosity and orbital period reported in the literature suggests that has an orbital period longer than @xmath010 hours .
both the radio and x ray light curve show evidence for flares . the radio x ray correlation can be well described by a power law with index @xmath00.18 .
this is much lower than the index of @xmath00.60.7 found for the decay of several black hole transients before .
the radio spectral index measured during one of the radio flares while the source is in the low hard state , is [email protected] , which indicates that the radio emission is optically thin .
this is unlike what has been found before in black hole sources in the low hard state .
we attribute the radio flares and the low index for the radio x ray correlation to the presence of shocks downstream the jet flow , triggered by ejection events earlier in the outburst .
we find no evidence for a change in x ray power law spectral index during the decay , although the relatively high extinction of @xmath3 limits the detected number of soft photons and thus the accuracy of the spectral fits .
stars : individual ( ) accretion : accretion discs stars : binaries x - rays : binaries |
there is solid observational evidence for the existence of black holes ( bhs ) in two very different regimes .
stellar - mass bhs , with masses in the range @xmath2 m@xmath3 , are the end products of normal high - mass stellar evolution .
solid mass estimates for these bhs come from measurements of the mass function in several tens of galactic x - ray binaries .
supermassive bhs ( smbhs ) , at the other extreme , with masses between @xmath4 and @xmath5 m@xmath3 , have been reliably weighed in galactic centers based on the motions of their surrounding stars and gas .
a tight correlation between the masses of smbhs and the velocity dispersion of their host spheroid ( the @xmath6 relation ) evidences that an intimate link must exist between these bhs and the processes relevant to galaxy formation , and highlights the importance of understanding the seeds and growth mechanisms of smbhs .
imbhs , with masses in between those of the above two regimes , are likely to play a key role in these questions but they are yet to be convincingly detected .
existing theoretical work provides arguments for and against the presence of imbhs at the centers of globular clusters ( gcs ) .
several possible channels of formation have been proposed ( see van der marel 2004 for a review ) , and , interestingly , all those mechanisms lead to imbh masses consistent with an extrapolation of the @xmath7 relation down to dispersions typical of gcs . on the other hand , it has been argued that it may be difficult for gcs , given their relatively shallow potential wells , to retain growing massive black holes at their centers for too long ( baker et al .
2008 ; holley
bockelmann et al .
2008 ; moody & sigurdsson 2009 ) .
this current debate , however , should not prevent astronomers to set out plans to look for the existence of these objects . to the question of whether imbhs are possible or not in gcs , the right attitude should be one that , although we do not know the answer at this moment , we will also act independently from theoretical prejudices and observationally probe for their presence anyway .
any tight upper limits to the masses of central compact objects in a number of gcs are equally important for the field as any solid mass measurements eventually confirming their existence .
certainly , any of these alternatives would clearly constitute an important improvement over today s situation .
as in the case of smbhs in active galactic nuclei , one could argue that the most unambiguous signature of the presence of a massive compact object would be the detection of central , unresolved radio and/or x - ray luminosity generated by the accretion of matter onto the central object . in the case of star clusters , however , one must make sure that this signature is not consistent with an accumulation of more common stellar - mass accreting objects ( cataclysmic variables , x - ray binaries , stellar - mass black holes ) or pulsars .
this danger is well illustrated by the case of a gc associated with the giant elliptical galaxy ngc 4472 in the virgo cluster , where maccarone et al .
( 2007 ) reported a strong and highly variable x - ray signature `` which rules out any object other than a black hole in such an old stellar population '' , then estimating about 400 m@xmath3 for the mass of this object .
however , as noted by those authors , such mass estimate was necessarily based on several assumptions regarding highly uncertain variables of accretion physics ( e.g. , innermost stable orbit , accretion rate , disk and continuum model , unabsorbed luminosity , etc . ) , and can hardly be considered a mass measurement
. indeed , they later determined , based on optical spectroscopy , that the object in question is most likely a stellar - mass black hole ( zepf et al .
2008 ) . as a consequence , while the observation of central , unresolved accretion signatures originating from star clusters certainly constitutes evidence for the possible presence of imbhs , and , moreover , may be considered a prime way to search for imbh candidates , it is nevertheless not adequate for the purposes of a solid , unambiguous mass measurement .
the same considerations apply to the possibility of catching an event of stellar disruption by an imbh ( irwin et al .
2009 ) .
therefore , for unambiguous mass measurements , we need to rely on dynamics , and there have been various observational efforts in this direction . based on dynamical modeling of measured line - of - sight ( los ) kinematics , evidence for the presence of imbhs has been claimed in three gcs : m15 and @xmath8 cen in the milky way ( gerssen et al .
2002 and noyola et al .
2008 , respectively ) , and g1 in m31 ( gebhardt et al .
2002 , 2005 ) . in the cases of m15 and g1 , however , the available data are also consistent with nonequilibrium models in which mass segregation occurs over time , but which do not posses an imbh .
( baumgardt et al .
2003a , b ) . moreover ,
those measurements are restricted to the luminous red giants in the clusters .
not being very numerous , any analysis based on the kinematics of the giants will be more affected by shot noise and projection effects ( i.e. , not knowing whether the tracers contributing to most of the velocity signal are actually probing the imbh s sphere of influence , or whether they are instead located in front of or behind the cluster center ) than if the data set were dominated by the more numerous , although fainter , main - sequence stars .
the cases of g1 and @xmath8 cen are unique in their own way .
we start by pointing out that , since the subject of this meeting is star clusters , and although the presence of an imbh does not have anything to do with this fact , it is nevertheless appropriate to recall that @xmath8 cen and g1 are both likely to be the nuclei of stripped dwarf galaxies rather than regular gcs .
however , as this is yet to be considered an established result , we will discuss them here anyway and , instead of having to specify their uncertain nature by calling them ` stellar spheroids of low central velocity dispersion ' , we will refer to them loosely as clusters . with respect to g1 , one can argue that this remains the best case so far for an imbh in a gc .
this is because , on top of the dynamical evidence quoted above , this cluster is also a source of radio and x - ray emission ( ulvestad et al . 2007 ; kong 2007 ) , as one would expect for an accreting compact object .
the problem with this additional accretion evidence , however , is one of angular resolution , since both the radio and x - ray source positions are too uncertain and they could still be the result of a number of less exotic objects , such as x - ray binaries or cvs distributed throughout the cluster . of course , this situation is expected to improve when performing higher - angular - resolution observations . in the case of @xmath8 cen
, the problem has been again related to the effect of bright giants , in this case the effect that their shot noise has on the determination of the cluster center , a key step in the determination of light and velocity - dispersion profiles .
briefly , anderson & van der marel ( 2009 ) used high - angular - resolution _ hst _ images to measure the proper motions of cluster members , most still on the main sequence , and found a cluster center about 12 arcsec away from that determined by noyola et al .
the light and ( 2d ) velocity - dispersion profiles computed with respect to the new center did not show the signatures that noyola et al .
( 2008 ) atributed to a 40000 m@xmath3 imbh .
next , van der marel & anderson ( 2009 ) constructed dynamical models with the proper - motion data and determined that models both with and without a black hole fit the data equally well , reporting a tight upper limit of 12000 m@xmath3 in case an imbh is present .
until just a few years ago , the generalized intuition indicated that the best places to look for imbhs were the centers of very dense stellar systems such as the cores of galactic globular clusters with very steep luminosity profiles .
solid theoretical basis for this expectation came from the seminal work of bahcall & wolf ( 1976 , 1977 ) , which predicted the development of stellar density cusps surrounding massive central bhs .
in well - relaxed clusters , groups of stars of different mass formed cusps with correspondingly different slopes , reflecting the well - known phenomenon of mass segregation .
this prompted the selection of galactic globular clusters with dense , collapsed cores for observational studies that would look for the dynamical signature of an imbh at their centers .
however , drawing conclusions from the simulations as to what to expect observationally has never been an easy task , given that while the theoretical results are unambiguous regarding the mass of the particles , the observations can only track the luminous stars , and conversion from one to the other depends on many assumptions .
more recently , it was realized that an imbh at the center of a star cluster constitutes an additional source of heat acting against the rapid collapse of the cluster s core .
baumgardt et al .
( 2004a , b ) found that the imbh not only prevents the collapse of the core but also produces an expansion of the cluster , making it unlikely that dense core - collapsed clusters could harbor imbhs .
therefore , a completely new picture of where it would be best to look for imbhs has emerged : unlike collapsed cores and dense central regions , the most likely clusters to harbor an imbh would be those whose projected surface brightness profile are well fit by regular king - like models with an extended core and intermediate concentrations ( @xmath9 ) , with the surface brightness rising slightly only in the very inner regions of this core , forming a shallow , barely perceptible density cusp in the form of a power law with slope @xmath10 ( baumgardt et al .
2005 ; miocchi 2007 ; trenti et al .
in section2 we started by arguing that the observation of accretion signatures , while great for the detection of imbh candidates , was not adequate for a reliable mass measurement of satisfactory precision , for which we had to rely on dynamical analyses .
the current status of the evidence for imbhs in gcs , however , proves that dynamics has its own problems .
a large part of these , as summarized in section2 , have to do with the limited amount of data available for those studies , which did not provide enough kinematic information to , say , distinguish the effect of a single massive central object from that of a more extended mass distribution .
but this constitutes just one aspect of the potential problems associated with dynamics , the other side having to do with the subsequent handling of the available data .
indeed , in the process of the modeling of kinematic data , two of the most important dangers are ( i ) the inadequate exploitation of the full information content of a given dataset and ( ii ) the adoption of simplifying assumptions that are too restrictive and thus explore only a limited region of the space of possible solutions .
the management of these latter points can have profound implications for the inferred mass distribution ( arguably the main goal of stellar dynamics ) , thus creating a delicate balance between the available observations for a given problem and the complexity of the models chosen to make use of such data .
this balance was nicely exemplified by the initial debate regarding the dark - matter content of some intermediate - luminosity elliptical galaxies , during which los velocity measurements of planetary nebulae in the halos of these galaxies showing a keplerian - like decline of velocity dispersion with radius were interpreted as evidence of the presence of little , if any , dark matter in those galactic halos ( romanowsky et al .
although this remains a subject of current discussion , it was later shown that such conclusions may well be the result of too restrictive assumptions regarding both the geometry of the underlying potential and the anisotropy of the orbits of the planetary nebulae around those galaxies ( dekel et al .
2005 ; de lorenzi et al .
2009 ) .
the key for an unambiguous assessment of the presence of imbhs in gcs is availability of many stars with well - measured velocities probing the gc s core , and containing enough kinematic information to permit models to be very general ( section4.2 ) .
this demands : * _ high spatial resolution : _ stellar velocities need to be measured inside the sphere of influence of the putative bh , @xmath11 .
thus , for an imbh of @xmath12 m@xmath3 , many stars have to be well resolved inside @xmath13 . * _ at least two velocity components : _ a well - known degeneracy between mass and anisotropy ( binney & mamon 1982 ) is the main obstacle for a reliable measure of the mass distribution in stellar systems based on line - of - sight velocities .
proper motions provide two components of the space velocity , allowing to actually measure the anisotropy of the stellar orbits instead of assuming it . * _ optimization of observing ( wavelength ) window : _ high levels of astrometric accuracy require minimizing as much as possible the red light from nearby , luminous giants and the unresolved background of low - mass stars populating the crowded cores of gcs .
thus , observations must be conducted at short enough wavelengths , from the blue to the near - ultraviolet .
even with the largest - aperture telescopes and assisted by adaptive optics , ground - based observations are unable to produce radial velocities and/or proper motions for many stars in the inner arcsecond of gcs ( e.g. , gebhardt et al . 2000 ) . built exactly to overcome these problems , _ hst _ is by design the _ only _ instrument that , by largely satisfying the three above requirements , will be able to resolve , within the next decade , the question of the existence of imbhs in gcs .
appropriate data represent only half of what is needed , and highly specialized models are required to avoid the shortcomings of more simplified methods that were developed for other circumstances .
first , special machinery is necessary to handle data of a discrete nature ( i.e. , individual 2d velocities of a number of resolved stars ) .
the most typical approaches for dynamical analysis ( the jeans equations and orbit - based models designed to work with observed velocity profiles from integrated light ) must necessarily bin the data to produce averaged profiles of velocity dispersions ( see figure 13 in van de ven et al .
clearly , binning amounts to not exploiting the entire information content of the dataset , and , since only a limited number of stars are available inside the bh s sphere of influence , this can only have the effect of degrading and possibly erasing the signature of an imbh .
second , given that the signature of any central bh would be imprinted on the detailed orbital structure of the nearest stars , overly restrictive assumptions regarding the form of the stellar distribution function , most crucially the degree of isotropy of the orbits of the tracer stars , must be avoided .
this point becomes obvious when considering that , for m15 for example , _ hst _ los velocities indicate that there is a clear increase in the net rotation of the stars in the inner arcseconds of the cluster ( see figure 9 of gerssen et al .
this fact stands in stark contradiction to the standard expectation that relaxation should rapidly drive the velocity distribution at the centers of gcs towards isotropy , thus illustrating the necessity of using the most general models possible to fit these kinematics .
third , the fact that there is flattening and substantial rotation seen in some gcs makes the use of axisymmetric models , rather than spherical ones , a must .
starting around 2004 , our team embarked on a comprehensive program to use accurate _ hst _ proper - motion measurements of the stars in the central regions of a number of galactic gcs and search for evidence of imbhs . reflecting the theoretical expectations of that time , all gcs included in these initial _ hst _ programs
were selected to be among those with high central densities ( several of them being core - collapsed systems ) . through the following years
, target clusters spanning a variety of structural properties were included in the sample , and today we have accumulated data , from our own programs as well as by exploiting the _ hst _ archive , for a total of nine gcs with two or more astrometric epochs , producing time baselines between 2 and 6 years . reflecting the history of _ hst _ of the last few years , the data were obtained using those instruments providing the highest angular resolution possible at the time of the observations .
table 1 summarizes our current dataset .
we are currently in the process of reducing these data , and preliminary stellar proper - motion catalogs for some of these clusters are being generated ( bruursema et al .
, in prep . ) .
for a review of this intensive process , see the proper - motion catalog of @xmath8 cen of anderson & van der marel ( 2009 ) .
we plan to perform the dynamical analysis of all these data sets by making use of schwarzschild models specifically designed to exploit data of a discrete nature , such as that of our proper - motion catalogs .
schwarzschild s technique is arguably the most developed and well - tested method available for constraining the detailed mass distribution of equilibrium stellar systems , and is based on the simple idea of finding the best combination ( or superposition ) of all possible orbits that , allowed by some previously specified potential , reproduces both the spatial distribution of the tracers ( i.e. , the light distribution ) and the measured kinematics .
the simplicity and success of the method therefore rely on two aspects : ( i ) that the overall stellar system can be considered to be in equilibrium ( a safe assumption in the case of old galactic gcs ) and ( ii ) whether or not the set of orbits considered for the superposition is really comprehensive . as long as
these two conditions are satisfied , the method is very general and free from most assumptions . as of today ,
the only orbit - based dynamical tool available that fulfills all requirements outlined in section4.2 is the discrete schwarzschild code we recently developed and tested ( chanam et al .
2008 ) , and which we will be employing in this program .
jc acknowledges the support of the american astronomical society and the international astronomical union in the form of travel grants to attend the iau general assembly , as well as the support from nasa through hubble fellowship grant hf-01216.01-a , awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 . | establishing or ruling out , either through solid mass measurements or upper limits , the presence of intermediate - mass black holes ( imbhs ; with masses of @xmath0 m@xmath1 ) at the centers of star clusters would profoundly impact our understanding of problems ranging from the formation and long - term dynamical evolution of stellar systems , to the nature of the seeds and the growth mechanisms of supermassive black holes .
while there are sound theoretical arguments both for and against their presence in today s clusters , observational studies have so far not yielded truly conclusive imbh detections nor upper limits .
we argue that the most promising approach to solving this issue is provided by the combination of measurements of the proper motions of stars at the centers of galactic globular clusters and dynamical models able to take full advantage of this type of data set .
we present a program based on _ hst _ observations and recently developed tools for dynamical analysis designed to do just that . |
since the original discovery of millisecond pulsars ( msps ; * ? ? ?
* ) the prevailing theory of their formation has centered on `` recycling '' by transfer of matter and angular momentum from a close companion star during a low - mass x - ray binary ( lmxb ) phase @xcite .
the discovery of the first accreting x - ray msp , sax j1808.43658 @xcite , provided compelling evidence for this evolutionary sequence ( see also * ? ? ? * for a review ) .
additional observational support in favor of this hypothesis came with the discovery of the `` missing link '' radio pulsar psr j1023 + 0038 ( also known as first j102347.6 + 003841 ) .
optical observations revealed an accretion disk in the system in 2001 @xcite , which was absent after 2003 @xcite and at the time of the radio pulsar discovery in 2009 @xcite .
the long - suspected evolutionary connection between lxmbs and radio msps was conclusively established when psr j18242452i in the globular cluster m28 was seen to switch between rotation - powered ( radio ) and high - luminosity accretion - powered ( x - ray ) pulsations @xcite .
recent radio , x - ray , and @xmath1-ray observations revealed that in 2013 june psr j1023 + 0038 had undergone another transformation @xcite .
complete cessation of radio pulsations was observed , accompanied by an extraordinary five - fold increase in the _ fermi _ large area telescope ( lat ) flux and enhancement in uv and x - ray brightness @xcite .
these recent findings have revealed that rather than a one - time change from a lmxb to a permanent radio msp state , for some systems this phase of neutron star compact binary evolution involves recurrent switching between the two states . in the absence of an accretion disk ,
psrs j1023 + 0038 and psr j18242452i belong to the family of so - called `` redbacks '' ( see * ? ? ? * and references therein ) , namely , eclipsing radio msps with relatively massive ( @xmath20.2 m@xmath3 ) non - degenerate companions that are ( nearly ) roche - lobe filling .
these objects are distinct from `` black widow '' systems , which are bound to very low mass degenerate stars being ablated by the pulsar wind @xcite .
extensive x - ray studies of redbacks in globular clusters @xcite and the field of the galaxy @xcite have revealed that the dominant source of x - rays from these binaries is non - thermal radiation with @xmath4 originating in an intra - binary shock @xcite generated by the interaction of the pulsar wind and matter from the companion star .
the x - ray emission is strongly modulated at the orbital period in all redback systems , with a decline in flux when the secondary star is between the pulsar and the observer .
this orbital - phase - dependent x - ray variability appears to be a distinguishing characteristic of these peculiar msp binaries and provides a convenient way to identify additional redbacks even in the absence of radio pulsations .
the galactic x - ray source xss j122704859 ( 1rxs j122758.8485343 ) has posed a mystery since its discovery . although initially classified as a cataclysmic variable ( masetti et al .
2006 ; butters et al .
2008 ) , subsequent multi - wavelength studies ( pretorius 2009 ; de martino et al . 2010 ; hill et al . 2011 ; de martino et al .
2013 ; papitto et al .
2014 ) revealed that this system closely resembles a quiescent lmxb .
xss j122704859 is exceptional in that it was the first lmxb putatively associated with a @xmath1-ray source , 1fgl j1227.94852 @xcite , now known as 2fgl j1227.74853 @xcite .
deep searches have failed to detect radio pulsation from xss j122704859 during its high optical / x - ray flux state @xcite .
prior to 2012 november / december , the x - ray spectrum of xss j122704859 was well described by a pure powerlaw with @xmath5 and a 0.110 kev luminosity of @xmath6@xmath7 erg s@xmath8 @xcite .
the source exhibited occasional intense flares and peculiar , frequent drops in x - ray flux that are not correlated with orbital phase .
this behavior is reminiscent of the erratic flux variations observed in psrs j1023 + 0038 ( patruno et al .
2014 ; tendulkar et al .
2014 ; s. bogdanov in prep . ) and j18242452i ( papitto et al . 2013 ; linares et al . 2014 ) in their accretion disk states , hinting at a connection with these systems .
recent optical and x - ray observations revealed that xss j122704859 had undergone a substantial decline in brightness and no longer exhibits evidence for an accretion disk ( bassa et al .
the abrupt change to a disk - free system appears to have occured between 2012 november 14 and 2012 december 21 .
given the similarities with psr j1023 + 0038 in both states , xss j122704859 is presently consistent with hosting a rotation - powered millisecond pulsar .
this prediction was confirmed with the recent detection of 1.69-millisecond radio pulsations from the system ( roy et al .
2014 ) . here
we present recently acquired _ xmm - newton _
european photon imaging camera ( epic ) and _ chandra x - ray observatory _ advanced ccd imaging spectrometer ( acis ) observations of xss j122704859 in its new low flux state .
these observations confirm that this source has undergone a metamorphosis to a redback , i.e. a compact binary containing a rotation - powered millisecond pulsar .
the analysis is presented as follows . in 2 ,
we provide details on the observations and data reduction procedures . in 3 , we present an x - ray variability study . in
4 we summarize the phase - averaged and phase - resolved spectroscopy , while in 5 we report on an imaging analysis .
we provide a discussion and conclusions in 6 .
motivated by the sudden decline in x - ray and optical brightness of xss j122704859 @xcite , we obtained a target of opportunity _ xmm - newton _ observation on 2013 december 29 ( obsid 0727961401 ) for a total exposure time of 38-ks . the pn @xcite and two mos @xcite epic instruments used the thin optical blocking filters and were set up in full window mode .
the data were processed using the _ xmm - newton _ science analysis software ( sas ) version xmmsas_20130501_1901 - 13.0.0 . after applying the recommended flag , pattern , and pulse invariant filters ,
the event lists were screened for instances of strong background flares , which resulted in effective exposures of 31.2 , 31.2 , and 26.9 ks for the mos1 , mos2 , and pn , respectively . for the variability and spectral analyses presented in 3 and 4 ,
source counts were extracted from each detector within a circular region of radius 40@xmath9 , which contains @xmath690% of the point source energy .
the background counts were extracted from three source - free regions on the same detector chip as the target . to study the system s x - ray variability , the photon arrival times
were barycentred using the barycen sas command with the de405 solar system ephemeris and assuming the source j2000 coordinates ra=12@xmath1027@xmath1158@xmath12748 , dec=@xmath1348@xmath1453@xmath042@xmath1588 .
owing to the coarse read - out time of the epic instruments in full window mode ( 0.73 s and 2.6 s for the pn and mos , respectively ) , it was not possible to search for pulsations in the millisecond range .
the simultaneous fast photometry u band filter data obtained with the _ xmm - newton _ optical monitor ( om ) were presented in @xcite .
+ to look for extended emission around xss j122704859 , a target of opportunity _ chandra _ acis - s observation was conducted on 2014 january 11 ( obsid 16561 ) for a duration of 30.2-ks .
the best known position of xss j122704859 was placed at the nominal aim point of the acis s3 chip , which was configured in faint telemetry mode .
the data processing and analysis were performed with ciao 4.6 @xcite and caldb 4.5.9 using standard procedures . for the variability and spectroscopic studies , we extracted photons within 2@xmath16 of the source . for the variability study
the event times were translated to the solar system barycenter assuming the de405 jpl solar system ephemeris and the same position for xss j122704859 used for the _ xmm - newton _ data .
the 3.2-s read - out for acis - s prevented a search for pulsations . to produce spectra suitable for analysis
, the extracted source counts in the 0.38 kev range were grouped so as to ensure at least 15 counts per energy bin .
the background was taken from three source - free circular regions near the target .
based on the refined binary orbital period ( @xmath17 hours ) and reference epoch of the ascending node ( hjd @xmath18 corresponding to orbital phase @xmath19 ) of xss j122704859 obtained in @xcite we have folded the barycentered _ xmm - newton _ and _ chandra _ point source x - rays at the binary period .
both observations span slightly more than one binary orbit .
large - amplitude variability correlated with binary phase is immediately evident from figure 1 .
this behavior is in stark contrast with the _ xmm - newton _
observations from 2009 and 2011 @xcite where no orbital dependence on the x - ray flux is seen and only random , aperiodic variability is present . to formally establish the statistical significance of the variability
, we use a kuiper test @xcite , which accounts for the non - uniform exposure across the orbit of the folded lightcurves and is not dependent on the choice of binning of the lightcurve , making it a better - suited choice than the kolmogorov - smirnov and @xmath20 tests for this purpose . applying this test to the folded , unbinned lightcurves , yields @xmath21 ( 13.3@xmath22 ) and @xmath23 ( 7.7@xmath22 ) probabilities for the _ xmm - newton _ and _ chandra _ data , respectively , that photons drawn from a constant distribution would exhibit this level of non - uniformity . for comparison , the folded lightcurve of psr j1023 + 0038 from a 83-ks observation obtained on 2010 march 24 with _
acis - s ( obsid 11075 ) , when the binary was in an accretion disk - free state ( see * ? ? ?
* for further details ) , is shown in the bottom panel of figure 1 .
the close similarities between the lightcurves both in morphology and phase alignment are quite striking , further indication that the two systems are close analogs .
lccc * pl * & & & + @xmath24 ( @xmath25 @xmath26 ) & @xmath27 & @xmath28 & @xmath29 + @xmath30 & @xmath31 & @xmath32 & @xmath33 + @xmath34 ( @xmath35 kev ) & @xmath36 & @xmath37 & @xmath38 + @xmath39/dof & @xmath40 & @xmath41 & @xmath42 + * pl + nsa * & & & + @xmath24 ( @xmath25 @xmath26 ) & @xmath43 & @xmath44 & @xmath45 + @xmath30 & @xmath46 & @xmath47 & @xmath48 + @xmath49 ( @xmath50 k ) & @xmath51 & @xmath52 & @xmath53 + @xmath54 ( km ) & @xmath55 & @xmath56 & @xmath57 + thermal fraction & @xmath58 & @xmath59 & @xmath60 + @xmath34 ( @xmath35 kev ) & @xmath61 & @xmath37 & @xmath62 + @xmath63dof & @xmath64 & @xmath65 & @xmath66 ( _ middle _ ) and @xmath67 ( _ bottom _ ) , fitted with an absorbed power - law .
the smaller panels show the best - fit residuals in terms of @xmath22 with error bars of size one . for the best fit parameters ,
see text and table 1.,title="fig:",scaledwidth=45.0% ] + ( _ middle _ ) and @xmath67 ( _ bottom _ ) , fitted with an absorbed power - law .
the smaller panels show the best - fit residuals in terms of @xmath22 with error bars of size one .
for the best fit parameters , see text and table 1.,title="fig:",scaledwidth=45.0% ] + ( _ middle _ ) and @xmath67 ( _ bottom _ ) , fitted with an absorbed power - law .
the smaller panels show the best - fit residuals in terms of @xmath22 with error bars of size one . for the best fit parameters , see text and table 1.,title="fig:",scaledwidth=45.0% ]
the spectroscopic analysis was carried out in xspec 12.7.1 @xcite .
preliminary fits of the individual _ xmm - newton _ and _ chandra _ spectra produced consistent results so all fits were performed jointly on both data sets to obtain the best constraints on the x - ray spectral properties of xss j122704859 .
fitting the phase - averaged spectra with an absorbed power - law results in a statistically acceptable fit ( see table 1 and top panel of figure 2 ) , with @xmath68 for 216 d.o.f . , while a single or even a two - temperature neutron star atmosphere thermal spectrum does not yield a satisfactory fit , with @xmath69 for 216 d.o.f . and @xmath70 for 214 d.o.f .
, respectively .
since xss j122704859 currently hosts an active rotation - powered pulsar , x - ray - emitting magnetic polar caps heated by a back flow of particles from the pulsar magnetosphere may be present , as seen in many radio msps ( see , e.g. , * ? ? ? * ; * ? ? ?
therefore , in addition to the dominant non - thermal emission there may be a non - negligible contribution from the pulsar polar caps . including the non - magnetic neutron star hydrogen atmosphere nsa model @xcite component results in an acceptable fit but the parameters of the thermal component are poorly constrained .
the thermal flux contribution is consistent with zero even at a 1@xmath22 confidence level and the upper limit on the thermal fraction is @xmath719% . for both the power - law and composite models , the derived hydrogen column densities along the line of sight , @xmath24 , are in agreement with the value measured through the galaxy of @xmath6@xmath72 @xmath26 in the direction of xss j122704859 @xcite .
assuming the distance range of @xmath73 kpc @xcite , the derived time - averaged unabsorbed fluxes imply an x - ray luminosity of @xmath74 erg s@xmath8 in the 0.110 kev interval , an order of magnitude fainter than the average luminosity in the accretion disk dominated state .
is centered on the position of xss j122704859 .
the 95% confidence error ellipse of 2fgl j1227.74853 is shown with the dashed blue line .
the grayscale shows intensity increasing logarithmically from white to black .
north is up and east is to the left.,scaledwidth=47.0% ] to look for changes in the x - ray spectrum as a function of the binary orbital phase , we divided the data in two equal phase portions : @xmath75 ( encompassing the x - ray minimum ) and @xmath67 based on the orbital ephemeris from @xcite .
the spectral fits are summarized in columns 3 and 4 of table 1 .
an absorbed power - law model yields good fits for both phase intervals ( figure 2 ) with similar values for the spectral photon index @xmath30 , implying no appreciable spectral variability over the orbit despite the factor of 1.51.7 difference in flux . as with the phase - averaged analysis
, the addition of an nsa model component also results in acceptable fits . for both orbital phase intervals
any thermal emission contributes @xmath7610% to the @xmath35 kev flux .
the nsa temperature and emission radii are poorly constrained but are in general agreement with those observed for thermally - emitting msps ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
even in its current low state , xss j122704859 still remains the brightest x - ray source in its field .
the substantially lower flux allows a close examination of the spatial distibution of emission around the binary for the first time . for xss j122704859 ,
this is especially interesting since a currently active rotation - powered pulsar wind may , in principle interact with circum - binary material that was expelled during the lmxb state via outflows .
moreover , the resulting wind nebula may only be visible for a brief period after the state transition since presumably the material would eventually be swept up by the pulsar wind . in the most recent images , a diffuse source situated @xmath61.5@xmath0 east - north - east of xss j122704859 is clearly visible ( figures 3 and 4 ) , which can only barely be seen in the _ xmm - newton _
epic mos1 + 2 image from the high state ( 2011 january 1 ) due to the much brighter wings of the point spread function of xss j122704859 .
the centroid of the diffuse source , which is @xmath6@xmath77@xmath16 in extent , is at approximately ra=12@xmath7828@xmath1107@xmath155 and dec=@xmath1348@xmath1453@xmath024@xmath9 ( j2000 ) . to establish if the two objects are associated , we looked for emission connecting xss j122704859 with the extended region by comparing the count rates around xss j122704859 in the direction towards and away from the diffuse source .
this was accomplished by considering a set of adjacent 8@xmath9 by 16@xmath9 rectangular regions aligned along the axis connecting the centroid of the diffuse source and xss j122704859 ( see figure 4 ) .
the size was chosen so as to ensure that a sufficient number of photons are collected in each region . comparing
the distribution of counts towards and away from the diffuse region ( figure 5 ) shows that the emission from the diffuse source does not extend to the position of xss j122704859 , suggesting no physical connection between the two x - ray sources .
in addition , if the two objects are in fact associated , the angular separation of @xmath61.5@xmath0 would translate to a physical separation of @xmath79 pc , and the size of the diffuse region to @xmath80 pc , for an assumed distance range of @xmath73 kpc .
these physical scales are much larger than what is observed in other accreting systems with outflows , such as the microquasars xte j1550564 @xcite and h1743322 @xcite .
therefore , it is unlikely that the extended emission is a product of an outflow from the binary .
likewise , a pulsar wind nebula ( pwn ) can be ruled out because xss j122704859 is not embedded in the nebular emission , which also does not resemble a bow shock or a trailing tail ( see , e.g. , * ? ? ? * for an overview of x - ray pwne ) .
these factors indicate that an association between the binary and the diffuse source is highly unlikely .
it is more likely that the extended object is simply a chance superposition and is unrelated to the binary .
examination of vlt / fors2 images ( @xmath81s exposure in @xmath82-band obtained on 2010 february 10 ) show a bright source with morphology resembling that of a galaxy located at ra@xmath8312@xmath1028@xmath1107@xmath12427 , dec.@xmath83@xmath1348@xmath1453@xmath023@xmath15561 ( see bottom panel of figure 4 ) , which is coincident with the centroid of the diffuse x - ray region and is surrounded by several fainter objects .
therefore , this is likely the central galaxy of a previously uncatalogued galaxy cluster , which we designate as j122807.448532 , with the diffuse x - ray emission originating from the intra - cluster medium @xcite .
additional optical observations are required to conclusively establish this scenario . .
the green @xmath84 marks the position of xss j122704859 , the green diamond is centered on the optically identified galaxy ( see text ) , and the green circles mark the atca radio sources 1 and 3 reported by hill et al .
the 95% confidence error ellipse of 2fgl j1227.74853 is shown with the dashed red line .
the grayscale shows intensity increasing logarithmically from white to black .
_ middle _ : the same image binned at the intrinsic detector resolution of 0.5@xmath9 .
the red contour levels were obtained from the smoothed image in the top panel .
the blue strip of rectangles shows the regions used to extract the count rates shown in figure 5 .
pixel randomization has been removed from the pipeline processing .
_ bottom _ : vlt / fors2 image of the field around xss j122704859 with the same x - ray contours shown in the middle panel . note the extended optical source near the middle of the diffuse x - ray region surrounded by a number of fainter sources.,title="fig:",scaledwidth=46.0% ] .
the green @xmath84 marks the position of xss j122704859 , the green diamond is centered on the optically identified galaxy ( see text ) , and the green circles mark the atca radio sources 1 and 3 reported by hill et al .
the 95% confidence error ellipse of 2fgl j1227.74853 is shown with the dashed red line .
the grayscale shows intensity increasing logarithmically from white to black .
_ middle _ : the same image binned at the intrinsic detector resolution of 0.5@xmath9 .
the red contour levels were obtained from the smoothed image in the top panel .
the blue strip of rectangles shows the regions used to extract the count rates shown in figure 5 .
pixel randomization has been removed from the pipeline processing . _
bottom _ : vlt / fors2 image of the field around xss j122704859 with the same x - ray contours shown in the middle panel .
note the extended optical source near the middle of the diffuse x - ray region surrounded by a number of fainter sources.,title="fig:",scaledwidth=45.5% ] the 0.5@xmath9 angular resolution of the acis - s image enables an investigation of any nebular emission in the immediate vicinity of the binary ( within @xmath620@xmath9 ) .
for this purpose , we have simulated 10 observations of xss j122704859 using chart and marx 4.5 assuming the parameters of the acis - s exposure and the best - fit non - thermal model reported in 4 .
the mean of the simulated observations was subtracted from the real data to look for excess emission around the binary . however , as apparent from figure 6 , the radial profile of the x - ray emission from xss j122704859 is fully consistent with that of a point source .
we note that this diffuse x - ray source has no radio counterpart in the atca image presented by @xcite . on the other hand , the eastern component of the radio source designated by @xcite as 1 , as well as their source 3 , are coincindent with very faint x - ray sources ( marked in figure 4 ) .
since these sources are at the threshold of detection in the _ chandra _ image , their spectral properties can not be well constrained .
as reported in bassa et al .
( 2014 ) , the recent decline in optical and x - ray brightness and the disappearance of the previously prominent optical emission lines are indicative of the disappearance of the accretion disk in xss j122704859 .
the _ xmm - newton _ and _ chandra _ observations of xss j122704859 in its new low flux state presented herein reveal further information regarding this binary : these x - ray properties , as well as the optical properties described in @xcite , are identical to what is observed from `` redback '' radio msps , including psrs j1023 + 0038 @xcite , j17232837 @xcite , j2215 + 5135 @xcite in the field of the galaxy , as well as psr j00247204w in the globular cluster 47 tuc @xcite , psr j17405340 in ngc 6397 @xcite , and psr j18242452h in m28 @xcite .
therefore , the results obtained herein , and the recent radio pulsar discovery by @xcite further demonstrate that xss j122704859 is a redback , i.e. a compact binary with an active rotation - powered pulsar and no accretion disk .
the non - thermal x - rays from redback systems are generated in an intra - binary shock formed by the interaction of the pulsar wind with material from the close companion star @xcite .
as shown in @xcite , the decline in x - ray flux at @xmath87 seen in psr j1023 + 0038 can be reproduced by a simple geometric model in which the intra - binary shock is partially occulted by the bloated secondary star . in this scenario ,
the depth and phase extent of the x - ray eclipse require that the shock be situated at or very near the surface of the secondary star on the side exposed to the pulsar wind .
there is no indication of extended x - ray emission associated with xss j122704859 that could arise due to outflows in the lmxb phase or the interaction of the pulsar wind with the ambient medium .
an adjacent diffuse source is likely not associated with the binary and is probably a background galaxy cluster .
given that only two msps , psrs b1957 + 21 @xcite and j21243358 @xcite show evidence for x - ray bow shocks , the absence of nebular emission associated with xss j122704859 does not imply the lack of a pulsar wind . the lack of knowledge of the orbital separation , the companion mass , and the pulsar spin - down luminosity ( @xmath88 ) for xss j122704859 , do not allow meaningful constraints on the physics of the intrabinary shock .
nevertheless , given the nearly identical x - ray properties compared to other redbacks it is possible to place crude constraints on the energetics of the msp in xss j122704859 .
in particular , if we consider that the shock luminosity at x - ray maximum is @xmath6@xmath89 for psrs j1023 + 0038 @xcite and j17232837 @xcite , the implied spin - down of the pulsar in xss j122704859 is @xmath90 erg s@xmath8 assuming a distance range @xmath73 kpc .
this suggests that the pulsar in xss j122704859 may be among the small number of energetic msps with @xmath91 erg s@xmath8 .
the 0.1100 gev @xmath1-ray luminosity of 2fgl j1227.74853 , the putative @xmath1-ray counterpart of xss j122704859 , of @xmath92 erg s@xmath8 provides an independent hard lower limit on the @xmath88 of the pulsar , if we assume a @xmath1-ray conversion efficiency of 100% . for efficiencies typical of the msp population ( @xmath7650%
; see table 10 in * ? ? ?
* ) ) , this yields @xmath91 erg s@xmath8 , in general agreement with the limit obtained from the x - ray luminosity . continued timing of the radio pulsar would allow a determination of the true @xmath88 of this pulsar .
the redback nature of the xss j122704859 binary determined from the x - ray analysis presented above implies that the pulsar should undergo radio eclipses of frequency - dependent duration around superior conjunction .
these eclipses are caused by intra - binary plasma , presumably emanating from the companion star .
for most redback and black widow systems that are detectable in radio pulsations , the radio emission is eclipsed for @xmath7650% of the orbit , although there is indication that for some systems this fraction can be close to 100% ( see * ? ? ?
* for the case of psr j13113430 ) .
the non - detections with the parkes telescope @xcite and only a single , brief detection with the gmrt @xcite despite deep pulsation searches , are indicative of a high radio eclipse fraction for xss j122704859 that likely varies from orbit to orbit .
the large @xmath88 estimated above is consistent with this scenario , since an energetic wind would drive off material from the surface of companion at a higher rate than an msp with a lower @xmath88 .
this may result in occasional severe enshrouding by this stripped material , causing the pulsar to be eclipsed at radio frequencies for a larger fraction of the orbit .
this would make xss j122704859 a so - called `` hidden '' msp , as postulated by @xcite . the x - ray properties of xss j122704859 in both the present disk - free radio pulsar state and the accretion disk - dominated state ( see * ? ? ?
* ; * ? ? ?
* ) are virtually identical to those of psrs j1023 + 0038 @xcite and j18242452i @xcite .
this remarkable consistency in x - ray characteristics provides a powerful discriminant for identifying more such msp binaries , especially ones that are heavily enshrouded and hence can not be identified via radio pulsation searches .
a.m.a . and j.w.t.h .
acknowledge support from a vrije competitie grant from nwo .
c.b . acknowledges support from erc advanced grant `` leap '' ( 227947 , pi : michael kramer ) .
j.w.t.h . and a.p .
acknowledge support from nwo vidi grants .
j.w.t.h . also acknowledges funding from an erc starting grant `` dragnet '' ( 337062 ) .
a portion of the results presented was based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and nasa .
the scientific results reported in this article are based in part on observations made by the _ chandra x - ray observatory_. this research has made use of the nasa astrophysics data system ( ads ) , software provided by the chandra x - ray center ( cxc ) in the application package ciao .
hessels , j. w. t. , roberts , m. s. e. , mclaughlin , m. a. , et al .
2011 , in aip conf .
1357 , radio pulsars : an astrophysical key to unlock the secrets of the universe , ed .
m. burgay et al .
( melville , ny : aip ) , 40 roberts , m. s. e. 2011 , in aip conf . ser .
1357 , radio pulsars : an astrophysical key to unlock the secrets of the universe , ed .
m. burgay , n. damico , p. esposito , a. pellizzoni , & a. possenti ( melville , ny : aip ) , 127 | we present _ xmm - newton _ and _ chandra _ observations of the low - mass x - ray binary xss j122704859 , which experienced a dramatic decline in optical / x - ray brightness at the end of 2012 , indicative of the disappearance of its accretion disk . in this new state ,
the system exhibits previously absent orbital - phase - dependent , large - amplitude x - ray modulations with a decline in flux at superior conjunction .
the x - ray emission remains predominantly non - thermal but with an order of magnitude lower mean luminosity and significantly harder spectrum relative to the previous high flux state .
this phenomenology is identical to the behavior of the radio millisecond pulsar binary psr j1023 + 0038 in the absence of an accretion disk , where the x - ray emission is produced in an intra - binary shock driven by the pulsar wind .
this further demonstrates that xss j122704859 no longer has an accretion disk and has transformed to a full - fledged eclipsing `` redback '' system that hosts an active rotation - powered millisecond pulsar .
there is no evidence for diffuse x - ray emission associated with the binary that may arise due to outflows or a wind nebula .
an extended source situated 1.5@xmath0 from xss j122704859 is unlikely to be associated , and is probably a previously uncatalogued galaxy cluster . |
in contrast to electrodynamics , the dynamics of yang - mills fields is inherently nonlinear and chaotic at any density of energy . this assumption was confirmed analytically and numerically @xcite .
further analysis of spatially homogeneous field configurations @xcite showed that inclusion of higgs field leads to order - chaos transition at some density of energy of classical gauge fields @xcite .
classical higgs field regularizes chaotic dynamics of classical gauge fields below critical energy density and leads to the emergence of order - chaos transition .
chaos in yang - mills fields @xcite and vacuum state instability in nonperturbative qcd models @xcite are also considered in connection with confinement .
it has been also shown recently that interaction of the constant chromo - magnetic field with axial field could generate confinement @xcite .
these results indicate the importance of nonperturbative background fields . in our previous paper @xcite we have investigated the stability of yang - mills - higgs fields and described analytically the regions of chaotic and stable motion . in this work yang - mills fields
are considered on the background of the homogeneous ( anti- ) self - dual field @xcite . as the dynamics of arbitrary field configurations is too complicated ,
we follow @xcite and reduce our model to spatially homogeneous fields which depend on time . after that we are left with only finite number of degrees of freedom ( two in our case ) which allows us to investigate the dynamics of the system using conventional methods developed for mechanical systems . in this work
one more mechanism of the chaos suppression in the yang - mills fields models is proposed .
homogeneous ( anti-)self - dual field eliminates chaoticity of yang - mills dynamics below critical energy density .
in this paper classical dynamics of su(2 ) model gauge fields system is considered on the background of the homogeneous ( anti-)self - dual field .
various properties of this solution of the yang - mills equations in su(2 ) theory were investigated originally by other authors @xcite .
it was demonstrated that self - dual homogeneous field provides the wilson confinement criterion @xcite .
therefore this field is at least a possible source of confinement in qcd if it is a dominant configuration in the qcd functional integral .
homogeneous self - dual field is defined by the following expressions @xcite : @xmath0 @xmath1 where @xmath2 - value of the field strength , vector @xmath3 and tensor @xmath4 characterize the direction of the field , respectively , in color space and in space - time .
the latter has the following properties @xmath5 @xmath6 where positive and negative signs in last expression correspond , respectively , to self - dual and anti - self - dual cases .
as the directions of the background field in color space and in space - time can be chosen arbitrarily , we will assume that the gauge field has color components @xmath7 and space - time components @xmath8 .
the lagrangian of su(2 ) gauge theory in euclidean metrics is @xmath9 where @xmath10 is a field tensor which has the following form : @xmath11 in last expression @xmath12 are the three non - abelian yang - mills fields and @xmath13 denotes the coupling constant of these fields .
we consider the fluctuations around background homogeneous self - dual field .
self - dual field is regarded as external one and it is taken into account by substituting modified vector potential in the yang - mills lagrangian @xmath14 where @xmath15 is the fluctuation to the background field @xmath16 .
we use the gauge : @xmath17 and consider spatially homogeneous field configurations @xcite @xmath18 our model of yang - mills fields in ( anti-)self - dual field is constructed in euclidean space . in order to analyze the model by using analytical and numerical methods we should switch to minkowski space .
we consider chromo - magnetic model .
thus we put chromo - electric field is equal to zero .
if @xmath19 , @xmath20 and the other components of the perturbative yang - mills fields are equal to zero , the potential of the model is : @xmath21 where @xmath22 - chromo - magnetic background field strength , @xmath23 and @xmath24 - coordinates which play the role of the parameters , @xmath25 and @xmath26 - field variables .
at first , stability of the model is investigated using well known technique based on the toda criterion of local instability @xcite which allows us to obtain the value of the critical energy density of order - chaos transition in the system .
this energy and minimum of the energy as the functions of the model parameter @xmath27 are shown on the fig.[grecvmin ] .
critical and minimal energies are close to each other for @xmath28 .
this behavior indicates the absence of the region of regular motion in the system . in other case
@xmath29 , the critical energy density is much larger than minimal one and the system is regular up to this energy .
these results will be checked using numerical methods in next subsection .
the system is investigated using poincare sections and lyapunov exponents for wide range of model parameter values .
these numerical methods could indicate global regular regimes of motion whereas toda criterion reveals only the local chaotic properties of the trajectories @xcite .
thus numerical methods are more precise for stability analysis .
results of the numerical calculations for the system with model parameter @xmath30 are shown on the fig.[grps0 ] and fig.[grle0 ] .
contrary to toda criterion , system is regular at small energies below the energy of the background field @xmath31 .
there are only regular regimes of motion for small energies ( e.g. fig.[grps0]a and fig.[grle0]a ) and two types of motion for energies above @xmath32 ( fig.[grps0]b and fig.[grle0]b ) .
we have the following poincare sections ( fig.[grps25 ] ) for large model parameter values .
it is seen that system is fully regular ( fig.[grps25]a ) for high values of energy @xmath33 as it was revealed by toda criterion .
all trajectories have zero maximal lyapunov exponents for this case .
two types of trajectories ( chaotic and regular ) are present in the system with energies above the critical one ( fig.[grps25]b ) .
it is seen that toda criterion rather good describes the region of the large values of model parameter @xmath34 and fails for @xmath35 .
numerical calculations have shown that there is a region of regular motion at low densities of energy in our system at any value of the model parameter .
therefore , homogeneous ( anti-)self - dual field regularizes chaotic dynamics of yang - mills fields system .
the existence of the nonperturbative component of the yang - mills field is crucial for the confinement phenomenon . on the other hand , classical dynamics of yang - mills field is chaotic at any density of energy in the absence of background fields and can be regularized only if some other fields , for example , the higgs field , are included in the model . in this work
we have demonstrated that homogeneous ( anti-)self - dual background field has similar properties .
yang - mills field on such background has the region of regular motion at low densities of energy .
there is order - chaos transition in the system at any values of model parameters .
the critical density of energy of this transition is equal to the background field energy for small parameters and is much larger for large parameter values .
this work is supported by belarusian republican foundation for fundamental research . _
berman g. p. , mankov y. i. , sadreyev a. f. _ stochastic instability of classical homogeneous su(2)@xmath36u(1 ) fields with spontaneously broken symmetry //zh .
1985 . v. 88 , no . 3 .
p. 705714 .
_ efimov g. v. , kalloniatis a. c. , nedelko s. n. _ confining properties of the homogeneous self - dual field and the effective potential in su(2 ) yang - mills theory //phys.rev .
p. 014026
, hep - th/9806165 . | stability of yang - mills fields system in the background field is investigated basing on toda criterion , poincare sections and the values of the maximal lyapunov exponents .
the existence of the region of regular motion at low densities of energy is demonstrated .
critical energy density of the order - chaos transition is analyzed for the different values of the model parameter .
* stability of yang - mills fields system + in the homogeneous ( anti-)self - dual background field * + + joint institute for power and nuclear research , minsk , belarus |
due to the central limit theorem a great deal of phenomena can be described by gaussian statistics .
this also guides our perception of the risks of large deviations from an expectation value .
consequently , the occurence of any aggravated probability of extreme events is always cause for concern and subject of intense research interest . in a large variety of systems
where heavy - tailed distributions are observed , gaussian statistics holds only locally the parameters of the distribution are changing , either in time or in space .
thus , to describe the sample statistics for the whole system , one has to average the parametric distribution over the distribution of the ( shape ) parameter .
this construction is known as _ compounding _ or _ mixture _ @xcite in the mathematics and as _ superstatistics _
@xcite in the physics literature .
an important example for parameter distribution functions is the k - distribution mentioned 1978 for the first time by jakeman and pusey @xcite .
it was introduced in the context of intensity distributions , and their significance for scattering processes of a wide range of length scales was stressed .
moreover the distribution is known to be an equilibrium solution for the population in a simple birth - death - immigration process which was already applied in the description of eddy evolution in a turbulent medium .
the underlying picture of turbulence assumes that large eddies are spontaneously created and then give birth to generations of children eddies , which terminates when the smallest eddies die out due to viscous dissipation . in @xcite jakeman and pusey
use the k - distribution for fitting data of microwave sea echo , which turned out to be highly non - rayleigh .
the k - distribution is also found as a special case of a full statistical - mechanical formulation for non - gaussian compound markov process , developed in @xcite .
field and tough find k - distributed noise for the diffusion process in electromagnetic scattering @xcite .
experimentally the k - distribution appeared in the contexts of irradiance fluctuations of a multipass laser beam propagating through atmospheric turbulence @xcite , synthetic aperture radar data @xcite , ultrasonic scattering from tissues @xcite and mesoscopic systems @xcite .
also in our study we will encounter the k - distribution for one of the systems under consideration .
compounded distributions can be applied to very different empirical situations : they can describe aggregated statistics for many time series , where each time series obeys stationary gaussian statistics , the parameters of which vary only between time series . in this case
it is straightforward to estimate the parameter distribution .
the situation is more difficult when we consider the statistics of single long time series with time - varying parameters . in this non - stationary case ,
one often makes an ad hoc assumption about the analytical form of the parameter distribution , and only the compounded distribution is compared to empirical findings . in this paper
we address the problem of determining the parameter distribution empirically for univariate non - stationary time series .
specifically we consider the case of gaussian statistics with time - varying variance . in this endeavor
we encounter several problems : if the variance for each time point is purely random , as the compouding ansatz would suggest , we have no way of determining the variance distribution from empirical data . a prerequisite for an empirical approach to the parameter distribution
is a time series which is quasi - stationary on short time intervals . in other words ,
the variance should vary only slowly compared to the time scale of fluctuations in the signal .
the estimation noise for the local variances competes with the variance distribution itself .
therefore the time interval on which quasi - stationarity holds , should not be too short .
furthermore , we have to heed possible autocorrelations in the time series themselves , since they might lead to an estimation bias for the local variances .
our aim is to test the validity of the compounding approach on two different data sets .
the paper is organized as follows : in section [ sec2 ] we give a short summary of the compounding approach and present two recent applications where the k - distribution comes into play . in section [ sec3 ]
we introduce the two systems we are going to analyse , a table top experiment on air turbulence and the empirical time series of exchange rates between us dollar and euro . in section [ sec4 ] we address the problem of estimating non - stationary variances in univariate time series .
our empirical results are presented in section [ sec5 ] .
we consider a distribution @xmath0 of @xmath1 random variables , ordered in the vector @xmath2 .
it is also a function of a parameter @xmath3 that determines the shape or other features of the distribution , e.g. the variance of a gaussian .
if , in a given data set , the parameter @xmath3 varies in an interval @xmath4 , one can try to construct the distribution of @xmath2 as the linear superposition @xmath5 of all distributions @xmath0 with @xmath6 . here , @xmath7 is the weight function determining the contribution of each value of @xmath3 in the superposition .
since @xmath3 itself typically is a random variable , we assume that the function @xmath7 is a proper distribution . in particular , it is positive semidefinite . as each @xmath0 and
the resulting @xmath8 have to be normalized with respect to the random vector @xmath2 , eq .
( [ comp1 ] ) implies the normalization @xmath9 the physics reasons for the variation of the parameter @xmath3 can be very different . in non
equilibrium thermodynamics , @xmath3 might be the locally fluctuating temperature .
although our systems are not of a thermodynamic kind , we also have in mind non stationarities . in recent experiments
@xcite we studied the propagation of microwaves through an arrangement of disordered scatterers in a cavity .
the distribution of the electric fields was measured at fixed frequencies as a function of position .
then time - dependent wave fields were generated by superposition of @xmath10 patterns , @xmath11 here @xmath12 is the wave pattern at frequency @xmath13 , and @xmath14 is a random phase . for fixed positions ,
always a rayleigh distribution was found in the time sequence @xmath15 for the distribution of intensities @xmath16 , @xmath17 this is nothing but a manifestation of the central limit theorem . the variance , _
i.e. _ , the averaged @xmath18 depends on the position .
the large amount of data made it possible to extract the distribution of the parameter @xmath18 .
in good approximation , it turned out to be a @xmath19 distribution , see fig .
[ fig : freak ] . the number @xmath20 of degrees of freedom was related to the number of independent field components and took a value of @xmath21 .
the authors of refs .
@xcite then used the compounding ansatz eq .
( [ comp1 ] ) in the form @xmath22 the integral can be done and yields @xmath23 where @xmath24 is the modified bessel function of degree @xmath25 .
this is the k - distribution introduced in the introduction . omitting local regions with extremely high amplitudes ,
so called `` hot spots '' , the intensity distributions could be perfectly well interpreted in terms of k - distributions , see fig .
[ fig : freak ] . found for the 780 pixels of our measurement .
the inset shows the same data using a semi - logarithmic scale .
the solid curve is a @xmath26 distribution with @xmath27 degrees of freedom .
( bottom ) intensity distribution for the time - dependent wave patterns generated by eq .
( [ eq : pulse ] ) .
the solid ( red ) line is given by eq .
( [ eq : kbess]).,title="fig:",scaledwidth=46.0% ] + found for the 780 pixels of our measurement .
the inset shows the same data using a semi - logarithmic scale .
the solid curve is a @xmath26 distribution with @xmath27 degrees of freedom .
( bottom ) intensity distribution for the time - dependent wave patterns generated by eq .
( [ eq : pulse ] ) .
the solid ( red ) line is given by eq .
( [ eq : kbess]).,title="fig:",scaledwidth=46.0% ] we briefly sketch another example which stems from finance .
details can be found in ref .
the conceptual difference to the previous example is that the compounding formula ( [ comp1 ] ) appears as an intermediate result , not as an ansatz .
we consider a selection of @xmath28 stocks with prices @xmath29 belonging to the same market .
one is interested in the distribution of the relative price changes over a fixed time interval @xmath30 , also referred to as returns @xmath31 as the companies operate on the same market , there are correlations between the stocks which have to be measured over a time window @xmath32 much larger than @xmath30 .
the corresponding covariances form the @xmath33 matrix @xmath34 . the business relations between the companies as well as the market expectations of
the stock market traders change in time .
thus , the market is non stationary and the correlation coefficients fluctuate in time . only for time windows @xmath32 of a month or less
, @xmath34 is approximately constant .
the multivariate distribution of the returns @xmath35 ordered in the @xmath28 component vector @xmath36 at a given time @xmath37 is well described by the gaussian @xmath38 the non stationarity over longer time windows @xmath32 can be modeled by observing that the ensemble of the fluctuating covariance matrices @xmath34 can be approximated by an ensemble of random matrices . in ref .
@xcite a wishart distribution was assumed for this ensemble .
the ensemble average of the distribution ( [ multivar ] ) over the wishart ensemble yields @xmath39 where @xmath40 is the sample - averaged covariance matrix over the entire time window @xmath32 .
as @xmath40 is fixed , this result has the form of the compounding ansatz ( [ comp1 ] ) .
furthermore , it closely matches the result ( [ comp3 ] ) found in the context of microwave scattering .
the number @xmath41 of degrees of freedom in the @xmath42 distribution determines the variance in the distribution of the random covariance matrices .
the role of the locally averaged intensity @xmath18 is now played by an effective parameter @xmath43 which fully accounts for the ensemble average . again , a k - distribution follows , @xmath44 in which the bilinear form @xmath45 takes the place of the intensity @xmath16 in eq .
( [ eq : kbess ] ) . in both examples ,
averages over fluctuating quantities produce heavy - tailed distributions which describe the vast majority of the large events .
the first data set is obtained by measuring the noise generated by a turbulent air flow . for the turbulence generation we used a standard fan with a rotor frequency of 18.44hz .
we restricted ourselves to standard audio technique handling frequencies up to 20khz and standard sampling rates of 48khz offering reliable quality at an attractive price .
the microphone for the sound recording is a e 614 by sennheiser with a frequency response of 40hz - 20khz , a good directional characteristic and a small diameter of 20 mm .
it guarantees a broadband frequency resolution and a point - like measuring position .
an external sound card with matching properties was necessary to use the full capacity of the quality of microphone and to minimize the influence of the intrinsic noise of the pc .
[ fig : sound ] shows a photograph or the used setup .
a microphone has been placed in front of a fan running continuously and generating a highly turbulent air flow .
the microphone records the time signal of the sound waves excited by the turbulence .
the details about the analysed time series will be discussed in section [ sec4 ] .
the foreign exchange markets have the peculiar feature of all - day continuous trading .
this is in contrast to stock markets ,
where the trading hours of different stock exchanges vary due to time zones , with partial overlap of different markets and very peculiar trading behavior at the beginning and the end of each trading day . therefore foreign exchange rates are particularly suited for the study of long time series .
we consider the time series of hourly exchange rates between euro and us dollar in the time period from january 2001 to may 2013 .
the empirical data were obtained from ` www.fxhistoricaldata.com ` .
we denote the time series of exchange rates by @xmath46 . from these
we calculate the time series of returns , i.e. , the relative changes in the exchange rates on time intervals @xmath30 , @xmath47 since we work with hourly data , the smallest possible value for @xmath30 is one hour .
however , as we will see later on , a return interval of one trading day , @xmath48 , is preferable for the variance estimation .
note that foreign exchange rates are typically modeled by a multiplicative random process , such as a geometric brownian motion , see , _
e.g. _ , @xcite .
therefore we consider the relative changes of the exchange rates instead of the exchange rates themselves .
while the latter resemble at least locally - a lognormal distribution , the returns are approximately gaussian , conditioned on the local variance , that is .
we consider the problem of univariate time series with time - dependent variance . more specifically , we consider time series where the variance is changing , but exhibits a slowly decaying autocorrelation function .
this point is crucial , because otherwise it is not possible to make meaningful estimates of the local variances .
time series with this feature show extended periods of large fluctuations interupted by periods with moderate or small fluctuations .
this is illustrated in fig .
[ fig : signal ] for the two data sets we are studying in this paper . in the top plot of fig .
[ fig : signal ] , we show the sound signal for the ventilator measurement . in the bottom plot , the time series of daily returns for the foreign exchange data is plotted . in both cases
we observe the same qualitative behavior , which is well - known in the finance literature as volatility clustering .
the compounding ansatz for univariate time series assumes a normal distribution on short time horizons , where the local variance is nearly stationary .
however , we wish to determine the distribution of the local variances empirically , since it is a critical part in the compounding ansatz .
if the variances were fluctuating without a noticable time - lagged correlation , this would not be feasible .
still , we need to establish the right time horizon on which to estimate the local variances .
@xcite introduced a method to locally normalize time series with autocorrelated variances . to this end
, a local average was subtracted from the data and the result was divided by a local standard deviation . in this spirit , we determine the time horizon on which this local normalization yields normal distributed values and analyse the corresponding local variances . another aspect we need to take into account is a possible bias in the variance estimation which occurs for correlated events . in fig .
[ fig : acf ] we show the autocorrelation function ( acf ) of the measured sound signal , as well as the autocorrelation function of the absolute value of hourly returns . both plots hint at possible problems for the variance estimation . due to the high sampling frequency , the sound signal is highly correlated . in other words ,
the sampling time scale is much shorter than the time scale on which the turbulent air flow changes . after 2500 data points , or about 52 ms ,
the autocorrelation function has decayed to zero .
consequently , we consider only every 2500th data point for our local variance estimation . to improve statistics , we repeat the variance estimation starting with an offset of 1 to 2499 .
the results are presented in the following section . in the case of the foreign exchange data
we are confronted with a different problem .
the consecutive hourly returns are not correlated .
while local trends may always exist , it is unpredictable when a positive trend switches to a negative one , and _
vice versa _ , see ref .
@xcite . however , the autocorrelation of the absolute values shows a rich structure which is due to characteristic intraday variability .
this would lead to a biased variance estimation and , consequently , to a distortion of the variance distribution .
therefore we consider returns between consecutive trading days at the same hour of the day .
put differently , we consider @xmath49 for the returns and get 24 different time series , one for each hour of the day as starting point . + +
we first discuss the results for the turbulent air flow . as described in the previous section , we sliced the single measurement time series into 2500 time series with lower sampling rate , taking only every 2500th measurement point .
this is necessary to avoid a bias in the estimation of the local variances . before proceeding ,
each of these time series is globally normalized to mean zero and standard deviation one .
figure [ fig : venti ] shows the empirical results for the distribution of local variances and the compounded distribution .
the local variances are rather well described by a @xmath26 distribution with @xmath41 degrees of freedom .
we find @xmath50 to provide the best fit to the data .
the distribution of the empirical sound amplitudes is well described by a k - distribution with the same @xmath41 which fits the variance distribution .
hence , we arrive at a consistent picture , which supports our compounding ansatz for this measurement .
-distribution with @xmath41 degrees of freedom .
( bottom ) distribution of the sound amplitudes , compared to the k - distribution with parameter @xmath50 . , title="fig:",scaledwidth=46.0% ] + -distribution with @xmath41 degrees of freedom . ( bottom )
distribution of the sound amplitudes , compared to the k - distribution with parameter @xmath50 .
, title="fig:",scaledwidth=46.0% ] the results for the daily returns of eur - usd foreign exchange rates are shown in fig . [
fig : forex ] . as outlined in section [ sec4 ]
, we calculated the daily returns as the relative changes of the exchange rate between consecutive trading days with respect to the same hour of each day . this procedure yields 24 time series of daily returns .
we normalize each time series to mean zero and standard deviation one .
this allows us to produce a single aggregated statistics . in the top plot of fig .
[ fig : forex ] we show the histogram of local variances , i.e. the variances estimated on 13-day intervals . in accordance with the finance literature , the empirical variances follow a lognormal distribution over almost three orders of magnitude , with only some deviations in the tail .
the histogram of the daily returns is shown in the bottom plot of fig .
[ fig : forex ] .
the empirical result agrees rather well with the normal - lognormal compounded distribution .
it is important to note , however , that we only achieve this consistent picture of variance and compounded return distribution because we have taken into account all the pitfalls of variance estimation , which we described in section [ sec4 ] .
we applied the compounding approach to two different systems , a ventilator setup generating turbulent air flow and foreign exchange rates .
both systems are characterized by univariate time series with non - stationary variances .
our main objective was to empirically determine the distribution of variances and thus arrive at a consistent picture .
the estimation of variances from a single , non - stationary time series presents several pitfalls , which have to be taken into account carefully .
first of all , we have to avoid serial correlations in the signal itself .
these might otherwise lead to an estimation bias . for the sound measurement
, we had to reduce the sampling rate of the data to achieve this .
the foreign exchange data presented another obstacle for variance estimation : we observed a characteristic intraday variability which had to be taken into account .
last but not least , it is a prerequisite that the non - stationary variances are not purely stochastic , but exhibit a slowly decaying autocorrelation . otherwise we would not be able to determine a reasonable variance distribution for the compounding ansatz .
when we take all these aspects into account , we arrive at the correct variance distribution . in good approximation
we found a @xmath26 distribution in the case of ventilator turbulence , which leads to a k - distribution for the compounded statistics .
for the foreign exchange returns we observe lognormal distributed variances ; and the normal - lognormal compounded distribution fit the return histogram well .
a central assumption in the compounding ansatz is the stationarity of the variance distribution .
this assumption might not always be satisfied and lead to deviations from the compounded distribution .
the sound measurements have been supported by the deutsche forschungsgemeinschaft via the forschergruppe 760 `` scattering systems with complex dynamics '' .
j. s. lee , d. l. schuler , r. h. lang , and k. j. ranson , in _ geoscience and remote sensing symposium , 1994 .
surface and atmospheric remote sensing : technologies , data analysis and interpretation .
, international _ ( ieee , new york , 1994 ) , vol . 4 , pp . | a defining feature of non - stationary systems is the time dependence of their statistical parameters .
measured time series may exhibit gaussian statistics on short time horizons , due to the central limit theorem .
the sample statistics for long time horizons , however , averages over the time - dependent parameters . to model the long - term statistical behavior , we compound the local distribution with the distribution of its parameters . here
we consider two concrete , but diverse examples of such non - stationary systems , the turbulent air flow of a fan and a time series of foreign exchange rates .
our main focus is to empirically determine the appropriate parameter distribution for the compounding approach . to this end
we have to estimate the parameter distribution for univariate time series in a highly non - stationary situation . |
gamma ray bursts ( grbs ) are brief , intense flashes of high energy gamma rays originating at cosmological distances and often associated with subsequent radiation emitted at longer wavelengths from x - ray to radio waves on times scales of minutes to days after the initial gamma ray burst . in the standard model a typical long duration grb
is thought to be formed by the explosion of a compact source that generates an expanding relativistic fireball ( rees @xmath4 mszros 1992 ) .
if the central engine remains active for some time several expanding shells with different speeds ( different lorentz factors , @xmath5 ) can be produced .
the collisions between these shells power the @xmath6-ray prompt emission itself ( internal shocks ) , while the interactions of the relativistic flow with the surrounding medium ( external forward shocks ) generate the so - called afterglow emission that dominates at longer wavelengths and is more long - lived than the prompt emission ( mszros 2002 , piran 1999 ) . assuming that the shock - accelerated electrons producing the radiation have a power - law spectral energy distribution , the afterglow synchrotron emission is expected to exhibit a standard form of spectrum , with two characteristic break frequencies : the typical synchrotron frequency @xmath7 and the cooling frequency @xmath8 ( sari , piran @xmath4 narayan 1998 ) . when the forward shock is formed , a reverse shock that propagates backwards into the ejecta is also generated .
the brightness of the reverse shock emission decays very rapidly compared to the decay of the forward component .
it is predicted that at early time the reverse shock can produce extremely bright optical flashes while at late time the optical flux is completely dominated by the forward shock emission ( mszros @xmath4 rees 1997 , sari @xmath4 piran 1999 ) . in reality , the resultant light curve is a complex , time - dependent mixture of these components and unravelling them provides important insight into the physics and energetics of the explosion .
the study of pre-_swift _ bursts allowed the understanding of late time multiband properties of the afterglows confirming many predictions of the fireball model .
well - sampled pre-_swift _ optical light curves were mostly obtained at late times , typically after few hours from the burst event , and exhibited relatively smooth light curves with simple power law decays , showing breaks at late time ( hints of a jet evolution ) and making clear a strong connection with supernova emission and , thus , the death of massive stars ( e.g. woosley @xmath4 bloom 2006 , malesani et al .
2004 , stanek et al . 2003 ) . the advent of the _ swift _ satellite ( gehrels et al .
2004 ) has opened up a new observational window at early times , revealing more complex light curve behaviour than previously known .
it is now accepted that the x - ray temporal decay of grbs observed by _
swift _ is well described by a canonical light curve ( nousek et al .
2006 ; zhang et al .
2006 ) ; combining the @xmath6-ray ( burst alert telescope , bat , * ? ? ?
* ) and x - ray ( x - ray telescope , xrt , burrows et al .
2005 ) data the initial x - ray emission ( rapid fall - off for the first hundred seconds ) is consistent with the tail of the @xmath6-ray emission ( tagliaferri et al .
2005 , obrien et al .
2006 ) and can be modelled by two components that have exactly the same functional form ( willingale et al .
these functions are completely empirical and do not do not provide a physical explanation for the x - ray flares seen in many bursts , but the majority of _ swift _ bursts seem to follow this behaviour .
it was clear that rapid response to obtain early - time optical observations with ground based robotic telescopes was required to pin down the open issue of the emission mechanism for the grb itself and its afterglow .
however , the small number of prompt optical observations simultaneous to the grb @xmath6-ray emission do not yet allow a firm conclusion to be drawn about the grb emission models ( yost et al .
the statistics of these events remain small due to the fact that the grbs detected by _ swift _
satellite are fainter and located at higher redshift ( @xmath9 , le @xmath4 dermer 2007 , jakobsson et al .
2006 ) than those detected by previous missions .
large 2-m robotic telescopes , such as the liverpool and faulkes ( north and south ) telescopes , responding within 1 - 3 minutes of the burst offer a unique tool for probing early - time light curves over a wide range of brightnesses , allowing the extension of the analysis of grb properties to unprecedented depth and time coverage . based on the theoretical predictions of the forward and reverse shock emission theories , optical light curves at early times should show different shapes depending on the relative contribution of the two components ( kobayashi @xmath4 zhang 2003 , zhang , kobayashi @xmath4 mszros 2003 ) ; possible light curve shapes are illustrated in figure [ figc1 ] .
in particular , if the optical observations start when the reverse shock component still dominates the optical flux , the shape of the optical light curve will appear as case 1 or case 2 in figure [ figc1 ] . in the first case
the light curve will show a transition from steep to shallow power law decay index . in the second case
it will show a re - brightening , but if observations do not start early enough the first steep segment ( thick dashed line on figure [ figc1 ] ) of the light curve should be missing and it will be visible only in the rise of the forward component .
the observed peak in case 2 will correspond to the passage of the @xmath7 in the observing band .
if there is energy injection to the forward shock emission then the light curve should appear as case 3 in figure [ figc1 ] .
this behaviour can be explained by long - lived central engine activity or large dispersion in the distribution of lorentz factors ( zhang et al .
2006 ) or related to the time when the energy is transferred from the fireball to the ambient medium ( kobayashi @xmath4 zhang 2007 ) . within this theoretical framework , we present the analysis of a sample of 63 grbs observed with the network of three 2-m telescopes _ robonet-1.0 _
( gomboc at al .
2006 ) , formed by the liverpool telescope ( lt , la palma , canary islands ) , the faulkes telescope north ( ftn , haleakala , hawaii ) and the faulkes telescope south ( fts , siding spring , australia ) .
for those grbs with detected optical counterparst , we discuss their light curve properties , compare optical and x - ray data and analyse the intrinsic rest frame properties of those bursts with known spectroscopic redshift .
an analysis of the bursts for which no afterglow was detected is also presented and discussed within the standard fireball model . throughout
we use the following conventions : the power law flux is given as @xmath10 , where @xmath11 is the temporal decay index and @xmath12 is the spectral slope ; we assume a standard cosmology with @xmath13 km s@xmath14 mpc@xmath14 , @xmath15 and @xmath16 ; all errors and uncertainties are quoted at the @xmath17 confidence level .
all three telescopes enable rapid response ( the typical mean reaction time is @xmath18 minutes after the trigger ) and deep observations ( @xmath19 21 at @xmath20 minutes after the trigger ) to grb alerts , which are crucial in the case of faint or optically dark bursts .
each telescope operates in a fully robotic mode , responding automatically to a grb satellite alert by immediately over - riding the current observing programme , then obtaining , analysing and interpreting optical images of the grb field using the specially designed , sophisticated pipeline ( lt - trap ) ; subsequent robotic followup observations are then optimised and driven by the automatically derived properties of the afterglow ( see guidorzi et al .
up to september 2007 the network robotically reacted to 63 grbs with 24 optical afterglow detections and 39 upper limits ( figure [ figall ] ) . in this paper ,
the optical photometry of each optical afterglow was performed using starlink / gaia photometry tools .
each field of view was calibrated with the best data available : 1 ) standards stars observed during the night , if the night was photometric ; 2 ) pre - burst calibration fields ( both for sdss or bessell filters ) reported in gcn circulars ; 3 ) differential photometry with catalogued stars in the field of view ( usnob1.0 catalogue ) .
this procedure has been followed for all the sample data .
all the calibrated magnitudes in the sdss - r filter ( for all the bursts observed with the lt telescope ) were then transformed to the @xmath21 band using the filter transformations given by smith et al .
( 2002 ) ; the colour term @xmath22 was then derived for the selected stars used to calibrate the field and finally applied to the estimated magnitude .
data calibration is also discussed in guidorzi et al .
( 2005b , 2007 ) , monfardini et al .
( 2006 ) and mundell et al .
( 2007b ) .
the observed optical light curves of the detected afterglows in our sample are shown in figure [ figall ] ( left panel ) . in the right panel of the same figure
all the deep upper limits for the undetected afterglows are shown , using the same scale on the y - axis to emphasize the large range in brightness and time covered by our observations .
the individual light curves of the detected afterglows are then shown in figure [ figxopt ] , which shows @xmath0 band observations together with the corresponding _ swift _ x - ray light curves .
temporal and spectral properties are summarised in table 1 .
we give both temporal and spectral information for each burst , when available .
we fitted the data either with a single or with a broken power law ; the values reported in the table are the best fitting results .
the values reported for @xmath23 , @xmath24 and @xmath25 are retrieved from the literature . in the last two columns of the table we report the redshift ( @xmath26 ) and the derived isotropic gamma - ray energy ( @xmath27 ) of the burst .
for all the bursts detected by the _
swift _ satellite the value of @xmath27 reported in the table is taken from butler et al .
( 2007 ) . for the no-_swift
_ bursts we calculated @xmath27 assuming a standard cosmology as reported at the end of section 1 , with the following formula : @xmath28 , where @xmath29 is the luminosity distance and _ f _ is the gamma ray fluence of the burst .
some of the sample bursts have been discussed in previous dedicated papers , but are presented here for completeness . for each burst in our sample
we summarise its key properties , together with references to more detailed work where relevant : * grb 041006 : this hete burst was observed with chandra between 16.8 and 42.5 hours after the burst event , showing a single temporal power law decay , whereas in the optical a break in the light curve at early time was clearly visible . at late time ( @xmath30 s ) the contribution of the underlying supernova emerged ( stanek et al .
2005 , soderberg et al .
the spectral indices in the two bands are consistent , with a slope @xmath31 ( butler at al .
2005 ; garnavich et al .
chandra x - ray data are not shown in our plot ; the results of x - ray analysis are taken from butler et al . 2005 .
* grb 041218 : for this long integral burst there is only a late xrt observation .
we do not have information for the x - ray temporal decay but in the optical we see a steepening of the decay after @xmath32 days .
no spectral information is available for this burst .
* grb 050502a : for this long integral burst the xrt provided only an upper limit for the x - ray flux at @xmath33 days . in the optical band
the light curve can be described by a single power law with a hint of a bump at 0.02 days . in the x - ray band ,
temporal and spectral parameters are consistent with similar behaviour to that of the optical ( see guidorzi et al .
2005b for the detailed analysis of this burst ) .
* grb 050713a : xmm observations performed between 5.8 and 13.9 hours after the burst , show a break in the x - ray light curve at @xmath34 s. that break is marginally detected in the x - ray data acquired by _ swift _ , while the optical light curve is well described by a shallow power law at all times ( guetta et al . , 2006 ) .
xmm data are not reported in our plot ; an accurate spectral analysis has been done by morris et al .
( 2007 ) . *
grb 050730 : both x - ray and optical light curves of this _ swift _ burst show a steepening of the initial temporal power law decay at 0.1 days , more pronounced in the x - ray band .
the derived spectral slopes from x - ray and optical data are statistically in agreement ( pandey et al . 2006 ) . *
grb 051111 : the optical light curve is well fitted by a broken power law with an early break around 12 minutes , while in the x - ray no early data were acquired and no break is visible .
a comprehensive multi - wavelength temporal and spectral study of this burst has been done by guidorzi et al .
( 2007 ) , yost et al .
( 2007 ) and butler at al .
* grb 060108 : this burst has a faint , but relatively blue , optical afterglow that was identified by deep rapid ftn observation , without which it would have been classified as dark burst ( @xmath35 and no afterglow was detected by the _
swift _ uvot within the first 300 s ) .
it shows a canonical x - ray light curve while in the optical band the flux may show a similar behaviour but , due to its optical faintness the resultant sparse sampling with co - added images provides a light curve that is consistent wit a single power law .
the darkness of this burst can be explained by a combination of an intrinsically optical faintness , an hard optical to x - ray spectrum and a moderate extinction in the host galaxy ( oates et al .
* grb 060203 : after an initial rise , the optical light curve follows a shallow power law decay ( typical example of the case 2 reported in figure [ figc1 ] ) . due to observing constraints ,
x - ray observations started only @xmath36 s after the trigger and the light curve shows evidence of a power law decay roughly consistent with the optical one .
the spectral fit gives a value of @xmath37 and shows a value of n@xmath38 in excess of the galactic one ( morris et al .
* grb 060204b : the early x - ray light curve is dominated by flaring activity and only after @xmath39 s does the decay appear to be a single power law ( falcone et al .
2006a ) . in the optical
the afterglow is faint and the light curve is a simple power law , shallower than the x - ray decay .
* grb 060206 : the detailed analysis of the optical light curve showed dramatic energy injection at early time , with at least 3 episodes , with the largest one also visible in the x - ray light curve . at late times
a break is evident in the optical but not in the x - ray band ; the chromatic nature of the break is not consistent with the possible jet - break interpretation and can be ascribed to a change in the circumburst density profile ( monfardini et al .
in a recent work curran et al .
( 2007 ) found that the data are also consistent with an achromatic break , even if the break in the x - ray is less pronounced .
however , the complete x - ray data set acquired with xrt remains consistent with a single unbroken power law decay up to @xmath40 seconds ( morris et al .
2006b , burrows @xmath4 racusin 2007 ) , and the first explanation for the break seems to be the more plausible .
* grb 060210 : the optical light curve exhibits a power law decay after an initial flat phase ( @xmath41 , case 3 in figure [ figc1 ] ) .
there is marginal evidence for a break at late times ( @xmath42 @xmath43 days ) but there are not enough optical data to support that
. instead a break is visible in the x - ray light curve , but the late time break is not simultaneous ( @xmath44 days ) . we can not exclude the possible achromatic nature of that break due to the very complex behaviour of x - ray light curve .
* grb 060418 : in the optical and infrared bands ( see molinari et al . , 2007 for the infrared analysis ) an initial rise is visible , followed by a straight power law decay ( case 2 in figure [ figc1 ] ) . fitting the early x - ray data it is possible to see a change in the slope , but , accounting for the presence of a large flare , the data are also consistent with a single power law decay .
there is no evidence of temporal break up to @xmath45 s in both optical and x - ray bands .
the spectrum of the underlying x - ray afterglow can be described by a simple absorbed power law with @xmath46 ( falcone et al .
the low degree of polarization of the optical light at early time ( at @xmath47 s after the event ) ruled out the presence of a large scale ordered magnetic field in the emitting region ( mundell et al .
* grb 060510b : the temporal behaviour in the x - ray seems to be the canonical steep - shallow - steep decay with superimposed flares , even if the light curve at late times is poorly sampled . looking at the flat light curve at early times , possibly the prompt gamma - ray component is detected in xrt ( t@xmath48 s ) . due to the high redshift ,
the optical flux is suppressed by the ly@xmath11 absorption and our early optical data are consistent with a single power law decay .
no optical observations at late time were available to confirm the possible achromatic nature of the break observed in the x - ray band .
the early time x - ray spectrum is well fitted with an absorbed power law with @xmath49 ( perri et al .
this is the value of @xmath24 reported in table1 because estimated at a time consistent with the time of our optical observations . at late times , taking into account the intrinsic absorption at the redshift of the burst , the spectrum is well fitted by an absorbed power law with @xmath50 ( campana et al . , 2006 ) .
grb 060512 : optical data are consistent with a single power law . in the x - ray a clear flare is visible at early times and the decay after @xmath36 s is a simple power law . at late time
, the x - ray spectrum has a slope @xmath51 ( godet et al .
* grb 060927 : due to the high redshift ( @xmath52 ) , the optical light curve is highly affected by ly@xmath11 suppression , particularly in the @xmath0 filter .
the light curve in the @xmath53 band shows evidence of a possible extra component at early times , but after 500 s the decay is a single power law ( @xmath54 ) . in the x - ray
a change of slope is clearly visible and the spectrum is well modelled with an absorbed power law with @xmath51 ( ruiz - velasco et al . , 2007 ) .
* grb 061007 : the optical light curve exhibits an early peak followed by an unprecedented straight power law decay up to ( and likely beyond ) @xmath45 s after the burst ( a good example of case 2 in figure [ figc1 ] ) , perfectly mirrored in the x - ray band .
the peak at early times can be explained in the context of the fireball model : no optical flash is seen because the typical frequency of the reverse shock emission lies in the radio band at early time and the optical afterglow is dominated by forward shock emission ( mundell et al .
the broad band optical to @xmath6-ray spectral energy distribution is well described by an absorbed power law with @xmath55 ( mundell et al . 2007b , schady et al .
* grb 061110b : this burst is intrinsically faint and displays a simple power law decay both in the optical and in the x - ray bands .
it showed a typical grb afterglow spectrum with @xmath56 ( grupe et al .
* grb 061121 : this is a perfect example of canonical light curve in the x - ray band not replicated in the optical .
the light curve follows a simple power law decay even though the observations began when the initial steep x - ray phase was still ongoing ( page et al . , 2007 ) .
* grb 061126 : the optical light curve shows a steep to shallow transition at about 13 minutes after the trigger .
the early , steep component can be interpreted as due to the reverse shock while the later slowly fading component is coming from the forward shock ( clear example of case 1 in figure [ figc1 ] ) .
x - ray observations started after the transition in optical , and show the x - ray afterglow decaying with @xmath57 ( faster than the optical afterglow ) to the end of observations at @xmath40 s ( for more details see gomboc et al .
* grb 070208 : after an initial rising phase , the x - ray light curve shows a power law decay .
the optical data cover the same time interval as the x - ray data but no rising phase is detected and the optical light curve is well described by a simple power law , after an initial flat phase ( case 3 in figure [ figc1 ] ) .
the optical decay index after the flat phase remains shallower than the decay in the x - ray band .
* grb 070411 : in this case the temporal decay in the two bands is very similar , but in the optical band an initial rising phase is detected that is not visible in the x - ray band , probably due to the poor sampling at early times in that band .
the optical light curve seems to be another example of case 2 reported in figure [ figc1 ] but the shape is not smooth and clear , with significant scatter ( possibly variability ) around the power law decay . *
grb 070419a : different observations in the optical band show that at early time the afterglow brightened before starting a shallow decay phase that lasts up to @xmath40 s. although the x - ray temporal decay at late time ( @xmath58 = @xmath59 ) agrees very well with the optical decay ( @xmath60 = @xmath61 ) , at early times the x - ray light curve shows a rapid decay , with no hints of any flare activity .
the shape observed in the optical band does not fit the three cases reported in figure [ figc1 ] and is probably the result of an episode of energy injection . *
grb 070420 : another example of canonical light curve in the x - ray band not replicated in the optical .
however , in this case the optical light curve is less well sampled than grb 061121 , so it is more difficult to constrain the optical behaviour in the entire time interval covered by xrt observations . *
grb 070714b : the x - ray data show a steep - shallow - steep decay , not well constrained at late times and with possible small flaring activity .
the optical counterpart is very faint but appears to show a shallower power law decay compared with the x - ray decay at late times . in table 2
we report the optical upper limits for all the bursts observed with the liverpool and the faulkes telescopes for which we did not detect any afterglow candidate . for each grb we specified when
the xrt position for the x - ray afterglow was found .
the duration , the bat fluence @xmath62 ( 15 - 150 kev ) , the xrt early flux f@xmath63 ( 0.3 - 10 kev ) , the time of x - rays observations ( @xmath64 ) , the temporal decay ( @xmath65 ) and the spectral slope ( @xmath24 ) in the x - ray band , are the values reported in the gcns or taken from the _ swift _ general table .
@xmath66 are the values of our optical upper limits at @xmath67 minutes from the trigger , where @xmath68 is the starting time of our observations .
@xmath69 are the values of our optical upper limits at @xmath70 minutes from the trigger ( mean time ) for the coadded frames with a total integration time of @xmath71 minutes .
the columns ot and @xmath72 show when an optical ( o ) or infrared ( ir ) afterglow for that burst had been detected by other facilities and the extinction for the @xmath0 band in the direction of the burst .
the last two columns on the table are the value of the x - ray temporal decay inferred from our fit of the xrt light curves ( @xmath73 ) and the estimate of the x - ray flux ( f@xmath63 ) at the time of our optical upper limit @xmath70 .
x - ray light curves are given in figure [ figul2 ] together with our optical upper limit . the best fit for the x - ray light curves
the temporal decay indices in the x - ray band ( @xmath74 ) reported in table 2 refer to the segment of the light curve contemporaneous with our optical limit .
only two of the bursts listed in the table were detected by uvot in the optical bands ( grb 070721a and grb 070721b ) ; we did not detect the optical counterpart for these two bursts because our observations were performed at late time due to observational constraints . of the 39 non - detections ,
10 were detected by other facilities , primarily at infrared wavelengths or using larger aperture optical telescopes ; the other 29 remain as non - detections .
details are summarised below : * grb 050124 : an infrared candidate was detected and confirmed by two keck observations performed in the @xmath75 band about 24.6 and 47.8 hours after the burst ( @xmath76 , berger et al .
2005a , berger et al .
no afterglow was detected in the optical bands even by the uvot telescope ( @xmath77 hours after the trigger , lin et al .
2005 , hunsberger et al .
our observations were performed manually in the @xmath0 band only after 14.7 hours for observational constraints and no optical counterpart was detected .
* grb 050716 : our observations in the optical band began 3.8 minutes after the burst with the ftn telescope but no optical candidate was found down to a limit of r@xmath320 mag ( guidorzi et al .
a potential infrared counterpart ( @xmath78 ) was found in ukirt observations just outside the xrt error circle ( tanvir et al . 2005 ) . at the position of that candidate we did not clearly detect any source in the @xmath0 and @xmath53 bands but we found an excess flux which suggests that the afterglow is probably reddened rather than at very high redshift .
a broad - band analysis found that @xmath79 2 is a good estimate for the redshift of this event and that a host galaxy extinction of a@xmath80 can account for the relatively faint optical / infrared afterglow observed ( rol et al .
* grb 060116 : also for this burst an infrared candidate was detected in ukirt observations ( kocevski et al .
2006a , kocevski et al .
the afterglow of that burst was detected with an unfiltered magnitude @xmath320 ( swan et al .
2006 ) and the very red colours of this afterglow ( @xmath81 , @xmath82 , malesani et al .
2006a ) suggest the source to be a highly dust extincted ( a@xmath83 ) but not at high redshift ( tanvir at al .
* grb 060121 : the sxc error circle for this hete - ii short burst was partially covered by our first lt observation starting 0.83 hours after the burst . only in the second observation starting 2.44 hours after the event
the xrt error circle was entirely inside our field of view .
the detected optical / infrared afterglow ( levan et al
. 2006a , malesani et al . 2006b ,
hearty et al . 2006a , hearty et al .
2006b ) is not seen in our images down to a limiting magnitude of @xmath84 at @xmath77 hours after the burst .
subsequent hst observations revealed the presence of a very faint ( both in optical and infrared ) red galaxy , probably an edge - on disk , close to the position of the afterglow ( levan et al .
2006 ) . in this case
this would favour a higher redshift for this burst than has been measured for most short bursts to date . *
grb 060319 : an infrared candidate was found in wht observations ( @xmath85 , tanvir et al .
2006b ) but no claims about variability have been made .
our observations set an upper limit of @xmath86 mag after 10 minutes of the burst and no other optical observation detected any possible counterpart down to @xmath87 mag ( davanzo et al .
2006a , lipunov et al . 2006 ) .
* grb 060602a : a very faint possible optical candidate was detected about 15 minutes after the burst ( @xmath88 , jensen et al .
the transient nature of the source was difficult to assess but was not visible in the sdss pre - burst image .
our observations were affected by the bright moon and only an upper limit of @xmath89 mag was set about 36 minutes after the burst .
* grb 060923a : an infrared afterglow was detected in the @xmath90-band but undetected in @xmath53 and @xmath91 bands ( tanvir et al .
2008 , fox et al .
2006a , fox 2006 ) , suggesting again a highly extinguished or high redshift afterglow for this burst .
the subsequent detection of a faint host galaxy ( @xmath92 , tanvir et al .
2008 ) set an upper limit for the redshift of this burst of @xmath93 , leaving the extinction as the most likely cause of the extremely red colours measured at early times .
* grb 060923c : the afterglow of that burst was detected and confirmed in the infrared bands ( covino et al .
2006 , fox et al .
2006b , davanzo et al .
no indication of any optical identification means again that the possible explanation is a high redshift or high extinction origin for that burst . *
grb 061006 : a source was found to vary with @xmath94 mag between 0.6 and 1.6 days after the burst in the @xmath53-band ( malesani et al .
2006c , malesani et al .
this source was identified with the afterglow of this short - hard burst .
however the inferred power - law decay slope was quite shallow ( @xmath95 ) and in the second observation the source was extended so the detection was contaminated by the host galaxy ( malesani et al .
* grb 070223 : the afterglow of that burst was confirmed in the infrared band ( castro - tirado et al .
2007 , rol et al . 2007 ) and found to be very faint in the optical band , close to the deep limit of our observations . in summary , of the 10 afterglows discussed above
, 7 were detected in the infrared bands with very red colours and 3 in the optical band .
of those latter 3 optical afterglows , two ( grb 060602a and grb 061006 ) had a magnitude below our limiting magnitude in the same band at the same time .
for the remaining one ( grb 060121 ) our observations performed @xmath96 hours after the burst support a probable high - redshift nature for this event .
for a simple visual comparison we show in figure [ figxopt ] the x - ray light curves ( where data were available , from evans et al .
2007 ) , together with the optical light curves , including all published data . when optical and x - ray data cover the same time interval we superimposed on the data a simple power - law fit , to better understand the temporal decay behaviour . as can be seen
, the light curves in the two bands do not follow the same temporal decay for all the grbs .
for the majority of the bursts the behaviour in the x - ray and optical bands is different , especially at early times where in the x - ray band the temporal decay is steep , showing the hints of large flare activity . in the standard fireball model ,
observed afterglow emission is synchrotron radiation from a quasi - spherical relativistic blast wave ( forward shock ) that propagates into the homogeneous or wind - like ambient medium .
the model can give clear predictions of the shape of light curves at different frequencies .
here we test this model comparing the theoretical expectations with the observed light curves properties .
we take into account simple modifications to the standard model if needed .
the modifications are energy injection into the blast wave in the afterglow phase ( @xmath97 ) and a generalized wind environment ( @xmath98 ) .
the injection would modify the bast wave dynamics as long as @xmath99 and @xmath100 corresponds to the case without energy injection .
the possible theoretical values for the temporal decay index @xmath11 and the spectral index @xmath12 as functions of the electron spectral index @xmath101 are summarized in table 3 .
traditionally , optical data from ground based telescopes alone have been used to establish the presence of achromatic breaks since the number of pre-_swift _ bursts with good simultaneous x - ray and optical data at any time ( early and late ) was very small .
however , recent studies of _ swift _ bursts have shown that many _
swift _ grbs exhibit a well defined steepening of the x - ray light curve while the optical decay continues to be described by a single un - broken power law ( panaitescu et al .
2006 ) . in some cases
the decay is a straight power law at all times with no breaks either in the optical or in the x - ray band ( mundell et al .
2007b ) , while in some bursts the break is observed only in the optical but not in the x - ray band . in this paper
we study breaks in x - ray and optical light curves in the decay phase . immediately after the prompt emission , some light curves show a peak or flare features which are likely to be due to central engine activity or reverse shock emission .
we concentrate our discussion on the study of simpler forward shock emission .
the bursts in our sample ( see figure [ figxopt ] ) can be divided into four main classes , depending on the presence or not of a break in the optical curve ( see figure [ figc2 ] ) : * class a : no break in the optical or in the x - ray band ; * class b : no break in the optical band , break in the x - ray band ; * class c : break in the optical band , no break in the x - ray band ; * class d : break in the optical and in the x - ray band . in the cartoon shown in figure [ figc2 ]
we show the shape of the four classes .
this classification is based on the available data in the two bands : _
swift_-xrt for the x - ray light curve and our telescopes , gcns and published data for optical light curves .
there is clearly the possibility of additional breaks in the period not covered by the observations both in the optical and in the x - ray band .
the possible mechanisms that can produce a break in the observed light curves in the decay phase can be summarised as follow : 1 .
the cooling break @xmath102 chromatic break ; 2 . cessation of energy injection @xmath102 achromatic break ; 3 .
jet break @xmath102 achromatic break ; 4 .
change in the ambient distribution @xmath102 chromatic or achromatic break ; 5 . additional emission component ( reverse shock , late central engine activities , sn - component , host galaxy contamination ) @xmath102 chromatic break ; in the epoch under investigation in this paper ( @xmath103 s ) , the emission process is in the slow cooling regime .
one of the most natural explanations for a break in a light curve is the cooling break ( i ) : on the passage of the cooling frequency through the observation band , the light curve steepens by @xmath104 .
the steepening happens in the x - ray band first and in the optical band later for a homogeneous medium , while it occurs in the optical band first and in the x - ray band later for a wind - like medium .
note that the cooling frequency increases in time for the wind - like medium .
the cessation of energy injection into the blast wave ( ii ) or a jet break ( iii ) causes a change in the hydrodynamics of the blast wave , producing achromatic breaks in the light curves . in the post - break phase
, the optical and x - ray decay indices could be the same or different by @xmath105 for mechanism ( ii ) .
the decay indices in the two bands should be the same for a jet break ( iii ) .
it should be noted that for the mechanism ( iv ) the afterglow emission above the cooling frequency does not depend on the ambient matter density .
the break is achromatic if the cooling frequency is located above the x - ray band , while it is chromatic if the cooling frequency lies between the two bands .
the mechanisms ( v ) , especially reverse shock emission and late time internal shocks , are generally believed to be relevant only at early times and their contribution to the shape of the light curves becomes negligible at late times .
temporal and spectral properties of all the bursts in the sample are reported in table 1 . assuming the standard fireball model and a homogeneous or wind - like circumburst medium it is possible to derive the closure relations between the temporal decay index ( @xmath11 ) and the spectral slope ( @xmath12 ) in order to satisfy the models ( e.g. zhang et al .
2006 ) . in figure
[ figab ] the comparison between observed properties and model expectations is shown in the optical ( left panel ) or in the x - ray ( right panel ) band .
the two bursts indicated on the left panel of figure [ figab ] ( grb 060108 and grb 060210 ) are the ones that , based on the optical data , deviate the most from the standard model . in the case of grb 060108 the spectral optical analysis ( without accounting for extinction ) gives a steep value for @xmath23 , hard to explain for the standard model .
when extinction is included a shallower value is found ( @xmath106 ) more in agreement with the observed spectral energy distribution ( oates et al .
2006 ) . for grb 060210 , the spectral analysis reveals a large difference between @xmath24 , @xmath23 and @xmath25 , probably due to a large amount of extinction which is difficult to evaluate .
if again we consider the value found from the fit of the multiband spectral energy distribution a value in agreement with the model is found ( see curran et al .
2007 for detailed analysis of this burst ) .
using temporal and spectral information ( when available ) , it is now possible to constraint the value of electron spectral index ( @xmath101 ) for different bursts and study each burst belonging to our four classes in the context of the standard fireball model .
for 10 bursts ( grb 050502a , grb 060203 , grb 060204b , grb 060418 , grb 060512 , grb 061007 , grb 061110b , grb 070208 , grb 070411 and grb 070419a ) no break is observed in the optical or x - ray bands , but in general , the decay indices of the optical and the x - ray light curves are different .
the simplest explanation for the difference is that the cooling frequency is situated between the optical and the x - ray band .
if this is the case , the difference is @xmath104 as we have discussed in the previous section .
if there is energy injected into a blast wave @xmath97 , the difference is given by @xmath107 ( see table 3 ) .
the x - ray light curve is steeper by @xmath108 than the optical in the homogeneous medium , while the optical one is steeper by @xmath109 in the wind - like medium . in the generalized wind - like medium case @xmath98 ( with no energy injection )
, optical afterglow decays faster by @xmath110 than the x - ray afterglow ( monfardini et al .
2006 ) .
* grb 050502a : the decay index of the x - ray afterglow is not well determined , and late time _ swift _
observations give only a lower limit for @xmath65 .
since the optical decay is shallower , @xmath8 should lie between the two bands and the homogeneous ambient medium is favored .
the electron spectral index is given by @xmath111 .
the theoretical values @xmath112 and @xmath113 are in good agreement with the observations ( guidorzi et al .
* grb 060203 : the x - ray light curve is steeper by @xmath114 than the optical light curve .
it indicates the uniform ambient medium and @xmath115 during the power law decay phase .
the theoretical estimates @xmath116 , @xmath117 and @xmath56 ( @xmath118 ) can explain well the observations ( @xmath119 , @xmath120 and @xmath121 ) .
* grb 060204b : the x - ray afterglow decays faster by @xmath122 than the optical afterglow .
the difference @xmath108 is larger than the value @xmath104 for a simple model in which the optical and x - ray bands are in different spectral domain ( i.e. @xmath115 ) .
since x - ray light curve is steeper than the optical light curve , wind - like medium model does not work .
although constant energy injection ( q=0 ) gives a close value @xmath123 we do not expect the observed steep decay @xmath124 and @xmath125 for such significant energy injection . in the early phase flares
are noticeable in the x - ray light curve .
late internal shock emission might dominate x - ray band at later times as well .
superposed flares might steepen the x - ray light curve .
another possible explanation is that fluctuations in the ambient medium produce bumps in the late optical light curve ( this interpretation was already proposed by guidorzi et al . 2006 to explain the bump in the light curve observed for grb 050502a ) .
if the cooling frequency lies between the two bands , the bumps are produced only in the optical light curve .
since the optical observations are very sparse for this afterglow , a bump in the optical light curve might take the decay index shallower than the real value .
* grb 060418 : the x - ray light curve is steeper by @xmath126 than the optical light curve in the late decay phase .
this indicates an uniform ambient medium with @xmath127 during that phase .
temporal observed values ( @xmath128 , @xmath129 ) are in agreement with the theoretical expectation of a simple model with a value for the spectral index of @xmath130 ( @xmath131 , @xmath132 ) . * grb 060512 : the x - ray emission decays faster by @xmath133 than the optical emission .
the difference @xmath108 is not consistent with the simplest scenario ( @xmath108 = 1/4 ) .
the energy injection ( q=0.48 ) in homogeneous ambient can account for the large value of @xmath108 .
however , even with @xmath134 , the maximum value allowed from the observed spectral index @xmath135 , the expected decay indices @xmath136/4 = 0.22 $ ] and @xmath137/4 = 0.60 $ ] ( zhang et al .
2006 ) are much shallower than the observed values .
x - ray flares might make the x - ray decay index larger than the real decay index of the blast wave emission .
* grb 061007 : the afterglow of this burst is very bright in optical and x - ray , the decay is a straight power law from early time ( there is the hint of a rise in the optical not mirrored in the x - ray ) since late times .
a comprehensive multiwavelength analysis of this burst is presented in mundell et al .
( 2007b ) : the evolution of the afterglow can be explained in the context of the fireball model with @xmath138 for the entire @xmath45 s period covered by the observations .
* grb 061110b : for that burst the optical light curve decays faster by @xmath139 than the x - ray , indicating that during the observations @xmath8 is located between the two bands in a wind - like medium .
the x - ray data imply a value of @xmath140 and the expected value for the optical temporal decay @xmath141 is in good agreement with that observed ( @xmath142 ) .
* grb 070208 : the x - ray afterglow decays faster by @xmath143 than the optical afterglow .
again this difference is much larger than the value @xmath104 for a simple model , similar to the case of grb 060204b .
also the energy injection model can not account for the larger value @xmath144 . beyond the standard model , possible explanations of such a large difference
are x - ray flares ( late time internal shocks ) which make the x - ray steeper coupled with energy injection which makes the optical decay shallower .
* grb 070411 : the x - ray afterglow decays faster by @xmath145 than the optical afterglow .
it indicates the uniform ambient medium and @xmath115 during the power law decay phase .
if this is the case , the value of p derived from the x - ray data ( @xmath146 ) imply a value of the decay indices @xmath147 and @xmath148 , well in agreement with the observed values ( @xmath149 and @xmath150 ) .
* grb 070419a : for this burst the temporal decay in the two bands is again very similar , but this is true only at late times ( @xmath151 ) , while at early times the shape in the two bands is very different : a very steep decay in the x - ray ( @xmath152 ) and a possible broad re - brightening in the optical . even in the late power law phase ,
no closure relations for the simple models can reconcile the observed value @xmath153 .
assuming @xmath154 ( then the emission does not depend on the ambient medium ) we obtain @xmath155 from @xmath24 .
the observed decay indices in the two bands ( @xmath156 ) and the closure relation could be explained if there is significant energy injection ( @xmath157 ) .
the total injected energy increases by a factor @xmath158 between the break time @xmath159 days@xmath160 s and the end of the observations @xmath161
s. this could contradict with the energy budget of the central engine ( solar mass scale ) .
a steepening in the x - ray decay slope is observed in eight bursts of our sample ( grb 050713a , grb 060108 , grb 060210 , grb 060510b , grb 060927 , grb 061121 , grb 070420 and grb 070714b ) , while the rate of the optical decay remains constant .
a simple explanation of this behaviour could be the passage of the cooling frequency through the x - ray band . for a homogeneous ambient medium
, the decay indices of the optical and x - ray light curves should be the same in the pre - break phase , with only the x - ray light curve steepening due to the passage of @xmath8 .
in contrast , for wind - like medium the optical light curve is steeper than the x - ray light curve in the pre - break phase and the decay indices in the two bands become the same after a break in the x - ray light curve . for the eight bursts in that class ,
the observed steepening @xmath108 is always larger than the value @xmath104 expected in the simplest scenario .
* grb 050713a : the x - ray emission decays faster by @xmath162 than the optical in the post - break phase .
the decay indices in the pre - break phase are also significantly different from each other ( @xmath163 ) .
we can not explain the break ( and the behaviour of the light curves in the two bands ) by the cooling break even if the energy injection and generalized wind - like medium are assumed .
the cessation of the energy injection and a jet break also can not account for the observed break because of their achromatic nature .
the most likely explanation is that x - ray flares due to late internal shocks shapes the x - ray light curve ( there are notable fluctuations in the x - ray light curve ) .
although the optical light curve is poorly sampled , the rather shallow observed decay ( @xmath164 ) might indicate energy injection into the blast wave . * grb 060108 : the decay indices of the optical and x - ray light curves are almost the same in the pre - break phase , but the steepening in the x - ray band ( @xmath165 ) is too large to be explained in a simple cooling break model ( also reported by oates et al .
since there are no optical observations at late times , achromatic break mechanisms are also applicable to this event .
the x - ray decay after the break ( @xmath166 ) is too shallow for the jet break model in which the electron spectral index should be equal to the post - break decay index ( @xmath167 in that case ) . considering the shallow pre - break decay in the x - ray and optical
, the probable explanation is the energy injection ceasing .
* grb 060210 : the optical light curve is complex
. the early flat portion might be due to energy injection or to the passage of the typical frequency through the optical band , and there is a hint of a late break .
since we are interested in breaks in the decay phase , we discuss the intermediate power law part of the optical light curve together with x - ray observations . before the x - ray break the optical afterglow decays faster ( @xmath168 ) than the x - ray afterglow
. this could indicate that the cooling frequency lies between the two bands and that the ambient medium is wind - like .
since after the break the x - ray light curve becomes steeper than the optical , the steepening is not explained by a cooling break even if we consider energy injection or a generalized wind - like medium .
the post - break x - ray index ( @xmath169 ) is too shallow to be explained in the jet break model . in the model of cessation of energy injection
the difference of the decay indices in the two bands ( @xmath115 ) is @xmath170 in the pre - break phase ( during the energy injection ) and the steepening in the x - ray light curve ( ceasing of energy injection ) is given by @xmath171 .
the observed difference @xmath168 is smaller than the expected value in the energy injection model ( @xmath172 should be smaller than unity ) .
the x - ray spectrum @xmath173 and the observed steepening @xmath174 require @xmath175 and significant energy injection ( @xmath176 ) , for which the decay indices are expected to be @xmath177 , @xmath178 , @xmath179 .
this could marginally explain the observations .
a more plausible explanation is that x - ray flares due to late central engine activity shape the x - ray light curve .
* grb 060510b : in the pre - break phase , the decay indices in the two bands are the same if the prompt emission and x - ray flares at early times are ignored .
considering the shallow pre - break decay in the x - ray and optical bands , a possible explanation for the x - ray break is the cessation of energy injection .
if this is the case , the optical light curve should have a break at the same time , although there are no optical observations at late times . *
grb 060927 : the optical light curve in the r - band does not show a simple decay ; @xmath180 and @xmath181 .
there is a possible flat phase in the optical between @xmath182 and @xmath183 s. a possible explanation could be late energy injection but this behaviour is not mirrored in the x - ray band .
the observed characteristics of this burst are difficult to explain in the context of the standard model .
* grb 061121 : page et al .
( 2007 ) found that the decay indices in the optical and x - ray bands are consistent , within the uncertainties , with the ( possible ) presence of an achromatic break .
however , the optical data after the break , as reported in our figure [ figxopt3 ] , are in very good agreement with the same power law decay index observed before the break ( @xmath184 ) , without requiring any break in the optical .
the chromatic nature of the break excludes the possibilities of a jet break and cessation of energy injection for the explanation of the break in the x - ray light curve .
since the x - ray decay index evolves from almost zero ( flat ) to a steep value ( @xmath185 ) , which is much larger than the optical decay index ( @xmath184 ) , neither cooling break models nor change of the ambient density distribution ( e.g. change from ism medium to wind - like medium during the propagation of the blast wave ) can explain the evolution of the x - ray light curve .
the emission from late central engine activities could mask the x - ray radiation from the forward shock .
the darkness @xmath186 might be due to the bright additional x - ray emission from the central engine .
* grb 070420 : the observed data are not consistent with any closure relation for the standard model .
the behaviour of the x - ray and the optical light curves looks quite different .
however , the sparse observations can not exclude the presence of a flat phase in the optical light curve as visible in the x - ray .
if this is the case , the ceasing of energy injection could explain the x - ray and optical observations .
* grb 070714b : similarly to grb 070420 , the behaviour or the x - ray and optical light curves can not be explained in the standard model if a single power law is assumed for the decay of the optical light curve .
although the optical data are too sparse to firmly constrain the behaviour in that band , but a possible initial flat phase may be present ( like in the x - ray band ) followed be a standard power law decay . like the previous case , the ceasing of energy injection could explain the x - ray and optical behaviours .
in five cases a change of decay index is detected only in the optical light curve and not in the x - ray band ( grb 041006 , grb 041218 , grb 051111 , grb 060206 and grb 061126 ) .
a possible explanation for these breaks is the passage of @xmath187 through the optical band .
if a homogeneous medium is assumed , the decay indices of the optical and the x - ray light curves should be the same in the post - break phase , with only the optical light curve steepening with the passage of @xmath8 .
if a wind - like ambient medium is assumed , the decay indices in the two bands are the same in the pre - break phase and the optical light curve is steeper than the x - ray after the break . * grb 041006 : counting the rather large error in the value of @xmath65 , the x - ray decay and the post - break optical decay indices could be considered as almost the same . if the cooling break model is assumed to explain the steepening of the optical light curve , a homogeneous ambient medium is favored , because the decay indices in the two bands are the same in the post - break phase .
the drastic steepening in the optical light curve ( @xmath188 ) requires almost constant energy injection ( @xmath189 ) .
it is hard to achieve the steep decay in the post - break phase ( @xmath190 ) with such a massive energy injection .
the cooling break model does not work .
since there are no x - ray observations at early times , achromatic break models are acceptable .
the post - break optical index is too shallow to consider a jet break .
+ next we consider the cessation of energy injection . from the observed spectral indices ( @xmath191 )
, the two bands should be in the same spectral domain and we obtain @xmath192 for @xmath193 or @xmath194 for @xmath195 .
the closure relations are marginally satisfied in both cases .
the ceasing of energy injection causes steepening in a light curve .
if the observation band is above @xmath8 the flux does not depend on the ambient medium and the steepening is @xmath196 . if @xmath197 the steepening is @xmath198 for a homogeneous ambient medium and @xmath199 for a wind - like ambient medium .
for the combination of @xmath193 and homogeneous ambient medium , the initial energy injection is mildest @xmath200 and the expected values @xmath201 , @xmath202 and @xmath203 could be consistent with the observations .
* grb 041218 : there is only a late time observation for the x - ray band , neither @xmath65 nor @xmath24 are constrained from the observations .
the optical spectral index @xmath23 is also not available .
the break in the optical light curve ( @xmath204 ) could be explained in many models including the cooling break .
* grb 051111 : since the x - ray decay index @xmath205 is incoincident with both the pre - break ( @xmath206 ) and the post - break ( @xmath207 ) optical decay index , cooling break models can not account for the optical break even if energy injection is considered .
the fact that the x - ray emission decays faster than the optical emission rules out a wind - like ambient medium scenario ( and scenarios related to the wind medium ) .
no x - ray observations are available before the optical break and the break might be achromatic .
the jet break is unlikely because the afterglow decays with significantly different rates in the two bands after the break .
the difference of the spectral indices in the two bands indicates that the two bands are in different spectral domains ( i.e. @xmath208 ) . this highlighted by the spectral energy distribution analysis presented in guidorzi et al .
2007 . using the observed @xmath209 ( for which the error is smaller than in @xmath24 )
, we obtain @xmath210 .
if the energy injection rate changes from @xmath211 to @xmath212 at the break , the steepening of the optical decay is given by @xmath213 and the difference of the decay indices is @xmath214 in the post - break phase .
the resulting @xmath211 and @xmath212 are negative and unphysical , with which the predicted temporal and spectral indices deviate largely from the observed values
. the energy injection model also does not work .
* grb 060206 : the x - ray light curve is consistent with a single unbroken power law , achromatic break models ( i.e. jet break and ceasing of energy injection ) are ruled out .
since the x - ray decay index ( @xmath215 ) is different from both the pre - preak ( @xmath216 ) and the post - break ( @xmath217 ) optical decay indices , the cooling break models can not account for the optical break even if the energy injection or generic wind - like ambient medium is considered .
the observed behaviour of the light curves might be due to a transition in the ambient matter distribution ( monfardini et al .
2006 ) : a blast wave initially propagates into a constant medium and then it breaks out into a wind - like medium .
note that x - ray emission does not depend on the ambient matter density ( and its distribution ) as long as the x - ray band is above the cooling frequency , and that the optical emission ( below @xmath8 ) reflects the change in the ambient matter distribution .
* grb 061126 : for this burst no break is visible in the x - ray band , while in the optical the transition is from steeper to shallower decay index .
this behaviour at early times can be explained as the contribution from the reverse shock component . the detailed study by gomboc et al .
( 2008 ) shows its inconsistency with the standard fireball model ( the steeper decay in the x - ray band and the large ratio of x - ray to optical flux ) .
the bursts belonging to this class are those showing a break in both the optical and the x - ray light curves in their decay phases . in general , breaks in the two bands occur at different times , both chromatic and achromatic breaks are considered .
surprisingly , for only one burst do we observe a break in both bands , although grb 060210 might be classified in this case if we take seriously the last optical data points .
* grb 050730 : if a jet break is responsible for the steep x - ray decay , the electron energy index is @xmath218 . given that x - ray band is below the cooling frequency , the observed x - ray spectral index @xmath219 is consistent with the model prediction @xmath220 .
however , the much shallower optical decay ( @xmath221 ) is inconsistent with the jet break model . as pointed out by pandey et al .
( 2006 ) , the possibility of a contribution from the host galaxy or an associated sn to the late time optical afterglow can be ruled out considering the high redshift of the burst ( @xmath222 ) . since after a jet
break a forward shock emits photons practically at a constant radius ( the exponential slowing down ; sari , piran @xmath4 halpern 1999 ) , fluctuations in the ambient medium do not seem to affect the decay rate of the emission . even if there is an effect , both light curves should become shallower ( or steeper ) in the same way because both optical and x - ray bands are in the same spectral domain ( @xmath223 ) . energy injection into a forward shock also can not explain the shallow optical decay because of the same reason .
additional emission components , e.g. the two component jet model or late time internal shocks might make the optical decay slower . the early shallow decay phase observed in the optical and x - ray light curves
could be explained by energy injection ( pandey et al .
2006 ) , though the ceasing of the injection should happen around the time of the jet break . from this analysis of the optical and x - ray light curves
we conclude that in our sample of 24 optical grb afterglows : 15 bursts can be explained in the context of the standard fireball model ( with modifications : energy injection or variation in the ambient matter distribution ) ; while for the remaining 9 bursts : * class a : grb 060204b , grb 060512 , grb 070208 , grb 070419a * class b : grb 050713a , grb 060927 , grb 061121 * class c : grb 051111 , grb 061126 the observed data are inconsistent with the predictions of the standard model . in figure [ figall2 ]
we translate the observed magnitudes of the optical afterglows into the rest - frame luminosity ( left panel ) .
the subscript @xmath224 used in this section refers to the time in the rest frame of the grbs .
we assumed a standard cosmology ( defined in section 1 ) and we include any correction expected from the distance of the event ( @xmath26 ) and its spectral properties ( @xmath12 ) in order to report all the observed quantities in the rest frame of the grb .
we corrected the optical magnitude for galactic extinction using the reddening maps of schlegel et al .
( 1998 ) and we applied the _
k_-correction to take into account the fact that sources are observed at different redshifts ( @xmath225 ) .
we do not correct for the host galaxy dust absorption . from this analysis we excluded :
1 ) the bursts for which no spectroscopic redshift was available ( grb 041218 , grb 050713a , grb 060108 , grb 060203 and grb 060204b ) ; 2 ) grb 060510b , for which there is a value of the spectroscopic redshift but the optical light curve is sparsely sampled . after this selection our sub - sample totals 16 objects . even in the rest frame of the burst , starting the observations about 0.5 minutes after the burst event , a difference of about 4 orders of magnitude in luminosity is evident , particularly at early time .
this spread in intrinsic luminosity remains after including all the available early and late time public data , although there is a hint of a convergence at later times .
it should be noted that our analysis does not take into account any beaming effect .
this collimation correction is known to reduce the observed spread in luminosity for bursts ( frail et al .
2001 , panaitescu @xmath4 kumar 2001 ) but requires a correct determination of the jet opening angle for each burst based on an unambiguous identification of a jet break .
identification of such break times is clearly important but , as discussed in section 4.1.1 is non - trivial due to complex light curve properties and requires well - sampled optical and x - ray light curves from the earliest to the latest possible times . in a previous study of optical afterglow light curves , liang @xmath4 zhang ( 2006 , hereafter lz06 )
suggested that the optical luminosity at t=1 day ( source frame time ) after the burst shows a bimodal distribution , with a separation at @xmath226 erg s@xmath14 .
the majority of the bursts in their sample ( 44 bursts in total ) fall into the luminous group ( 34 bursts with @xmath227 ) .
kann et al .
2006 and nardini et al .
2006 ( hereafter k06 and n06 respectively ) found a similar result .
lz06 selected t=1 day as the reference time because at this time the light curves of their sample were better sampled .
moreover they selected this late time because they were concerned about the possible contribution of the reverse shock component or additional energy injection at early times .
the result of n06 was obtained in the same way but extrapolating the luminosity at t=0.5 days ( source frame time ) .
as our observations have good coverage starting at earlier time ( between 1 and 20 minutes in the grb rest - frame ) we have estimated the intrinsic optical luminosity at three different times : 10 minutes , 0.5 days and 1 day ( source frame time ) . in the cases presented here
we have confirmed that the reverse shock component does not affect our analysis .
as discussed in the previous section , in only one case ( grb 061126 ) do we detect the possible contribution of the reverse shock at early times .
our early observations at t=10 minutes are more directly related to the explosion energy during the prompt emission phase . in table 4
we report the mean values for the rest - frame luminosity calculated at different times .
the two classes defined by lz06 ( dim and lum ) are not consistent with a single population , as clear from their figure 2 . our data ( at any times )
are consistent within the uncertainties with a single population rather than with two separate classes . in fig [ figlum12h ]
we show the observed luminosity distribution of our sample extrapolated at 12 hours.the distribution is well fitted with a single log - normal function with an average of @xmath228 and a @xmath229 of @xmath230 .
figure [ figall2 ] ( right panel ) shows the luminosity - redshift distribution for the bursts in our sample . in this figure
we over plot the separation line between the two classes and the highest values for the redshift of the members of the two groups in the lz06 sample .
as pointed out by lz06 , a possible reason for the lack of high redshift members in their dim group is a lack of deep and rapid followup observations .
the burst population detected by _
swift _ has a larger mean redshift and fainter brightness distribution that previous missions ( @xmath231 , @xmath232 , le @xmath4 dermer 2007 ) .
this could explain the results of lz06 , whose sample was based on bursts detected up to august 2005 , thus containing only 7 _ swift _ bursts .
the study presented by k06 of a sample of 16 pre-_swift _ bursts similarly probed the bright end of the grb luminosity function and found similar conclusions to lz06 .
k06 found that on average low - redshift afterglows are less luminous than high - redshift ones , suggesting a bimodal luminosity distribution .
strong selection effects due to observational bias against intrinsically faint afterglows at higher redshifts is a likely explanation for this result .
this observational bias is greatly reduced in our sample thanks to the rapid response and use of red filters on our ground based telescopes to _ swift _ triggers (
14/16 objects used for this analysis were detected by _ swift _ ) .
our results show that faint _
swift _ grbs at higher redshift are readily detected with such rapid , deep optical observations in red filters ( figure [ figall2 ] , right panel ) , a region of parameter space not accessible in the samples of lz06 or k06 .
the population that is not prevalent in the right panel of figure [ figall2 ] is bright bursts at low redshifts , probably due to the grb luminosity function such that very luminous bursts rare and a large survey volume is therefore required to detect them .
other authors discuss the possibility of two separate luminosity functions for luminous and underluminous grbs ( nardini et al .
2007 , kann et al .
2007 , liang et al .
2007 , chapman et al . 2007 ) but given the many complex instrumental selection effects inherent in grb discovery and followup , significantly larger samples are required to draw robust conclusions .
swift _ trigger thresholds may provide the basis for such samples in the future .
we consider as an optically dark event , or `` dark burst '' to be a one that satisfies the definition of jakobsson et al .
( 2004 ) , i.e. that the slope of the spectral energy distribution between the optical and the x - ray band or spectral index @xmath25 , is @xmath2 0.5 .
even optically detected bursts may be classified as `` dark '' , providing that the optical flux is much fainter than expected from scaling the x - ray flux ( e.g. grb060108 ) .
for all the bursts reported in table 2 no optical counterpart was detected by our telescopes .
apart from a few cases in which our observations were performed at late times , the majority of grbs in our sample were observed by our telescopes reacting rapidly and performing deep , early - time optical observations ( @xmath233 at 5 minutes after the trigger , for co - added images ) . a late response or poor sensitivity
are therefore ruled out as explanations for non - detections in most cases . what , therefore , is the explanation for the lack ( or faintness ) of the optical afterglow for these dark bursts ? to understand this we analyse the x - ray light curve of the bursts in our sub - sample observed by the _
swift_-xrt . using the decay inferred from the fit of the light curves the x - ray and optical fluxes
are extrapolated to a common time .
we assume that the optical light curves for those undetected bursts follow a power law decay with a slope equal to the mean temporal decay of the detected afterglows analysed in section 3 ( @xmath234 ) . alternatively ,
in the x - ray band we use the value derived from the fit of the light curve and reported on table 2 ( @xmath73 ) .
three characteristic times t@xmath235 , t@xmath236 and t@xmath237 are used here . the time t@xmath237 = 11 hours is chosen for consistency with the dark burst classification of jakobsson et al .
( 2004 ) , while fluxes extrapolated to t@xmath236 = 1 hour and t@xmath235 = 10 minutes exploit our early - time data without compromising observing sensitivity . as can be seen in figure [ figuljako ] , the majority of the bursts are located close to the dark bursts region ( @xmath238 ) independent of the selected time , ruling out late observation time as an explanation for the apparent darkness of most bursts . for almost all the bursts ,
the evolution of the optical and x - ray flux follows a line almost parallel to the lines of constant @xmath25 ( bottom right panel ) .
this behaviour can be seen , in the optical band , as a consequence of the assumption of an average decay when making the extrapolation .
any change in the temporal decay would clearly modify this behaviour .
however , it seems that bursts that are classified as `` normal '' after 10 minutes ( bottom left panel ) remain in the same class also after 11 hours ( top left panel ) . at the same time those bursts that are optically dark soon after 10 minutes belong to the class of the so - called `` dark bursts '' also after 11 hours . in only five cases ,
does the classification of the burst depend on the time for the extrapolation of the flux .
for these five bursts we report also the errors ( including the uncertainties on the flux and on the @xmath11 used for the extrapolation ) in the bottom right panel of figure [ figuljako ] .
three bursts are classified as normal bursts if we extrapolate at t = t@xmath237 but belong to the class of `` dark bursts '' if we extrapolate at t = t@xmath235 ( grb 050124 , grb 060901 and grb 070721b , red circles on figure [ figuljako ] ) ; in two cases it is the contrary ( grb 050504 and grb 070219 , blue circles ) . in the latter cases it is clear that the optical flux significantly suppressed compared to the x - ray flux .
this may be due to spectral evolution of those bursts , but within the uncertainties we can not rule out the possibility that nothing changes also for these five bursts . for 10/39 bursts , as reported in table 2 , an optical / infrared counterpart was identified by larger optical or ir telescopes , with counterparts detected primarily at nir wavelengths .
our observations rule out the large population of bright optical counterparts that were predicted , pre-_swift _ , to exist and be observable with suitably rapid followup observations .
explanations for dark bursts in the era of rapid followup remain : extinction caused by dust ( galactic or host ) , high redshift origin , or both . in some cases ( @xmath239 )
the galactic absorption along the light of sight in the observing band ( a@xmath240 ) for our bursts is significant and may explain the undetected optical counterpart ; however , it is interesting to note that no ir detections have been reported for these grbs .
the effect of ly-@xmath11 absorption due to an high - redshift event ( @xmath241 ) is difficult to evaluate but again this effect could be responsible at least for a fraction of our non - detections ( roming et al .
the possibility of a rapid temporal decay seems to be the most unlikely : in fact , assuming that the undetected bursts of table 2 have a temporal behaviour similar to the detected afterglows of table 1 than the temporal optical decay at early times appears to be shallow , not steep .
another possibility could be an excess of x - ray emission at late time ; if late - time central engine activity is responsible for the production of the early x - ray afterglow in some cases , the additional emission will mask the forward shock x - ray emission and the total flux in the x - ray band would be higher than the value expected for the forward shock emission alone .
this might explain some dark bursts and their distribution on the log @xmath242-log @xmath243 diagram .
a combination of these mechanisms and others ( i.e. intrinsic optical faintness , low density circumburst medium ) may combine to explain the high number of bursts that remain undetected at optical wavelengths ( @xmath244 in our sample ) .
* we have classified our afterglows sample into four groups based on breaks in the optical and the x - ray afterglow light curves during the decay phase .
we have used the temporal and spectral properties of the x - ray and optical afterglows to investigate the blastwave physics around the break times within the framework of the standard fireball model ( the synchrotron shock model ) .
the majority of the bursts in our sample ( 15 out of 24 ) are consistent with the standard model . however , for a significant fraction of our sample ( 9 bursts : grb 050713a , grb 051111 , grb 060204b , grb 060512 , grb 060927 , grb 061121 , grb 061126 , grb 070208 and grb 070419a ) , the data can not be explained by the standard model , even if modifications to the simple model are made ( i.e. energy injection or variation in the ambient matter ) .
a possible explanation beyond the standard model is that the early x - ray afterglow is not due to forward shock emission but is instead produced by late - time central engine activity ( e.g. ghisellini et al .
* we have derived the light curves of the optical afterglows in the source rest frame for those bursts with spectroscopically confirmed redshifts ( i.e. not merely assuming a fixed redshift z=1 for all bursts ) .
the optical luminosity function measured at @xmath245 mins and the corresponding distributions for light curves extrapolated to @xmath246 hours and 1 day are uni - modal , showing no evidence for the bi - modality suggested by previous authors .
a fit of the distribution at 10 minutes with a single log - normal yields an average and a sigma values of @xmath247(erg s@xmath14)@xmath248 and @xmath249 , respectively .
liang @xmath4 zhang ( 2006 ) reported a bimodal distribution of optical luminosity at t= 1 day .
two recent studies on the afterglows of _ swift_-era grbs ( kann et al .
2007 and nardini et al . 2007 ) also suggested a clustering of optical afterglow luminosities at one day and 12 hours ( already found by kann et al .
2006 , and nardini et al .
2006 ) , showing again a bi - modality in the luminosity distribution .
this discrepancy may be explained by our ability to detect fainter grbs at high redshift ; in future , larger samples covering a wide range of grb luminosities ( possibly facilitated by lower triggering thresholds on _ swift _ ) will provide stronger tests for the existence of separate classes of grbs .
* by comparing x - ray flux densities and optical upper limits , we have shown that the majority of non - detections in our sample should be classified as dark bursts .
the rapid response of our telescopes to real - time localisations from _ swift _ show that there remains a significant number of genuinely dark grb afterglows and rapid optical temporal decay at early time is ruled out as an explanation for failure to detect optical afterglows at later time . of our 39 non - detections ,
ten afterglows were identified by other facilities , primarily at nir wavelengths , demostrating a small population of bursts in high density host environments .
the lack of optical / ir afterglows for the remaining 29 bursts may be due to effects such as high levels of extinction ( galactic or host ) , circumburst absorption , ly-@xmath11 absorption due to high - redshift or low - density environments suppressing production of optical sychrotron ( or a combination of effects ) .
alternatively , we suggest that if late - time central engine activity is responsible for the production of the early x - ray afterglow emission in some cases , the additional emission will mask the simultaneous , but fainter forward shock x - ray emission and result in an observed x - ray flux that is larger than expected from forward shock emission alone .
this might explain some dark bursts .
am acknowledge founding from the particle physics and astronomy research council ( pparc ) .
cgm acknowledges financial support from the royal society and research councils uk .
the liverpool telescope is operated by liverpool john moores university at the observatorio del roque de los muchachos of the instituto de astrofisica de canarias .
the faulkes telescopes , now owned by las cumbres observatory , are operated with support from the dill faulkes educational trust .
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 .
barthelmy s.d .
2005 , space science reviews , 120 , 143 berger e. @xmath4 kulkarni s.r . , 2005a , gcn circ .
2978 berger e. @xmath4 kulkarni s.r . , 2005b , gcn circ .
2983 bloom j.s . ,
foley r.j .
, koceveki d. @xmath4 perley d. , 2006a , gcn circ .
5217 bloom j.s .
, perley d.a .
@xmath4 chen h.w .
, 2006b , gcn circ .
5826 burrows d.n .
et al . , 2005 , space science reviews , 120 , 165 burrows d.n . @xmath4 racusin j. , arxiv : astro - ph/0702.633 butler n. et al .
, 2005 , apj , 629 , 908 butler n. et al
. , 2006 , apj , 652 , 1390 butler n. , kocevski d. , bloom j.s .
@xmath4 curtis j.l . , 2007 ,
apj , 671 , 656 campana s. @xmath4 de luca a. , 2006 , gcn circ .
5157 castro - tirado a. j. et al . , 2007 ,
6168 cenko s.b . , gezari s. , small t. , fox d.b .
@xmath4 berger e. , 2007 , gcn circ .
6322 chapman r. , tanvir n.r .
, priddey r.s . ,
levan a.j . , 2007 , mnras , 382 , 21 chen h. , thompson i. , prochaska j.x .
@xmath4 josh bloom j.s . , 2005 ,
3709 conciatore m.l . ,
stratta g. , perri m. , sato g. @xmath4 burrows d. , 2007 , gcn circ .
6085 covino s. , malesani d. @xmath4 tagliaferri g. , 2006 , gcn circ .
5604 cucchiara a. , fox d.b .
@xmath4 berger e. , 2006 , gcn circ .
4729 cucchiara a. , fox d.b . , cenko s.b .
@xmath4 price d.a . , 2007 ,
6083 curran p. et al .
, 2007 , a@xmath4a , 467 , 1049 davanzo p.
@xmath4 israel g.l . , 2006a , gcn circ .
4893 davanzo p. , covino s. , malesani d. @xmath4 tagliaferri g. , 2006b , gcn circ .
5609 dupree a.k . ,
falco e. , prochaska j.x .
, chen h.w .
@xmath4 bloom j.s . , 2006 , gcn circ
4969 evans p.a .
et al . , 2007 ,
a@xmath4a , 469 , 379 falcone a.d . ,
burrows d.n .
, morris d.c .
@xmath4 gehrels n. , 2006a , gcn circ .
4669 falcone a. et al . , 2006b , gcn circ .
5009 fox d.b . , 2006 , gcn circ .
5605 fox d.b . , rau a. @xmath4 ofek e.o . , 2006a , gcn circ .
5597 fox d.b . , 2006b , gcn circ .
5607 frail d. a. et al . , 2001 , apj , 562 , l55 fugazza d. et al . , 2004 ,
2782 fynbo j.p.u .
et al . , 2006c , gcn circ .
5809 garnavich p. , zhao x. @xmath4 pimenova t. , 2004 , gcn circ .
2792 gehrels n. et al . , 2004 , apj , 611 , 1005 ghisellini g. , celotti a. , ghirlanda g. , firmani c. @xmath4 nava l. , 2007 , mnras , 382 , l72 godet o. , page k.l .
@xmath4 burrows d.n . , 2006 ,
5128 gomboc a. et al . , 2006 , il nuovo cimento , 121b , 1303 gomboc a. et al . , 2008 ,
submitted to apj graham j.f .
et al . , 2007 ,
6836 grupe d. , fox d.b . , racusin j. @xmath4 kennea j. , 2006 , gcn circ .
5811 guetta d. et al . , 2007 ,
a@xmath4a , 461 , 95 guidorzi c. et al .
, 2005a , gcn circ .
3625 guidorzi c. et al .
, 2005b , apj , 630 , l121 guidorzi c. et al . , 2006 , pasp , 118 , 288 guidorzi c. et al . , 2007 , a@xmath4a , 463 , 539 hearty f. et al . , 2006a , gcn circ .
4604 hearty f. et al . , 2006b , gcn circ .
4611 hill g. , prochaska j.x , fox d.b .
schaefer b. @xmath4 reed m. , 2005 , gcn circ .
4255 hunsberger s. et al . , 2005 ,
2982 jakobsson p.
et al . , 2004 , apj , 617 , l21 jakobsson p.
et al . , 2006 ,
a@xmath4a , 447 , 897 jakobsson p. et al . , 2007 ,
6283 jensen b.l . ,
hjorth j. , fynbo j. @xmath4 n@xmath250r@xmath250nen j. , 2006 , gcn circ .
5203 kann d.a . , klose s. @xmath4 zeh a. , 2006 , apj , 641 , 993 kann d.a .
et al . , 2007 , astro - ph/0712.2186 kobayashi s. @xmath4 zhang b. , 2003 , apj , 582 , l75 kobayashi s. @xmath4 zhang b. , 2007 , apj , 655 , 973 kocevski d. , bloom j.s .
@xmath4 mcgrath e.j . , 2006a , gcn circ .
4528 kocevski d. , bloom j.s .
@xmath4 mcgrath e.j .
, 2006b , gcn circ .
4540 le t. @xmath4 dermer c. d. , 2007 , apj , 661 , 394 levan a.j .
et al . , 2006 ,
apj , 648 , l9 liang e. @xmath4 zhang b. , 2006 , apj , 638 , l67 liang e. , zhang b. , virgili f. @xmath4 dai z.g . , 2007 , apj , 662 , 1111 lin z.y .
et al . , 2005 ,
2976 lipunov v. et al . , 2006 ,
4892 malesani d. et al . , 2004 ,
apj , 609 , l5 malesani d. et al . , 2006a , gcn circ .
4541 malesani d. et al . , 2006b , gcn circ .
4561 malesani d. et al . , 2006c , gcn circ .
5705 malesani d. et al . , 2006d , gcn circ .
5718 mszros p.
@xmath4 rees m.j . , 1999 ,
mnras , 306 , l39 mszros p. , 2002 ,
ara@xmath4a , 40 , 137 molinari e. et al . , 2007 , a@xmath4a , 469 , l13 monfardini a. et al . , 2006 , apj , 648 , 1125 moretti a. , romano p. @xmath4 guidorzi c. , 2007 , gcn circ .
6286 morris d.c .
, burrows d.n . , gehrels n. , boyd p. @xmath4 voges w. , 2006a , gcn circ .
4651 morris d.c .
, pagani c. , burrows d.n .
, kennea j.a .
@xmath4 page k.l . , 2006b , gcn circ .
4764 morris d.c .
et al . , 2007 ,
apj , 654 , 413 mundell c. et al .
, 2007a , science , 315 , 1822 mundell c. et al . , 2007b , apj , 660 , 489 nardini m. et al . , 2006 , a@xmath4a , 451 , 821 nardini m. , ghisellini g. @xmath4 ghirlanda g. , 2007 , astro - ph/0801.4759 nousek j.a .
et al . , 2006 ,
apj , 642 , 389 obrien , p.t .
et al . , 2006 ,
apj , 647 1213 oates s. et al . , 2006 , mnras , 372 , 327 osip d. , chen h.w .
@xmath4 prochaska j.x .
, 2007 , gcn circ .
5715 page k.l .
et al . , 2007 , apj , 663 , 1125 panaitescu a. @xmath4 kumar p. , 2001 ,
apj , 560 , l49 panaitescu a. et al .
, 2006 , mnras , 369 , 2059 pandey s.b .
et al . , 2006 ,
a@xmath4a , 460 , 415 perri m. et al . , 2006 ,
5110 perri m. et al . , 2007 , gcn circ .
6333 perley d.a .
et al . , 2007 , astro - ph/0703.538 piran t. , 1999 , physics reports , 314 , 575 price p.a . , 2006 ,
5104 prochaska j.x .
, ellison s. , foley r.j .
, bloom j.s .
@xmath4 chen h.w . , 2005 ,
3332 prochaska j.x .
et al . , 2006 ,
4701 racusin j. , kennea j. , pagani c. , vetere l. @xmath4 evans p. , 2007
, gcn circ .
6627 rees m.j .
@xmath4 mszros p. , 1992 ,
mnras , 258 , 41 rol e. et al . , 2007 , gcn circ .
6221 rol e. et al . , 2007b , mnras , 374 , 1078 ruiz - velasco a.e .
et al . , 2007 , apj , 669 , 1 sari r. , piran t. @xmath4 narayan r. , 1998 , apj , 497 , l17 sari r. , piran t. @xmath4 halpern j.p . , 1999 , apj , 519 , l17 sari r. @xmath4 piran t. , 1999 , apj , 520 , 641 schady p.
et al . , 2007 , mnras , 380 , 1041 schlegel d. et al . , 1998 , apj , 500 , 525 smith j.a .
et al . , 2002 ,
aj , 123 , 2121 soderberg a.m. et al . , 2006 , apj , 636 , 391 stanek k.z .
et al . , 2003 , apj , 591 , l17 stanek k.z .
et al . , 2005 ,
apj , 626 , l5 stratta g. , perri m. , burrows d.n .
@xmath4 stamatikos m. , 2007 , gcn circ .
6337 swan h. , akerlof c. , rykoff e. , yost s. @xmath4 smith i. , 2006 , gcn circ .
4568 tagliaferri g. et al . , 2005 ,
nature , 436 , 985 tanvir n.r .
et al . , 2005 ,
3632 tanvir n.r . ,
levan a.j .
, priddey r.s .
, fruchter a.s .
@xmath4 hjorth j. , 2006a , gcn circ .
4602 tanvir n.r . ,
rol e. , wiersema k. , starling r. @xmath4 omahoney n. , 2006b , gcn circ .
4897 tanvir n.r .
et al . , 2008 , astro - ph/0803.4100 yost s.a .
et al . , 2006 , apj , 636 , 959 yost s.a .
et al . , 2007 ,
apj , 657 , 925 willingale , r. et al . , 2007 , apj , 662 , 1093 woosley s.e .
@xmath4 bloom j.s . , 2006 , ara@xmath4a , 44 , 507 wreeswijk p. @xmath4 jaunsen a. , 2006 , gcn circ .
4974 zhang b. , kobayashi s. @xmath4 mszros p. , 2003 ,
apj , 595 , 950 zhang b. et al . , 2006 , apj , 642 , 354 -band ) light curves for the 24 grb afterglows detected by our telescopes from october 2004 to september 2007 . for each burst
we show the x - ray flux density in @xmath251jy ( black filled squares ) and optical flux density in mjy ( red filled circles for our observations and red open circles for published data , when available ) .
we show also the value of the spectroscopic redshift when available .
x - ray data of _ swift_/xrt are from evans et al .
we superimpose simple power - law fit segments to each curve ( the details of the fit are reported on table 1).,title="fig:",width=566,height=359 ] -band ) light curves for the 24 grb afterglows detected by our telescopes from october 2004 to september 2007 . for each burst
we show the x - ray flux density in @xmath251jy ( black filled squares ) and optical flux density in mjy ( red filled circles for our observations and red open circles for published data , when available ) .
we show also the value of the spectroscopic redshift when available .
x - ray data of _ swift_/xrt are from evans et al .
we superimpose simple power - law fit segments to each curve ( the details of the fit are reported on table 1).,title="fig:",width=566,height=359 ] @cccccccccccccc [ tabsample ] 041006 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & & @xmath256 & & & @xmath257 & @xmath258 & @xmath259 & 0.716 & @xmath260 + 041218 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & & & & & & & & & + 050502a & @xmath265 & & & @xmath266 & @xmath267 & & & & @xmath268 & @xmath269 & @xmath270 & 3.793 & @xmath271 + 050713a & @xmath272 & & & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & & @xmath278 & @xmath279 & & + 050730 & @xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath263 & @xmath286 & @xmath287 & @xmath288 & & 3.967 & @xmath289 + 051111 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & & @xmath294 & & @xmath295 & @xmath296 & @xmath297 & @xmath298 & 1.55 & @xmath299 + 060108 & @xmath300 & & & @xmath301 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & @xmath310 + 060203 & @xmath311 & & & @xmath312 & @xmath313 & & & @xmath314 & & @xmath315 & & & + 060204b & @xmath316 & & & @xmath317 & & @xmath318 & & @xmath319 & & @xmath258 & & & + 060206 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath324 & & & @xmath325 & @xmath326 & @xmath327 & @xmath320 & 4.048 & @xmath328 + 060210 & @xmath329 & ( @xmath330 ) & ( @xmath331 ) & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & 3.91 & @xmath340 + 060418 & @xmath341 & & & @xmath342 & @xmath343 & & & @xmath344 & & @xmath345 & & 1.489 & @xmath346 + 060510b & @xmath347 & & & @xmath348 & @xmath349 & @xmath350 & @xmath351 & @xmath352 & & @xmath353 & & 4.9 & @xmath354 + 060512 & @xmath355 & & & @xmath356 & @xmath357 & & & @xmath358 & & @xmath359 & & 0.4428 & @xmath360 + 060927 & @xmath361 & & & @xmath362 & @xmath363 & @xmath364 & @xmath365 & @xmath366 & & @xmath367 & & 5.467 & @xmath368 + 061007 & @xmath369 & & & @xmath370 & @xmath371 & & & @xmath372 & & @xmath373 & @xmath374 & 1.261 & @xmath375 + 061110b & @xmath376 & & & @xmath377 & @xmath378 & & & @xmath379 & & @xmath380 & & 3.44 & @xmath381 + 061121 & @xmath382 & & & @xmath383 & @xmath384 & @xmath385 & @xmath386 & @xmath387 & - & @xmath388 & @xmath389 & 1.314 & @xmath390 + 061126 & @xmath391 & @xmath392 & @xmath393 & @xmath394 & & @xmath395 & & @xmath396 & @xmath397 & @xmath398 & @xmath399 & 1.158 & @xmath400 + 070208 & @xmath401 & & & @xmath402 & & @xmath403 & & @xmath404 & & @xmath405 & & 1.165 & @xmath406 + 070411 & @xmath407 & & & @xmath408 & @xmath409 & & & @xmath410 & & @xmath411 & & 2.954 & @xmath412 + 070419a & & @xmath61 & & @xmath413 & @xmath414 & @xmath59 & @xmath415 & @xmath416 & & @xmath417 & & 0.97 & @xmath418 + 070420 & @xmath419 & & & @xmath317 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & & @xmath257 & & & + 070714b & @xmath424 & & & @xmath425 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & & @xmath430 & & 0.92 & @xmath431 + @ccccccccrrrrrcccc [ tabul ] 041211 & no & 30.2 & 100 & & & & & 197.20 & 19.18 & 242.94 & 20.86 & 4.5 & & 0.45 & & + 050124 & yes & 4.1 & 12.3 & 6.9 & 185.2 & & 0.3 & 885.64 & 19.00 & 886.02 & 19.20 & 0.5 & ir & 0.09 & @xmath432 & 0.103 + 050128 & yes & 13.8 & 51.7 & 24.0 & 3.62 & & & 697.20 & 21.13 & 788.4 & 21.85 & 20.0 & & 0.21 & @xmath433 & 0.296 + 050412 & yes & 26.0 & 5.66 & 0.39 & 166.7 & 1.35 & 0.4 & 2.5 & 18.7 & 3.83 & 20.82 & 11.5 & & 0.05 & @xmath434 & 6.790 + 050504 & yes & 80.0 & 15.0 & & 326.8 & & & 3.7 & 19.00 & 17.39 & 20.33 & 17.67 & & 0.03 & @xmath435 & 0.043 + 050520 & yes & 80.0 & 24.0 & 0.01 & 127.7 & 1.4 & & 4.5 & 16.60 & 8.34 & 19.4 & 2.5 & & 0.04 & & + 050528 & no & 10.8 & 4.40 & & 849.0 & & & 2.5 & 17.2 & 3.88 & 17.96 & 1.0 & & 0.43 & & + 050713b & yes & 75.0 & 45.7 & 90.2 & 2.27 & 2.88 & 0.7 & 3.3 & 18.2 & 3.80 & 19.32 & 0.5 & & 1.25 & @xmath436 & 52.947 + 050716 & yes & 69.0 & 63.2 & 70.2 & 3.83 & 1.68 & 0.32 & 3.8 & 19.8 & 4.29 & 20.61 & 8.5 & ir & 0.29 & @xmath437 & 58.970 + 050925 & no & 0.068 & 0.75 & & 1.66 & & & 3.3 & 19.0 & 3.69 & 21.12 & 2.0 & & 5.69 & & + 051211a & no & 4.2 & 9.2 & & & & & 353.4 & 20.86 & 413.4 & 21.72 & 60.5 & & 0.32 & & + 051211b & yes & 80.0 & 20.0 & 0.13 & 179.7 & 1.16 & 1.0 & 66.1 & 16.5 & 67.4 & 17.0 & 1.0 & & 1.26 & @xmath438 & 0.322 + 060114 & no & 100.0 & 13.0 & & & & & 2.2 & 19.0 & 37.59 & 20.70 & 15.5 & & 0.09 & & + 060116 & yes & 113.0 & 26.0 & 0.9 & 2.57 & 0.95 & 1.1 & 18.72 & 18.58 & 39.98 & 20.0 & 8.5 & ir & 0.69 & @xmath439 & 0.990 + 060121 & yes & 2.0 & 43.0 & 0.46 & 176.70 & 1.08 & 1.07 & 50.23 & 19.5 & 175.56 & 22.22 & 20.83 & o & 0.04 & @xmath265 & 0.644 + 060204c & no & 60.0 & 3.5 & 0.001 & 2.6 & & & 6.42 & 18.7 & 6.89 & 19.39 & 1.0 & & 0.49 & & + 060319 & yes & 12.0 & 2.7 & 2.2 & 2.88 & 1.02 & 1.10 & 7.0 & 19.0 & 9.90 & 21.63 & 3.0 & ir & 0.06 & @xmath440 & 4.559 + 060602a & yes & 60.0 & 16.0 & & & & & 7.72 & 15.0 & 36.08 & 16.83 & 21.0 & o & 0.07 & @xmath441 & 1.595 + 060602b & yes & 9.0 & 1.8 & 0.32 & 1.38 & 1.05 & 2.1 & 19.2 & 18.0 & 38.20 & 20.85 & 2.5 & & 95.63 & & + 060825 & yes & 8.1 & 9.8 & 3.57 & 1.1 & 0.87 & 0.64 & 4.43 & 18.7 & 7.31 & 19.47 & 3.0 & & 1.55 & @xmath442 & 1.626 + 060901 & yes & 20.0 & 7.0 & 0.26 & 226.0 & 1.7 & 1.1 & 142.8 & 21.0 & 177.6 & 22.10 & 20.0 & & 1.85 & @xmath443 & 0.553 + 060923a & yes & 51.7 & 8.7 & 4.9 & 1.35 & 2.7 & 1.1 & 2.8 & 19.0 & 8.88 & 19.90 & 2.0 &
ir & 0.16 & @xmath444 & 1.732 + 060923c & yes & 76.0 & 16.0 & 85.0 & 3.38 & 3.4 & 0.85 & 4.22 & 19.0 & 14.50 & 20.3 & 3.67 & ir & 0.17 & @xmath445 & 0.700 + 060929 & yes & 12.4 & 2.8 & 0.53 & 1.53 & 0.79 & 1.3 & 21.13 & 19.0 & 25.88 & 20.36 & 3.0 & & 0.13 & @xmath446 & 0.396 + 060930 & no & 20.0 & 2.5 & & & & & 1.98 & 19.5 & 6.45 & 20.85 & 1.5 & & 0.22 & & + 061006 & yes & 0.5 & 14.3 & 0.19 & 2.38 & 2.26 & 0.7 & 22.61 & 18.0 & 23.05 & 18.20 & 0.5 & o & 0.85 & @xmath447 & 0.306 + 061210 & yes & 0.2 & 11.0 & & & & & 2.39 & 17.0 & 2.8 & 18.20 & 0.5 & & 0.09 & @xmath448 & 118664.2 + 070103 & yes & 19.0 & 3.4 & 0.38 & 1.15 & 1.4 & 1.3 & 23.7 & 19.0 & 25.20 & 19.45 & 1.0 & & 0.18 & @xmath449 & 3.00 + 070219 & yes & 17.0 & 3.2 & 0.12 & 1.37 & 2.2 & 1.0 & 51.4 & 19.9 & 59.09 & 20.64 & 5.0 & & 0.09 & @xmath450 & 0.104 + 070220 & yes & 129.0 & 106.0 & 16.5 & 1.32 & 1.76 & 0.55 & 1.93 & 19.5 & 34.01 & 20.47 & 10.0 & & 2.41 & @xmath451 & 3.834 + 070223 & yes & 89.0 & 17.0 & 92.0 & 1.83 & 2.3 & 0.7 & 18.7 & 21.4 & 38.84 & 22.29 & 13.0 & ir & 0.04 & @xmath452 & 0.407 + 070412 & yes & 34.0 & 4.8 & 33.0 & 1.02 & 0.98 & 1.2 & 14.1 & 21.0 & 35.68 & 21.89 & 13.0 & & 0.06 & @xmath453 & 1.34 + 070521 & yes & 37.9 & 80.0 & 3.2 & 1.28 & 0.5 & 1.11 & 2.35 & 19.3 & 31.77 & 22.70 & 12.0 & & 0.07 & @xmath454 & 8.02 + 070531 & yes & 44.0 & 11.0 & & 2.13 & & & 11.9 & 18.2 & 14.64 & 19.45 & 3.0 & & 1.00 & @xmath455 & 1.01 + 070704 & yes & 380.0 & 59.0 & 285.0 & 2.55 & 0.87 & 0.85 & 239.5 & 21.1 & 254.34 & 22.11 & 30.0 & & 5.01 & @xmath456 & 0.150 + 070721a & yes & 3.4 & 0.71 & 0.823 & 1.43 & 2.97 & 1.24 & 229.2 & 19.0 & 261.72 & 20.08 & 15.5 & & 0.04 & @xmath457 & 0.089 + 070721b & yes & 340.0 & 36.0 & 24.5 & 1.53 & 0.9 & 0.48 & 327.0 & 18.5 & 364.44 & 19.31 & 15.17 & & 0.08 & @xmath458 & 0.195 + 070808 & yes & 32.0 & 12.0 & 1.0 & 1.9 & 3.5 & 1.8 & 2.35 & 19.7 & 36.69 & 20.42 & 16.0 & & 0.06 & @xmath459 & 0.283 + 070810b & yes & 80.0 & 0.12 & & 1.03 & & & 2.80 & 20.0 & 40.63 & 21.15 & 14.0 & & 0.14 & & + | we present a multiwavelength analysis of 63 gamma - ray bursts observed with the world s three largest robotic optical telescopes , the liverpool and faulkes telescopes ( north and south ) .
optical emission was detected for 24 grbs with brightnesses ranging from @xmath0 = 10 to 22 mag in the first 10 minutes after the burst . by comparing optical and x - ray light curves from t = 100 to @xmath1 seconds
, we introduce four main classes , defined by the presence or absence of temporal breaks at optical and/or x - ray wavelengths .
while 15/24 grbs can be modelled with the forward - shock model , explanation of the remaining nine is very challenging in the standard framework even with the introduction of energy injection or an ambient density gradient .
early x - ray afterglows , even segments of light curves described by a power - law , may be due to additional emission from the central engine .
39 grbs in our sample were not detected and have deep upper limits ( @xmath0 @xmath2 22 mag ) at early time .
of these , only ten were identified by other facilities , primarily at near infrared wavelengths , resulting in a dark burst fraction of @xmath350% .
additional emission in the early time x - ray afterglow due to late - time central engine activity may also explain some dark bursts by making the bursts brighter than expected in the x - ray band compared to the optical band .
[ firstpage ] |
the timed automata of alur and dill @xcite are a popular model for describing timed behaviours .
while these automata have the plus point of being very expressive and having a decidable emptiness problem , they are neither determinisable nor closed under complementation .
this is a drawback from a couple of points of view .
firstly , one can not carry out model checking in the framework where a system is modelled as a timed transition system @xmath0 and a specification of timed behaviours as a timed automaton @xmath1 , and where one asks `` is @xmath2 ? '' .
this would normally involve _ complementing _ @xmath1 and then checking if its intersection with @xmath0 is non - empty .
one can get around this problem to some extent by using determinisable specifications , or specifying directly the negation of the required property .
a second reason why lack of closure properties may concern us is that it precludes the existence of an unrestricted logical characterisation of the class of languages accepted by timed automata .
the existence of a monadic second order logic ( mso ) characterisation of a class of languages is a strong endorsement of the `` regularity '' of the class .
it also helps in identifying expressively complete temporal logics , which are natural to use as specification languages and have relatively efficient model checking algorithms .
the event clock automata of @xcite was one of the first steps towards identifying a subclass of timed automata with the required closure properties .
they were shown to be determinisable in @xcite , and later to admit a robust logical framework in terms of an mso characterisation and an expressively complete timed temporal logic @xcite .
similar results were shown in @xcite , @xcite and @xcite .
a common technique used in all these results was the idea of `` implicit '' clocks , whose values are determined solely by the timed word being read .
for example the event recording clock @xmath3 records the time since the last @xmath4 action w.r.t . the current position in a timed word , and is thus implicitly reset with each @xmath4 action .
the truth of a guard over these clocks at a point in a timed word is thus completely determined by the word itself , unlike in a timed automaton where the value of a clock depends on the path taken in the automaton . in this paper
we generalise the notion of an implicit clock to that of an _ input determined operator_. an input determined operator @xmath5 identifies for a given timed word and position in it , a set of intervals in which it is `` satisfied '' .
the guard @xmath6 is then satisfied at a point in a timed word if the set of intervals identified by @xmath5 contains @xmath7 .
for example , the event recording clock @xmath3 can be modelled as an input determined operator @xmath8 which identifies at a given point in a timed word , the ( infinite ) set of intervals containing the distance to the last @xmath4 action .
the guard @xmath9 now translates to @xmath10 . as an example to show that this framework is more general than implicit clocks , consider the input determined operator @xmath11 inspired by the metric temporal logic ( mtl ) of @xcite
this operator identifies the set of all intervals @xmath7 for which there is a future occurrence of an @xmath4 at a distance which lies in @xmath7 .
the guard @xmath12 is now true iff there is a future occurrence of an @xmath4 action , at a distance which lies in @xmath7
. timed automata which use guards based on a set of input determined operators are what we call _ input determined automata_. we show that input determined automata form a robust class of timed languages , in that they are ( a ) determinisable , ( b ) effectively closed under boolean operations , ( c ) admit a logical characterisation via an unrestricted mso , and ( d ) identify a natural expressively complete timed temporal logic .
we then go over to a more expressive framework using the idea of _ recursive _ event clocks from @xcite . in the recursive version of our input determined operator ,
the operators now expect a third parameter ( apart from the timed word and a position in it ) which identifies a set of positions in the timed word
. this argument could be ( recursively ) another input determined automaton , or as is better illustrated , a temporal logic formula @xmath13 .
the formula @xmath13 naturally identifies a set of positions in a timed word where the formula is satisfied .
thus a recursive operator @xmath5 along with the formula @xmath13 , written @xmath14 , behaves like an input determined operator above , and the guard @xmath15 is true iff the set of intervals identified by @xmath14 contains @xmath7 .
these recursive input determined automata are also shown to admit similar robust logical properties above
. we should be careful to point out here that , firstly , these results hold in the _ pointwise _ semantics , where formulas are evaluated only at the `` action points '' in a timed word ( used e.g. in @xcite ) , and not at arbitrary points in between actions in a timed word as allowed in the _ continuous _ semantics of @xcite .
secondly , we make no claims about the existence of _ decision procedures _ for these automata and logics .
in fact it can be seen the operator @xmath11 above takes us out of the class of timed automata as we can define the language of timed sequences of @xmath4 s in which no two @xmath4 s are a distance 1 apart , with a single state input determined automaton which has the guard @xmath16 \in { \ensuremath{\mbox{\boldmath $ \diamond$}}}_a)$ ] .
similar versions can be seen to have undecidable emptiness problems and correspondingly undecidable logics @xcite .
thus the contribution of this paper should be seen more in terms of a general framework for displaying logical characterisations of timed automata , and proving expressive completeness of temporal logics related to these automata . many of the results along these lines from @xcite and some in the pointwise semantics from @xcite follow from the results in this paper . as a new application of this framework
, we provide an expressive completeness result for mitl in the pointwise semantics , by showing that it is expressively equivalent to the first order fragment of an mso based on recursive operators .
this answers an open question from @xcite , apart from identifying an interesting class of timed automata .
the techniques used in this paper essentially build on those from @xcite and @xcite which use the notion of _ proper _ symbolic alphabets and factor through the results of bchi @xcite and kamp @xcite .
the idea of using recursive operators comes from @xcite , who show a variety of expressiveness results , including an expressive completeness for mitl in the continuous semantics .
their result for mitl is more interesting in that it uses event - clock modalities , while we use essentially the same modalities as mitl .
however , our mso is more natural as unlike the mso in @xcite it has unrestricted second order quantification .
we use @xmath17 to denote the set of natural numbers @xmath18 , and @xmath19 and @xmath20 to denote the set of non - negative reals and rationals respectively .
the set of finite and infinite words over an alphabet @xmath21 will be denoted by @xmath22 and @xmath23 respectively .
we use the notation @xmath24 to denote the set of functions from @xmath25 to @xmath26 .
an _ ( infinite ) timed word _ over an alphabet @xmath27 is an element @xmath28 of @xmath29 satisfying the following conditions .
let @xmath30 . then : 1 .
( _ monotonicity _ ) for each @xmath31 , @xmath32 , 2 .
( _ progressiveness _ ) for each @xmath33 there exists @xmath31 such that @xmath34 .
let @xmath35 denote the set of infinite timed words over @xmath27 .
where convenient , we will use the representation of @xmath28 as @xmath36 where @xmath37 and @xmath38 is a time sequence satisfying the conditions above .
we will use rational bounded intervals to specify timing constraints .
these intervals can be open or closed , and we allow @xmath39 as an open right end .
these intervals denote a subset of reals in the usual manner for example @xmath40 denotes the set @xmath41 .
the set of all such intervals is denoted @xmath42 .
our input determined automata will use guards of the form `` @xmath43 '' , where @xmath7 is an interval and @xmath5 is an operator which determines for a given timed word @xmath28 and a position @xmath44 in it , a set of intervals `` satisfying '' it at that point .
we then say that @xmath28 at position @xmath44 satisfies the guard `` @xmath45 '' if @xmath7 belongs to the set of intervals identified by @xmath5 . by a `` position '' in the timed word we mean one of the `` action points '' or instants given by the time - stamp sequence , and use natural numbers @xmath44 ( instead of the time @xmath46 ) to denote these positions .
more formally , an input determined operator @xmath5 ( w.r.t . the alphabet @xmath27 ) has a semantic function @xmath47 .
the guard @xmath6 is satisfied at position @xmath44 in @xmath48 iff @xmath49 .
the transitions of our input determined automata are labelled by symbolic actions of the form @xmath50 where @xmath4 is an action , and @xmath51 is a guard which is a boolean combination of atomic guards of the form @xmath6 .
the set of guards over a finite set of input determined operators @xmath52 is denoted by @xmath53 and given by the syntax @xmath54 .
the satisfaction of a guard @xmath51 in a timed word @xmath28 at position @xmath44 , written @xmath55 , is given in the expected way : we have @xmath56 always , @xmath57 as above , and the boolean operators @xmath58 , @xmath59 , and @xmath60 interpreted as usual . a
_ symbolic alphabet _
@xmath61 based on @xmath62 is a finite subset of @xmath63 .
an infinite word @xmath64 in @xmath65 specifies in a natural way a subset of timed words @xmath66 defined as follows .
let @xmath67 for each @xmath31 .
let @xmath68 with @xmath69 for each @xmath31 .
then @xmath70 iff for each @xmath71 , @xmath72 and @xmath73 .
we extend the map @xmath74 to work on subsets of @xmath75 in the natural way .
thus , for @xmath76 , we define @xmath77 .
finally , we denote the vocabulary of intervals mentioned in @xmath61 by @xmath78 .
recall that a bchi automaton over an alphabet @xmath21 is a structure @xmath79 where @xmath80 is a finite set of states , @xmath81 is an initial state , @xmath82 is the transition relation , and @xmath83 is a set of accepting states . let @xmath84 .
a run of @xmath1 over @xmath85 is a map @xmath86 which satisfies : @xmath87 and @xmath88 for every @xmath31 .
we say @xmath89 is an _ accepting _ run of @xmath1 on @xmath85 if @xmath90 for infinitely many @xmath31 .
the set of words accepted by @xmath1 , denoted here as @xmath91 ( for the `` symbolic '' language accepted by @xmath1 ) , is defined to be the set of words in @xmath23 on which @xmath1 has an accepting run .
we are now in a position to define an input determined automaton .
an _ input determined automaton _
( ida for short ) over an alphabet @xmath27 and a set of operators @xmath52 , is simply a bchi automaton over a symbolic alphabet based on @xmath92 . viewed as a bchi automaton over a symbolic alphabet @xmath61
, an input determined automaton @xmath1 accepts the language @xmath93 which we call the symbolic language accepted by @xmath1 .
however , we will be more interested in the timed language accepted by @xmath1 : this is denoted @xmath94 and is defined to be @xmath95 . to give a concrete illustration of input determined automata , we show how
the event clock automata of @xcite can be realized in the above framework .
take @xmath52 to be the set of operators @xmath96 , where the operators @xmath8 and @xmath97 essentially record the time since the last @xmath4 action , and the time to the next @xmath4 action .
the operator @xmath8 ( and similarly @xmath97 ) can be defined here by setting @xmath98 to be @xmath99 as another example which we will use later in the paper , consider the operator @xmath11 related to mtl @xcite .
the guard @xmath100 is meant to be true in a word @xmath28 at time @xmath44 iff there is a future instant @xmath101 labelled @xmath4 and the distance to it lies in @xmath7 i.e. @xmath102 . the guard @xmath103 makes a similar assertion about the
past _ of @xmath28 w.r.t . the current position .
an input determined automaton based on these operators can be defined by taking @xmath104 , and where , for example , @xmath105 .
we now want to show that the class of timed languages accepted by input determined automata ( for a given choice of @xmath27 and @xmath52 ) is closed under boolean operations .
the notion of a _ proper _ symbolic alphabet will play an important role here and subsequently .
a _ proper symbolic alphabet _ based on @xmath62 is of the form @xmath106 where @xmath107 is a finite subset of @xmath42 .
an element of @xmath61 is thus of the form @xmath108 , where the set of intervals specified by @xmath109 is interpreted as the _ exact _ subset of intervals in @xmath78 which are satisfied by @xmath5 .
this is formalised in the following definition of @xmath110 for a proper symbolic alphabet @xmath61 .
let @xmath111 with @xmath112 .
let @xmath68 with @xmath69 .
then @xmath113 iff for each @xmath71 : @xmath72 and for each @xmath114 , @xmath115 .
let @xmath61 be a proper symbolic alphabet based on @xmath62 .
then a bchi automaton @xmath1 over @xmath61 , which we call a _ proper _ ida over @xmath116 , determines a timed language over @xmath27 given by @xmath117 .
the class of timed languages defined by ida s and proper ida s over @xmath62 coincide .
an ida over a symbolic alphabet @xmath61 can be converted to an equivalent one ( in terms of the timed language they define ) over a proper symbolic alphabet @xmath118 .
firstly , each transition label @xmath50 in @xmath61 can be written in a disjunctive normal form @xmath119 , with each @xmath120 being a conjunction of literals @xmath6 or @xmath121 .
thus each transition labelled @xmath50 can be replaced by a set of transitions labelled @xmath122 , one for each @xmath44 .
now each transition labelled @xmath123 , with @xmath124 a conjunct guard , can be replaced by a set of transitions @xmath108 , one for each @xmath125 `` consistent '' with @xmath124 : i.e. @xmath125 should satisfy the condition that if @xmath6 is one of the conjuncts in @xmath124 then @xmath126 , and if @xmath127 is one of the conjuncts in @xmath124 then @xmath128 . in the other direction , to go from a proper ida to an ida , a label @xmath108 of a proper symbolic alphabet can be replaced by the guard @xmath129 the following property of proper symbolic alphabets will play a crucial role . [
lemma : proper1 ] let @xmath61 be a proper symbolic alphabet based on @xmath27 .
then for any @xmath68 there is a _
unique _ symbolic word @xmath64 in @xmath75 such that @xmath130 .
let @xmath131 .
then the only possible symbolic word @xmath64 we can use must be given by @xmath112 , where @xmath115 .
in the light of lemma [ lemma : proper1 ] , going from a symbolic alphabet to a proper one can be viewed as a step towards determinising the automaton with respect to its timed language . from here
one can simply use classical automata theoretic techniques to determinise the automaton w.r.t
. its _ symbolic _ language .
( of course , since we deal with infinite words we will need to go from a bchi to a muller or rabin acceptance condition @xcite ) .
[ thm : closure1 ] the class of ida s over @xmath62 are effectively closed under the boolean operations of union , intersection , and complement .
it is sufficient to address union and complementation .
given automata @xmath1 and @xmath132 over symbolic alphabets @xmath61 and @xmath133 respectively , we can simply construct an automaton over @xmath134 which accepts the union of the two symbolic languages . for complementing the timed language of @xmath1
, we can go over to an equivalent proper ida @xmath135 over a proper symbolic alphabet @xmath136 , and now simply complement the symbolic language accepted by @xmath135 to get an automaton @xmath137 .
it is easy to verify , using the uniqueness property of proper alphabets given in lemma [ lemma : proper1 ] , that @xmath138 . in the constructions above we have made use of the closure properties of @xmath139-regular languages @xcite .
we now show that input determined automata admit a natural characterisation via a timed mso in the spirit of @xcite .
recall that for an alphabet @xmath21 , bchi s monadic second order logic ( denoted here by @xmath140 ) is given as follows : @xmath141 the logic is interpreted over a word @xmath84 , along with an interpretation @xmath142 which assigns individual variables @xmath143 a position in @xmath85 ( i.e. an @xmath31 ) , and to set variables @xmath25 a set of positions @xmath144 .
the relation @xmath145 is interpreted as the usual ordering of natural numbers , and the predicate @xmath146 ( one for each @xmath147 ) as the set of positions in @xmath85 labelled @xmath4 .
the formal semantics of the logic is given below .
for an interpretation @xmath142 we use the notation @xmath148 $ ] to denote the interpretation which sends @xmath143 to @xmath44 and agrees with @xmath142 on all other variables .
similarly , @xmath149 $ ] denotes the modification of @xmath142 which maps the set variable @xmath25 to a subset @xmath150 of @xmath17 .
later we will also use the notation @xmath151 $ ] to denote the interpretation with sends @xmath143 to @xmath44 when the rest of the interpretation is irrelevant .
@xmath152 \models \varphi.\\ \alpha , \mathbb{i } \models \exists x \varphi & \mathrm{\ iff \ } & \mathrm { there \ exists \ } s \subseteq { \ensuremath{\mathbb{n}}}\mathrm { \ such \ that \ } \sigma,\mathbb{i}[s / x ] \models \varphi . \end{array}\ ] ] for a sentence @xmath153 ( i.e. a formula without free variables ) in @xmath140 we set @xmath154 bchi s result then states that a language @xmath155 is accepted by a bchi automaton over @xmath21 iff @xmath156 for a sentence @xmath153 in @xmath140 .
we define a timed mso called @xmath157 , parameterised by the alphabet @xmath27 and set of input determined operators @xmath52 , whose syntax is given by : @xmath158 in the predicate `` @xmath159 '' , @xmath7 is an interval in @xmath42 , @xmath114 , and @xmath143 is a variable .
the logic is interpreted in a similar manner to @xmath160 , except that models are now timed words over @xmath27 . in particular , for a timed word @xmath161 , we have : @xmath162 given a sentence @xmath153 in @xmath163 we define @xmath164 .
[ thm : msoba1 ] a timed language @xmath165 is accepted by an input determined automaton over @xmath62 iff @xmath156 for some sentence @xmath153 in @xmath166 . given an ida @xmath1 over @xmath62 we can give a @xmath167 sentence @xmath153 which describes the existence of an accepting run of @xmath1 on a timed word .
following @xcite , for @xmath168 with @xmath169 , we can take @xmath153 to be the sentence @xmath170 here @xmath171 denotes the formula obtained by replacing each @xmath43 in @xmath51 by @xmath159 . further
, `` @xmath172 '' abbreviates @xmath173 where @xmath174 in turn stands for @xmath175 .
similarly @xmath176 can be expressed via @xmath177 , where @xmath178 is the formula @xmath179 . in the converse direction
we take the route used in @xcite as it will be useful in the sequel .
let @xmath153 be a formula in @xmath180 , and let @xmath61 be a _ proper _ symbolic alphabet with the same interval vocabulary as @xmath153 .
we give a way of translating @xmath153 to a formula @xmath181 in @xmath182 in such a way that the timed languages are preserved .
the translation @xmath183 is done with respect to @xmath61 and simply replaces each occurrence of @xmath184 the translation preserves the timed models of a formula @xmath153 in the following sense : let @xmath68 , @xmath111 , and @xmath113 .
let @xmath142 be an interpretation for variables
. then @xmath185 iff @xmath186 .
the lemma is easy to prove using induction on the structure of the formula @xmath153 and making use of the properties of proper symbolic alphabets . from the lemma it immediately follows now that for a sentence @xmath153 in @xmath157 , we have @xmath187 , and this is the sense in which the translation preserves timed languages
. we can now argue the converse direction of theorem [ thm : msoba1 ] using this translation and factoring through bchi s theorem .
let @xmath153 be a sentence in @xmath180 and let @xmath188 .
then by bchi s theorem we have an automaton @xmath1 over @xmath61 which recognises exactly @xmath189 .
thus @xmath1 is our required proper ida since @xmath190 .
in this section we identify a natural , expressively complete , timed temporal logic based on input determined operators .
the logic is denoted @xmath192 , parameterised by the alphabet @xmath27 and set of input determined operators @xmath52 .
the formulas of @xmath192 are given by : @xmath193 here we require @xmath194 , @xmath195 , and @xmath114 . the models for @xmath196 formulas are timed words over @xmath27 .
let @xmath68 , with @xmath161 , and let @xmath31 .
then the satisfaction relation @xmath197 is given by @xmath198 we define @xmath199 .
let us denote by @xmath200 the first - order fragment of @xmath180 ( i.e. the fragment we get by disallowing quantification over set variables ) .
the logics @xmath201 and @xmath202 are _ expressively equivalent _ in the following sense : [ thm : ltl1 ] a timed language @xmath165 is definable by a @xmath196 formula @xmath13 iff it is definable by a sentence @xmath153 in @xmath203 . given a @xmath196 formula @xmath13 we can associate an @xmath200 formula @xmath153 which has a single free variable @xmath143 , and satisfies the property that @xmath204 iff @xmath205 \models \varphi$ ] .
this can be done in a straightforward inductive manner as follows . for the atomic formulas @xmath4 and @xmath6 we can take @xmath153 to be @xmath206 and @xmath159 respectively . in the inductive step , assuming we have already translated @xmath13 and @xmath207 into @xmath153 and @xmath208 respectively , we can translate @xmath209 into @xmath210 { \ensuremath{\wedge}}\forall z ( ( x \leq z { \ensuremath{\wedge}}z
\leq y ) { \ensuremath{\ , \rightarrow \,}}\varphi[z / x])).\ ] ] here @xmath211 $ ] denotes the standard renaming of the free variable @xmath143 to @xmath212 in @xmath208 .
the remaining modalities are handled in a similar way , and we can verify that if @xmath153 is the above translation of @xmath13 then @xmath204 iff @xmath205 \models \varphi$ ] .
it also follows that @xmath213 satisfies @xmath13 iff @xmath28 satisfies the sentence @xmath214 given by @xmath215 .
hence we have that @xmath216 . in the converse direction a more transparent proof is obtained by factoring through kamp s result for classical ltl .
recall that the syntax of @xmath217 is given by : @xmath218 where @xmath147 .
the semantics is given in a similar manner to @xmath201 , except that models are words in @xmath23 .
in particular the satisfaction relation @xmath219 for the atomic formula @xmath4 is given by : @xmath220 iff @xmath221 .
let @xmath222 denote the first - order fragment of @xmath140 .
then the result due to kamp @xcite states that : [ thm : kamp1 ] @xmath217 is expressively equivalent to @xmath222 .
consider now a proper symbolic alphabet @xmath61 based on @xmath62 . we can define a timed language preserving translation of an @xmath223 formula @xmath224 to a formula @xmath225 in @xmath196 . in the translation @xmath226
we replace subformulas @xmath108 by @xmath227 it is easy to argue along the lines of lemma [ lemma : proper1 ] that [ lemma : stot - ltl ] let @xmath68 and @xmath111 with @xmath113
. then @xmath228 iff @xmath229 .
hence we have @xmath230 .
we can now translate a sentence @xmath153 in @xmath200 to an equivalent @xmath196 formula @xmath13 as follows .
let @xmath61 be the proper symbolic alphabet based on @xmath62 with the same interval vocabulary as @xmath153 .
let @xmath231 be the @xmath232 formula @xmath181 .
note that the translation @xmath226 preserves first - orderness and hence @xmath231 belongs to @xmath232 .
now by theorem [ thm : kamp1 ] , we have a formula @xmath224 in @xmath223 which is equivalent to @xmath231 .
we now use the translation @xmath183 on the formula @xmath224 to get a @xmath196 formula @xmath13 .
@xmath13 is our required @xmath196 formula .
observe that firstly @xmath233 by the property of the translation @xmath226 .
next , by kamp s theorem we have that @xmath234 and hence @xmath235 .
but by the property of the translation @xmath183 applied to @xmath153 , we have @xmath236 , and hence we can conclude that @xmath237 .
this completes the proof of theorem [ thm : ltl1 ] .
we point out here that the past temporal operators of @xmath238 ( `` previous '' ) and @xmath239 ( `` since '' ) can be dropped from our logic without affecting the expressiveness of the logic .
this follows since it is shown in @xcite that theorem [ thm : kamp1 ] holds for the future fragment of @xmath191 .
the reason we retain the past operators is because they are needed when we consider a recursive version of the logic in section [ section : rltl ] .
we now consider `` recursive '' input determined operators . the main motivation is to increase the expressive power of our automata , as well as to characterise the expressiveness of recursive temporal logics which occur naturally in the real - time setting .
to introduce recursion in our operators , we need to consider _ parameterised _ ( or _ recursive _ ) input determined operators .
these operators , which we continue to denote by @xmath5 , have a semantic function @xmath240 , whose first argument is a subset of positions @xmath25 .
thus @xmath5 with the parameter @xmath25 determines an input determined operator of the type introduced earlier , whose semantic function is given by the map @xmath241 .
the set of positions @xmath25 will typically be specified by a temporal logic formula or a `` floating '' automaton , in the sense that given a timed word @xmath28 , the formula ( resp .
automaton ) will identify a set of positions in @xmath28 where the formula is satisfied ( resp .
automaton accepts ) .
these ideas will soon be made more precise .
we first recall the idea of a `` floating '' automaton introduced in @xcite . these are automata which accept pairs of the form @xmath242 with @xmath28 a timed word , and @xmath44 a position ( i.e. @xmath243 .
we will represent a `` floating '' word @xmath244 as a timed word over @xmath245 .
thus a timed word @xmath246 over @xmath245 represents the floating word @xmath242 , iff @xmath247 , with @xmath248 with a _
@xmath249 in the @xmath44-th position , and @xmath161 .
we use @xmath250 to denote the ( partial ) map which given a timed word @xmath246 over @xmath245 returns the floating word @xmath251 corresponding to @xmath246 , and extend it to apply to timed languages over @xmath245 in the natural way .
let @xmath52 be a set of input determined operators w.r.t . @xmath27 . then
a _ floating ida _ over @xmath62 is an idaover @xmath252 , where the set of operators @xmath253 w.r.t .
@xmath245 is defined to be @xmath254 where @xmath255 is a timed word over @xmath245 , with @xmath256 and @xmath257 .
thus the operator @xmath258 simply ignores the @xmath259 component of @xmath255 and behaves like @xmath5 on the @xmath27 component . a floating ida @xmath132 accepts the floating timed language @xmath260 .
we now give a more precise definition of recursive input determined automata , denoted rec - ida , and their floating counterparts frec - ida .
let @xmath261 be a finite set of recursive input determined operators .
then the class of rec - ida s over @xmath262 , and the timed languages they accept , are defined as follows .
* every ida @xmath1 over @xmath27 that uses only the guard @xmath263 is a rec - ida over @xmath264 , and accepts the timed language @xmath94 . + similarly , every floating ida @xmath132 over @xmath27 which uses only the guard @xmath263 is a frec - ida over @xmath262 , and accepts the floating language @xmath265 .
* let @xmath266 be a finite collection of frec - ida s over @xmath262 .
let @xmath52 be the set of input determined operators @xmath267 , where the semantic function of each @xmath268 is given as follows .
let @xmath269 denote the set of positions @xmath44 such that @xmath270
. then @xmath271 . +
then any ida @xmath1 over @xmath62 is a rec - ida over @xmath262 , and accepts the timed language @xmath94 ( defined in section [ section : idta ] ) .
+ similarly every floating ida @xmath132 over @xmath62 is a frec - ida over @xmath262 , and accepts the floating language @xmath265 .
recursive automata fall into a natural `` level '' based on the level of nesting of operators they use
. a rec - ida is of _ level _ 0 if the only guard it uses is @xmath263 .
similarly a frec - ida is of level 0 , if the only guard it uses is @xmath263
. a rec - ida is of _ level _
( i+1 ) if it uses an operator @xmath268 , with @xmath272 and @xmath132 a frec - ida of level @xmath44 , and no operator @xmath273 with @xmath274 and @xmath137 of level greater than @xmath44 .
a similar definition of level applies to frec - ida s .
as an example consider the level 1 rec - ida @xmath1 over the alphabet @xmath275 below .
the floating automaton @xmath132 accepts a floating word @xmath242 iff the position @xmath44 is labelled @xmath276 and the previous and next positions are labelled @xmath4 .
the recursive input determined operator @xmath277 is defined formally in sec .
[ section : mitl ] .
the rec - ida @xmath1 thus recognises the set of timed words @xmath28 over @xmath275 which begin with an @xmath4 and have an occurrence of @xmath276 with @xmath4 s on its left and right exactly 1 time unit later .
[ thm : closure2 ] the class of rec - ida s over @xmath262 is closed under boolean operations . in fact , for each @xmath44 , the class of level @xmath44 rec - ida s is closed under boolean operations .
let @xmath1 and @xmath135 be two rec - ida s of level @xmath44 .
let @xmath52 be the union of operators used in @xmath1 and @xmath135 .
then both @xmath1 and @xmath135 are ida s over @xmath62 , and hence by theorem [ thm : closure1 ] there exists an ida @xmath132 over @xmath62 which accepts @xmath278 .
similarly there exists an ida @xmath137 over @xmath62 , which accepts the language @xmath279 .
notice that @xmath132 and @xmath137 use the same set of operators @xmath52 , and hence are also level @xmath44 automata .
we note that ida s over @xmath62 are a special case of level 1 rec - ida s over @xmath262 , where the set of recursive operators @xmath261 is taken to be @xmath280 with @xmath281 .
thus each guard @xmath6 in an ida over @xmath62 can be replaced by the guard @xmath282 , for any `` dummy '' level 0 frec - ida @xmath132 .
we now introduce a recursive version of @xmath167 which will characterise the class of timed languages defined by rec - ida s .
the logic is parameterised by an alphabet @xmath27 and set of recursive input determined operators @xmath261 , and denoted @xmath283 .
the syntax of the logic is given by @xmath284 in the predicate @xmath285 , we have @xmath195 , @xmath286 , and @xmath208 a @xmath283 formula with a single free variable @xmath287 .
the logic is interpreted over timed words in @xmath35 .
its semantics is similar to @xmath167 except for the predicate `` @xmath288 '' which is defined inductively as follows . if @xmath208 is a formula which uses no @xmath5 predicates , then the satisfaction relation @xmath289 is defined as for @xmath167 .
inductively , assuming the semantics of @xmath208 has already been defined , @xmath290 is interpreted as an input determined operator as follows .
let @xmath291 denote the set of interpretations for @xmath287 that make @xmath208 true in the timed word @xmath28
i.e. @xmath292 \models \psi\}$ ]
. then @xmath293 thus we have @xmath294 note that the variable @xmath287 , which is free in @xmath208 , is _ not _ free in the formula @xmath288 .
a sentence @xmath153 in @xmath283 defines the language @xmath295 , and a @xmath283 formula @xmath208 with one free variable @xmath287 defines a floating language @xmath296 \models \psi \}$ ] .
we note that each @xmath283 formula @xmath153 can be viewed as a @xmath180 formula , for a suitably defined set of input determined operators @xmath52 .
we say an operator @xmath290 has a _ top - level _ occurrence in @xmath153 if there is an occurrence of @xmath290 in @xmath153 which is _ not _ in the scope of any @xmath258 operator .
we can now take @xmath52 to be the set of all top - level operators @xmath297 in @xmath153 .
analogous to the notion of level for rec - ida s we can define the _ level _ of an @xmath298 formula @xmath153 .
the level of @xmath153 is 0 , if @xmath153 uses no @xmath5 predicates .
@xmath153 has level @xmath299 if it uses a predicate of the form @xmath288 with @xmath208 a level @xmath44 formula , and _ no _ predicate of the form @xmath300 with @xmath301 of level greater than @xmath44 . as an example the level 1 sentence @xmath153 below defines the same timed language as the level 1 rec - ida @xmath1 defined in section [ section : idta ] .
we can take @xmath153 to be @xmath302 \in { \ensuremath{\mbox{\boldmath $ \diamond$}}}_{\psi}(x ) ) ) ) , $ ] where @xmath208 is the level 0 formula @xmath303 .
[ thm : msoba2 ] @xmath165 is accepted by a rec - ida over @xmath304 iff @xmath305 is definable by a @xmath283 sentence .
in fact , we will show that for each @xmath44 , the class of rec - ida s of level @xmath44 correspond to the sentences of @xmath283 of level @xmath44 . but first it will be useful to state a characterisation of floating languages along the lines of theorem [ thm : msoba1 ] .
[ thm : msobaf1 ] let @xmath305 be a a floating language over @xmath27 .
then @xmath306 for some floating ida over @xmath62 iff @xmath307 , for some @xmath180 formula @xmath208 with one free variable .
let @xmath132 be a floating ida over @xmath62 .
keeping in mind that @xmath132 runs over the alphabet @xmath308 , we define a formula @xmath208 with one free variable @xmath287 as follows .
@xmath208 is the formula @xmath153 given in the proof of theorem [ thm : msoba1 ] , except for the clause ( * ) which we replace by @xmath309 the formula @xmath208 satisfies @xmath310 iff @xmath311 \models \psi$ ] . in the converse direction , let @xmath312 denote a @xmath180 formula with free variables @xmath313 .
an interpretation @xmath142 for these variables is encoded ( along with @xmath28 ) as a timed word over @xmath314 .
we extend the definition of a floating ida to an ida which works over such an alphabet , where , in particular , the @xmath5 operators apply only to the @xmath27 component of the timed word
. then we can inductively associate with @xmath312 a floating ida @xmath132 over @xmath245 such that @xmath315 . in the inductive step for @xmath316
we make use of the fact that the class of languages accepted by floating ida s over @xmath62 are closed under the restricted renaming operation required in this case .
the reader is referred to @xcite for a similar argument .
returning now to the proof of theorem [ thm : msoba2 ] , we use induction on the level of automata and formulas to argue that 1 .
@xmath165 is accepted by a level @xmath44 rec - ida over @xmath262 iff @xmath305 is definable by a level @xmath44 @xmath283 sentence @xmath153 . and 2 .
[ item : float ] a floating language @xmath305 over @xmath27 is accepted by a level @xmath44 frec - ida over @xmath262 iff @xmath305 is definable by a level @xmath44 @xmath283 formula @xmath208 with one free variable . for the base case we consider level 0 automata and sentences . since level 0 automata only make use of the guard @xmath263 , they are simply bchi automata over @xmath27 .
similarly , level 0 sentences do nt use any @xmath5 predicates and hence they are simply @xmath317 sentences . by bchi
s theorem , we have that level 0 automata and sentences are expressively equivalent .
for the base case for the second part of the claim , given a level 0 floating automaton @xmath132 we can apply the construction in the proof of theorem [ thm : msobaf1 ] to get a @xmath163 formula @xmath208 with one free variable . since the construction preserves the guards used , @xmath208 has no @xmath5 operators , and hence is a level 0 @xmath283 formula .
conversely , for a level 0 formula @xmath208 we can apply the construction of theorem [ thm : msobaf1 ] to obtain a floating automaton @xmath132 such that @xmath318 .
the construction preserves the @xmath5 operators used , and hence @xmath132 is a level 0 automaton .
turning now to the induction step , let @xmath1 be a level @xmath299 automaton over @xmath262 .
let @xmath52 be the set of top - level @xmath5 operators in @xmath1 .
now since @xmath1 is an ida over @xmath62 , by theorem [ thm : msoba1 ] , we have a @xmath180 sentence @xmath153 such that @xmath319 .
now for each @xmath268 in @xmath52 , @xmath132 is of level @xmath44 or lower , and by our induction hypothesis there is a corresponding @xmath283 formula @xmath208 with one free variable , of the same level as @xmath132 , with @xmath318 .
hence for each @xmath268 we have a semantically equivalent operator @xmath290 .
this is because @xmath318 , which implies @xmath320 , which in turn implies @xmath321 .
we can now simply replace each occurrence of @xmath268 in @xmath153 to get an equivalent sentence @xmath322 which is in @xmath283 .
further , by construction it follows that @xmath322 is also of level @xmath299 .
conversely , let @xmath153 be a level @xmath299 sentence in @xmath283 .
let @xmath52 be the set of top level @xmath5 operators in @xmath153 .
then @xmath153 is a @xmath180 sentence , and hence by theorem [ thm : msoba1 ] we have an equivalent input determined automaton @xmath1 over @xmath62 . once again
, for each @xmath290 in @xmath52 , the formula @xmath208 is of level @xmath44 or lower , and hence by induction hypothesis we have a frec - ida @xmath132 over @xmath262 , of the same level as @xmath208 , and accepting the same floating language .
the operators @xmath290 and @xmath268 are now equivalent , and we can replace each @xmath290 in @xmath1 by the corresponding @xmath268 to get a language equivalent input determined automaton .
this automaton is now the required level @xmath299 rec - ida over @xmath262 which accepts the same language as @xmath323 .
the induction step for part [ item : float ] is proved similarly , making use of theorem [ thm : msobaf1 ] and the induction hypothesis .
this completes the proof of theorem [ thm : msoba2 ] .
we now define a recursive timed temporal logic along the lines of @xcite .
the logic is similar to the logic @xmath201 defined in sec .
[ section : tltl ] .
it is parameterised by an alphabet @xmath27 and a set of recursive input determined operators @xmath261 , and denoted @xmath325 .
the syntax of the logic is given by @xmath326 where @xmath194 , and @xmath272 .
the logic is interpreted over timed words in a similar manner to @xmath201 .
the predicate @xmath15 is interpreted as follows .
if @xmath13 does not use a @xmath5 predicate , then the satisfaction relation @xmath204 is defined as for @xmath201 . inductively assuming the semantics of a @xmath325 formula @xmath13 has been defined , and setting @xmath327 , the operator @xmath14 is interpreted as an input determined operator with the semantic function @xmath328 the satisfaction relation @xmath329 is then defined as in @xmath201 . once again , since @xmath14 behaves like an input determined operator , each @xmath325 formula is also a @xmath330 formula , for an appropriately chosen set of input determined operators @xmath52 , containing operators of the form @xmath14 .
a @xmath325 formula @xmath13 naturally defines both a timed language @xmath331 and a floating language @xmath332 . as an example ,
the formula @xmath333 \in { \ensuremath{\mbox{\boldmath $ \diamond$}}}_{\theta})$ ] where @xmath334 , restates the property expressed by the @xmath298 formula in sec .
[ section : mso2 ] .
let us denote by @xmath335 the first - order fragment of the logic @xmath283 .
then we have the following expressive completeness result : [ thm : ltl2 ] @xmath325 is expressively equivalent to @xmath335 . as before we show that formulas in the logics are equivalent level - wise ( the level of a @xmath336 formula being defined analogous to @xmath298 ) .
we show by induction on @xmath44 that 1 .
[ item : ltl2:1 ] a timed language @xmath165 is definable by a level @xmath44 @xmath325 formula iff it is definable by a level @xmath44 @xmath335 sentence .
[ item : ltl2:2 ] a floating timed language over @xmath27 is definable by a level @xmath44 @xmath325 formula iff it is definable by a level @xmath44 @xmath335 formula with one free variable .
the base case for part [ item : ltl2:1 ] follows from theorem [ thm : kamp1 ] , since level 0 formulas are simply untimed @xmath337 and @xmath338 formulas . for the base case for part [ item : ltl2:2 ] , a level 0 @xmath325 formula @xmath13 can be translated to a level 0 @xmath335 formula @xmath208 with one free variable @xmath287 using the translation given in the proof of theorem [ thm : ltl1 ] .
the formula @xmath208 satisfies @xmath311 \models \psi$ ] iff @xmath204 .
the converse direction follows immediately from the following version of kamp s result : [ thm : kamp2 ] for any @xmath222 formula @xmath208 with one free variable @xmath287 , there is a @xmath217 formula @xmath13 s.t . for each @xmath339 and @xmath31 , @xmath340 \models \psi$ ] iff @xmath219 .
turning now to the induction step , let @xmath13 be a level @xmath299 @xmath325 formula .
let @xmath52 be the set of top - level @xmath5 operators used in @xmath13 .
then @xmath13 is a @xmath196 formula , and hence by theorem [ thm : ltl1 ] we have an equivalent @xmath200 sentence @xmath153 ( i.e. with @xmath341 ) .
now each operator in @xmath52 is of the form @xmath342 where @xmath207 is a level @xmath44 or less @xmath325 formula , and hence by the induction hypothesis we have an equivalent @xmath335 formula @xmath208 with one free variable , such that @xmath343 .
it now follows that the input determined operators @xmath342 and @xmath290 are semantically equivalent , and hence we can replace each @xmath342 by @xmath290 in @xmath153 to get an equivalent @xmath335 sentence @xmath322 . by construction , the sentence @xmath322 is also of level @xmath299 .
the converse direction is argued in a very similar manner , once again factoring through theorem [ thm : ltl1 ] . for part [ item : ltl2:2 ]
, a level @xmath299 @xmath325 formula @xmath13 is a @xmath196 formula , for the set of operators @xmath52 defined above . now using the translation given in the proof of theorem [ thm : ltl1 ] we obtain a @xmath200 formula @xmath208 with a one free variable , satisfying @xmath344 .
again , by the induction hypothesis , we can replace each @xmath342 in @xmath52 with an equivalent @xmath345 , to get an equivalent @xmath335 with the required properties .
in the converse direction , let @xmath208 be a level @xmath299 @xmath335 formula with one free variable @xmath287 .
let @xmath52 be set of top - level @xmath5 operators in @xmath208 .
then @xmath208 is also a formula in @xmath200 .
let @xmath61 be the proper symbolic alphabet induced by @xmath208 .
then we can use the translation @xmath183 ( cf .
sec [ section : tmso ] ) on @xmath208 ( w.r.t .
@xmath61 ) to get a formula @xmath346 in @xmath232 with one free variable @xmath287 which preserves timed models . by kamp s theorem above
, we have an equivalent @xmath223 formula @xmath224 which preserves the floating language accepted .
finally we can apply the translation @xmath226 on @xmath224 to get a @xmath196 formula @xmath13 which preserves timed models ( cf .
[ section : tltl ] ) .
the formula @xmath13 satisfies the property that @xmath347 . now using the induction hypothesis each operator @xmath345 in @xmath13
can be replaced by an equivalent @xmath342 operator , with @xmath207 a @xmath196 formula , to get an equivalent level @xmath299 @xmath325 formula @xmath348 .
this ends the proof of theorem [ thm : ltl2 ] .
as an application of the results in this paper we show that the logic @xmath349 introduced in @xcite is expressively equivalent to @xmath350 for a suitably defined set of recursive input determined operators .
we point out here that this result is shown for the pointwise semantics of @xmath349 given below .
we begin with the logic @xmath351 which has the following syntax @xcite : @xmath352 here @xmath7 is an interval in @xmath42 . when @xmath7 is restricted to be _ non - singular _
( i.e. not of the form @xmath353 $ ] ) then we get the logic @xmath354 .
the logic is interpreted over timed words in @xmath35 similarly to @xmath201 .
the modalities @xmath355 and @xmath356 are interpreted as follows , for a timed word @xmath161 .
@xmath357 we first observe that @xmath351 is expressively equivalent to its sublogic @xmath358 in which the modalities @xmath359 and @xmath360 are replaced by the modalities @xmath361 , @xmath150 , @xmath362 and @xmath363 , where @xmath361 and @xmath150 are as usual and @xmath364 and @xmath365 .
this is because the formula @xmath366 ( and dually @xmath367 ) can be translated as follows . here ` @xmath368 ' denotes either a ` @xmath369 $ ] ' or ` @xmath370 ' interval bracket .
@xmath371}(\theta { \ensuremath{u}}(\theta { \ensuremath{\wedge}}{\ensuremath{o}}{\eta } ) ) & \mathrm{if \ }
i = ( a , b\rangle , a>0 \\ { \ensuremath{\diamond}}_{i } \eta { \ensuremath{\wedge}}(\theta { \ensuremath{u}}\eta ) & \mathrm{if \ }
i = [ 0,b\rangle \\ { \ensuremath{\diamond}}_{i } \eta { \ensuremath{\wedge}}(\theta { \ensuremath{u}}(\theta { \ensuremath{\wedge}}{\ensuremath{o}}{\eta } ) ) & \mathrm{if \ } i = ( 0,b\rangle .
\end{array } \right . \end{array}\ ] ] next we consider the logic @xmath372 where the semantics of the recursive input determined operators @xmath277 and @xmath373 are given below ( as usual @xmath68 with @xmath161 ) .
@xmath374 the logic @xmath358 is clearly expressively equivalent to @xmath375 since the predicates @xmath376 and @xmath377 are equivalent . using theorem [ thm : ltl2 ] we can now conclude that let @xmath379 denote the restriction of @xmath350 to non - singular intervals .
then since the translation of @xmath380 to @xmath381 does not introduce any singular intervals , and the constructions in theorem [ thm : ltl2 ] preserve the interval vocabulary of the formulas , we conclude that the logics @xmath354 and @xmath382 are expressively equivalent . | we consider a general notion of timed automata with _ input - determined _ guards and show that they admit a robust logical framework along the lines of @xcite , in terms of a monadic second order logic characterisation and an expressively complete timed temporal logic .
we then generalise these automata using the notion of recursive operators introduced by henzinger , raskin , and schobbens @xcite , and show that they admit a similar logical framework .
these results hold in the `` pointwise '' semantics .
we finally use this framework to show that the real - time logic mitl of alur et al @xcite is expressively complete with respect to an mso corresponding to an appropriate input - determined operator .
keywords : timed automata , monadic second - order logic , real - time temporal logics |
to date , over 50 exoplanetary systems with more than one planet have been discovered , including : the extraordinary detections of the first exoplanets orbiting the pulsar psr b1257 + 12 @xcite ; the imaged system orbiting hr 8799 ; those discovered during the microlensing event ogle-2006-blg-109l @xcite ; several systems discovered by transit , including four or five _ multiply _ transiting systems from the kepler mission @xcite ; and 43 systems discovered by radial velocity ( rv ) searches @xcite .
the rv systems include the four - planet systems @xmath1 ara @xcite , gj 581 @xcite and gj 876 @xcite and the five - planet system orbiting 55 cancri @xcite . of all these multplanet systems , only four are known to host three or more giant ] planets with well - determined orbital paramaters : @xmath2 and @xcite , hip 14810 @xcite , @xmath1 ara @xcite and hd 37124 @xcite .
hd 37124 ( hip 26381 ) is a 0.85 @xmath3 metal - poor ( @xmath4$]=-0.44 , * ? ? ?
* ) g4 dwarf ( v=7.7 ) .
@xcite announced the a jovian , @xmath5 d planet orbiting hd 37124 from hires data taken at keck observatory as part of the california and carnegie planet search .
further monitoring of the star revealed substantial long - term residuals .
@xcite fit these residuals with an eccentric , 1940 d planet , but noted that the solution was not unique ( and @xcite showed that this fit was , in fact , unstable . ) after collecting two more years of data , @xcite was able to report the detection of a third planet in the system , though with an ambiguity : while the @xmath6 and @xmath7 components had clearly defined periods , the @xmath8 component could be fit nearly equally well with periods of either 2300 d or 29.32 d , the latter likely being an alias due to the lunar cycle .
@xcite reported that recent keck velocities had resolved the ambiguity qualitatively in favor of the longer orbital period .
@xcite explored the many possible dynamical configurations consistent with the @xcite velocities , including many resonant solutions .
@xcite used the system to demonstrate a fast mengo algorithm , but they did not explore the 2:1 resonance , as the data did not seem to favor it at the time .
we present new keck observations , and these data provide a unique orbital solution for the outer planet .
the outer planet period we find is more consistent with the original period reported by @xcite than the refined orbit of @xcite component , despite the prior 1940 d fit of @xcite , because that prior fit was so speculative , and because their new fit put the very existence of a 1940 d periodicity in some doubt . ]
( though we find a much lower eccentricity ) . herein , we present the entire history of keck velocities obtained for this star , and present self - consistent orbital solutions showing that the outer two planets are in or very near a 2:1 mean - motion resonance ( mmr )
. this is the 20th exoplanetary system to be found near an mmr , and only the tenth system with an apparent 2:1 commensurability .
period commensurabilities ( pcs ) represent important dynamical indicators in the solar system and have been linked with observables and formation mechanisms @xcite .
the near-5:2 pc of jupiter and saturn , also known as the great inequality " , might be the remnant of a divergent resonant crossing that produced the current architecture of the outer solar system , the late heavy bombardment , and the trojan asteroids @xcite .
the populations of the asteroid belt and the kuiper belt , exemplified by the pc and near - pc - populated kirkwood gaps ( e.g. * ? ? ?
* ) the plutinos ( 3:2 pcs with pluto and neptune ) and the twotinos ( 2:1 pcs with pluto and neptune ; e.g. * ? ? ?
* ; * ? ? ?
* ) , have implications for the migratory history of jupiter and neptune and the prospect of , e.g. secular resonant sweeping ( e.g. , * ? ? ?
near - pcs found in satellite and ring systems have had direct observational consequences ; the saturnian satellite pandora was @xmath9 behind its predicted orbital longitude in a 1995 ring plane crossing @xcite due to its 121:118 pc with neighboring satellite prometheus . by extension
, we may anticipate similar importance in the growing number of exoplanetary systems exhibiting pcs .
in extrasolar systems , mean motion resonances ( mmrs ) have been interpreted as the indication of convergent migration in multi - planet systems , ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
several subsequent studies @xcite exploring convergent migration for a variety of masses , separations and disk properties have found many regions of mass and orbital element phase space in which planets are easily captured through this mechanism .
table [ velocities ] contains radial velocity measurements for hd 37124 from the hires spectrograph @xcite at keck observatory obtained by the california planet survey consortium using the iodine technique @xcite .
note that the quoted errors are our internal ( random ) errors , with no `` jitter '' included @xcite .
these velocities supersede our previously published velocities for this star , as we continue to refine our data reduction pipeline . our ever - evolving radial velocity pipeline is descended in spirit and form from that described in @xcite , but includes many small and large technical improvements , a thorough discussion of which is beyond the scope of this manuscript .
some details can be found in 4.1 of @xcite , 3 of @xcite , and in batalha et
al.2011 ( apj , accepted ) . on issue of instant relevance
is that in august 2004 the hires ccd detector was upgraded to a ccd mosaic .
the old tektronix 2048 eb2 engineering - grade ccd displayed a variable instrumental profile asymmetry due to a charge transfer inefficiency which manifested itself as small changes in a star s measured radial velocity as a function of exposure time ( i.e. raw counts on the chip . )
we apply an emperical , spectral - type dependent model to correct this effect for velocities measured prior to the detector change .
the new ccd mosaic shows no evidence of this effect , but as a consequence of the switch there is a small velocity offset between data sets that span the two detector sets similar to the detector - to - detector offsets discussed in @xcite .
these offsets could , in principle , be different for every target .
analysis of rv standards and known planetary systems show that such an offset is usually small of order 5 m / s and very often consistent with zero . as a result ,
we report two independent data sets for this system in table [ velocities ] , one from each of the two detectors .
we solve for the detector offset as an unconstrained free parameter .
the times of observation are given in jd-2440000 .
we fitted the data using the publicly available multi - planet rv - fitting idl package ` rv_fit_mp ` , described in @xcite . in table
[ orbital ] we present our 3-planet keplerian ( kinematic ) fit , which yields r.m.s .
residuals of 4.4 m@xmath10 , and we plot the fit and velocities in figure [ fig ] . we find a best - fit offset between ccds of @xmath11 m@xmath10 .
the orbital parameters and their uncertainties were determined from 100 bootstrapped trials ( as described in * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the orbital fits and dynamical analysis herein are put forth under the assumption that the velocities are not detectably influenced by additional , unmodeled planets in the system .
we have integrated these orbital parameters for 10 myr using the methods described in [ newton ] assuming coplanarity , and found them to yield a stable configuration .
cccc 10420.04655 & 54.81 & 1.5 & 1 + 10546.73646 & 28.74 & 1.2 & 1 + 10837.76625 & 6.32 & 1.6 & 1 + 10838.94868 & 6.38 & 1.7 & 1 + 10861.80464 & 17.89 & 1.4 & 1 + 11069.03617 & -3.99 & 1.4 & 1 + 11070.13190 & -1.66 & 1.3 & 1 + 11071.11494 & 1.28 & 1.6 & 1 + 11072.12947 & -11.58 & 1.5 & 1 + 11073.02962 & -8.72 & 1.3 & 1 + 11172.89571 & 39.40 & 1.6 & 1 + 11226.78065 & -0.48 & 1.4 & 1 + 11227.78167 & -2.05 & 1.4 & 1 + 11228.74293 & -11.01 & 1.3 & 1 + 11412.14161 & -33.76 & 1.6 & 1 + 11543.98278 & -31.36 & 1.4 & 1 + 11550.94262 & -44.57 & 1.4 & 1 + 11551.94008 & -46.20 & 1.5 & 1 + 11552.89162 & -47.97 & 1.5 & 1 + 11580.76121 & -36.28 & 1.8 & 1 + 11581.83559 & -36.17 & 1.4 & 1 + 11582.78849 & -37.36 & 1.4 & 1 + 11583.72387 & -35.69 & 1.5 & 1 + 11884.04436 & -14.73 & 1.6 & 1 + 11900.03518 & 3.01 & 1.4 & 1 + 11974.80019 & 39.56 & 1.4 & 1 + 12007.74522 & 4.48 & 1.5 & 1 + 12242.99064 & 48.20 & 1.5 & 1 + 12333.94545 & -19.13 & 1.7 & 1 + 12334.78556 & -12.82 & 1.7 & 1 + 12536.12848 & 10.64 & 1.8 & 1 + 12537.08597 & 10.28 & 1.7 & 1 + 12573.03767 & 27.38 & 1.6 & 1 + 12574.99934 & 27.48 & 1.7 & 1 + 12576.02212 & 23.58 & 1.6 & 1 + 12600.99996 & 8.06 & 1.7 & 1 + 12602.03213 & 7.02 & 1.7 & 1 + 12925.01639 & 13.03 & 1.7 & 1 + 13044.77359 & 36.56 & 1.6 & 1 + 13045.74638 & 32.78 & 1.5 & 1 + 13072.85948 & -2.85 & 1.7 & 1 + 13240.13983 & -34.52 & 1.5 & 2 + 13302.13193 & -4.28 & 1.6 & 2 + 13302.97959 & -5.69 & 1.4 & 2 + 13338.96446 & 15.77 & 1.2 & 2 + 13340.09520 & 16.45 & 1.5 & 2 + 13368.88930 & -6.51 & 1.0 & 2 + 13369.78156 & -6.60 & 1.0 & 2 + 13425.87197 & -39.98 & 1.4 & 2 + 13426.82980 & -39.12 & 1.3 & 2 + 13428.77030 & -36.85 & 1.3 & 2 + 13483.72749 & 19.49 & 1.0 & 2 + 13723.90630 & -7.28 & 1.6 & 2 + 13841.76427 & 34.09 & 1.4 & 2 + 14544.83023 & 25.40 & 1.6 & 2 + 14545.78169 & 26.99 & 1.4 & 2 + 14546.78977 & 24.28 & 1.3 & 2 + 14718.08322 & 43.60 & 1.6 & 2 + 14806.91704 & -7.74 & 1.5 & 2 + 14810.89031 & -9.50 & 1.7 & 2 + 14838.94681 & 0.39 & 1.8 & 2 + 14864.95362 & 29.45 & 1.8 & 2 + 14929.76349 & -5.49 & 1.7 & 2 + 15135.00085 & -21.29 & 1.6 & 2 + 15172.99171 & 12.60 & 1.6 & 2 + 15229.78574 & -12.32 & 1.6 & 2 + cccc @xmath12 ( d ) & 154.378 @xmath13 0.089 & 885.5 @xmath13 5.1 & 1862 @xmath13 38 + @xmath14 ( jd-2440000 ) & 10305 @xmath13 11 & 9534 @xmath13 11 & 8558 @xmath13 11 + @xmath15 & 0.054 @xmath13 0.028 & 0.125 @xmath13 0.055 & 0.16 @xmath13 0.14 + @xmath16 ( degrees ) & 130 & 53 @xmath13 17 & 0 + @xmath17 ( m@xmath10 ) & 28.50 @xmath13 0.78 & 15.4 @xmath13 1.2 & 12.8 @xmath13 1.3 + @xmath18(@xmath19 ) & 0.675 @xmath13 0.017 & 0.652 @xmath13 0.052 & 0.696 @xmath13 0.059 + a ( au ) & 0.53364 @xmath13 0.00020 & 1.7100 @xmath13 0.0065 & 2.807 @xmath13 0.038 + r.m.s . & + jitter & + @xmath20 & + @xmath21 & + the residuals to this fit have an rms 4.03 m / s and show with no significant periodogram peak at any period .
the tallest peak is at 3.81 days .
we have run a monte carlo fap analysis on these residuals of our best fit for this tallest peak , and find similarly good fits in 50% of velocity - scrambled trials , consistent with noise .
we thus conclude that our model is sufficient to explain the data and that there are no other statistically significant planetary signals detected .
we studied the dynamical stability of hd 37124 by combining the radial velocity data with markov chain monte carlo ( mcmc ) analyses to obtain ensembles of masses , semimajor axes , eccentricities , and orbital angles consistent with the rv data .
these ensembles were generated without regard to dynamical stability considerations .
we then imposed line - of - sight and relative inclination distributions on these sets of parameters . by incorporating the unknown inclination parameters with observation - derived parameters , we sampled the entire phase space of orbital parameters .
we subsequently ran n - body simulations on each element in these ensembles in order to assess each system s stability and resonant evolution .
our treatment follows that of @xcite . in particular , we calculated 5 markov chains , each containing over @xmath22 states .
each state includes the orbital period ( @xmath12 ) , velocity amplitude ( @xmath17 ) , eccentricity ( @xmath15 ) , argument of pericenter measured from the plane of the sky ( @xmath16 ) , and mean anomaly at a given epoch ( @xmath23 ) for planets b , c and d. the mcmc uses a standard gaussian random walk proposal distribution and the metropolis - hastings algorithm for accepting or rejecting each proposal for all model parameters except @xmath24 and @xmath25 .
since the radial velocity signature is only weakly dependant on these values , @xmath24 and @xmath25 were drawn randomly for each state .
this can still be considered a markov chain , as the procedure satisfies the markov condition , i.e. that a trial state not depend on states other than the current state , as well as the other conditions ( time - homogeneous , irreducible , aperiodic ) to prove that the markov chain will ( eventually ) converge to the posterior distribution .
we imposed an isotropic distribution of line - of - sight inclinations ( @xmath26 ) and a uniform sample of longitude of ascending nodes ( @xmath25 ) on our mcmc - derived initial conditions .
the planet masses , @xmath27 , and semimajor axes , @xmath28 , were obtained from each set of ( @xmath29 ) values from relations derived with a jacobi coordinate system @xcite .
the approximate range of minimum masses obtained , in jupiter masses , were : @xmath30 , @xmath31 , and @xmath32 .
we treated the both the offset between the chips and the jitter as free parameters .
the 5@xmath33 percentile , median , and 95@xmath33 percentile offsets between the chips in our ensemble were 3.16 , 3.78 , 4.62 , and the median jitter we find to be 4 m / s .
c + we integrated 850 sets of initial conditions in the coplanar case with all three planets in prograde orbits by using the burlish - stoer integrator of mercury @xcite for @xmath34 yr with an output interval of @xmath35 yr .
we also incorporated the effects of general relativity in the code , which could have profound consequences for multi - planet system stability @xcite , although the effect is likely to be negligible in this system .
we classified systems as `` unstable '' if , for any planet , @xmath36 , where @xmath37 , @xmath38 and @xmath39 represent the maximum , minimum and initial values of the semimajor axis , and @xmath40 .
one may visualize a representative architecture of hd 37124 by comparing the semimajor axis and eccentricity ranges of all three planets .
figure [ mcmc ] plots the observed eccentricity vs. derived semimajor axis for all planets in the prograde coplanar state .
black dots indicate unstable systems while green squares and red crosses indicate stable systems , and red crosses indicate systems which are in a @xmath41:@xmath42 mean motion resonance ( mmr ) between planets c and d , according to our definition below .
the figure indicates i ) a closely packed system , with the inner and outer planets separated by no more than six times the innermost planet s semimajor axis .
ii ) a relatively circular innermost planet ( with @xmath43 in most cases ) that is likely too far from the parent star to be classified as a `` hot jupiter '' , iii ) the greater the number of orbital periods sampled by rv , the greater the constraint on the planet s likely semimajor axes and eccentricities , iv ) most ( @xmath44 ) current orbital fits predict an unstable system , v ) the majority of initial conditions which produce stable orbits contain an outer planet with a low ( @xmath45 ) eccentricity , and a middle planet with a semimajor axis @xmath46 au and eccentricity less than about 0.2 . ,
vi ) systems containing a @xmath41:@xmath42 resonance occur only when @xmath47 au @xmath48 au . we emphasize that this approximate semimajor axis range appears to be necessary but not sufficient for resonance to occur .
the figure demonstrates that other mcmc fits with outer planet periods in the resonant range are either unstable , or stable but non - resonant .
the architecture of these systems ( as defined by , e.g. , the mean longitude and longitude of pericenter ) do not allow them to settle into resonance even though the outer planet period might favor resonance . because of finite sampling , our definition of `` resonance '' in this analysis comes from consideration of the rms deviation of each resonant angle about each of @xmath49 , which includes common libration centers .
we flag systems as `` resonant '' if at least one of these angles has rms under @xmath50 for @xmath51 myr , the entire duration of our simulations .
below , we refer to this value as a `` libration rms '' .
hd 37124 presents a clear initial choice of angles to test for libration . as indicated by fig .
[ mcmc ] , the semimajor axis ratio of planets c and d is suggestive of a @xmath41:@xmath42 mmr .
therefore , we sampled the following angles for libration : @xmath52 @xmath53 and found that @xmath54 librates in @xmath55 of cases , while @xmath56 librates in @xmath57 of cases .
further , the systems for which @xmath56 is resonant are a subset of those for which @xmath54 is resonant . if we tighten our definition of resonance to include only those systems with rms resonant angles under @xmath58 , then no @xmath56 arguments are resonant . under this stricter definition ,
the @xmath54 arguments are only resonant in @xmath59 of cases , and if we further tighten the libration criterion to an rms of or @xmath60 , then this number decreases to @xmath61 .
the lowest libration rms detected is @xmath62 .
all rms s under @xmath63 were for a libration center of @xmath64 .
figure [ res23 ] illustrates three examples of `` resonant '' systems from this , each with a different libration rms .
ccc + + + cc + + ( -40,-60 ) cc + + ( -40,-60 ) we additionally sampled all 3 pairs of apsidal angles ( the difference between two longitudes of pericenter ) in the coplanar prograde state , and found only two instances of libration , both at high ( @xmath65 ) libration rms s and around the `` asymmetric '' centers @xmath50 and @xmath66 for the inner and outer planet apsidal angle .
inspection reveals , however , that these instances of libration are more indicative of long period ( @xmath67 myr ) circulation . additionally , the semimajor axis ratio of planets b and c could indicate the presence of a @xmath68:@xmath42 mmr .
therefore , we sampled all angles of the form @xmath69 , where @xmath70 . none of the coplanar prograde systems exhibited libration of any of the @xmath68:@xmath42 angles between planets `` b '' and `` c '' over 10 myr . however , preliminary sampling of these angles over intervals of @xmath41 myr does occasionally exhibit libration rms s close to @xmath50 . because the period ratios between planets c and d may skirt the @xmath71:@xmath72 pc , we also tested the @xmath73 angles , where @xmath74 , but found no instances of libration .
having analyzed the coplanar prograde bin , we can now consider the case where the planets have nonzero mutual inclinations .
we used rejection sampling to obtain triplets of @xmath26 values such that the system is placed into one of 144 bins according to the relative inclination between planets @xmath6 and @xmath7 ( @xmath75 ) and planets @xmath7 and @xmath8 ( @xmath76 ) . in no two bins were the same ensemble of initial conditions used .
we binned relative inclinations in intervals of @xmath77 , and used stratified sampling in order to obtain a uniform number of samples per bin .
we initially sampled 100 initial system states per bin .
for those bins where we found more than one system to be stable , we added 200 additional ensembles of initial conditions by considering the fraction of stable systems in the non - coplanar cases , we can obtain a broader dynamical portrait of this system .
[ stab ] illustrates the fraction of stable systems in each bin overall ( top panel ) and with respect to all systems for which the initial @xmath78 ( bottom panel ) .
this cutoff was motivated by the rightmost panel in fig .
[ mcmc ] and could suggest a constraint on the orbital properties of the system in order to ensure that it remains stable .
[ stab ] shows that the system _ must be roughly coplanar _ , with relative inclinations less than @xmath79 , in order to be stable .
this constraint allows various pairs of planets to harbor retrograde orbits .
we also performed limited resonant testing for systems in these bins .
the fraction of total systems which exhibit libration of @xmath54 and @xmath56 under @xmath50 for 10 myr is given by fig .
[ restables ] . @xmath41:@xmath42
resonant systems occur , therefore , generally at the few percent level , and most likely when all planets are coplanar with prograde orbits .
as @xcite showed for the gj 876 system , even the most well - established and deepest mean - motion resonances can prove illusory if additional planets are found in the system ( although in that case it appears that the resonance still present , albiet considerably shallower and more complex than previously thought ) .
even for truly resonant systems , a demonstration of resonance can be difficult .
for instance , triple - planet systems may feature two planets with a mostly - librating resonant argument that occasionally circulates due to interactions with the third planet . near separatrix behavior ( as in the case of @xmath2 and ; * ? ? ? * ; * ? ? ?
* ) can also make libration and circulation essentially indistinguishable .
we note that near - resonant behavior can itself dynamically interesting : the 5:2 near - resonance of jupiter and saturn ( the great inequality ) has major consequences for the dynamics of the solar system . given the above - mentioned difficulties in _ proving _ that a resonant argument for a given system of planets satisfies some precisely specified definition of libration given the typical uncertainties in radial velocity measurements , we suggest that studies of resonant interactions would benefit from identifying systems that appear to be in or near resonance ( _ apparent period commensuribilities _ ) . with that in mind , we note that in addition to hd 37124 , there are 21 other systems in the exoplanet orbit database ( wright et al .
2010 , pasp submitted ) of peer - reviewed literature with well - established apparent pcs , which we present in table [ mmr ] .
this list includes all pairs of planets for which the period ratio @xmath80 is less than 6 and within 0.05 of an integer or half integer ( neglecting uncertainties in periods ) , and other exoplanetary pairs whose likely mmrs are discussed in the literature .
lccl gj 876 & @xmath81 & 4:2:1 & @xcite + hd 82943 & @xmath82 & 2:1 & @xcite + hd 37124 & @xmath83 & 2:1 & @xcite , this work + hd 128311 & @xmath84 & 2:1 & @xcite , others + hd 73526 & @xmath84 & 2:1 & @xcite + @xmath1 ara & @xmath85 & 2:1 & @xcite + koi 152 & 2,3 & 2:1 & @xcite + koi 877 & 2,1 & 2:1 & @xcite + 24 sex & @xmath84 & 2:1 & @xcite + kepler-9 & @xmath84 & 2:1 & @xcite + psr b1257 + 12 & @xmath86 & 3:2 & @xcite + hd 45364 & @xmath84 & 3:2 & @xcite + hd 200964 & @xmath84 & 4:3 & @xcite + 55 cnc & @xmath84 & 3:1 & @xcite + hd 10180 & @xmath87 & 3:1 & lovis et al .
2010 , a&a accepted + hd 60532 & @xmath84 & 3:1 & @xcite + hd 108874 & @xmath84 & 4:1 & @xcite + solar & @xmath88 & 5:2 & + hd 10180 & @xmath89 & 5:2 & lovis et al .
2010 , a&a accepted + koi 896 & 1,2 & 5:2 & @xcite + hd 202206 & @xmath84 & 5:1 & @xcite + the fraction of known multiplanet systems exhibiting at least one apparent pc is high . of the 43 well determined multiplanet systems discovered by radial velocities around normal stars , 15 appear in table [ mmr ] , or 35% , including 9 of the 30 apparent double - planet systems 30% . to determine
if this is more than would be expected simply by chance , we have performed two tests . in the first test , we randomly drew pairs of periods from the 340 rv - discovered planets in the exoplanet orbit database ( eod ) ( * ? ? ?
* pasp submitted ) , and rejected those pairs with period ratios @xmath90 ( corresponding to the smallest @xmath80 among real multiplanet systems ) .
we counted the fraction of remaining systems with @xmath80 within 0.05 of an integer or half integer @xmath91 ( corresponding to the largest apparent pc in table [ mmr ] ) .
we found that only 4% of our random pairs satisfy our apparent pc criterion , far smaller than the 30% of double - planet systems actually found in apparent pcs . in the second test
, we included the effects of triple and higher - multiple systems by randomly assigning periods from the eod to all planets in real multiple systems ( again subject to the constraint that no pair of planets in the system have @xmath90 ) .
we found 16% of these artificial systems passed our apparent pc criterion , reflecting the higher number of planet pairs available to test per star compared to our first test . despite this inflation
, the actual value of 33% among all multiplanet systems is still significantly higher .
these results underscore that the orbital periods of the population of planets known to be in multiplanet systems is inconsistent with the apparently - singleton sample @xcite .
this apparently high percentage of known multi - planet systems with an apparent pc might favor particular formation mechanisms .
planet - planet scattering , planetesimal disk migration , and gas disk migration have all been shown to produce systems with at least one pair of planets that not only are commensurate in period , but also resonant .
@xcite found that planet - planet scattering produced mmrs in roughly 5% of the systems that they simulated , and @xcite discovered that between 50% - 80% of systems undergoing planetesimal disk migration yielded resonant capture .
convergent gas disk migration , the thrust of the numerous papers cited in [ intro ] can occur with near 100% efficiency for certain initial planetary and disk parameters .
as observed by @xcite and @xcite , the inclination may be excited as well as the eccentricity in many resonant cases .
if a resonance exists in hd 37124 , it could have been produced by any of these methods . if the rms libration of such a resonance is representative of the bottom panel of fig .
[ res23 ] and has a value that approaches 90 degrees , then planet - planet scattering is a likely origin of this resonance .
alternatively , disk or gas migration would likely produce a system that is `` deeper '' in resonance , with a smaller variation in resonant angle , similar to the top panel of fig .
[ res23 ] .
resonant librating angles need not involve the eccentricities and pericenters , but instead the inclinations and longitudes of ascending nodes , similar to the @xmath92:@xmath41 mimas - tethys resonance in the saturnian system @xcite .
we have resolved the period ambiguity of hd 37124 @xmath8 from @xcite and find that hd 37124 @xmath7 and @xmath8 are in an apparent 2:1 period commensurability .
our numerical integrations show that both resonant and non - resonant configurations are consistent with the radial velocity data , and that stability requires a nearly circular orbit ( @xmath93 ) for the @xmath8 component .
our stability analysis shows that the system must be nearly coplanar , and that the three planets have identical minimum masses within the errors ( of 310% ) .
we show that the roughly one in three well - characterized multiplanet systems shows an apparent period commensurability , which is more than a nave estimate based on randomly drawing periods from the known exoplanet population would suggest .
this offers evidence for some particular proposed scattering and migration mechanisms , and we suggest that the statistics of multiplanet systems may now be sufficiently robust to provide a test and comparison of models of exoplanetary dynamical evolution .
j.t.w . received support from nasa origins of solar systems grant nnx10ai52 g .
d.v . and e.b.f .
were partially supported by nasa origins of solar systems grant nnx09ab35 g . a.w.h .
gratefully acknowledges support from a townes postdoctoral fellowship at the uc berkeley space sciences laboratory .
this work was partially supported by funding from the center for exoplanets and habitable worlds , which is supported by the pennsylvania state university , the eberly college of science , and the pennsylvania space grant consortium .
the work herein is based on observations obtained at the w. m. keck observatory , which is operated jointly by the university of california and the california institute of technology .
the keck observatory was made possible by the generous financial support of the w.m .
keck foundation .
we wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community .
we are most fortunate to have the opportunity to conduct observations from this mountain . , e. b. 2006 , in new horizons in astronomy : frank n. bash symposium , s. j. kannappan , s. redfield , j. e. kessler - silacci , m. landriau , and n. drory , eds . ,
volume 352 of _ astronomical society of the pacific conference series _ , pp .
15+ , b. s. , bennett , d. p. , udalski , a. , gould , a. , christie , g. w. , maoz , d. , dong , s. , mccormick , j. , szymaski , m. k. , tristram , p. j. , nikolaev , s. , paczyski , b. , kubiak , m. , pietrzyski , g. , soszyski , i. , szewczyk , o. , ulaczyk , k. , wyrzykowski , . ,
depoy , d. l. , han , c. , kaspi , s. , lee , c .- u . , mallia , f. , natusch , t. , pogge , r. w. , park , b .-
, abe , f. , bond , i. a. , botzler , c. s. , fukui , a. , hearnshaw , j. b. , itow , y. , kamiya , k. , korpela , a. v. , kilmartin , p. m. , lin , w. , masuda , k. , matsubara , y. , motomura , m. , muraki , y. , nakamura , s. , okumura , t. , ohnishi , k. , rattenbury , n. j. , sako , t. , saito , t. , sato , s. , skuljan , l. , sullivan , d. j. , sumi , t. , sweatman , w. l. , yock , p. c. m. , albrow , m. d. , allan , a. , beaulieu , j .-
p . , burgdorf , m. j. , cook , k. h. , coutures , c. , dominik , m. , dieters , s. , fouqu , p. , greenhill , j. , horne , k. , steele , i. , tsapras , y. , chaboyer , b. , crocker , a. , frank , s. , & macintosh , b. 2008 , science , 319 , 927 , m. j. , fabrycky , d. c. , ragozzine , d. , ford , e. b. , steffen , j. h. , welsh , w. f. , lissauer , j. j. , latham , d. w. , marcy , g. w. , walkowicz , l. m. , batalha , n. m. , jenkins , j. m. , rowe , j. f. , cochran , w. d. , fressin , f. , torres , g. , buchhave , l. a. , sasselov , d. d. , borucki , w. j. , koch , d. g. , basri , g. , brown , t. m. , caldwell , d. a. , charbonneau , d. , dunham , e. w. , gautier , t. n. , geary , j. c. , gilliland , r. l. , haas , m. r. , howell , s. b. , ciardi , d. r. , endl , m. , fischer , d. , frsz , g. , hartman , j. d. , isaacson , h. , johnson , j. a. , macqueen , p. j. , moorhead , a. v. , morehead , r. c. , & orosz , j. a. 2010 , science , 330 , 51 , a. w. , johnson , j. a. , marcy , g. w. , fischer , d. a. , wright , j. t. , bernat , d. , henry , g. w. , peek , k. m. g. , isaacson , h. , apps , k. , endl , m. , cochran , w. d. , valenti , j. a. , anderson , j. , & piskunov , n. e. 2010 , , 721 , 14671481 , f. , correia , a. c. m. , mayor , m. , tamuz , o. , couetdic , j. , benz , w. , bertaux , j .- l . , bouchy , f. , laskar , j. , lovis , c. , naef , d. , queloz , d. , santos , n. c. , sivan , j .- p . , sosnowska , d. , & udry , s. 2007 , , 462 , 769776 , n. c. , bouchy , f. , mayor , m. , pepe , f. , queloz , d. , udry , s. , lovis , c. , bazot , m. , benz , w. , bertaux , j .- l . , lo curto , g. , delfosse , x. , mordasini , c. , naef , d. , sivan , j .-
, & vauclair , s. 2004 , , 426 , l19l23 , s. s. , allen , s. l. , bigelow , b. c. , bresee , l. , brown , b. , cantrall , t. , conrad , a. , couture , m. , delaney , c. , epps , h. w. , hilyard , d. , hilyard , d. f. , horn , e. , jern , n. , kanto , d. , keane , m. j. , kibrick , r. i. , lewis , j. w. , osborne , j. , pardeilhan , g. h. , pfister , t. , ricketts , t. , robinson , l. b. , stover , r. j. , tucker , d. , ward , j. , & wei , m. z. 1994 , in proc .
spie instrumentation in astronomy viii , david l. crawford ; eric r. craine ; eds . , volume 2198 , p. 362 | we present new radial velocities from keck observatory and both newtonian and keplerian solutions for the triple - planet system orbiting hd 37124 .
the orbital solution for the system has improved dramatically since the third planet was first reported in @xcite with an ambiguous orbital period .
we have resolved this ambiguity , and show that the outer two planets have an apparent period commensurability of 2:1 .
a dynamical analysis finds both resonant and non - resonant configurations consistent with the radial velocity data , and constrains the mutual inclinations of the planets to be @xmath0 .
we discuss hd 37124 in the context of the other 19 exoplanetary systems with apparent period commenserabilities , which we summarize in a table .
we show that roughly one in three well - characterized multiplanet systems has a apparent low - order period commensuribility , which is more than would navely be expected if the periods of exoplanets in known multiplanet systems were drawn randomly from the observed distribution of planetary orbital periods . |
the dynamics of systems of particles subjected to nonpotential interactions remains poorly understood .
the absence of a hamiltonian for such systems , which generally are far from equilibrium , hampers applying the machinery of statistical mechanics based on the liouville equation .
many attempts have been made to investigate these systems using discrete algorithms to model this behavior . in nature
there are many examples of such systems @xcite .
since the discrete algorithms are hard to describe analytically it is natural also to consider continuum models of a hydrodynamic type . in standard hydrodynamics
the relation between microscopic kinetics ( boltzmann - type equations ) and navier - stokes equations is a standard topic of research @xcite . for the systems of interest
the construction of corresponding kinetic equations based on the specific dynamic rules and their connection with the hydrodynamics equations seems to be unknown so far and is worth studying .
one can expect that the continuous description of the collective behavior like swarming and flocking leads to quite unusual hydrodynamics . in our first paper @xcite
we proposed a hydrodynamic model which can be considered to be the continuum analogue of the discrete dynamic automaton proposed by vicsek et al .
@xcite for a system of self propelling particles .
it uses the continuity equation @xmath0 which implies that the total number of particles @xmath1 is constant .
the kinetic energy of a co - moving volume @xmath2 is also conserved @xmath3 using eq . and
eq . ( [ totenrg ] ) it can be shown that a field @xmath4 exists such that the eulerian velocity , @xmath5 satisfies : @xmath6 this equation can be considered as the continuous analogue of the conservative dynamic rule used by vicsek et al .
@xcite .
we proposed the following minimal model for the field of the angular velocity @xmath7 which is linear in spatial gradients of the fields @xmath8 or @xmath9 : @xmath10 @xmath4 has the proper pseudovector character .
the averaging kernels @xmath11 and @xmath12 should naturally decrease with the distance in realistic models .
they sample the density and the velocity around @xmath13 in order to determine @xmath7 . in the first paper we concentrated on @xmath14 .
the detailed derivation of the above equations from the discrete models based on the automaton proposed by vicsek et al .
@xcite will be the subject of a future paper .
note that the models based on eqs .
- allow solutions of uniform motion in the form of a solitary packet : @xmath15 with @xmath16 independent of position and time .
the contribution to @xmath17 due to @xmath14 is zero for an arbitrary density distribution @xmath18 .
the contribution due to @xmath19 is zero for density distributions @xmath18 which only depend on the position in the @xmath16 direction . in this second case
it follows from the continuity equation that @xmath18 should be everywhere constant .
the density distribution @xmath18 should be chosen such that the number of particles , and , correspondingly , the total kinetic energy are finite . the solutions of such type also were found analytically in @xcite and observed in simulations @xcite .
note that such solutions exist not only in nonlocal case like in @xcite but also for the local model which we consider below . within the first order of perturbation theory on small deviation of density and velocity fields the solitary solution given by eq
. shows neutral stability ; i.e. the perturbations grows linearly for small @xmath20 .
we restrict our discussion to the simple case of averaging kernels , which are @xmath21-functions : @xmath22 we will call this the local hydrodynamic model ( lhm ) . in the first paper , where we only considered @xmath14 , we scaled @xmath14 by dividing by @xmath23 and the density @xmath24 by multiplying with @xmath23 .
this made it then possible to restrict the discussion to @xmath25 is plus or minus one .
the disadvantage of this scaling procedure is that it changes the dimensionality of @xmath26 and @xmath24 . for two kernels it becomes impractical .
we note that @xmath27 is given by : @xmath28 for the local model eq.([1.4 ] ) reduces to : @xmath29 and eq . for the velocity
becomes @xmath30 note that the second term on the right - hand - side of eq .
corresponds to the rotor chemotaxis force when the number density is proportional to the attractant density introduced in @xcite . in the following section
we will show that the only stationary solutions for lhm with @xmath31 field given by eq .
are either the solutions of uniform motion ( see eq . ) or the radially symmetric planar solution which will be considered in detail in the following section . in the second section
we investigate the properties of the stationary radially symmetric solutions of the local hydrodynamic model for some special cases .
conclusions are given in the last section .
the equations of motion to be solved are eqs . and .
in order to find a class of 2d stationary solutions we consider this problem in a generalized curvilinear orthogonal coordinate system @xmath32 , which can be obtained from the cartesian one @xmath33 by some conformal transformation of the following form : @xmath34 where @xmath35 is an arbitrary analytical function of @xmath36 . in a curvilinear orthogonal
coordinate system the fundamental tensor has a diagonal form , @xmath37 , where the indices @xmath38 are either @xmath39 or @xmath40 .
the square of linear element in conformal coordinates is @xmath41 where @xmath42 is the jacobian of the inverse conformal transformation from the arbitrary curvilinear orthogonal to cartesian coordinates , @xmath43 .
furthermore @xmath44 is the determinant of the metric tensor . for a conformal transformation @xmath45 .
the differential operations are given by the following expressions @xcite : @xmath46 , \label{3.5b } \\
\mathop{\rm rot}\nolimits\mathbf{a } & = & d\left [ \frac{\partial } { \partial u}\left ( \frac{a_{v}}{\sqrt{d}}\right ) -\frac{\partial } { \partial v}\left ( \frac{a_{u}}{\sqrt{d}}\right ) \right ] , \label{3.5c } \\
\delta \phi & = & d\left [ \frac{\partial ^{2}\phi } { \partial u^{2}}+\frac{\partial ^{2}\phi } { \partial v^{2}}\right]\ , .
\label{3.5d}\end{aligned}\ ] ] here @xmath47 and @xmath48 are orthonormal base vectors in the directions of increasing @xmath39 and @xmath40 respectively .
these base vectors are functions of the coordinates @xmath39 and @xmath40 .
the projections of the vectorfield @xmath49on these directions are @xmath50 and @xmath51 . using eqs . - for the velocity field given by @xmath52 , one obtains : @xmath53 , \label{3.6a}\ ] ] @xmath54 ( -\mathrm{v}_{v}\mathbf{e}_{u}+\mathrm{v}_{u}\mathbf{e}_{v } ) , \label{3.6b}\ ] ] @xmath55 \left ( -\mathrm{v}_{v}\mathbf{e}_{u}+\mathrm{v}_{u}\mathbf{e}_{v}\right ) , \label{3.6c}\ ] ] @xmath56 = 0 . \label{3.6d}\ ] ] substituting eqs . - into eqs .
and we obtain the following system of equations , which determines all possible stationary flows for the lhm : @xmath57 @xmath58 @xmath59 where @xmath60 now let us consider the case of coordinate flows
, when the flow is directed along one of the families of coordinate lines @xmath61 for example along @xmath39-coordinate lines and is given by @xmath62 , the density distribution is @xmath63 .
the case of a velocity field @xmath64 is equivalent ( just interchange @xmath39 and @xmath40 ) . from eq .
we have : @xmath65\,,\ ] ] where @xmath66 equations and take the form : @xmath67 and lead to @xmath68 where @xmath69 is an arbitrary function of @xmath39 .
taking into account that : @xmath70 and eq .
we obtain : @xmath71\,.\ ] ] note that as it follows from eq .
@xmath72 for the integrand in eq.([vv ] ) this implies that @xmath73 therefore from eqs .
and we can conclude that the function @xmath74 , which determines the coordinate system , depends only on one variable , @xmath75 . in the case of conformal coordinates , defined by the metrics in eq .
, the gaussian curvature of the surface is given by @xcite : @xmath76 for a planar flows the condition @xmath77 leads to the following @xmath78 using the expression for the laplacian in the conformal representation ( see eqs . ) and taking into account a fact that @xmath79 one finds for : @xmath80 , \label{3.22}\ ] ] where @xmath81 are arbitrary constants .
the case @xmath82 determines a cartesian coordinate system , which related to a linear class of stationary flow .
the case @xmath83 determines a polar coordinate system @xcite , which corresponds to an axially symmetric ( or vortical ) type of flow .
finally the velocity field for the lhm , with @xmath84 takes the form : @xmath85\,du\right]\ , , \label{vvfin}\ ] ] thus it is proved that for the case @xmath84 the only stationary solutions are those either with planar or axial symmetry of the flow .
the case @xmath86 ( the lm2 ) is specific because as it follows from eqs . and
the velocity field @xmath87 is arbitrary while the density takes the form : @xmath88 the statement about the symmetry of the stationary solutions for such a model is the same as that proved above for the case @xmath89 .
note that the parameter @xmath90 can be considered as the weight factor of the rotor chemotaxis contribution .
in this section we investigate the stationary radially symmetrical solutions for different cases of the local hydrodynamical models . in our first paper we considered the @xmath91 case , which we called the local model one ( lm1 ) .
other models correspond to cases @xmath86 and @xmath92 which we will call the lm2 and the lm12 respectively .
we consider the finite and infinite flocking stationary states for these models .
it is natural to differentiate between these two cases by the finiteness of two integrals of motion - the total number of particles eq . and the kinetic energy eq . .
the infinite flocking is associated with @xmath93 infinite but finite @xmath94 , while finite flocking naturally corresponds to both @xmath93 and @xmath94 finite .
note that in the finite flocking behavior one may consider two cases with respect to the compactness of the @xmath95 .
compactness means that the density has some upper cut - off beyond which it can be put zero . in our previous paper @xcite
we considered the stationary solutions in lm1 and obtained the for velocity field profile @xmath96 here @xmath97 is the radial coordinate and @xmath98 is the angular one , @xmath99 is some radius , which for vortex - like solutions plays the role of lower cut - off radius of the vortex and determines its core . for vortex - like solutions the constant @xmath100 in eq.([2.1 ] ) is determined by the circulation of the core @xmath101 the spatial character of the solution given in eq.([2.1 ] ) strongly depends on the sign of the parameter @xmath25 .
the finiteness of integrals of motion eq . and
. is guaranteed by either the fast enough decrease of the density @xmath95 at @xmath102 or its compactness ( @xmath95 as a function has finite support ) .
let us consider the finite flocking behavior ( ffb ) , which is characterized by both @xmath93 and @xmath94 finite .
if at @xmath102 asymptotically @xmath103 , where @xmath104 the total number of particles @xmath93 is finite .
then at such a behavior of @xmath95 the total kinetic energy is finite only if @xmath105 . in a case
@xmath106 the condition of finiteness for the kinetic energy and the total number of particles is fulfilled only if @xmath95 has finite support . as an example we may give : @xmath107 where @xmath108 is the upper cut - off radius . substituting eq.([2.3 ] ) into eq.([2.1 ] ) one obtains : @xmath109 .
\label{2.4}\ ] ] the corresponding profiles for the velocity @xmath110 at different ratio @xmath111 are shown in fig .
[ dannye ] .
note that for the case considered at @xmath112 we get the monotonic profiles of the velocity which are similar to those observed in experiments @xcite . the infinite flocking behavior ( ifb )
is characterized by the @xmath93 infinite and @xmath94 finite . for the case
@xmath113 no physical solutions exist with such a behavior .
for the case @xmath105 slowly decaying density distributions @xmath114 at @xmath115 with @xmath116 are consistent with the finiteness of @xmath94 .
these statements are summed up in table [ tab1 ] .
.properties of the stationary solutions of the lm1 . [ cols="^,^,^",options="header " , ] the third case which is expedient to consider is @xmath117 . in that case , according to eq . , the field @xmath4 is coupled to the number density flux @xmath118 : @xmath119 so that eq . for
the velocity is @xmath120 for a radially symmetric stationary planar solution this gives @xmath121 .
\label{2.13}\ ] ] with @xmath122\ , .
\label{2.14}\ ] ] as a solution .
the constant @xmath123 is determined by the circulation of the core @xmath124 the properties of finite and infinite flocking behavior for this model are the same as those for the lm1 ( see table [ tab1 ] ) .
in this paper we consider the properties of the stationary 2d solutions of the lhm proposed in @xcite .
we established that the only possible stationary solutions in the model are those with translational or axial symmetry .
the cases of finite and infinite flocking behavior are considered for different specific types of the lhm .
it is shown that the case @xmath86 ( lm2 ) is specific in a sense that there is only one density distribution , for which many velocity profiles can be realized .
in general case ( @xmath84 ) one is free to choose axially symmetric density distribution which the velocity profile depends on ( eqs . and ) .
note that in this respect the general case is similar to the lm1 considered earlier . | in a first paper we proposed a continuum model for the dynamics of systems of self propelling particles with kinematic constraints on the velocities . the model aims to be analogous to a discrete algorithm used in works by t. vicsek et al . in this paper
we prove that the only types of the stationary planar solutions in the model are either of translational or axial symmetry of the flow . within the proposed model we differentiate between finite and infinite flocking behavior by the finiteness of the kinetic energy functional .
self - propelling particles ; finite - flocking behavior ; vortex . |
consider the adimensionalized reaction - diffusion partial differential equation @xmath0 of fisher - kolmogorov type , where @xmath1 is a smooth function satisfying certain hypotheses .
the usual fisher - kolmogorov equation corresponds to @xmath2 and models the spreading of biological populations , see @xcite .
other well - known cases are the newell - whitehead - segel equation , @xmath3 for describing rayleigh - benard convection , see @xcite , and the zeldovich equation , @xmath4 with @xmath5 that appears in combustion theory , see @xcite .
see also @xcite .
it is known that the traveling wave solutions @xmath6 of , satisfying @xmath7 appear when @xmath8 is a special solution of the second order equation @xmath9 where the dot indicates the derivative with respect to @xmath10 .
this solution can be seen as the heteroclinic orbit @xmath11 of the planar system @xmath12 that connects the origin with the saddle point at @xmath13 , see figure [ nfig1 ] .
for instance for the classical fisher - kolmogorov equation it exists only when @xmath14 nfig1 ( 30,5)@xmath15 ( 100,24)@xmath16 ( 24.5,36)@xmath17 the goal of this paper is to give analytic upper and lower bounds of the heteroclinic orbit @xmath11 as well as of their time parametrization .
we will approach to this question with similar tools to those introduced in @xcite . a key point consists in using the local behavior of the separatrices of the critical points to guess global algebraic bounds for the actual orbits .
first we prove a general result for system .
it is illustrated in figure [ nfig2 ] .
[ thm:1.1 ] consider system , with @xmath18 satisfying @xmath19 @xmath20 @xmath21 @xmath22 for all @xmath23 and @xmath24 .
let @xmath11 be its heteroclinic orbit and define @xmath25 then @xmath11 can be parametrized as @xmath26\}$ ] and for all @xmath27 @xmath28 moreover , if @xmath29 is the parametrization of @xmath11 such that @xmath30 it holds that @xmath31 where @xmath32 denotes the smallest closed interval containing @xmath33 and @xmath34 and @xmath35 is the solution of the cauchy problem @xmath36 u(0)=1/2 . \end{array } \right.\ ] ] cc nfig2i ( 33,18)@xmath37 ( 62,3)@xmath38 ( 94,27)@xmath16 ( 13,39)@xmath17 ( 30,-7)@xmath39 & nfig2ii ( 35,18)@xmath40 ( 78,10)@xmath41 ( 39.5,33)@xmath42 ( 85,33.5)@xmath43 ( 95,-2)@xmath10 ( 51,40)@xmath16 ( 30,-7)@xmath44 for instance when we particularize the above theorem to the classical fisher - kolmogorov system we obtain that @xmath45 the results obtained for system can be improved when we study this case .
we prove : [ thm:1.2 ] let @xmath46\}$ ] be a parametrization of the heteroclinic solution of system when @xmath2 . for @xmath47 and @xmath48
it holds that @xmath49 where @xmath50 is the taylor polynomial of degree @xmath51 of the unstable separatrix of the saddle point @xmath52 of the system and each @xmath53 is a rational function whose numerator has degree @xmath54 and its denominator degree @xmath55 , constructed from the pad approximants of @xmath56 .
see figure [ nfig3 ] .
nfig3 ( 33,16.5)@xmath57 ( 62,4.5)@xmath58 ( 93,27)@xmath16 ( 13,39)@xmath17 for a given value of @xmath59 it is also possible to study the maximum distance between two of the above functions .
for instance when @xmath60 , we can prove that @xmath61 it is also possible to go further in the computations .
for example we get that @xmath62 for all @xmath63 $ ] and the maximum error is at @xmath64 for bigger values of @xmath59 it is needed to compute approximations of higher degree to arrive to similar bounds of the error .
the @xmath10-parametrization of @xmath11 can also be obtained with more accuracy for the classical fisher - kolmogorov case .
we only present here a first result .
sharper approximations are detailed in section [ se:4 ] .
[ thm:1.3 ] let @xmath65 be the time - parametrization of the heteroclinic orbit @xmath66 of system with @xmath2 and such that @xmath67 .
define @xmath68 then : a. when @xmath69 it holds that @xmath70 b. when @xmath71 it holds that @xmath72 c. when @xmath73 it holds that @xmath74 the above inequalities improve the bounds given in .
notice also that when @xmath71 we have obtained the exact expression of the traveling wave solution of the classical fisher - kolmogorov equation , @xmath75 @xmath76 which coincides with the one given in @xcite .
the novelty of our result is that a similar expression gives a bound of the traveling wave for all the values of @xmath59 .
the methods developed to study the classical fisher - kolmogorov equation can also be applied for the newell - whitehead - segel and the zeldovich equations .
for computational reasons it is more convenient to locate the saddle point of system at the origin .
so we introduce the new variables @xmath77 @xmath78 and @xmath79 and it writes as @xmath80 where @xmath81 and the prime denotes derivative with respect to @xmath82 .
observe that the variable @xmath83 introduced above does not coincide with the one used in equation .
notice that @xmath84 satisfies the following set of hypotheses @xmath85 @xmath86 and @xmath87 for all @xmath88 .
the above system has only two critical points @xmath89 and @xmath52 .
the origin is a saddle point with eigenvalues @xmath90 and corresponding eigenvectors @xmath91 and @xmath92 .
similarly , when @xmath93 the point @xmath52 is an attracting node and its eigenvalues are @xmath94 a sketch of the phase portrait of system is given in figure [ nfig4 ] .
there we can see the heteroclinic connection that we are interested to locate .
indeed it is given by ones of the branches of the unstable separatrix of the saddle point .
in the new coordinates we will call it @xmath95 .
nfig4 ( 30,-1)@xmath96 ( 94,23.5)@xmath83 ( 13,36)@xmath97 [ prop:2.1 ] consider system , with @xmath84 satisfying hypotheses @xmath98 and @xmath99 .
set @xmath100 let @xmath95 be its heteroclinic orbit .
then @xmath95 can be parametrized as @xmath101\}$ ] and it holds for all @xmath102 that @xmath103 consider the 1-parameter family of maps @xmath104 .
we compute @xmath105 since on @xmath106 it holds that @xmath107 , if we choose @xmath108 such that @xmath109 does not vanish on the same interval we will have that the corresponding curve @xmath110 is without contact . notice that @xmath111 and so for @xmath112 , the hypotheses @xmath98 imply that the function @xmath109 is increasing . therefore : a. if for some @xmath113 , @xmath114 then @xmath115 for all @xmath116 b. if for some @xmath113 , @xmath117 then @xmath118 for all @xmath119 .
the conditions @xmath120 for @xmath121 , write as @xmath122 their solutions are @xmath123 it is easy to prove that @xmath124 therefore , taking @xmath125 and @xmath126 as an upper and lower boundary , respectively , we have constructed a subset of the strip @xmath127 , that contains the heteroclinic orbit @xmath95 , see figure [ nfig5 ] . then , on it , @xmath128 , @xmath95 can be parametrized as a function of @xmath83 , say @xmath129 , and the inequalities of the statement follow .
nfig5 ( 33,18)@xmath130 ( 62,3)@xmath131 ( 93,27)@xmath83 ( 13,39)@xmath97 notice that @xmath132 and therefore this curve is tangent to the unstable separatrix of the saddle point . using the above proposition we can also approach the parametrization of @xmath95 with respect the actual time @xmath133 we need to introduce some new functions . given the cauchy problem @xmath134 x(0)=1/2 , \end{array } \right.\ ] ] we will denote by @xmath135 its corresponding solution .
[ thm:2.2 ] under the same hypotheses and notations of proposition [ prop:2.1 ] , if @xmath136 is the parametrization of @xmath95 such that @xmath137 it holds that @xmath138 where @xmath139 is defined in , see figure [ nfig6 ] .
nfig6 ( 52,19)@xmath40 ( 42,33)@xmath140 ( 7,10)@xmath141 ( 96,5)@xmath82 ( 5,33)@xmath142 ( 51,40)@xmath83 by proposition [ prop:2.1 ] we know that @xmath103 recall that @xmath95 is parametrized as @xmath129 and therefore since in system , @xmath143 , it holds that @xmath144 .
therefore @xmath145 since @xmath146 , @xmath147 satisfies the differential inequality @xmath148 and @xmath149 the equality , it holds that @xmath150 for all @xmath151 as we wanted to see .
the other cases follow similarly .
[ cor:2.3 ] if in system , @xmath152 @xmath153 then @xmath154{1+(2^{m-1}-1 ) e^{\phantom{\overline{\lambda}}\!\!\!-(m-1)\underline{\lambda}\,t } } } , \frac{1}{\sqrt[m-1]{1+(2^{m-1}-1)e^{-(m-1)\overline{\lambda}\,t}}}\right\rangle.\ ] ] in particular for the fisher - kolmogorov case , @xmath155 , @xmath156 it suffices to solve the cauchy problem and then apply theorem [ thm:2.2 ] . when @xmath157 we obtain @xmath158{1+(2^{m-1}-1)e^{-(m-1){\lambda}t}}}\ ] ] and so the result follows
. theorem [ thm:1.1 ] is a reformulation for system of the results of this section obtained for system .
this section will be devoted to find sharper upper and lower bounds for @xmath95 in the fisher - kolmogov system when @xmath159 .
we will use dynamical tools inspired in @xcite .
one of the key points will be to find algebraic curves constructed by imposing that these curves coincide as much as possible with the unstable separatrix of the saddle point . to avoid the appearance of square roots during the computations it is convenient to include a new parameter @xmath160 in such a way that @xmath161 then system writes as @xmath162 .
\end{array } \right.\ ] ] notice that the condition on @xmath160 implies that @xmath163 .
one advantage of introducing this new parameter is that the eigenvalues of the saddle are now @xmath164 . in the notation of the previous section @xmath165 hence , if we denote @xmath166 the searched heteroclinic trajectory , from proposition [ prop:2.1 ] we know that @xmath167 is an upper bound for @xmath168 first we need to know the local expansion of the unstable manifold of the saddle point .
[ lem:3.1 ] the local unstable manifold of the origin of system writes as the analytic function @xmath169 and the subsequent terms can be determined recurrently and are positive for @xmath170 .
let @xmath171 be the local expression of any of the separatrices of the saddle point , being @xmath172 an analytic function at zero .
then @xmath173 or equivalently , @xmath174 writing @xmath175 and plugging this expression in the above one we get that @xmath176 so we choose @xmath177 then the right hand identity in writes as @xmath178 so @xmath179 , which implies @xmath180 in general , for @xmath181 it holds that @xmath182 therefore , @xmath183 and by induction @xmath184 , for @xmath181 , as we wanted to prove .
nfig7 ( 33,17.5)@xmath185 ( 62,3)@xmath186 ( 93,27)@xmath83 ( 13,39)@xmath97 [ prop:3.2 ] let @xmath187\ } , $ ] be a parametrization of the heteroclinic solution of system . then for all @xmath102 it holds that @xmath188 where @xmath189 is the taylor polynomial of degree @xmath51 in powers of @xmath83 of the function @xmath190 defined in lemma [ lem:3.1 ] .
see figure [ nfig7 ]
. we will give first the details for @xmath191 .
it is evident that for @xmath192 sufficiently small we have @xmath193 .
first we evaluate the polynomial @xmath194 at @xmath142 .
we obtain @xmath195 this quantity is negative for @xmath170 and it is a monotonous decreasing function of @xmath160 .
its value at @xmath196 is approximately @xmath197 .
moreover , @xmath198 , @xmath199 for @xmath192 and @xmath200 for @xmath192 and sufficiently small . then we conclude that @xmath200 for all @xmath201 $ ] .
we will show now that the flow of the vector field @xmath202 associated to system crosses the curve @xmath203 upwards .
we compute @xmath204 we obtain @xmath205 this expression is positive for @xmath192 .
all the other cases can be studied by using the same method .
the key point is that all the monomials in the corresponding expression of @xmath206 are positive for @xmath116 in proposition [ prop:3.2 ] , we have chosen to stop at the value 100 because the computation of the function @xmath207 used in its proof together with the testing that all its coefficients are positive takes more than four hours of cpu time in our computer . in any case , for practical uses , it suffices to consider small values of @xmath51 .
[ rem:3.3 ] it is clear that if we could prove that the radius of convergence of the series given in lemma [ lem:3.1 ] , when @xmath208 $ ] , is 1 we would have obtained an infinite monotonous sequence of polynomials tending to the actual separatrix .
unfortunately we have not succeeded in our attempts . in any case , computing an approximation of the radius of convergence , by using several hundreds of terms of the series , seems to show that the result is true .
we give now upper bounds for @xmath209 .
before to state the next proposition we introduce some definitions .
consider the rational function @xmath210 we define the sequence of rational functions @xmath211 as follows : @xmath212 where the @xmath213 are the pad approximants of order @xmath214 of the function @xmath215 .
for instance @xmath216 recall that the pad approximants @xmath213 are rational functions whose numerators and denominators are polynomials of degree @xmath51 and their taylor expansions in powers of @xmath83 are the same that the taylor expansion of @xmath215 up to order @xmath217 .
we write @xmath218 the rational functions @xmath211 vanish at @xmath219 and @xmath142 and their taylor expansions coincide with the taylor expansion of @xmath190 up to order @xmath217 .
as we will see in the proof of next proposition they are well defined for all @xmath220 $ ] because @xmath221 .
[ prop:3.4 ] let @xmath187\ } , $ ] be a parametrization of the heteroclinic solution of system . for @xmath102 and @xmath208 $ ] it holds that @xmath222 see figure [ nfig7 ] .
we will show that for each @xmath223 @xmath224 the flow of the vector field @xmath202 associated to system crosses each curve @xmath225 forwards .
we compute @xmath226 we obtain @xmath227 where @xmath228 and @xmath229 are polynomials of degree @xmath51 in @xmath83 .
the coefficients of @xmath228 and @xmath229 are polynomials in @xmath160 .
we want to prove that for @xmath170 both polynomials @xmath228 and @xmath229 are positive for @xmath102 .
we will approach the problem varying the parameter @xmath160 and studying which are the possible bifurcations for the number of zeros of these polynomials when @xmath102 .
the values of @xmath230 , @xmath231 , @xmath232 and @xmath233 , are polynomials in @xmath160 whose coefficients are natural numbers .
then we have @xmath234 , @xmath235 , @xmath236 and @xmath237 for @xmath170 . for @xmath238
, we evaluate now the discriminants of the polynomials @xmath228 and @xmath229 , @xmath239 and @xmath240 .
we obtain a polynomial in @xmath160 of degree @xmath241 with positive integers numbers as coefficients for @xmath239 and a polynomial in @xmath160 of degree @xmath242 with also positive integer numbers as coefficients for @xmath243 , where the degrees @xmath241 and @xmath242 are given in table [ ta:1 ] .
then , the two discriminants do not vanish for @xmath170 . in consequence , the number of real roots in @xmath106 of each polynomial , @xmath228 or @xmath229 , does not change for @xmath170 .
picking a concrete value of @xmath160 , for instance @xmath244 , we found , by applying the sturm algorithm , that they have no roots in @xmath106 .
then , we deduce that @xmath245 and @xmath246 for @xmath102 and @xmath170 .
therefore , we have proved that for @xmath247 @xmath248 for @xmath102 and @xmath170 , as we wanted to see .
the case @xmath249 is much easier because we do not need to compute the discriminants and we omit the details . [
cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] degrees of the numerator and the denominator of the resultants computed in the proof of proposition [ prop:3.4 ] . [ ta:1 ] finally for @xmath250 we obtain by a direct computation @xmath251 where @xmath252 is a polynomial with positive coefficients . hence @xmath253 as we wanted to show . as a corollary of propositions [ prop:3.2 ] and [ prop:3.4 ]
we obtain the following result : [ thm:3.5 ] let @xmath254\ } , $ ] be a parametrization of the heteroclinic solution of system .
for @xmath102 and @xmath208 $ ] it holds that @xmath255 if we were interested in obtaining more precise upper bounds for the heteroclinic orbit , we could simply increase @xmath51 and apply the same procedure .
for instance for @xmath244 we have performed all the computations and proved that for @xmath102 , @xmath256 where the functions @xmath257 @xmath258 are defined similarly to the ones given in the above theorem .
we remark that in all the cases the maximum error is at @xmath259 it is also important to notice that for bigger values of @xmath160 we need bigger values of @xmath51 in @xmath211 and @xmath189 to obtain similar bounds .
theorem [ thm:1.2 ] is simply a reformulation of the above theorem .
by using normal forms theory it is well - known that in a neighborhood of the origin of system its unstable manifold can be parametrized as @xmath260 for some analytic function @xmath261 . so it seems natural to find bounds of the actual heteroclinic orbit which are rational functions of @xmath262 .
as far as we know this idea is new .
first , we consider the family of rational functions @xmath263 to try to approximate the function @xmath264 , where we denote by @xmath265 the time - parametrization of the heteroclinic orbit @xmath209 .
notice that the parameters are taken in such a way that @xmath266 properties that are also satisfied by @xmath264 .
we do not care about the @xmath97 component because @xmath267 .
recall that the vector field associated to system is @xmath268 to study the behavior of the flow of this system on the curve @xmath269 where @xmath270 we compute @xmath271 if we introduce the compact notation @xmath272 we obtain that @xmath273 where @xmath274 is a polynomial of degree 3 with coefficients depending also polynomially on @xmath275 and @xmath276 . in order to simplify the expression of @xmath277 we take @xmath278 .
then @xmath279 finally , taking @xmath280 the numerator and the denominator of the above fraction have a common zero .
then it writes as @xmath281 in short we have proved the following : [ thm:4.1 ] let @xmath265 be the time - parametrization of the heteroclinic orbit @xmath209 of system that satisfies @xmath282 .
define @xmath283 then : a. when @xmath284 it holds that @xmath285 b. when @xmath286 it holds that @xmath287 c. when @xmath288 it holds that @xmath289 theorem [ thm:1.3 ] follows from the above result , simply using that @xmath290 and @xmath291 .
theorem [ thm:3.5 ] can also be used to obtain an explicit bound for @xmath292 since @xmath293 , where @xmath294 is the parametrization of @xmath209 , solving the cauchy problem @xmath295 we obtain a function @xmath296 such that @xmath297 the other bounds given in theorem [ thm:3.5 ] give rise to implicit bounds of the form @xmath298 of the curve @xmath299 .
this subsection improves the results of the previous section when @xmath284 .
fixed @xmath301 and given any natural @xmath51 , following similar techniques that the ones used to prove lemma [ lem:3.1 ] , we can compute a function @xmath302 , with @xmath303 and @xmath304 , that coincides with the solution @xmath264 until order @xmath217 in @xmath305 .
for instance @xmath306 looking @xmath305 as an independent variable , we compute the associated pad approximant of @xmath307 of order @xmath214 , obtaining @xmath308 where @xmath309 and @xmath310 are polynomials on @xmath311 for instance @xmath312 notice that it satisfies that @xmath313 but we do not impose neither that @xmath314 nor that @xmath315 as in the proof of theorem [ thm:4.1 ] , to study the behavior of the flow associated to system on the curve @xmath316 we compute the corresponding function .
it writes as @xmath317 where @xmath318 and @xmath319 are polynomials of degree @xmath320 with coefficients depending also polynomially on @xmath160 .
it can be seen that all these coefficients , as functions of @xmath160 , take positive values on @xmath321 we prove this fact introducing a new variable @xmath322 satisfying @xmath323 in each of the coefficients and then applying the descartes theorem to the resulting polynomials .
then @xmath324 for all @xmath82 and for all @xmath160 in @xmath321 moreover , this also shows that @xmath325 is well defined because @xmath326 since we want that @xmath314 we need to modify @xmath327 to do this , for @xmath328 we define the new family of functions @xmath329 it is not difficult to see that following the above procedure the same results hold .
moreover it can be proved that there exists a unique value @xmath330 such that @xmath331 so we define @xmath332 and it satisfies @xmath333 since the corresponding function is positive , for each @xmath334 it holds that @xmath335 moreover @xmath336 and it can be seen that these limits , which are rational functions in @xmath160 , satisfy @xmath337 furthermore @xmath338 decreases with @xmath320 .
for instance @xmath339 and @xmath340 .
in spite of the great interest of studying the traveling wave solutions of reaction - diffusion equations , @xmath341 there are no methods for obtaining explicit bounds for them . in this paper
we present an approach that allows to obtain this type of bounds in the general case .
moreover we introduce more elaborated tools , based on the control of the heteroclinic trajectories of an associated planar ordinary differential equation , that allow to improve the general results when we deal with a particular function @xmath18 .
we study with detail the classical fisher - kolmogorov case @xmath342 the methods developed can be easily adapted to treat the newell - whitehead - segel and the zeldovich equations a. kolmogorov , i. petrovskii and n. piscounov , _ a study of the diffusion equation with increase in the amount of substance , and its application to a biological problem _ , in v. m. tikhomirov , editor , _ selected works of a. n. kolmogorov i _ , 248270 .
kluwer 1991 .
translated by v. m. volosov from bull .
moscow univ . , math .
* 1 * ( 1937 ) , 125 .
y. b. zeldovich and d. a. frank - kamenetskii , _ a theory of thermal propagation of flame _ , acta physicochimica urss * 9 * ( 1938 ) , 341350 .
english translation : in dynamics of curved fronts , editor r. pelc , perspectives in physics series , academic press , new york ( 1988 ) , 131140 . | it is well - known that the existence of traveling wave solutions for reaction - diffusion partial differential equations can be proved by showing the existence of certain heteroclinic orbits for related autonomous planar differential equations .
we introduce a method for finding explicit upper and lower bounds of these heteroclinic orbits . in particular , for the classical fisher - kolmogorov equation we give rational upper and lower bounds which allow to locate these solutions analytically and with very high accuracy . |
recent theoretical @xcite and experimental @xcite studies have shown the possibility of creating novel types of microstructured materials that demonstrate the property of negative refraction . in particular , the composite materials created by arrays of wires and split - ring resonators were shown to possess both negative real parts of magnetic permeability and dielectric permittivity for microwaves .
these materials are often referred to as _ left - handed metamaterials _ , _
double - negative materials _ , _ negative - index materials _ , or _ materials with negative refraction_. properties of such materials were first analyzed theoretically by v. veselago a long time ago @xcite , but they were demonstrated experimentally only recently . as was shown by veselago @xcite ,
left - handed metamaterials possess a number of peculiar properties , including negative refraction for interface scattering , inverse light pressure , reverse doppler effect , etc . many suggested and demonstrated applications of negative - index metamaterials utilize unusual features of wave propagation and scattering at the interfaces .
in particular , the effect of negative refraction can be used to realize focusing with a flat slab , the so - called planar lens @xcite ; in a sharp contrast with the well - known properties of conventional lenses with a positive refractive index where curved surfaces are needed to form an image . moreover
, the resolution of the negative - index flat lens can be better than a wavelength due to the effect of amplification of evanescent modes @xcite .
direct numerical simulations provide the unique means for a design of microwave and optical devices based on the negative - index materials , however any realistic simulation should take into account metamaterial dispersion and losses @xcite as well as a nonlinear response @xcite .
such numerical simulations are often carried out within the framework of the effective medium approximation , when the metamaterial is characterized by the effective dielectric permittivity and magnetic permeability .
this simplification allows for modelling of large - scale wave dynamics using the well - known finite - difference time - domain ( fdtd ) numerical methods @xcite . in this paper
, we discuss the main features and major difficulties in applying the standard fdtd numerical schemes for simulating wave scattering by wedges and interfaces of finite - extend negative - index metamaterials , including a key issue of positioning of a discretization grid in the numerical yee scheme @xcite necessary to obtain consistent convergence in modeling surface waves at an interface between conventional dielectric and metamaterial with negative dielectric permittivity and negative magnetic permeability .
in particular , we demonstrate that , in the framework of the continuous - medium approximation , wave scattering on the wedge may result in a resonant excitation of surface waves with infinitely large spatial frequencies , leading to non - convergence of the simulation results that depend on the discretization step .
[ sect : model ] we consider a two - dimensional problem for the propagation of te - polarized electromagnetic waves in the plane @xmath0 ) , where the medium properties are isotropic and characterized by the dielectric permittivity @xmath1 and magnetic permeability @xmath2 . in the absence of losses , @xmath3 .
the response of negative - index materials is known to be strongly frequency dependent @xcite , however in the linear regime the wave propagation at different wavelengths can be described independently .
the stationary form of maxwell s equation for the complex wave envelopes is well - known @xmath4 where @xmath5 , @xmath6 , and @xmath7 are the components of the magnetic and electric fields , respectively , @xmath8 is angular frequency , and @xmath9 is the speed of light in vacuum .
the system of coupled equations ( [ eq : maxwell ] ) can be reduced to a single helmholtz - type equation for the electric field envelope , @xmath10 where @xmath11 , and @xmath12 is the refractive index of the medium .
[ sect : interface ] the concept of perfect sub - wavelength imaging of a point source through reconstitution of the evanescent waves by a flat lens has remained highly controversial @xcite because it is severely constrained by anisotropy and losses of the metamaterials .
nevertheless , several numerical studies showed that nearly - perfect imaging should be expected even under realistic conditions when both dispersion and losses are taken into account @xcite . in this section , we consider the numerical simulations of the wave scattering by a slab of the negative - index material , i.e. the problem close to that of the perfect lens imaging , and discuss the convergence of the yee numerical discretization scheme .
( a ) geometry of the scattering problem , the slab of the thickness @xmath13 is made of a negative - index metamaterial with both @xmath14 and @xmath15 negative .
( b ) discretization scheme in the yee method.,scaledwidth=65.0% ] we start our analysis by considering wave propagation through a slab made of the negative - index material , as schematically illustrated in fig .
[ fig : scheme1d ] , with homogeneous properties in the @xmath16 plane characterized by two functions , the electric permittivity @xmath17 and magnetic permeability @xmath18 . to solve this problem numerically ,
we employ the well - known numerical yee method @xcite and perform the discretization of the electric and magnetic fields on a square grid ( @xmath19 , @xmath20 ) presenting the fields in the form , @xmath21 where we use the notation @xmath22 then , we replace the continuum model by a closed set of the discrete equations for the field amplitudes obtained by averaging equations ( [ eq : maxwell ] ) over the cells of discretization mesh , taking into account the continuity of the tangential field components at the interface @xcite , @xmath23 { \displaystyle \frac{e_y|_{m , n+1 } - e_y|_{m , n}}{h } \langle{\mu^{-1}}\rangle_{m , n+1/2 } = - \frac{i \omega}{c } h_x|_{m , n+1/2 } , } \\*[12pt ] { \displaystyle \frac{e_y|_{m , n } - e_y|_{m+1,n}}{h } = \frac{i \omega}{c } \left\langle\mu\right\rangle_{m+1/2,n } h_z|_{m+1/2,n } . }
\end{array}\ ] ] whereas the general form of the discrete equations ( [ eq : maxwelld ] ) is well known @xcite , we point out a number of specific features arising in numerical simulations of the waves scattering at the interfaces with the negative - index media . since the real parts of both @xmath1 and @xmath2 change sign at these interfaces , the corresponding averaged values may become small or even vanish for a certain layer position with respect to the numerical grid .
in this case , eqs .
( [ eq : maxwelld ] ) may become ( almost ) singular , leading to poor convergence . in this paper
, we suggest that consistent convergence can be achieved by artificially shifting the layer boundary with respect to the grid in order to ensure that the averaged values do not vanish .
this shift will not exceed @xmath24 , assuring convergence as the step - size is decreased . because the tangential component @xmath7 of the electric field should be continuous at the interface
, is seems that a natural choice is to align the boundary position with the grid points @xmath25 , where @xmath26 is defined , and we use this configuration in the numerical simulations presented below .
however , we note that such a selection leads to singularities for averaged values if @xmath27 or @xmath28 , which coincides with the flat - lens condition . therefore , it is necessary to take into account losses in the metamaterial , described by nonzero imaginary parts of the complex values @xmath29 and @xmath15 , or to choose a different boundary alignment to the grid for the numerical simulations of perfect lenses @xcite .
( a ) spectrum of wavenumbers @xmath30 for a negative - index layer : exact ( circles ) and discrete ( crosses , for @xmath31 ) solutions .
( b ) absolute differences between the exact and discrete values of @xmath30 vs. the number of points ( @xmath32 ) along @xmath33 for the marked points 1,2,3 in ( a ) .
bottom : mode profiles marked 1,2,3 in ( a ) .
the computational domain is @xmath34 , @xmath35 is the width of the negative - index layer with @xmath36 and @xmath37 , and @xmath38 .
the wavenumber in vacuum is normalized as @xmath39 .
, scaledwidth=65.0% ] in order to illustrate the convergence of the proposed numerical scheme , we compare the solutions of discrete and continuous equations .
we note that wave scattering from an infinite layer is fully characterized by the properties of spatial modes , which wavevector components along the layer ( @xmath30 ) are conserved .
these modes have the form @xmath40 substituting eqs .
( [ eq : mode1d ] ) into eq .
( [ eq : maxwell ] ) and eq .
( [ eq : maxwelld ] ) , we obtain a set of corresponding continuous and discrete eigenmode equations . for every @xmath30
, the mode profiles can be determined analytically , e.g. , using the transfer - matrix method @xcite .
the wave spectrum can contain solutions corresponding to the guided modes of a negative - index layer @xcite , and extended ( or propagating ) modes that should also be taken into account as well , in order to describe scattering of arbitrary fields .
( a ) geometry of the scattering problem , the finite - extent slab is made of a negative - index metamaterials with both @xmath14 and @xmath15 negative .
( b ) discretization scheme in the yee method.,scaledwidth=65.0% ] dependence of ( a ) the cosine of the singularity parameter and ( b ) its real ( solid ) and imaginary ( dashed ) parts on @xmath41 according to eq .
( [ eq : corner ] ) .
shading marks the region with @xmath42.,scaledwidth=65.0% ] we solve the discrete eigenmode equations numerically for the slab geometry with periodic boundary conditions , and compare the spectrum of eigenvalues @xmath30 with exact solutions of the continuous model . in fig .
[ fig : spectr1d](a ) , we show a part of the spectrum of the discrete eigenvalues ( crosses ) , which indeed coincides with the exact values ( circles ) .
the rate of convergence can be judged from fig .
[ fig : spectr1d](b ) , where the differences between the approximate and exact solutions are shown in logarithmic scale .
[ sect : corner ] one of the fundamental problems in the theory of negative - index metamaterials is the wave scattering by wedges @xcite , where convergence of numerical methods can be slow due to the appearance of singularities @xcite . in this section , we demonstrate that the nature of such singularities has to be taken into account when performing fdtd numerical simulations .
[ sect : singularity ] ( a ) amplitudes of reflected plane waves vs. the number of points ; ( b , c ) electric field profiles for @xmath43 and @xmath44 .
parameters are the same as in fig .
[ fig : spectr1d ] , except @xmath45 .
, scaledwidth=65.0% ] ( a ) amplitudes of reflected plane waves vs. the number of points ; ( b , c ) electric field profiles for @xmath43 and @xmath44 .
parameters are the same as in fig .
[ fig : spectr1d ] . , scaledwidth=65.0% ] the behavior of the electric and magnetic fields at the wedges between homogeneous materials characterized by different values of @xmath1 and @xmath2 was described analytically in the pioneering paper of meixner @xcite and further refined in the subsequent studies ( see , e.g. , ref .
@xcite , and references therein ) . in the case of the te wave scattering by a negative - index wedge , as schematically illustrated in fig .
[ fig : wedge ] , the amplitudes of magnetic fields exhibit singular behavior at the wedge of the order of @xmath46 , where @xmath47 is the distance from the wedge . for a @xmath48 wedge angle ,
corresponding to a corner of a rectangular slab , the coefficient @xmath49 is found as @xcite @xmath50 where @xmath51 and @xmath15 are magnetic permeabilities of the two neighboring media . in the case of conventional dielectric or magnetic media with @xmath52 ,
( [ eq : corner ] ) has solutions with real @xmath49 .
however , when @xmath2 changes its sign at the interface with a negative - index medium , the coefficient @xmath49 becomes _ complex _ playing role of a singularity parameter .
this happens when @xmath53 so that @xmath54 , see fig .
[ fig : gamma ] . for real @xmath49 ( taking the solution with @xmath55 , according to ref .
@xcite ) , the field amplitudes decay monotonously away from the corner . in this case
, numerical simulations may be based on the simplest discretization , although the convergence rate can be improved by taking into account the singular behavior in the discrete equations .
however , the case of complex @xmath49 corresponds to the fields that _ oscillate infinitely fast _ near the corner because @xmath56.\ ] ] the second multiplier indicates the excitation of infinitely large spatial harmonics , however such a situation is unphysical because the effective - medium approximation of the negative - index materials is valid for slowly varying fields only .
therefore , in numerical simulations it is necessary to take into account the physical effects that suppress such oscillations , in particular we discuss the effect of losses in sec .
[ sect : losses ] below .
we now analyze convergence of the numerical finite - difference solutions for the problem of wave scattering by a finite - extent negative - index slab .
we align the boundaries of the negative - index domain with the grid , as shown schematically in fig .
[ fig : scheme1d](b ) . since the electric field components @xmath7 are continuous at the interfaces , it is possible to obtain the discrete equations that have the form of eqs .
( [ eq : maxwelld ] ) .
same as in fig .
[ fig : edge15 ] but for @xmath57 . ,
scaledwidth=65.0% ] in order to construct the full solutions for scattering problem , we decompose the field into a set of eigenmodes of the negative - index layer ( @xmath58 ) and free space ( @xmath59 ) .
more specifically , for @xmath58 we have @xmath60 where @xmath61 is the number of eigenmodes . here
the summation is performed over _ the propagating modes _ ( @xmath62 ) which transport energy away from the interface , and _ evanescently decaying modes _ with @xmath63 . in free space at @xmath59 , the field is composed of incident and reflected plane waves , @xmath64 here @xmath32 is the number of points in the @xmath33 direction , and the discrete wavenumber is @xmath65^{1/2 } 2/h$ ] @xcite .
the sign of @xmath66 is chosen with a proper wave asymptotic behavior , i.e. , we choose @xmath67 , if @xmath66 is real , and @xmath68 if @xmath66 is complex .
the magnetic field at @xmath59 is found from eq .
( [ eq : maxwelld ] ) , with homogeneous parameters @xmath69 , @xmath70 , @xmath71 .
then , we substitute eqs .
( [ eq : blochdiscr ] ) , ( [ eq : planediscr ] ) into the first of eq .
( [ eq : maxwelld ] ) , and using the condition of the continuity of the electric field , obtain a set of equations for all @xmath72 that are used to calculate the amplitudes @xmath73 and @xmath74 of the reflected and transmitted waves , @xmath75 { \displaystyle \sum_j a_j \left [ { \cal h}_x|_{m,-1/2}(k_z^{(j ) } ) + { \cal h}_x|_{m,1/2}(k_z^{(j ) } ) \right ] } \\*[12pt ] { \displaystyle \qquad = \frac{- 2 c}{\omega \mu_1 h } \sum_{j =- n/2 + 1}^{n/2 } ( f_j^{\rm ( in)}-f_j^{\rm ( refl ) } ) \sin(k^{(j ) }
h ) \exp ( i 2 \pi j m / n ) . } \end{array}\ ] ] these equations are solved using the standard linear algebra package .
same as in fig .
[ fig : edge15 ] but for @xmath76 . ,
scaledwidth=65.0% ] we now present results of our numerical simulations for the scattering of normally incident plane waves , with @xmath77 , @xmath78 .
first , we consider scattering by a negative - index slab with @xmath79 . in this case , we observe a steady convergence of numerical solutions , as shown in fig .
[ fig : edge35 ] .
this demonstrates that even simplest finite - difference numerical schemes can be successfully employed to model the scattering process when the sinularity parameter @xmath49 is real , is in a full agreement with earlier studies of wave scattering at dielectric wedges @xcite .
however , the situation changes dramatically when @xmath49 is complex , i.e. for @xmath37 . according to the analytical solution , in this case the magnetic field should oscillate infinitely fast in the vicinity of the corner , corresponding to excitation of infinitely large spatial harmonics .
however , such behavior can not be described by discrete equations , and we find that in this regime _ solutions of finite - difference equation do not converge _ , as demonstrated in fig .
[ fig : edge15 ] . [ sect : losses ] the analytical description of the edge singularities discussed above is only valid for lossless media , i.e. when all @xmath1 and @xmath2 are real .
however , the negative - index metamaterials always have non - vanishing losses , and we have studied whether this important physical effect can regularize the field oscillations at the corner . however , our results demonstrate that even substantial losses may not be sufficient enough to suppress such oscillations , as presented in the example of fig . [
fig : edge15loss ] . finally , we consider the problem of wave scattering from the corners of perfect lenses , where @xmath80 and @xmath81 ( we take @xmath38 ) .
this is a special case , where the type of singularity becomes indefinite if losses are neglected .
we find that introducing _
sufficiently large losses _ does indeed regularize the field oscillations at the corners , leading to convergence of numerical simulations , as demonstrated for the example of fig .
[ fig : edge101loss ] .
however , this only occurs when the value of losses exceeds a certain threshold ; if the losses are too weak then non - convergent behavior is again observed , as shown in fig .
[ fig : edge101losslow ] .
the threshold value of losses for achieving convergence of the numerical scheme is increased for larger @xmath82 .
we note that this is completely different from the temporal dynamics at an infinitely extended slab , where convergence to a steady state is eventually achieved with arbitrarily small losses @xcite .
same as in fig .
[ fig : edge15 ] but for @xmath83 . ,
we have discussed a number of specific features manifested in numerical simulations of wedges and interfaces of metamaterials , i.e. composite materials with negative dielectric permittivity and negative magnetic permeability . we have demonstrated that a numerical discretization grid in the yee method may have a dramatic effect on the convergence in numerical modelling of surface waves at interfaces and wedges . in the framework of the continuous - medium approximation , wave scattering on
the wedge may result in a resonant excitation of surface waves with infinitely large spatial frequencies , leading to non - convergence of the numerical simulation results that depend strongly on the value of the discretization step .
we find that sufficiently high losses may suppress oscillations and allow to obtain converging solutions to the scattering problem , however in the case of smaller losses it may be necessary to take into account the meta - material properties beyond the effective - medium approximation , such as the effect of spatial dispersion .
the authors thank alexander zharov and pavel belov for useful discussions and suggestions . this work has been supported by the australian research council . c. g. parazzoli , r. b. greegor , k. li , b. e. c. koltenbah , and m. tanielian , `` experimental verification and simulation of negative index of refraction using snell s law , '' phys .
* 90 , * 1074014 ( 2003 ) .
v. g. veselago , `` the electrodynamics of substances with simultaneously negative values of @xmath1 and @xmath2 , '' usp . fiz .
nauk * 92 , * 517526 ( 1967 ) ( in russian ) [ english translation : phys .
* 10 , * 509514 ( 1968 ) ] .
n. a. zharova , i. v. shadrivov , a. a. zharov , and yu . s. kivshar , `` nonlinear transmission and spatiotemporal solitons in metamaterials with negative refraction , '' optics express * 13 , * 12911298 ( 2005 ) . | we outline specific features of numerical simulations of metamaterial wedges and interfaces .
we study the effect of different positioning of a grid in the yee method , which is necessary to obtain consistent convergence in modeling of interfaces with metamaterials characterized by negative dielectric permittivity and negative magnetic permeability .
we demonstrate however that , in the framework of the continuous - medium approximation , wave scattering on the wedge may result in a resonant excitation of surface waves with infinitely large spatial frequencies , leading to non - convergence of the simulation results that depend on the discretization step . |
the sunyaev - zeldovich ( 1972 , sz ) effect , unlike optical and x - ray cluster surveys , is not affected by redshift because it measures the integrated line of sight intracluster gas pressure via its compton scattering of cosmic microwave background ( cmb ) photons . during the last 10 years
, interferometric techniques have been developed , which are providing firm detections of known clusters ( eg .
jones et al .
1993 , carlstrom , joy and grego 1996 ) .
the technology and expertise is now available to survey the sky to discover clusters .
the construction of a cmb telescope for the observation of the cmb on angular scales of one to several arcminutes and with a sensitivity of a few micro - kelvin , comparable to the planck surveyor , has been proposed in cambridge .
the arcminute microkelvin imager ( ami ) will consist of a small compact array of 4 meter dish antennas combined with the existing ryle telescope antennas in an extended array , with a new receiving system and novel correlator .
this design achieves optimal sensitivity to the cluster sz effect and a separation with other components such as the primary cmb and radio sources .
although the instrument is dedicated to the study of clusters , it will generally probe the structure of the cmb on sub - planck and super - alma scales .
therefore it is also sensitive to other phenomena such as , inhomogeneous ionisation , density - velocity correlations ( vishniac and ostriker effect ) , filaments and topological defects , if they exist , which are all of immense interest as well . in
the following however we demonstrate the ability of ami to discover clusters . to produce simulated sz cluster sky maps , the press - schechter expression ( 1974 ) is used to create a list of cluster masses and redshifts having an abundance consistent with the local cluster temperature function ( eke et al .
these clusters are placed at random angular positions within a 5@xmath3 @xmath4 5@xmath3 sky map with 40 arcsec pixels . to model the cluster sz signal ,
template maps have been created from the hydrodynamical simulations of eke , navarro and frenk ( 1998 ) , and these are pasted , suitably scaled , onto the cluster positions .
this procedure is performed for two cases , a low present density ( @xmath5 ) and a high density ( @xmath6 ) universe , both with a hubble constant of 70 km s@xmath7 mpc@xmath7 .
the gas fraction is fixed at 10 % in both cases rather than to the primordial nucleosynthesis value , which would have introduced an @xmath8 dependence , enhancing the differences .
our value for the gas fraction is at the low end of the value estimated from x - ray clusters ( ettori and fabian ( 1999 ) and mohr , mathiesen and evrard ( 1999 ) find it to be 0.1 - 0.25 at the 95 % cl ) , and the model sz number counts would increase with a less conservative choice .
we also produced corresponding x - ray maps in 0.5 - 2 and 2 - 10 kev bands , and have checked that they are complete to an x - ray flux limit of @xmath9 erg @xmath10 s@xmath7 [ 0.5 - 2 kev ] .
the sensitivity of future instruments , we use ami specific numbers , is high enough to allow a survey of the sky for sz clusters . to make our simulation of the observation process computationally feasible we simplified the observation by running a compensated beam with carefully chosen shape and amplitude over the cluster map .
this procedure had been gauged with detailed simulations of the interferometer response including the specific noise properties of a field observation and through the recovery process .
the resulting number counts are shown in figure 1 .
there is a turn - over at the confusion limit for a 4.5 arcmin beam . also shown in figure 1
are sensitivity lines each corresponding to a fixed observation time and displayed as the inverse of the survey area against the flux limit .
the ratio between the number counts and the sensitivity lines is the number of clusters detected in the respective time .
the maximum detection number lies at a limiting observed flux of about 100 @xmath11jy or a corresponding survey area of 5 deg@xmath12 and about 10 clusters for the high density case and tens of clusters for low density are expected for a 6 months observation . since the optimum is shallow it will also be possible to adopt different surveying strategies at a low cost of inefficiency .
this will probe the slope of the number counts which is a function of @xmath8 .
the parameters of the simulations , which have the largest effects on the number counts are listed with present uncertainties in table 1 .
their effect is considerable , but substantial improvements can be expected in the future , in particular from the new x - ray missions . due to the @xmath13 dimming of bolometric x - ray flux compared to the compton scattering process
the x - ray and sz flux limited cluster samples have very different redshift distributions , even for an x - ray limit of @xmath14 erg @xmath10 s@xmath7 [ 0.5 - 2 kev ] in the case of the currently favoured low density model .
with scaling laws we find that the ratio between sz to x - ray flux scales with @xmath15 and with only a weak dependence on @xmath8 and the cluster temperature .
when observing the same field in microwaves and x - rays this `` photometric '' redshift effect will immediately allow a crude measure of the redshift distribution ( see figure 2 ) . a limit on the redshift will be obtained for the sz clusters undetected in x - rays .
a major uncertainty in determining @xmath8 from this distribution comes from @xmath2 .
however when optical redshifts are obtained we find that for the flux limit reached with ami the median redshift of only 20 clusters will allow to distinguish between the two cases with more than 99 % confidence . going back to the x - ray and sz data with an estimate of @xmath8 , the gas physics for example @xmath16 can be studied .
ccc + parameter & change in percent & fractional change in @xmath17 + @xmath18 & 20 % & 1.3 + @xmath2 & 30 % & 1.5 + @xmath19 & 7 % ( 1@xmath20 ) & 1.5 + & 14 % ( 2@xmath20 ) & 3.2 +
instruments to survey the sky for sz clusters can be build .
the expected number of clusters is greater than 10 for a half year observation and depends strongly on the matter density .
the sz data in combination with x - ray and optical data can be used to constrain @xmath21 , @xmath19 , and from distance measurements , @xmath0 and most likely @xmath22 , if enough suitable clusters are present at high redshift .
sz cluster surveys will be particularly useful as pathfinders for future x - ray missions such as constellation - x and xeus , which will have small fields of view .
with the cluster sz effect a sample of massive clusters at high redshift can be provided as targets for a detailed study of the plasma physics in x - rays , interesting for the understanding of cluster formation and cosmology .
99 carlstrom j.e . ,
joy m. , grego l. , 1996 , , 456 , l75 eke v.r . , cole , s. , frenk c.s . , 1996 , 282 , 263 eke v.r . , navarro j.f .
, frenk c.s . , 1998 , 503 , 569 ettori s. , fabian a.c . , 1999 , 305 , 834 jones m.e . , et al . , 1993 , 365 , 320 mohr j.j.,mathiesen b. , evrard a.e . , 1999 , 517 , 627 press w.h . , schechter p. , 1974 , 187 , 425 sunyaev r.a .
, zeldovich ya.b . , 1972 , comm .
space phys . 4 , 173 | hydrodynamical simulations are used in combination with the press - schechter expression to simulate sunyaev - zeldovich ( sz ) galaxy cluster sky maps . these are used to gauge the ability of future sz observations to provide information about the cosmological parameters @xmath0 , @xmath1 , and the gas fraction @xmath2 in clusters .
this work concentrates on prospects for ami , the arcminute microkelvin imager , a new type of compact interferometric array currently proposed in cambridge .
the expectations are contrasted with those for x - ray missions , such as xmm , and the benefits of combining sz and x - ray data are highlighted . |
the propagation of two pulses in resonant interaction in a @xmath0- atomic medium has been widely studied during the last decade , in particular with application to quantum information processing ( see , e.g. , reviews @xcite ) .
the possibility of quantum information storage in the gas phase has been demonstrated in cold atoms @xcite and in hot ensembles @xcite . achieving information storage in solid - state systems , which are more attractive due to their higher density , compactness , and absence of diffusion ,
has been pursued @xcite .
the main drawbacks preventing the efficiency of storage for such solid - state materials are huge inhomogeneous broadenings and high rates of decoherence .
for instance , in crystalline films doped with rare - earth metals the rates of transverse relaxations amount to tens of ghz @xcite . in practice an efficient storage of information requires rather large optical length such that even a weak loss rate will ruin it @xcite . in order to reduce the inhomogeneous broadening ,
it was proposed in a number of works to use the so - called hole burning technique @xcite .
it has been shown that media prepared in such a way allow efficient coherent population transfer in @xmath0 systems @xcite .
however , the hole burning technique faces a so far unsolved problem that leads to the reduction of the optical length of the samples , which will be detrimental in general for the efficiency of the storage . in the present work
we show that , for the storage of optical information ( i.e. of classical fields ) , the recording length can be dramatically reduced with the use of intense pulses .
the possibility to store and retrieve optical information in resonant media has mainly been studied in the `` linear approximation '' with respect to the so - called probe field that has to be stored , i.e. for a weak probe pulse .
it was usually assumed that the control pulse propagates in a medium without pulse shape change .
numerical simulation of this problem without restrictions on the probe intensity was performed in @xcite .
analytical studies of the problem taking into account the group velocities of both pulses were performed in @xcite in the limit of pulses of duration much shorter than all the relaxation times , but sufficiently long to allow adiabatic evolution during the interaction .
it has been found that the length of information storage depends remarkably on the ratio between the oscillator strengths of the adjacent transitions that determine the group velocities of the pulses in a medium .
the present work aims at a complete analytical study of information storage in case of arbitrary relaxation times and intensities of the probe and control fields .
we show analytically and numerically that it is possible , for a proper choice of solid - state medium , to dramatically reduce the optical length needed to store intense short pulses .
this would in particular make more efficient the use of the hole burning technique .
the paper is organized as follows : we first describe the model , and analyze the propagation in the adiabatic limit .
we next derive the conditions of storage for short lengths of the medium , before concluding . -system
interacting with two laser pulses . ]
we consider a medium of @xmath0 three - level atoms with two ground states @xmath1 and a metastable excited state @xmath2 , of energy @xmath3 , @xmath4 , and two laser pulses referred to as the probe and control pulses , coupling the transition respectively @xmath5 and @xmath6 ( see fig .
the interaction hamiltonian in the rotating wave approximation can be written in the basis @xmath7 as @xmath8\ ] ] with @xmath9 the one - photon detuning , @xmath10 ( @xmath11 ) the probe ( coupling ) laser frequency , @xmath12 the rabi frequencies associated to the corresponding field amplitudes @xmath13 and to the dipole moments @xmath14 of the corresponding atomic transitions .
we consider an exact two - photon resonance .
the propagation of two laser pulses in the medium is described by the maxwell equations which , in the slowly - varying - amplitude approximation and in the moving coordinate system @xmath15 ( such that the original wave operator @xmath16 becomes @xmath17 ) , read @xcite @xmath18 here @xmath19 are the coupling factors with @xmath20 the density of atoms .
the density matrix elements @xmath21 are determined by the system of equations ( where the dot denotes the derivative @xmath22 , since @xmath23 ) [ syst ] @xmath24 with @xmath25 and @xmath26 the rates of spontaneous emission from state 3 to states 1 and 2 respectively , @xmath27 the dephasing rate between the two ground states , and @xmath28 the transverse relaxation rate .
we moreover assume pulses of durations much shorter than the characteristic time of decoherence of the ground states ( usually microseconds in solids ) such that @xmath27 will be neglected .
the envelopes of the fields @xmath29 , the rabi frequencies @xmath30 , and the density matrix components @xmath31 depend on the position @xmath32 $ ] in the medium of length @xmath33 and on the time @xmath34 .
the boundary conditions for the maxwell equations are the pulses @xmath35 , @xmath36 given at the entrance @xmath37 of the medium .
we assume that all atoms in the medium are in state 1 before the interaction with the laser fields .
it is well known that the use of a counterintuitive sequence of pulses in two - photon resonance , here the coupling pulse switched on before the probe pulse , leads , in the limit of slow evolution of the pulses , to an adiabatic passage along the dark state ( of null eigenvalue ) immune to decoherence .
the system of linear equations ( [ syst ] ) can be recast as @xmath38 with @xmath33 being the associated linear operator and @xmath39 the formal ( small ) adiabatic parameter ( we have omitted here the dependence on @xmath40 for simplification ) .
the change of variable @xmath41 leads to @xmath42 the components of the eigenvector @xmath43 of null eigenvalue satisfying @xmath44 read [ adiab ] @xmath45 with @xmath46 as @xmath47 in the adiabatic limit , there is thus no coherence between the ground states and the excited state , which induce no pulse distortion during the propagation @xcite . to determine non - adiabatic corrections , we make around the adiabatic solution @xmath48 the superadiabatic expansion : @xmath49 inserting this expansion in eq .
( [ lindblad_adiab ] ) and identifying the orders of @xmath39 , we obtain a system of linear equations that can be solved iteratively .
for the first order @xmath50 denoting @xmath51 , we get [ superadiab ] @xmath52,\\ \label{superadiabc } \rho^{(1)}_{31}&=&e^{i\varphi_p}\frac{\cos\theta}{\omega}\left(i\dot\theta -\frac{\dot\varphi}{2}\sin2\theta\right),\\ \label{superadiabd } \rho^{(1)}_{32}&=&e^{i\varphi_c}\frac{\sin\theta}{\omega}\left(-i\dot\theta + \frac{\dot\varphi}{2}\sin2\theta\right)\end{aligned}\ ] ] with @xmath53 equations ( [ superadiaba ] ) and ( [ superadiabb ] ) show that at the first superadiabatic order the excited state stays unpopulated and that the adiabatic dynamics along the dark state is subjected to a loss of order @xmath54 and to a pure dephasing loss of order @xmath55 with @xmath56 the characteristic time of interaction . equations ( [ superadiabc ] ) and ( [ superadiabd ] ) show the transient appearance of coherences between the ground states and the excited state , which depend on the first - order non - adiabatic loss @xmath57 ( and also on @xmath58 ) , but _ not _ on the relaxation loss
. these coherences will be responsible of pulse distortion during the propagation [ through the maxwell equations ( [ maxwell ] ) ] , which will be thus independent of the relaxation loss .
inspection of the solution ( [ superadiab ] ) allows one to identify the small parameter @xmath59 that permits the expansion ( [ expansion ] ) ( at least at the first order ) .
we get the conditions : @xmath60 the two first conditions ( large pulse area and small one - photon detuning ) are the standard condition for adiabatic passage for a non - lossy system
. the third condition can be interpreted as the requirement of a long interaction time and a large pulse area with respect to the decay of the dynamics along the dark state .
this third condition shows that a fast relaxation @xmath61 usually encounters in solids can be compensated by a strong pulse area satisfying @xmath62 we have added in eqs .
( [ conds ] ) the well known fourth condition considering an additional decoherence as a pure dephasing of rate @xmath27 between the two ground states .
it is frequently mentioned that the first condition of adiabaticity is not required and that the control pulse may have a rectangular shape ( see for instance @xcite ) .
it is indeed shown in this ref .
@xcite that in case of a rectangular control pulse an efficient storage of information is possible .
however , the parameter @xmath56 of the above conditions characterizes the time of interaction when the probe and the control pulses _ both _ interact , i.e. when they _ overlap_. one can thus consider a rectangular control pulse but turned on earlier and turned off later than the probe pulse if one preserves the adiabaticity during the field overlapping .
using the density matrix at first order ( [ expansion ] ) with ( [ adiab ] ) and ( [ superadiab ] ) , we can rewrite the maxwell equations ( [ maxwell ] ) as [ maxwell2 ] @xmath63 with @xmath64 respectively the two - photon transition strength and the group velocity , and the boundary condition @xmath65 denoting the photon fluxes ( @xmath66 ) @xmath67 eq .
( [ maxwell2b ] ) can be rewritten as @xmath68 and interpreted as the conservation of the photon fluxes during the propagation : @xmath69 , whereas at high relaxation rates the inequality ( [ cond_fast_relax ] ) should be fulfilled .
solutions of the system of equations ( [ maxwell2 ] ) have been obtained and studied for different regimes of propagation in a number of works ( see , e.g. , the review @xcite ) .
the method of characteristics , recalled in appendix [ app_characteristics ] , allows one to determine them .
in particular , equal transition strengths @xmath70 lead to a constant value @xmath71 , and from eq .
( [ maxwell2b ] ) to @xmath72 independent of @xmath40 , thus of shape invariance during the propagation : @xmath73 , as well as @xmath74 : @xmath75 with @xmath76 given at the entrance of the medium .
the pulses [ through their rabi frequencies @xmath77 and @xmath78 propagate in a complicated manner through @xmath79 and the mixing angle of solution [ see eq .
( [ ode_sol ] ) ] @xmath80 where @xmath81 is defined by ( [ xi ] ) ( where @xmath82 should be replaced by @xmath34 ) @xcite @xmath83 when @xmath79 is additionally constant , the excitation propagates at constant velocity satisfying @xmath84 .
this propagation regime is known as adiabaton @xcite . in the more general case of unequal transition strengths , eq .
( [ maxwell2a ] ) entails that the mixing angle still reads @xcite @xmath85 and from ( [ q ] ) that @xmath86 with @xmath87 , and @xmath81 determined from eq .
( [ xi ] ) ( where @xmath82 should be replaced by @xmath34 ) : @xmath88 @xmath89 propagates in a more complicated manner than in case of equal transition strength as @xmath90 .
since @xmath91 and @xmath92 do not propagate at the same velocity , @xmath93 is not any more shape - invariant during the propagation .
if we assume constant phases @xmath94 at the entrance of the medium , one can conclude from eq.([maxwell2c ] ) that they will thus not change during the propagation .
note that this statement is only valid at the exact two - photon resonance .
evolution of the phase self - modulation for a two - photon detuning different from zero has been analyzed in @xcite . in the linear approximation ( with respect to the probe field ) , i.e. , in the first order with respect to @xmath95 , one has @xmath96 , @xmath97 , @xmath98 .
we remark that in case of equal strengths of adjacent transitions @xmath70 , similar expressions are obtained with @xmath89 in place of @xmath99 .
thus , in case of equal transition strengths the nonlinearity in the probe field results in replacement of the rabi frequency of the control field by the generalized rabi frequency @xmath89 .
the nonlinearity is more important when the adjacent transition strengths are unequal and can lead to formation of shock - wave fronts and violation of the interaction adiabaticity , as shown in @xcite .
the storage of the probe pulse is achieved when its time profile is mapped into the spatial distribution of the atoms in the medium .
more precisely , the probe pulse can be encoded in the spatial distribution of the angle @xmath100 after the passage of the pulses through the medium , inducing for instance a storage in the coherence between the lower states ( in the adiabatic limit ) : @xmath101.\ ] ] an efficient storage requires that the excitation , through @xmath102 , propagates adiabatically until it eventually stops , which necessitates a long enough medium .
it is usually achieved for a strong control pulse switched on before and switched off after a weak probe .
we present below a novel pulse configuration with a strong probe and a weak control that allows one to shorten significantly the length of storage . $ ] , @xmath66 , at the entrance of the medium with @xmath103 and @xmath104 : a ) usual scheme ( shown here for @xmath105 , @xmath106 ) ; b ) proposed scheme with @xmath107 , @xmath108 .
, title="fig : " ] $ ] , @xmath66 , at the entrance of the medium with @xmath103 and @xmath104 : a ) usual scheme ( shown here for @xmath105 , @xmath106 ) ; b ) proposed scheme with @xmath107 , @xmath108 .
, title="fig : " ] the storage requires the stopping of the excitation , @xmath109 , for all the characteristics connected with the initial excitation at the entrance of the medium @xmath110 . we can estimate the length @xmath111 required to completely store the excitation from ( [ xi2 ] ) using the longest characteristic , i.e. @xmath112 and @xmath113 @xcite : [ xmax ] @xmath114 with @xmath115 and @xmath116 the duration of respectively the probe and the control , and @xmath117 , @xmath66 .
minimizing this quantity requires thus @xmath118 and a weak probe pulse , as is well known . on the other hand , adiabaticity of the interaction is required .
criterion for this adiabaticity has been analyzed in @xcite , where it has been shown that , under the certain condition adiabaticity is broken with formation of shock - wave fronts for distances exceeding the critical length @xmath119 estimated as @xmath120 the latter approximation holds for a probe pulse of duration @xmath115 shorter than the control .
this shows that reducing @xmath121 to lower @xmath111 will however also shorten @xmath119 .
preserving adiabaticity usually requires thus a strong control pulse and an efficient storage will take place when @xmath122 through their asymmetry in @xmath123 and @xmath124 with respect to @xmath121 , eqs .
( [ xmax ] ) and ( [ x0 ] ) show that we can shorten @xmath111 with the use of a _ weak control _ pulse in addition to using a small @xmath121 parameter , while preserving the condition of adiabaticity using a
_ strong probe _ ( see fig . 2 for the pulse scheme ) . for @xmath125 : a ) usual scheme of pulses with gaussian envelopes at the entrance of the medium displayed in fig .
2a with @xmath126 ; b ) novel scheme of pulses ( displayed in fig . 2b).,title="fig : " ] for @xmath125 : a ) usual scheme of pulses with gaussian envelopes at the entrance of the medium displayed in fig .
2a with @xmath126 ; b ) novel scheme of pulses ( displayed in fig . 2b).,title="fig : " ] + we have performed numerical calculations of the full set of the density - matrix and maxwell equations for both pulses taking into account a fast relaxation @xmath104 .
it indeed shows that the storage of the probe field in inhomogeneously broadened media is more efficient when we use atoms of different adjacent transition strengths .
3b presents the temporal evolution of the probe pulse at different propagation lengths ( normalized as @xmath127 ) for the novel scheme of the pulses at the entrance of the medium shown in fig .
the chosen parameters lead to @xmath128 , @xmath129 , and @xmath130 , which satisfy the adiabatic conditions ( [ conds ] ) and ( [ cond_fast_relax ] ) at the entrance of the medium ( here the time of interaction is @xmath131 ) . for the comparison we present in fig .
3a the dynamics for the usual scheme of pulses used for the optical information storage .
one can see from figs . 3a and 3b that in the case of the novel scheme the length of information storage is dramatically reduced as compared with that of the usual case .
namely , for the novel scheme at the propagation length @xmath132 of about 1 the probe pulse is completely absorbed by the medium while at the same length in the usual scheme the probe pulse remains still unabsorbed .
we have shown that the optical information storage of a probe in inhomogeneously broadened media of @xmath0 atoms occurs for short medium length when strong probe pulses , weak control pulses and very different strengths of adjacent transitions are used .
for such a scheme the optical length may be of the order of unity in contrast to the conventional scheme where the optical length must well exceed unity .
the proposed scheme would allow one to reduce the unavoidable losses caused by decoherence .
the required @xmath0 systems with very different strengths of adjacent transitions are expected to be found in , e.g. , rare - earth doped solid - state materials .
study to elaborate efficient schemes to retrieve the stored pulse in this strong regime is in progress .
we briefly recall here the characteristics method ( see for instance @xcite ) to solve the maxwell equation @xmath133 in the case of a non - linear group velocity of the form : @xmath134 .
we want to transform this linear first order partial differential equation into an ordinary differential equation along the characteristic curves @xmath135 . using the chain rule @xmath136
, we obtain @xmath137 if we set ( choosing @xmath138 and denoting @xmath139 ) equation ( [ ode ] ) leads to the solution @xmath141 which is thus a constant along the characteristics .
@xmath142 allows one to label a characteristics @xmath143 rewritten from eq .
( [ cond_odeb ] ) as ( since @xmath144 ) @xmath145 the non - linear velocity @xmath74 derived from ( [ cond_odec ] ) : @xmath146 corresponds to the instantaneous velocity along a characteristic .
we acknowledge support from intas 06 - 100001 - 9234 , the research project ansef ps - opt-1347 of the republic of armenia , the agence nationale de la recherche ( anr comoc ) and the conseil rgional de bourgogne .
lukin , d.f .
phillips , a. fleischhauer , a. mair , r.l .
walsworth , phys .
lett . * 86 * , 783 ( 2001 ) ; a.s .
zibrov , a.b .
matsko , o. kocharovskaya , y.v .
rostovtsev , g.r .
welch , m.o .
scully , phys.rev.lett . *
88 * , 103601 ( 2002 ) ; m.d .
eisaman , a. andre , f. massou , m. fleischhauer , a.s .
zibrov , m.d .
lukin , nature * 438 * , 837 , ( 2005 ) ; r. pugatch , m. shuker , o. firstenberg , a. ron , n. davidson , phys .
lett . * 98 * , 203601 ( 2007 ) ; p. k. vudyasetu , r. m. camacho , j. c. howell phys .
rev . lett . * 100 * , 123903 ( 2008 ) .
e. kuznetsova , o. kocharovskaya , ph .
hemmer , m.o .
scully , phys .
a * 66 * , 063802 ( 2002 ) ; a.v .
turukhin , v.s .
sudarshanam , m.s .
shahriar , j.a .
musser , b.s .
ham , p.r .
* 88 * , 023602 ( 2002 ) ; s.e .
yellin , p.r .
hemmer , phys .
a , * 66 * , 013803 ( 2002 ) ; l. alexander , j. j. longdell , m. j. sellars , and n. b.manson , phys .
lett . * 96 * , 043602 ( 2006 ) ; m. u. staudt , s. r. hastings - simon , m. nilsson , m. afzelius , v. scarani , r. ricken , h. suche , w. sohler , w. tittel , and n. gisin , phys . rev . lett . * 98 * , 113601 ( 2007 ) ; s@xmath153rensen , m. d. lukin , and r. l. walsworth , phys . rev . lett . * 98 * , 243602 ( 2007 ) .
i. novikova , a. v. gorshkov , d. f. phillips , a. s. s@xmath153rensen , m. d. lukin , r. l. walsworth , phys .
lett . * 98 * , 243602 ( 2007 ) ; a. v. gorshkov , a. andr , m. fleischhauer , a. s. s@xmath153rensen , m. d. lukin , phys .
lett . * 98 * , 123601 ( 2007 ) . | we propose a novel scheme of storage of intense pulses which allows a significant reduction of the storage length with respect to standard schemes . this scheme is particularly adapted to store optical information in media with fast relaxations . |
an interesting extension of the standard model ( sm ) proposed by manohar and wise @xcite consists of adding a color - octet electroweak doublet scalar .
the original motivation for this extension is that it consists of one of two scalar representations allowed by minimal flavor violation for scalars that do not transform under the flavor group . from this perspective
the most general renormalizable potential consistent with this field content is constructed resulting in generic studies for tev scale color octet scalars .
scalars such as these occur in specific models , for example in unification with @xmath4 @xcite or with @xmath5 @xcite .
the model has interesting phenomenology and many studies have been performed in the literature @xcite . a salient feature of the mw model being that it is very difficult to observe or exclude the new scalars at the lhc .
the model has a very large parameter space , but these studies have shown that it is possible to constrain it significantly by imposing custodial symmetry , partial wave unitarity and vacuum stability . along these lines , we have recently proposed considering the mw extension in the context of a two higgs doublet model ( 2hdm ) . in ref .
@xcite we introduced the model and constrained its parameter space by imposing tree level theoretical constraints arising from symmetries and from perturbative unitarity .
in addition we briefly discussed lhc phenomenology , concluding that the largest effects from this addition to the 2hdm would appear in corrections to one - loop couplings of the higgs boson to two gluons or photons .
previous studies of the mw model @xcite and of the 2hdm @xcite indicate that the viable parameter space is further constrained when one includes renormalization group corrections , and in this paper we apply this rationale to the mw extension of the 2hdm .
we first compute the beta functions for the couplings of the model .
we then consider the requirements that there be no landau poles ( lp ) below a certain high mass scale @xmath6 , that the scalar potential remains stable and that two - to - two scattering amplitudes remain perturbative at all scales below @xmath6 .
these conditions have a long history of being used as theoretical constraints from their early application to the sm higgs boson mass @xcite .
the construction of the model was described in ref .
the scalar sector of the sm is replaced by three @xmath7 scalar doublets : two color singlets ( @xmath8 and one color - octet @xmath9 . the most general renormalizable potential for the 2hdm sector ( @xmath8 is well known from the literature @xcite and we restrict ourselves to the case of a cp conserving potential with a discrete symmetry @xmath10 that is only violated softly ( although in the end we replace this with the requirement of mfv ) . to this known potential
we add the couplings between the color octet @xmath9 and the two color singlets ( @xmath8 as well as the color octet self interactions @xcite resulting in @xmath11 \nonumber\\ & + & 2m_s^2 { \rm tr}s^{\dag i}s_i + \mu_1 { \rm tr}s^{\dag i}s_i s^{\dag
j}s_j + \mu_2 { \rm tr } s^{\dag i}s_j s^{\dag j}s_i + \mu_3 { \rm tr } s^{\dag i}s_i { \rm tr}s^{\dag j } s_j\nonumber\\ & + & \mu_4 { \rm tr}s^{\dag i}s_j { \rm tr}s^{\dag j}s_i + \mu_5 { \rm tr}s_i s_j{\rm tr } s^{\dag i}s^{\dag j } + \mu_6 { \rm tr}s_i s_j s^{\dag j}s^{\dag i}\nonumber \\ & + & \nu_1 \phi_1^{\dag i}\phi_{1i}{\rm tr}s^{\dag j}s_j + \nu_2 \phi_1^{\dag i}\phi_{1j } { \rm tr}s^{\dag j}s_i\nonumber\\ & + & \left ( \nu_3 \phi_1^{\dag i}\phi_1^{\dag j}{\rm tr}s_i s_j + \nu_4 \phi_1^{\dag i}{\rm tr } s^{\dag j}s_j s_i + \nu_5 \phi_1^{\dag i}{\rm tr}s^{\dag j}s_i s_j + { \rm h.c . }
\right ) \nonumber \\ & + & \omega_1 \phi_2^{\dag i}\phi_{2i}{\rm tr}s^{\dag j}s_j + \omega_2 \phi_2^{\dag i}\phi_{2j } { \rm tr}s^{\dag j}s_i\nonumber\\ & + & \left ( \omega_3 \phi_2^{\dag i}\phi_2^{\dag j}{\rm tr}s_i s_j + \omega_4 \phi_2^{\dag i}{\rm tr } s^{\dag j}s_j s_i + \omega_5 \phi_2^{\dag i}{\rm tr}s^{\dag j}s_i s_j + { \rm h.c . } \right ) \nonumber\\ & + & \kappa_1 \phi_1^{\dag i}\phi_{2i}{\rm tr}s^{\dag j}s_j + \kappa_2 \phi_1^{\dag i}\phi_{2j}{\rm tr}s^{\dag j}s_i + \kappa_3 \phi_1^{\dag i}\phi_2^{\dag j}{\rm tr}s_j s_i + { \rm h.c . }
\label{potential}\end{aligned}\ ] ] in all the terms we have explicitly shown the @xmath7 indices @xmath12 , @xmath13 , and the trace is taken over color indices . as per our discussion in ref .
@xcite , in terms that are not part of the usual the 2hdm , we have allowed some that satisfy mfv but not the discrete symmetry mentioned above . after symmetry breaking some of these couplings are related to scalar masses , and these relations can be readily found in the literature @xcite .
the yukawa potential for this model in the flavour eigenstate basis is given by @xmath14 as is conventional , we use @xmath15 for all three scalar doublets @xmath16 , and @xmath17 are flavour indices . the large number of parameters present in eq .
[ potential ] and eq .
[ yukawas ] is reduced by the following theoretical considerations : * minimal flavour violation @xcite implies * * 2hdm type i : @xmath18 * * 2hdm type ii : @xmath19 + requiring mfv instead of a discrete symmetry to define the models allows quartic terms in the scalar potential that are odd in either of the doublets , such as @xmath20 , @xmath21 and @xmath22 .
it also allows additional terms in the pure 2hdm sector , but we do not consider those here . *
custodial symmetry @xcite .
as discussed in @xcite the least restrictive method to impose custodial symmetry results in all the @xmath23 s being real and in the relations @xmath24 these conditions imply mass degeneracies @xmath25 and @xmath26 . an alternative possibility , ` twisted ' custodial symmetry @xcite results instead in @xmath27 and @xmath28 @xcite . * perturbative unitarity @xcite + the two - to - two scattering matrix for this model in the neutral , color singlet channel is @xmath29 and in ref .
@xcite we diagonalize it numerically and illustrate the resulting constraints .
the numerical results can be roughly approximated by , + @xmath30 + in the present study we require the perturbative unitarity constraints , as obtained by diagonalizing the full @xmath29 matrix numerically , to be satisfied by the running couplings at all scales below @xmath6 . * stability , or having a positive definite higgs potential for the 2hdm @xcite implies that @xmath31 we will again require that this conditions be satisfied by the running couplings at all scales below @xmath6 .
a recent paper presents a way to extend the stability conditions to additional fields @xcite . here
we will only consider the effect of the color scalars through their one - loop contributions to the running of the parameters of the 2hdm , and require the conditions eq .
[ stabvac ] be satisfied at all scales .
* for our numerical analysis we will use color octet scalar masses near 1 tev as in ref .
@xcite , as masses at this scale are still allowed by lhc searches @xcite .
we now turn to the novel aspect of this paper : investigating the consequences of requiring the model to be valid up to some high energy scale .
the procedure is straightforward , we first derive the corresponding renormalization group equations ( rge ) for all the parameters in the model , a result we present in the appendix . for our numerical analysis
we restrict ourselves to the case with custodial symmetry and no cp violation , which reduces the rge to the following , @xmath32 in these equations @xmath33 , @xmath34 and @xmath35 are the @xmath36 , @xmath37 and @xmath38 couplings of the sm , @xmath39 is the top - quark yukawa coupling and we use the standard definition @xmath40 .
these equations have been checked against the known limits : the 2hdm @xcite ; the mw model @xcite .
when we include the effect of the coupling @xmath34 in the rge we end up with high scale couplings that no longer satisfy custodial symmetry .
the deviations from the symmetry limit are small as expected , proportional to @xmath34 , and we ignore them in our numerical analysis .
as we run the couplings between the electroweak and high scales we find three possible behaviors : well behaved couplings at all scales ; one or more of the couplings develop a lp ; or even though there are no lp for the scales considered , perturbative unitarity or stability are violated at some scale in the range .
we illustrate these three possibilities below .
figure [ f : running1 ] illustrates the case of well behaved couplings up to the planck scale .
figure [ f : running2 ] shows how it is possible to develop multiple lp at relatively low energy scales even when the couplings are small and perturbative at the electroweak scale .
finally , figure [ f : running3 ] illustrates a case where there are no lp below the planck scale , but perturbative unitarity is violated at some point below @xmath41 . in this case
, the violation is due to the presence of a lp just beyond @xmath41 . .
, title="fig : " ] .
, title="fig : " ] .
, title="fig : " ] .
, title="fig : " ] .
, title="fig : " ] .
, title="fig : " ] .
, title="fig : " ] .
, title="fig : " ] , but the unitarity and stability conditions are not satisfied for all scales @xmath42 .
, title="fig : " ] , but the unitarity and stability conditions are not satisfied for all scales @xmath42 . , title="fig : " ] , but the unitarity and stability conditions are not satisfied for all scales @xmath42 .
, title="fig : " ] , but the unitarity and stability conditions are not satisfied for all scales @xmath42 .
, title="fig : " ] for this purpose we use mathematica to solve the rge numerically with initial conditions at the electroweak scale .
the initial conditions used for the new couplings are points that satisfy perturbative unitarity and stability determined as in ref .
@xcite . for each point
we evolve all the couplings up to a high scale @xmath6 and discard the point if a lp is detected , or if the couplings at any scale below @xmath6 violate the perturbative unitarity or stability conditions .
this results in acceptable points satisfying a more stringent condition than the absence of lp , in the spirit of renormalization group improved unitarity bounds of ref .
the use of this condition in our numerical search makes it easier to find acceptable points than if we were to allow a lp at @xmath6 .
it would also be possible to constrain the parameters using higher order unitarity conditions , but we do not pursue this in this paper @xcite .
the parameter space is too large for a completely random scan to be efficient .
instead , we follow the approach described below . 1 .
we begin our study at an intermediate energy scale @xmath43 which , for the sake of computational efficiency .
we choose ( by trial and error ) it to be @xmath44 .
before running the rge for the whole model , we generate a large sample of points within the 2hdm subspace .
the sample is generated in such a way that a large portion of it is valid up to @xmath43 .
we then use this 2hdm data set as seeds , randomly assigning values to the new couplings within a proper range , to generate a sample for the whole parameter space . from this sample
we find a few hundred valid points and determine the hypercube which contains most of the solutions , a region somewhat smaller than that allowed by perturbative unitarity at the electroweak scale .
starting from these few hundred points we study nearby points to expand the allowed region .
we finally construct the region of parameter space where the full model is valid up to the scale @xmath43 by repeating step 4 recursively for a sufficiently long time .
points that are valid up to scales higher than @xmath43 , are generally inside a sub - region of the allowed region up to scale @xmath43 . therefore , to find the constraints for a higher scale , we select the seed points from step 5 and repeat step 4 to construct the new allowed region .
our results are illustrated in figure [ f : uni - gut ] for the scales @xmath43 and @xmath45 in representative two - dimensional projections .
the gut scale is chosen because of the existing @xmath4 @xcite or @xmath5 @xcite models which can have tev scale scalar color octets , but the figures illustrate the general trend as we require the model to be valid up to higher energy scales .
in general , for the new parameters involving the color - octet scalars , the allowed parameter space is now very significantly reduced with respect to that allowed by tree - level unitarity .
in addition , this procedure produces the first constraints on parameters like @xmath46 and @xmath47 which do not affect two to two processes at tree - level .
( red ) and @xmath45 ( blue ) .
the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) .
the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) .
the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) .
the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) . the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) .
the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) .
the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) .
the black points are the seed points used as initial values at the electroweak scale that satisfy both perturbative unitarity and stability .
[ f : uni - gut],title="fig : " ] validity up to the gut scale thus results in approximate one at a time constraints @xmath48 is closer to zero , as in the alignment limit ; and those for which @xmath49 is very close to @xmath50 as shown in figure [ f:2hdm - gut ] .
( red ) and @xmath45 ( blue ) .
[ f:2hdm - gut],title="fig : " ] ( red ) and @xmath45 ( blue ) .
[ f:2hdm - gut],title="fig : " ]
in this paper , we provide the renormalization group equations for all the couplings in the scalar potential of a 2hdm augmented with a color - octet . as an application
, we constrain the parameters of the model by requiring it to be valid up to some high scale .
the acceptable region of the parameter space that satisfies both unitarity and stability constraints without developing lp up to a high scale is determined numerically , and the resulting constraints are provided for the case @xmath51 . as expected , the allowed region is reduced as the scale increases ; looking at the 2hdm subspace , it contracts towards the alignment limit and mass - degeneracy of heavy neutral and charged higgses .
we thank dianne cook for help visualizing the multidimensional parameter space with the aid of the grand and guided tour @xcite with the software ggobi @xcite .
li cheng thanks david atwood and kerry whisnant for useful discussions .
the work of gv was supported in part by the australian government through the australian research council .
in this appendix we collect the general results for the rge , without the assumptions of custodial and cp symmetry used for our numerical study .
a. v. manohar and m. b. wise , phys .
d * 74 * , 035009 ( 2006 ) [ hep - ph/0606172 ] .
p. fileviez perez and c. murgui , arxiv:1604.03377 [ hep - ph ] .
i. dorsner and i. mocioiu , nucl .
b * 796 * , 123 ( 2008 ) doi:10.1016/j.nuclphysb.2007.12.004 [ arxiv:0708.3332 [ hep - ph ] ] .
s. bertolini , l. di luzio and m. malinsky , phys .
d * 87 * , no .
8 , 085020 ( 2013 ) doi:10.1103/physrevd.87.085020 [ arxiv:1302.3401 [ hep - ph ] ] . m. i. gresham and m. b. wise , phys .
d * 76 * , 075003 ( 2007 ) [ arxiv:0706.0909 [ hep - ph ] ] .
m. gerbush , t. j. khoo , d. j. phalen , a. pierce and d. tucker - smith , phys .
d * 77 * , 095003 ( 2008 ) [ arxiv:0710.3133 [ hep - ph ] ] .
x. -g .
he and g. valencia , phys .
b * 707 * , 381 ( 2012 ) [ arxiv:1108.0222 [ hep - ph ] ] .
b. a. dobrescu , g. d. kribs and a. martin , phys . rev .
d * 85 * , 074031 ( 2012 ) [ arxiv:1112.2208 [ hep - ph ] ] .
y. bai , j. fan and j. l. hewett , jhep * 1208 * , 014 ( 2012 ) [ arxiv:1112.1964 [ hep - ph ] ] .
j. m. arnold and b. fornal , phys .
d * 85 * , 055020 ( 2012 ) [ arxiv:1112.0003 [ hep - ph ] ] . g. d. kribs and a. martin , arxiv:1207.4496 [ hep - ph ] .
m. reece , arxiv:1208.1765 [ hep - ph ] .
j. cao , p. wan , j. m. yang and j. zhu , arxiv:1303.2426 [ hep - ph ] .
x. g. he , h. phoon , y. tang and g. valencia , jhep * 1305 * , 026 ( 2013 ) doi:10.1007/jhep05(2013)026 [ arxiv:1303.4848 [ hep - ph ] ] .
n. chakrabarty , u. k. dey and b. mukhopadhyaya , jhep * 1412 * , 166 ( 2014 ) doi:10.1007/jhep12(2014)166 [ arxiv:1407.2145 [ hep - ph ] ] .
p. ferreira , h. e. haber and e. santos , phys .
d * 92 * , 033003 ( 2015 ) erratum : [ phys .
d * 94 * , no . 5 , 059903 ( 2016 ) ] doi:10.1103/physrevd.92.033003 , 10.1103/physrevd.94.059903 [ arxiv:1505.04001 [ hep - ph ] ] .
d. j. e. callaway , phys .
rept . * 167 * , 241 ( 1988 ) .
doi:10.1016/0370 - 1573(88)90008 - 7 m. sher , phys .
rept . * 179 * , 273 ( 1989 ) .
doi:10.1016/0370 - 1573(89)90061 - 6 j. f. gunion , h. e. haber , g. l. kane and s. dawson , front . phys . * 80 * , 1 ( 2000 ) .
g. c. branco , p. m. ferreira , l. lavoura , m. n. rebelo , m. sher and j. p. silva , phys .
* 516 * , 1 ( 2012 ) [ arxiv:1106.0034 [ hep - ph ] ] .
r. s. chivukula and h. georgi , phys .
b * 188 * ( 1987 ) 99 .
g. dambrosio , g. f. giudice , g. isidori and a. strumia , nucl .
b * 645 * ( 2002 ) 155 [ hep - ph/0207036 ] .
p. sikivie , l. susskind , m. b. voloshin and v. i. zakharov , nucl .
b * 173 * , 189 ( 1980 ) .
doi:10.1016/0550 - 3213(80)90214-x a. pomarol and r. vega , nucl . phys .
b * 413 * , 3 ( 1994 ) doi:10.1016/0550 - 3213(94)90611 - 4 [ hep - ph/9305272 ] .
b. grzadkowski , m. maniatis and j. wudka , jhep * 1111 * , 030 ( 2011 ) doi:10.1007/jhep11(2011)030 [ arxiv:1011.5228 [ hep - ph ] ] .
gerard and m. herquet , phys .
rev . lett .
* 98 * , 251802 ( 2007 ) doi:10.1103/physrevlett.98.251802 [ hep - ph/0703051 [ hep - ph ] ] .
e. cervero and j. m. gerard , phys .
b * 712 * , 255 ( 2012 ) doi:10.1016/j.physletb.2012.05.010 [ arxiv:1202.1973 [ hep - ph ] ] .
b. w. lee , c. quigg and h. b. thacker , phys .
d * 16 * , 1519 ( 1977 ) .
s. kanemura , t. kubota and e. takasugi , phys .
b * 313 * , 155 ( 1993 ) [ hep - ph/9303263 ] .
j. horejsi and m. kladiva eur .
j. c * 46 * , 81 ( 2006 ) doi:10.1140/epjc / s2006 - 02472 - 3 [ hep - ph/0510154 ] . i. f. ginzburg and i. p. ivanov , phys .
d * 72 * , 115010 ( 2005 ) doi:10.1103/physrevd.72.115010 [ hep - ph/0508020 ] .
b. grinstein , c. w. murphy and p. uttayarat , arxiv:1512.04567 [ hep - ph ] .
n. g. deshpande and e. ma , phys .
d * 18 * , 2574 ( 1978 ) .
doi:10.1103/physrevd.18.2574 k. kannike , eur .
j. c * 76 * , no .
6 , 324 ( 2016 ) doi:10.1140/epjc / s10052 - 016 - 4160 - 3 [ arxiv:1603.02680 [ hep - ph ] ] .
a. hayreter and g. valencia , arxiv:1703.04164 [ hep - ph ] .
w. j. marciano , g. valencia and s. willenbrock , phys .
d * 40 * , 1725 ( 1989 ) .
v. cacchio , d. chowdhury , o. eberhardt and c. w. murphy , jhep * 1611 * , 026 ( 2016 ) doi:10.1007/jhep11(2016)026 [ arxiv:1609.01290 [ hep - ph ] ] . | we have recently studied theoretical constraints on the parameters of a 2hdm augmented with a color - octet scalar . in this paper
we consider the consequences of requiring the model to remain valid up to very high energy scales , such as the gut scale .
the acceptable region of parameter space is reduced when one insists on vacuum stability , perturbative unitarity and the absence of landau poles below a given scale .
as the scale to which we require the model to be valid is increased , the acceptable region of parameter space for the 2hdm sector is reduced in such a way that it approaches the alignment limit , @xmath0 , and the masses of @xmath1 , @xmath2 and @xmath3 are pushed closer to each other .
the parameters of the color octet sector are also restricted to an increasingly smaller region . |
alice @xcite is the dedicated heavy - ion experiment at the large hadron collider ( lhc ) .
the main physics goal of the experiment is the study of strongly - interacting matter in the conditions of high energy density ( @xmath0 ) and high temperature ( @xmath1 ) , expected to be reached in central collisions at @xmath2 . under these conditions , according to lattice qcd calculations , quark confinement into colourless hadrons should be removed and a deconfined quark
gluon plasma should be formed @xcite .
as we will detail in the next section , heavy - flavour particles are regarded as effective probes of the system conditions . in particular : open charm and beauty hadrons would be sensitive to the energy density , through the mechanism of in - medium energy loss of heavy quarks ; quarkonium states would be sensitive to the initial temperature of the system through their dissociation due to colour screening .
.expected @xmath3 yields at the lhc , from nlo pqcd @xcite . for , nuclear shadowing of the parton distribution functions
is included and binary scaling is applied . [ cols="^,^,^,^,^",options="header " , ] the expected yields for heavy - quark production in pp collisions at @xmath4 are reported in the first line of table [ tab : xsec ] @xcite , as obtained from the next - to - leading order perturbative qcd calculation implemented in the mnr program @xcite with _ reasonable _ parameter values ( quark masses and perturbative scales ) @xcite .
these yields have large uncertainties , of about a factor 2 , estimated by varying the parameters @xcite . in the table
we report also the @xmath5 and @xmath6 yields in central collisions at @xmath2 , obtained including in the calculation the nuclear modification of the parton distribution functions ( eks98 parametrization @xcite ) and applying a scaling of the hard yields with the mean number @xmath7 of binary nucleon
nucleon collisions in the collision .
experiments at the relativistic heavy ion collider ( rhic ) have shown that the nuclear modification factor of particles @xmath8 distributions , , is a sensitive observable for the study of the interaction of the hard partons with the medium produced in nucleus nucleus collisions .
heavy quark medium - induced quenching is one of the most captivating topics to be addressed in collisions at the lhc . due to the qcd nature of parton energy loss , quarks
are predicted to lose less energy than gluons ( that have a higher colour charge ) and , in addition , the ` dead - cone effect ' is expected to reduce the energy loss of massive quarks with respect to light partons @xcite .
therefore , one should observe a pattern of gradually decreasing @xmath9 suppression when going from the mostly gluon - originated light - flavour hadrons ( @xmath10 or @xmath11 ) to d and to b mesons @xcite : @xmath12 .
the enhancement , with respect to unity , of _ heavy - to - light @xmath9 ratios _ has been suggested @xcite as a sensitive observable to test the colour - charge ( @xmath13 ) and mass ( @xmath14 ) dependence of parton energy loss . in addition , the smaller predicted energy loss of b quarks may allow to overcome the main limitation of light partons as probes of the medium , namely their low sensitivity to the density of the medium . according to the theoretical models @xcite based on the bdmps framework @xcite ,
this so - called ` fragility ' became apparent already in collisions at rhic , and would be more pronounced in collisions at the lhc , where hard gluons would either be absorbed in the medium or escape the medium without really probing its density .
beauty quarks could instead provide stringent constraints on the energy density of the medium .
the measurement of d and b meson production cross sections will also serve as a baseline for the study of medium effects on quarkonia .
two of the most interesting items in the quarkonia sector at the lhc will be : ( a ) understanding the interplay between colour - screening - induced suppression and statistical regeneration for j/@xmath15 production in a medium containing on the order of 100 @xmath5 pairs ( see table [ tab : xsec ] ) ; ( b ) measuring for the first time medium effects on the bottomonia resonances , expected to be available with sufficient yields at the lhc . on this point ,
the predicted suppression pattern as a function of the plasma temperature is particularly interesting : the @xmath16 would melt at @xmath17 , a temperature that would be reached only at the lhc , while the @xmath18 would melt at the same temperature as the j/@xmath15 , @xmath19 .
it will thus be important for the experiments to be able to measure also the @xmath18 , because , at variance with the j/@xmath15 , it is expected to have a small regeneration probability and it would be very useful to disentangle j/@xmath15 suppression and regeneration . in summary , measuring the in - medium dissociation probability of the different quarkonium states at the lhc will provide an estimate of the initial medium temperature .
we note that , in order to study the medium effects on charmonia , it will be mandatory to measure the fraction of secondary charmonia from b decays , expected to be about 20% for the j/@xmath15 , in absence of medium - induced effects .
the desing of the alice apparatus @xcite will allow the detection of open heavy - flavour hadrons and quarkonia in the high - multiplicity environment of central collisions at lhc energy , where up to few thousand charged particles might be produced per unit of rapidity .
the heavy - flavour capability of the alice detector is provided by : tracking system ; the inner tracking system ( its ) , the time projection chamber ( tpc ) and the transition radiation detector ( trd ) , embedded in a magnetic field of @xmath20 t , allow track reconstruction in the pseudorapidity range @xmath21 with a momentum resolution better than 2% for @xmath22 and a transverse impact parameter , is defined as the distance of closest approach of the track to the interaction vertex , in the plane transverse to the beam direction . ]
resolution better than @xmath23 for @xmath24 ( the two innermost layers of the its are equipped with silicon pixel detectors ) .
particle identification ; charged hadrons ( @xmath25 , k , p ) are separated via @xmath26 in the tpc and in the its and via time - of - flight measurement in the time of flight ( tof ) detector ; electrons are separated from charged pions in the dedicated transition radiation detector ( trd ) , and in the tpc ; muons are identified in the forward muon spectrometer covering in acceptance the range @xmath27 .
simulation studies @xcite have shown that alice has good potential to carry out a rich heavy - flavour physics programme .
the main analyses in preparation are : open charm ( section [ open ] ) : fully reconstructed hadronic decays @xmath28 , @xmath29 , @xmath30 ( under study ) , @xmath31 ( under study ) in @xmath21 .
open beauty ( section [ open ] and ref .
@xcite ) : inclusive single leptons @xmath32 in @xmath21 and @xmath33 in @xmath27 ; inclusive displaced charmonia @xmath34 ( under study ) .
quarkonia ( section [ quarkonia ] ) : @xmath15 and @xmath16 states in the @xmath35 ( @xmath21 ) and @xmath36 ( @xmath27 ) channels . for all studies , a multiplicity of @xmath37@xmath38 was assumed for central collisions . in the following ,
we report the results corresponding to the expected statistics collected by alice per lhc year : @xmath39 central ( 05% @xmath40 ) events at @xmath41 and @xmath42 pp events at @xmath43 , in the barrel detectors ; the forward muon arm will collect about 40 times larger samples ( i.e. @xmath44 central events ) .
_ exclusive charm meson reconstruction_. among the most promising channels for open charm detection are the @xmath28 ( @xmath45 , branching ratio @xmath46 ) and @xmath29 ( @xmath47 , branching ratio @xmath48 ) decays . the detection strategy to cope with the large combinatorial background from the underlying event
is based on the selection of displaced - vertex topologies , i.e. separation from the primary vertex of the tracks from the secondary vertex and good alignment between the reconstructed d meson momentum and flight - line @xcite .
an invariant - mass analysis is used to extract the raw signal yield , to be then corrected for selection and reconstruction efficiency and for detector acceptance . as shown in fig .
[ fig : destat ] ( left ) , the accessible @xmath8 range for the @xmath49 is @xmath50@xmath51 in and @xmath20@xmath51 in pp , with statistical errors better than 1520% at high @xmath8 .
similar capability is expected for the @xmath52 ( right - hand panel ) , though at present the statistical errors are estimated only in the range @xmath53 .
the systematic errors ( acceptance and efficiency corrections , centrality selection for pb
pb ) are expected to be smaller than 20% . in the @xmath54 channel ( left ) and
@xmath55 in the @xmath56 channel ( right ) , in 05% central collisions and in pp collisions.,title="fig : " ] in the @xmath54 channel ( left ) and @xmath55 in the @xmath56 channel ( right ) , in 05% central collisions and in pp collisions.,title="fig : " ] _ beauty via single electrons . _
the main sources of background electrons are : decays of d mesons ; @xmath11 dalitz decays and decays of light vector mesons ( e.g.@xmath57 and @xmath58 ) ; conversions of photons in the beam pipe or in the inner detector layer and pions misidentified as electrons .
given that electrons from beauty have average impact parameter @xmath59 and a hard @xmath8 spectrum , it is possible to obtain a high - purity sample with a strategy that relies on : electron identification with a combined @xmath26 ( tpc ) and transition radiation ( trd ) selection ; impact parameter cut to reduce the charm - decay component and reject misidentified @xmath60 and @xmath61 from dalitz decays and @xmath62 conversions .
as an example , with @xmath63 and @xmath64 , the expected statistics of electrons from b decays is @xmath65 for @xmath39 central events , allowing the measurement of electron - level @xmath8-differential cross section in the range @xmath66 with statistical errors smaller than 15% at high @xmath8 .
similar performance figures are expected for pp collisions @xcite .
_ beauty via muons .
_ b production in collisions can be measured also in the alice muon spectrometer ( @xmath27 ) analyzing the single - muon @xmath8 distribution @xcite .
the main backgrounds to the ` beauty muon ' signal are @xmath60 , @xmath67 and charm decays .
the cut @xmath68 is applied to all reconstructed muons in order to increase the signal - to - background ratio .
then , a fit technique allows to extract a @xmath8 distribution of muons from b decays .
since only minimal cuts are applied , the statistical errors are expected to be smaller than 5% up to muon @xmath69
@xcite .
_ nuclear modification factors .
_ we investigated the possibility of using the described charm and beauty measurements to study the high-@xmath8 suppression induced by parton energy loss .
the sensitivity to @xmath70 and @xmath71 is presented in fig .
[ fig : raa ] . predictions @xcite with and without the effect of the heavy - quark mass , for a medium transport coefficient @xmath72 ( a measurement of the medium density ) in the range @xmath73@xmath74 , are also shown .
mesons ( left ) and for b - decay electrons ( right ) .
errors corresponding to the centre of the prediction bands for massive quarks are shown : bars = statistical , shaded area = systematic.,title="fig:",scaledwidth=52.0% ] mesons ( left ) and for b - decay electrons ( right ) .
errors corresponding to the centre of the prediction bands for massive quarks are shown : bars = statistical , shaded area = systematic.,title="fig:",scaledwidth=47.0% ]
alice can detect quarkonia in the dielectron channel at central rapidity ( @xmath75 ) and in the dimuon channel at forward rapidity ( @xmath76 ) . in both channels
the acceptance extends down to zero transverse momentum , since the minimum @xmath8 for e and @xmath77 identification is about @xmath78 .
the high @xmath8 reach is expected to be 10 ( 20 ) @xmath79 for the j/@xmath15 in @xmath80 ( @xmath36 ) , for a run of one month at nominal luminosity .
we emphasized the importance of separating the @xmath16 and @xmath18 , to probe the initial temperature of the medium ; given that the mass difference between the two bottomonium states is about @xmath81 , a mass resolution of order @xmath82 at @xmath83 , i.e. @xmath84 , is required .
this requirement is fulfilled for both dielectrons and dimuons , with a mass resolution of about @xmath85 . for illustration , in fig .
[ fig : upsilon ] we show the simulated dilepton mass spectra in the @xmath16 region after background subtraction @xcite .
simulation studies are in progress to prepare a measurement of the fraction of j/@xmath15 that feed - down from b decays .
such measurement can be performed by studying the separation of the dilepton pairs in the j/@xmath15 invariant mass region from the main interaction vertex .
the analysis is also expected to provide a measurement of the beauty @xmath8-differential cross section .
states in central collisions , as reconstructed by alice @xcite , in the dielectron and in the dimuon channel , in one month of data - taking.,scaledwidth=75.0% ]
we have discussed how heavy quarks , abundantly produced at lhc energies , will allow to address several issues at the heart of in heavy - ion physics .
they provide tools to probe the density ( via parton energy loss and its predicted mass dependence ) and the temperature ( via successive dissociation patterns of quarkonia ) of the high - density qcd medium formed in collisions .
the excellent tracking , vertexing and particle identification performance of alice will allow to fully explore this rich phenomenology . | the large hadron collider ( lhc ) , where lead nuclei will collide at the unprecedented c.m.s . energy of 5.5 tev per nucleon nucleon pair , will offer new and unique opportunities for the study of the properties of strongly interacting matter at high energy density over extended volumes .
we will briefly explain why heavy - flavour particles are well - suited tools for such a study and we will describe how the alice experiment is preparing to make use of these tools . |
[ [ mechanism - for - the - spectral - break ] ] mechanism for the spectral break + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the physics of the spectral break is very simple .
when a snr shock approaches a molecular cloud ( mc ) or a pre - supernova swept - up shell , confinement of accelerated particles generally deteriorates . due to the particle interaction with magnetic turbulence , confinement requires scales similar to the particle gyroradius @xcite .
while the waves are in a strongly ionized ( closer to the shock ) medium they propagate freely in a broad frequency range at the alfven speed @xmath5 with the frequencies @xmath6 . here
@xmath7 is the wave number ( assumed parallel to the local field @xmath8 ) and @xmath9 is the ion mass density .
as long as the alfven wave frequency is higher than the ion - neutral collision frequency @xmath10 , the waves are weakly damped .
when , on the other hand , the ion - neutral collision frequency is higher ( deeper into the cloud ) , neutrals are entrained by the oscillating plasma and the alfven waves are also able to propagate , albeit with a factor @xmath11 lower speed , where @xmath12 is the neutral density .
the propagation speed reduction occurs because every ion is now `` loaded '' with @xmath13 neutrals . in between these two regimes alfven waves are heavily damped and even disappear altogether for sufficiently small @xmath14 .
the evanescence wave number range is then bounded by @xmath15 and @xmath16 .
these phenomena have been studied in detail in @xcite , and specifically in the context of the dsa in @xcite .
now we turn to their impact on the particle confinement and emissivity . in the frame work of a quasilinear wave - particle interaction
the wave number @xmath7 is approximately related to the parallel ( to the magnetic field ) component of the particle momentum @xmath17 by the cyclotron resonance condition @xmath18 where the ( non - relativistic ) gyro - frequency @xmath19 .
therefore , the frequency range where the waves can not propagate may be conveniently translated into the parallel momentum range @xmath20 with @xmath21 that a spectral break must form at the photon energy corresponding to the particle momentum @xmath22 , can be readily understood from fig.[fig : scattering - zone - of ] .
the dead zones @xmath23 imply that particles with @xmath24 do not turn around ( while moving along the magnetic field ) and escape from the region of cr - dense gas collisions at a @xmath25 fraction of the speed of light .
this happens in the region of enhanced gamma radiation .
therefore , an initially isotropic distribution of accelerated particles fills only a slab in momentum space @xmath26 and becomes highly anisotropic ( a pancake distribution ) .
what matters for the integral emission , however , is a locally isotropic component @xmath27 of this new proton distribution .
it can be introduced by re - averaging the pancake ( @xmath26 ) distribution in pitch angle , @xmath28 , and is readily obtained assuming that particles remaining in the dense gas ( those with @xmath26 ) maintain their isotropic pitch - angle distribution , i.e. @xmath29 where @xmath30 is the initial ( isotropic ) distribution function and @xmath31 is the cosine of the pitch angle ( see fig.[fig : scattering - zone - of ] ) , @xmath32 . thus , the slope of the particle momentum distribution becomes steeper by exactly one power above @xmath33 .
in particular , any power - law distribution @xmath34 , upon entering an mc , turns into @xmath35 at @xmath36 , and preserves its form at @xmath37 .
now the question is whether particle escape from the mc can quench the acceleration process itself . in principle
it can , since also the particles from the interval @xmath26 can not return to the shock and continue acceleration . instead , by slowly leaking to the part of the momentum space with @xmath24 , they should ultimately escape as well .
this would certainly be the case if the mc were filling the entire shock precursor .
however , mcs are known to be clumpy @xcite , and may fill only a small fraction of the precursor . in this case
the acceleration process continues almost unimpeded , the accelerated protons illuminate the cloudlets and make them visible in @xmath38-rays due to the high density target material . [
[ break - momentum ] ] break momentum + + + + + + + + + + + + + + while the one power spectral break in the pitch - angle averaged particle distribution seems to be a robust environmental signature of a weakly ionized medium into which the accelerated particles propagate , the break momentum remains uncertain . according to eq.([eq : p12 ] ) , @xmath39 ( @xmath40 ) depends on the magnetic field strength and ion density as well as on the frequency of ion - neutral collisions , @xmath41 . here
@xmath42 is the product of the collision cross - section and collision velocity averaged over the thermal distribution . using an approximation of @xcite for @xmath42
, @xmath39 can be estimated as @xmath43 here the gas temperature @xmath44 is measured in the units of @xmath45 , magnetic field
@xmath46 -in microgauss , @xmath47 and @xmath48 ( number densities corresponding to the neutral / ion mass densities @xmath12 and @xmath9 ) -in @xmath49 . note that the numerical coefficient in the last expression may vary depending on the average ion and neutral masses and can be higher by a factor of a few for typical molecular cloud conditions @xcite than the estimate in eq.([eq : p1 ] ) suggests .
the remaining quantities in the last formula are also known too poorly to make an accurate independent prediction of the position of the break in the gamma ray emission region .
those are the regions near the blast wave where complicated physical processes unfold .
they include particle acceleration , strong mhd turbulence ( driven by particles and their interaction with ambient gas inhomogeneities ) , gas ionization by shock generated uv photons , turbulent plasma heating and even evaporation of magnetic cloudlets @xcite .
also important may be the ionization by the low energy crs accelerated at the blast wave .
however , as their diffusion length is shorter than that of the particles with @xmath50 , we may assume that they do not reach the mc .
pre - ionization by the uv photons can also be ignored for the column density @xmath51 ahead of the shock beyond which they are absorbed @xcite .
the authors or the ref .
@xcite , using the earlier data from @xcite have also analyzed the parameters involved in eq.([eq : p1 ] ) and found the above estimate of @xmath39 to be in a good agreement with the spectral break position measured by the _
nevertheless , we may run the argument in reverse and use the _ fermi _ observations @xcite of the gamma - ray spectrum of snr w44 to determine the break momentum in the parent particle spectrum and constrain the parameters in eq.([eq : p1 ] ) .
once we also know the amount of the slope variation @xmath52 , we can calculate the full spectrum up to the cut - off energy .
[ [ particle - spectra ] ] particle spectra + + + + + + + + + + + + + + + + to calculate the particle spectra , we need to determine the degree of nonlinear modification of the shock structure . in principle , it can be calculated consistently , given the shock parameters and the particle maximum momentum , @xmath53 .
in the case of a broken spectrum , @xmath39 plays the role of @xmath53 , since this is the momentum where the dominant contribution to the pressure of accelerated particles comes from , thus setting the scale of the modified shock precursor .
note that in the conventional nonlinear ( nl ) acceleration theory , the cut - off momentum @xmath53 plays this role , because the nonlinear spectra are sufficiently flat so as to make the pressure diverge with momentum .
the break in the photon spectrum is observed at about @xmath54 gev , which places the break on the proton distribution at about @xmath55 @xcite . for the strength of the break @xmath56 ,
the spectrum above it is clearly pressure converging , and we perform the calculation of the shock structure and the spectrum using this break momentum as the point of the maximum in the cr partial pressure .
once this point is set , we can use an analytic approach for a stationary nonlinear acceleration problem using @xmath39 as an input parameter .
apart from @xmath39 , the nonlinear solution depends on a number of other parameters , such as the injection rate of thermal particles into acceleration , mach number , the precursor heating rate and the shock velocity @xmath57 . of these parameters only the latter
is known accurately , @xmath58 , the other parameters are still difficult to ascertain .
fortunately , in sufficiently strong shocks the solution generally tends to either stay close to the test particle ( tp ) solution ( leaving the shock structure only weakly modified ) or else it transitions to a strongly modified nl - solution regime .
the tp regime typically establishes in cases of moderate mach numbers , low injection rates and low @xmath53 ( now @xmath39 ) , while the nl regime is unavoidable in the opposite part of the parameter space . in the tp regime
the spectrum is close to a power - law with the spectral index 2 throughout the supra - thermal energy range . in the nl regime , however , the spectrum develops a concave form , starting from a softer spectrum at the injection energy , with the index @xmath59 , where @xmath60 is the sub - shock compression ratio
. then it hardens , primarily in the region @xmath61 , where both the partial pressure and diffusivity of protons change their momentum scaling .
the slope reaches its minimum at the cut - off ( break ) energy , which , depending on the degree of nonlinearity , can be as low as 1.5 or even somewhat lower if the cut - off is abrupt .
the question now is into which of these two categories the w44 spectrum falls ?
generally , in cases of low maximum ( or , equivalently , low spectral break @xmath62 ) momentum , the shock modification is weak , so the spectrum is more likely to be in the only slightly nonlinear , almost tp regime . on the other hand
, there is a putative indication from the electron radio emission that their spectrum may be close to @xmath63 , which could be the signature of a moderately nonlinear acceleration process .
it should be remembered , however , that this is a global index across the w44 remnant .
there are resolved bright filaments where a canonical @xmath64 spectrum , corresponding precisely to the tp parent electron spectrum with @xmath65 is observed @xcite . moreover , there are regions with the positive indices @xmath66 which can not be indicative of a dsa process without corrections for subsequent spectral transformations such as an absorption by thermal electrons @xcite .
these regions can very well contribute to the overall spectral hardening discussed above , mimicking the acceleration nonlinearity .
finally , secondary electrons give rise to the flattening of the radio spectrum as well @xcite .
if the accelerated protons and electrons respond to the turbulence similarly , which is almost certainly the case in the ultra - relativistic regime , their spectra should have similar slopes there ( as long as the synchrotron losses are ignorable ) . in using the electron radio spectrum as a probe for the level of acceleration nonlinearity ,
the following two relations are useful .
first , there is a relation between the electron energy and the radio frequency @xmath67 .
the second relation , @xmath68 , links the spectral index of radio emission @xmath69 ( assuming the radio flux @xmath70 ) and the spectral index of the parent electrons @xmath71 ( assuming their energy spectrum @xmath72 ) .
once the global radio spectral index of w44 , @xmath73 @xcite is generated by freshly accelerated electrons in the frequency range @xmath74 mhz , the electrons should maintain their modified spectrum over the energy range spanning more than one order of magnitude .
for example , assuming @xmath75 @xcite , one sees that electrons must maintain an index @xmath63 between @xmath76 gev .
while the upper bound is acceptable given the spectral break proton energy inferred from the super gev emission measured by the fermi lat , the lower end is rather uncomfortable , since the nonlinear modification of both protons and electrons with the bohm ( or other similar for protons and electrons turbulent diffusivities ) starts ( slowly ) only at the proton rest energy .
the calculated nonlinear spectra are shown in fig.[fig : spectra - of - protons ] for the both species . at and below @xmath77gev ,
the electron spectrum is very close to the test particle solution , @xmath78 even though the proton spectrum may be somewhat steeper there , as we mentioned above .
[ [ photon - spectra ] ] photon spectra + + + + + + + + + + + + + + the above considerations somewhat weaken the radio data as a probe for the slope of the electron and ( more importantly ) for the proton spectrum .
therefore , the exact degree of nonlinearity of the acceleration remains unknown and we consider both the tp and weakly nl regimes in our calculations of the photon spectra , generated in @xmath79 collisions .
the results are shown in fig.[fig : gamma - emission - from ] .
the best fit to the fermi data is provided by a tp energy distribution ( @xmath0 ) below @xmath80gev / c with the spectrum steepening by exactly one power above it .
note that the small deviation of the computed spectrum from the lowest energy _
_ fermi-__point may also be corrected by including the bremsstrahlung of the secondary electrons @xcite .
the spectrum steepening is perfectly consistent with the proton partial escape described above ( with no parameters involved ) and shown in fig.[fig : scattering - zone - of ] . for comparison , a weakly nl spectrum , shown in fig.[fig : spectra - of - protons ] ,
is also used for these calculations ( dashed line in fig.[fig : gamma - emission - from ] ) , but its fit would require a stronger break ( @xmath81 ) or a low momentum cut - off , fig.[fig : gamma - emission - from ] , i.e. at least one additional free parameter .
to summarize the results , the previously suggested mechanism for a break on the spectrum of shock accelerated protons @xcite is found consistent with the recent @xcite fermi lat observations of the snr w44 .
the observed gamma ray spectrum most likely results from the decay of @xmath82-mesons which are born in @xmath79 collisions of shock accelerated protons with an ambient dense gas .
the parent proton spectrum is best represented by a classical test particle power law @xmath0 , steepening to @xmath1 at @xmath2 due to deteriorated particle confinement caused by the ion - neutral collisions and the resultant alfven wave evanescence .
the position of the break momentum in the particle spectrum may be estimated using eq.([eq : p1 ] ) , or conversely , the combination of parameters involved in this estimate can be inferred from the measured break momentum .
the cut - off momentum is not constrained in this scenario .
an alternative explanation , based on a different mechanism of the break , associated with the change of the particle transport in the cr shock precursor @xcite is also possible but is less definitive in the spectrum slope variation @xmath52 across the break ( see also @xcite for the most recent alternative suggestions ) .
in addition , the ref @xcite mechanism would imply a considerable nonlinearity , i.e. a stronger cr shock precursor than it seems appropriate for the observed low break momentum . still alternatively , assuming our `` environmental '' break mechanism is at work , i.e. @xmath56 , but the shock structure appreciably modified , we arrive at the @xmath83 spectrum below the break ( as the radio observations suggest for the electrons ) , and @xmath84 above the break
. a fit to the data is marginally possible , but it would require a relatively low cut - off momentum at about @xmath85 gev / c
. this possibility may be supported or ruled out once the data ( upper limit ) around this energy become available .
the most robust and attractive aspect of the suggested mechanism for the spectral break is the exact @xmath56 variation of the spectral index .
indeed , this change in the spectral slope is due to truncation of the particle momentum phase space and does not depend on any parameters . in a combination with the test particle regime operating below the break , which is consistent with the low value of the break momentum
, it provides a very good fit to the _ fermi _ lat data .
of three physically different types of spectral breaks suggested earlier @xcite , namely the current , `` environmental '' mechanism appears to be plausible in situations where dense target gas is present , as required for efficient @xmath82 production . in a more general context ,
a spectral break is a natural resolution of the well known but puzzling trend of the _ nonlinear _ ( i.e. supposedly improved ) dsa theory to develop spectra which are even harder than the simple test particle spectra , thus encountering more difficulties while accommodating the bulk of observations @xcite .
such a spectrum
i.e. , diverging in energy exhausts the shock energy available for the acceleration as the cut - off momentum grows .
note , that the spectrum of the rx j1713.7 - 3946 @xcite is also consistent with the same break mechanism but with @xmath86 and with naturally stronger acceleration nonlinearity @xcite .
however , it is difficult to make the case for hadronic origin of the gamma - ray emission of the rx j1713.7 - 3946 @xcite .
the fundamental role of the w44 for the problem of cr origin is that this particular remnant seems to rule out contaminating electron emission due to bremsstrahlung and the inverse compton scattering @xcite thus favoring the hadronic origin of the gamma emission .
[ [ acceleration - model - and - proton - spectrum ] ] acceleration model and proton spectrum + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + methods of calculation of the accelerated particle spectra , including the particle back - reaction on the shock structure , are now well developed .
we use the diffusion - convection equation for describing the distribution of high energy particles ( crs ) . to include the back - reaction ,
three further relations are used in a quasi - stationary acceleration regime .
these are the conservation of mass and momentum fluxes in the smooth part of the shock transition ( so called cr - precursor ) and the rankine - hugoniot relation for the shock compression at the gaseous discontinuity ( sub - shock ) .
the complete system of equations is then reduced to one nonlinear integral equation @xcite .
this equation is solved numerically and the results are shown in fig.[fig : spectra - of - protons ] .
the above method of calculation of the parent particle spectra is shown to be in excellent agreement @xcite with numerical simulations @xcite as well as with other semi - analytic approaches @xcite . once the parent proton spectrum is obtained , we calculate the @xmath82 production rate and the gamma - ray emissivity . in so doing ,
we adopt numerical recipe described in detail in @xcite .
the physical processes behind these calculations are @xcite ( i ) collisions of accelerated protons with the protons of the ambient gas resulting in the following spectrum of @xmath82-mesons : where , @xmath88 is the number density of protons in the gas , @xmath89 is the differential cross section for the @xmath82 production in collisions between accelerated protons of energy @xmath90 and gas protons , @xmath91 is the flux of accelerated protons , @xmath92 is the energy of @xmath82 mesons ; ( ii ) decay of @xmath82 resulting in the gamma emission spectrum we wold like to thank sabrina casanova for furnishing a detailed information about the code developed in ref.@xcite .
support by nasa under the grants nnx 07ag83 g and nnx09at94 g as well as by the department of energy , grant no .
de - fg02 - 04er54738 is gratefully acknowledged .
m.a.m . carried out the calculations of the spectra , p.h.d .
contributed to the wave propagation part , r.z.s .
contributed to the wave - particle interaction part .
the article was written by m.a.m and edited by all the co - authors
. - plane of momentum space .
pr in the stripes @xmath23 are not scattered by waves ( see text )
. therefore , particles from the domains @xmath95 maintain their propagation direction and promptly escape from the dense gas region .
[ fig : scattering - zone - of ] ] is steeper by two powers than the spectra in energy normalization , used in the text ) .
both spectra are multiplied by @xmath96 , so that the test particle distribution is flat .
shock parameters : acoustic mach number @xmath97 , shock velocity @xmath98 , the break momentum @xmath99 .
shock pre - compression ( flow compression across the cr precursor ) r=1.8 , injection parameter @xmath100 [ defined as @xmath101 , with @xmath102 and @xmath57 being the ambient gas density and the shock speed , respectively ] ; injection momentum @xmath103.[fig : spectra - of - protons ] ] decay and calculated for two different parent proton spectra compared against the fermi data ( open circles ) .
solid line : a test particle acceleration regime with the spectral index @xmath104 below the break and @xmath105 above the break at @xmath106/c . dashed line : a moderately nonlinear acceleration regime corresponding to the spectrum shown in fig.[fig : spectra - of - protons ] ( @xmath107 and @xmath108 below and above the break , respectively ) .
cut - offs are placed at 300 gev for tp- and 100 gev , for nl - spectrum .
[ fig : gamma - emission - from ] ] | recent observations of the supernova remnant w44 by the _ fermi _ spacecraft observatory strongly support the idea that the bulk of galactic cosmic rays is accelerated in such remnants by a fermi mechanism , also known as diffusive shock acceleration .
however , the w44 expands into weakly ionized dense gas , and so a significant revision of the mechanism is required . in this paper
we provide the necessary modifications and demonstrate that strong ion - neutral collisions in the remnant surrounding lead to the steepening of the energy spectrum of accelerated particles by _ exactly one power_. the spectral break is caused by alfven wave evanescence leading to the fractional particle losses . the gamma - ray spectrum generated in collisions of
the accelerated protons with the ambient gas is also calculated and successfully fitted to the fermi observatory data .
the parent proton spectrum is best represented by a classical test particle power law @xmath0 , steepening to @xmath1 at @xmath2 due to deteriorated particle confinement .
ongoing _ direct _ observations of a number of galactic supernova remnants ( snrs ) in the x- and gamma - ray bands supported by the radio , optical and uv data @xcite seem to rapidly close in on the century long problem of the cosmic ray ( cr ) origin .
overall , the observations favor the diffusive shock acceleration ( dsa , operating in snr shocks ) as a mechanism for the production of galactic crs . however , there are at least two critical questions that observations pose to the theory .
first , what exactly is the form of the spectrum that the theory predicts ?
second , when do we really see the proton ( i.e. , the primary cr component ) emission and when do we see just a contaminating but radiatively more efficient electron component ?
several decades of studies of the background galactic crs , preceding direct observations ( that became available only recently ) demonstrated that already a simple test particle version of the dsa ( leading to a power - law @xmath0 particle energy distribution ) reproduces the form of the energy spectrum reasonably well .
it is possible , however , that a somewhat steeper power - law would better accommodate the @xmath3 measured spectrum @xcite ( see also the discussion section below ) , even though this difference can be attributed to cr propagation losses ( also not completely certain ) .
nevertheless , the _ direct observations _ of emission coming from particles accelerated in snr shocks often indicate significantly steeper than @xmath4 spectra @xcite .
also the recent _ fermi _ large area telescope ( lat ) observations of the snr w44 suggest that the spectrum of the gamma ray producing protons is substantially steeper at high energies than the dsa predicts @xcite .
it should be noted that a similar discrepancy has already been found in the high energy gamma ray spectrum of the snr rxj 1713 measured by the cangaroo @xcite and hess @xcite atmospheric cerenkov telescopes .
the simple dsa prediction was even used as an argument against the hadronic origin of the gamma emission on the ground that if the high energy part of the spectrum was extrapolated ( with the dsa index 2 ) to lower energies , the emission there would be unacceptably high , e.g. @xcite . motivated by this inconsistency with the dsa theory in particular ,
we have previously published a suggestion of the mechanism for the spectral break @xcite , which can resolve the discrepancies .
we believe that a similar phenomenon is also applicable to the recent observations of the snr w44 .
in particular , we show below that ion - neutral collisions in a molecular cloud adjacent to the remnant steepen the energy spectrum of accelerated particles .
we calculate the gamma - ray spectrum resulting from the collisions of accelerated protons with the molecular cloud material and fit the result to the fermi observatory data . |
this work is supported in part by the dfg and the nsfc through funds provided to the sino - german crc 110 `` symmetries and the emergence of structure in qcd '' ( nsfc grant no .
11261130311 ) , by the eu integrated infrastructure initiative hadronphysics3 ( grant no .
283286 ) , and by nsfc ( grant no .
11165005 ) . | we study the production of the spin partner of the @xmath0 , which is a @xmath1 bound state with quantum numbers @xmath2 and named @xmath3 here , with the associated emission of a photon in electron positron collisions .
the results show that the ideal energy region to observe the @xmath4 in @xmath5 annihilations is from 4.4 gev to 4.5 gev , due to the presence of the @xmath6-wave @xmath7 and @xmath8 thresholds , respectively .
we also point out that it will be difficult to observe the @xmath9 at the @xmath5 center - of - mass energy around 4.26 gev . in the heavy quarkonium mass region , the so called @xmath10 states
have been observed , and many of these quarkonium - like states defy a conventional quark model interpretation .
they are therefore suggested to be exotic .
the @xmath0 , discovered by the belle collaboration @xcite , is the one of the most interesting exotic states .
as the mass of the @xmath0 is extremely close to the @xmath11 threshold , it is regarded as one especially promising candidate for a hadronic molecule .
effective field theory ( eft ) can cope with the interaction between heavy mesons in bound state systems at low energies .
for such a kind of systems , heavy quark symmetry is relevant due to the presence of the heavy quark / antiquark in the meson / antimeson .
this fact leads to predictions of new states as partners of the observed @xmath10 states in the hadron spectrum .
for example , with an eft description of the heavy mesonic molecules , the heavy quark symmetry can be used to predict the existence of the spin and bottom partners of the @xmath0 @xcite .
the spin partner of the @xmath0 , called @xmath4 hereafter , is predicted to exist as the @xmath6-wave bound state of @xmath1 with quantum numbers @xmath12 @xcite .
such a state was also expected to exist in other models , see refs .
@xcite .
it is different from the @xmath0 in several aspects : first , being an isoscalar state it should decay into the @xmath13 with a branching fraction much larger than that for the @xmath14 because the @xmath15 and @xmath16 thresholds are far below the mass of the @xmath4 ( very different to the case of the @xmath0 ) ; second , it is expected to decay dominantly into open charm mesons , @xmath17 , @xmath18 and @xmath19 , in a @xmath20-wave with a width of the order of a few mev @xcite ; third , its mass as set by the @xmath21 threshold is higher than the quark model prediction for the first radially excited @xmath22 @xcite .
the significance of the @xmath3 state is that its mass should be approximately given by the @xmath23 as dictated by heavy quark spin symmetry for heavy - flavor hadronic molecules @xcite .
notice that a state with the same quantum numbers @xmath12 was also predicted in the tetraquark model @xcite .
however , the fine splitting between the @xmath12 and @xmath24 tetraquarks , which was predicted to be 70 mev in ref .
@xcite , is not locked to that between the @xmath25 and @xmath20 .
similarly , the splitting between the @xmath26 @xmath27 states in the godfrey
isgur quark model is 30 mev @xcite , also much smaller than @xmath28 .
therefore , if a @xmath12 state will be observed in experiments with a mass around 4012 mev , the mass by itself would already be a strong support for the hadronic molecular nature of both the @xmath0 and the tensor state . as a result , searching for a @xmath12 state with a mass around 4012 mev is very important even for understanding the nature of the @xmath0 . however , although the @xmath0 has been observed by many other experiments after its discovery @xcite , no evidence for the existence of its spin partner has been reported . in ref .
@xcite , it is shown that the prompt production of the @xmath4 presents a significant discovery potential at hadron colliders . in this paper
, we will investigate the production of the @xmath4 associated with the photon radiation in electron positron collisions .
this work presents an extension of the study on the production of the @xmath0 as a @xmath29 molecule in charmonia radiative transitions reported in ref .
@xcite . in that paper
, it was shown that the favorite energy regions for the @xmath30 production are around the @xmath31 mass and 4.45 gev .
later on , the besiii collaboration observed events for the process @xmath32 @xcite , which may be regarded as a support of the dominantly molecular nature of the @xmath0 .
since the existence of the @xmath21 bound state , the @xmath4 , is the consequence of the heavy quark spin symmetry of the molecular nature of the @xmath0 , the production of the @xmath4 in @xmath5 collisions in the energy range of the besiii experiment @xcite thus provides an opportunity to search for new charmonium - like states on the one hand and can offer useful information towards understanding the @xmath0 on the other hand .
the production of the @xmath0 through the radiative decay of the @xmath33 charmonium is considered in ref .
@xcite using heavy hadron chiral perturbation theory along with the x - eft @xcite .
then , ref .
@xcite studied the @xmath0 production by considering the contribution from intermediate charmed meson loops , and it was argued that the dominant mechanism is as follows : the initial charmonium is coupled to a pair of charmed mesons with one being @xmath6-wave with @xmath34 , where @xmath35 is the total angular momentum of the light - flavor cloud in the charmed meson , and the other being @xmath36-wave with @xmath37 , and the @xmath36-wave charmed meson radiatively transits to a @xmath38 which coalesce with the other @xmath6-wave charmed meson , @xmath39 , into the @xmath0 . the spin partner of @xmath0 , the @xmath4 , can be produced by a similar mechanism as shown in fig .
[ fig : feynmandiagram ] . in vector charmonia
radiative decays .
the charge - conjugated diagrams are not shown here .
, scaledwidth=80.0% ] notice that the @xmath4 couples to @xmath40 instead of @xmath41 , as it is in the case of the @xmath0 .
we will only consider the neutral charmed mesons in the loops because the photonic coupling between the @xmath36-wave and @xmath6-wave charmed mesons for the neutral ones is much larger than that for the charged .
this is due to cancellation of contributions from the charm and down quarks in the charged mesons , see , e.g. @xcite . in the loops , the @xmath4 couples to the @xmath42 pair in an @xmath6-wave . with the quantum numbers being @xmath43 , the initial charmonium can couple to one @xmath36-wave and one @xmath6-wave charmed meson in either @xmath6- or @xmath20-wave .
since both the initial charmonium and the @xmath4 in the final state are close to the corresponding thresholds of the charmed - meson pairs , we are able to use a power counting in velocity of the intermediate mesons . following the power counting rules as detailed in ref .
@xcite and presented in the case of interest in refs .
@xcite , the dominant contribution comes from the case when the coupling of the initial charmonium to the charmed mesons is in an @xmath6-wave . in this case
, the initial charmonium should be a @xmath20-wave state in the heavy quark limit @xmath44 as a consequence of heavy quark spin symmetry @xcite .
the charmed mesons can be classified according to the total angular momentum of the light degrees of freedom @xmath35 and collected in doublets with total spin @xmath45 in the heavy quark limit .
the @xmath46 states correspond to charmed mesons in the doublet @xmath47 , here denoted as @xmath48 , whereas the @xmath49 states correspond to charmed mesons in the doublet @xmath50 , denoted as @xmath51 . to describe these heavy mesons
, we choose the two - component notation introduced in ref .
@xcite .
the notation uses @xmath52 matrix fields , and is convenient for nonrelativistic calculations .
the fields for the relevant heavy meson states are @xmath53 for the @xmath46 ( @xmath6-wave ) and @xmath54 ( @xmath36-wave ) heavy mesons , respectively , where @xmath55 are the pauli matrices , and @xmath56 is the flavor index for the light quarks . in eq . ,
@xmath57 and @xmath58 annihilate the pseudoscalar and vector heavy mesons , respectively , and @xmath59 and @xmath60 annihilate the excited axial - vector and tensor heavy mesons , respectively . under the same phase convention for charge conjugation specified in ref .
@xcite , the fields annihilating the mesons containing an anticharm quark are @xcite @xmath61 in nonrelativistic limit , the field for the @xmath20-wave @xmath43 charmonium state can be written as @xcite @xmath62 where @xmath63 annihilates the @xmath20-wave vector charmonium , and the spin-0 and spin-2 states irrelevant for our study are not shown . in order to calculate the triangle diagrams in fig .
[ fig : feynmandiagram ] , we need the lagrangian for coupling the @xmath20-wave charmonia to the @xmath64-@xmath65 charmed - meson pair as well as that for the e1 radiative transitions between the charmed mesons @xcite @xmath66 + \sum_a \frac{c_a}{2}\ , { \textrm{tr}}\left [ t_a^i h_a^\dag \right ] e^i + \text{h.c.},\end{aligned}\ ] ] where in the first term the einstein summation convention is used while for the latter we distinguish the coupling constants for different light flavors because there is no isospin symmetry in the electromagnetic interaction .
moreover ,
we parametrize the coupling of the @xmath4 to the pair of vector charm and anticharm mesons as @xmath67 with the above preparations , we can now proceed to calculate quantitatively the production of the @xmath68 in electron positron collisions .
although in the heavy quark limit the production of the @xmath20-wave vector heavy quarkonium or the pair of @xmath64 and @xmath65 heavy mesons are suppressed due to spin symmetry @xcite , we can expect a large spin symmetry breaking in the charmonium mass region above 4 gev .
this may be seen from similar values of electronic widths of the excited vector charmonia .
thus , we will assume that the production of the @xmath68 occurs through the @xmath20-wave charmonia or the @xmath20-wave components of excited vector charmonia .
without any detailed information about the values of the coupling constants , we can predict the energy regions with the maximal production cross sections .
-wave charmonium into @xmath69 on the mass of the charmonium .
the solid and dotted curves are obtained with and without taking into account the widths of the @xmath70 and @xmath71 , respectively . here ,
@xmath72 is used.,scaledwidth=60.0% ] in fig .
[ fig : psidxga ] , we show the dependence of the decay width of a @xmath20-wave charmonium into the @xmath68 , divided by @xmath73 , on the mass of the @xmath20-wave charmonium or the center - of - mass energy of the @xmath5 collisions . the value of the photonic coupling @xmath74 does not affect the shape of the dependence either .
nevertheless , we took @xmath72 which is a typical value evaluated from various quark model predictions for the decay widths @xmath75 @xcite . in the figure
, the dashed curve is obtained neglecting the widths of the @xmath76 and @xmath77 states , and the solid curve is the result of evaluating the triangle loop integrals with constant widths for the @xmath76 and @xmath77 as done in ref .
@xcite .
the maximum around 4.447 gev and the local minimum around 4.492 gev of the dashed curve are due to the presence of landau singularities @xcite of triangle diagrams in the complex plane at @xmath78 gev ( for the @xmath76 loop ) and @xmath79 gev ( for the @xmath77 loop ) , respectively ( for a discussion of the landau singularities in the triangle diagrams of heavy quarkonium transitions , we refer to ref .
@xcite ) .
the two cusps on both sides of the shoulders of the peak show up at the thresholds of the @xmath80 and @xmath81 . from the figure ,
it is clear that the ideal energy regions for producing the @xmath68 in @xmath5 collisions are around the @xmath80 and @xmath81 thresholds , i.e. between 4.4 gev and 4.5 gev .
it is also clear that the mass region of the @xmath82 is not good for the production of the @xmath68 , contrary to the case of the @xmath83 . in order to quantify the relative production rate of the @xmath68 with respect to the @xmath83
, we require the @xmath82 to couple to the @xmath64-@xmath65 meson pair as follows @xmath84 where we have assumed isospin symmetry in the couplings and the flavor index @xmath56 runs over up and down quarks .
notice that if the @xmath82 is a pure @xmath85 ( here and in the following the charge conjugated channels are dropped for simplicity ) molecule @xcite , it would not couple to the @xmath80 and @xmath81 as given by the @xmath86 and @xmath87 terms , and thus can not decay into the @xmath68 .
these two terms are included to allow the decay to occur . to have @xmath80 and @xmath81 components
was discussed in ref .
@xcite . ]
because the @xmath4 is the spin partner of the @xmath0 , for a rough estimate , we can assume that @xmath88 takes the same value as the coupling constant of the @xmath0 to the @xmath89 .
we also assume that the values of @xmath90 and @xmath91 are related to @xmath92 by a spin symmetry relation for @xmath20-wave charmoinia . comparing eq . with eq .
one obtains @xmath93 and @xmath94 .
then , the ratio of the partial decay widths of the @xmath82 to the @xmath68 and the @xmath83 can be estimated parameter - free , and is @xmath95 in the above ratio , whether or not to take into account the finite widths of the @xmath36-wave charmed mesons only results in a minor change of 2% .
it is clear that unless the @xmath82 couples to the @xmath80 and/or @xmath81 with a coupling much larger than that for the @xmath85 , which is less possible , the branching fraction of the @xmath96 is much smaller than that of the @xmath97 . given that the number of events for the latter process as observed at besiii is the order of 10 @xcite , it is unlikely to make an observation of the @xmath68 at an energy 4.26 gev at besiii .
to summarize , it is generally expected that the @xmath0 as a hadronic molecule has a spin partner close to the @xmath21 threshold . in this paper
, we have investigated the production of the @xmath68 in @xmath5 collisions . according to our calculation
, we strongly suggest to search for the @xmath4 associated with a photon in the energy region between 4.4 gev and 4.5 gev in @xmath5 collisions . besides , the width ratio of the @xmath82 decaying to @xmath69 and @xmath98 is quite small , at the order of @xmath99 .
thus observing the @xmath68 at an energy around 4.26 gev would be unlikely in the besiii experiment according to the current result of @xmath100 . |
galaxies like the draco dwarf spheroidal may be the building blocks of the universe ( koo 1994 ) .
draco is among the faintest known galaxies and indeed , at @xmath2 , ranks among the middling globular clusters in luminosity .
however , its very extended distribution of stars ( @xmath3 kpc ) clearly distinguishes it from the comparatively tightly condensed globulars .
moreover , assuming dynamical equilibrium leads one to very high estimates for the mass - to - light ratio ( aaronson 1983 ) , with recent multifiber spectroscopy yielding a remarkable @xmath4 ( armandroff , olszewski , & pryor 1995 ) .
if the dark matter is primarily stellar , as might be inferred from the findings of recent microlensing experiments , then the initial mass function ( imf ) in these galaxies might have been very different from the present - day imf in our own galaxy .
it is clearly important to try to piece together the star formation histories of such galaxies as this may have important consequences for ideas concerning the formation and evolution of more luminous galaxies like our own . the color - magnitude ( cm )
diagram of the draco dwarf was first studied by baade & swope ( 1961 ) , and found to be generally similar to the cm - diagrams of metal - poor globular clusters with the exception of having a rather broad giant branch .
baade & swope also first detected the `` anomalous '' cepheids , more luminous than the period - luminosity relation for type ii cepheids would predict , and now known to be characteristic of dwarf galaxies . another marked anomaly in the cm diagram is draco s red horizontal branch , which is incompatible with low metallicity in a population ii system ( the second parameter problem ) .
this phenomenon alone could be taken to suggest that draco is @xmath5 years younger than classic , metal - poor globular clusters like m92 ( lee , demarque , & zinn 1994 ) .
other hypotheses have been offered to explain the second parameter phenomenon ( renzini 1978 ) , and an increasing body of evidence suggests that age may be no more than a minor contributor to horizontal branch morphology ( stetson , vandenberg , & bolte 1996 ) . however , the debate is by no means settled , and the reader is referred to lee ( 1994 ) and chaboyer , demarque , & sarajedini ( 1996 ) for arguments in support of the age hypothesis .
the best ground - based ccd cm diagrams of the draco dwarf spheroidal to date have been presented by stetson , vandenberg , & mcclure ( 1985 ) and carney & seitzer ( 1986 , hereafter cs86 ) . like the ursa minor dwarf spheroidal ( the 2nd - faintest dwarf spheroidal known )
, draco appears to be as old as the oldest globular clusters , though cs86 and others discuss evidence from blue stragglers for an intermediate - age population as well .
the superb resolution afforded by the wide field / planetary camera 2 ( wfpc2 ) on the _ hubble space telescope _ permits a remarkable improvement in our ability to study the stellar content of the draco dwarf spheroidal as well as other galaxies in the local group .
here we describe the first results of an _ hst _ program to investigate the stellar population of this galaxy .
the images we presently discuss were taken on june 9 , 1995 with wfpc2 .
the region selected for wfpc2 imaging is shown overlaid on the digital sky survey in figure [ fig : plate1 ] ( plate 1 ) , and is roughly centered on field 1 of cs86 .
one 200s and two 1000s exposures were taken in f606w ( @xmath6 ) , while the f814w ( @xmath7 ) images comprise 200s , 1100s , and 1300s integration time .
f606w was chosen over f555w to give a somewhat higher signal - to - noise ratio . in the process of cosmic ray removal , the exposures in each filter were combined into single , composite images from which subsequent photometry was derived .
a mosaic of the coadded wfpc2 frames in f814w is shown in figure [ fig : plate2 ] ( plate 2 ) .
photometry on the three wide - field ( wf ) detectors ( 00996 pix@xmath8 ) was carried out using a combination of aperture photometry and allframe psf - fitting photometry ( stetson 1994 ) .
allframe is the latest development in the daophot series of photometry packages and differs from its predecessors primarily in its ability to use information from many individual frames simultaneously to better constrain the final fitted magnitudes .
allframe is applied to the data _ after _ object detection and aperture photometry is carried out using daophot ii ( stetson 1987 ) . to push the star detection limit to as faint a level as possible , we coadded all images taken through the _ same _ filter . coadding the f606w and f814w images
prior to object detection might in principle seem to offer an even fainter detection limit , but doing so would have severely compromised the reliability of subsequent completeness determinations .
two daophot detection passes were carried out , separated by photometry and subtraction of all stars found in the first pass .
faint stars hidden within the psf skirts of brighter companions were thereby revealed and subsequently added to the list of detected stars to be measured .
point - spread functions for each detector and each filter were constructed within daophot using the @xmath9 brightest and most isolated point sources in each frame .
allframe was then applied to the f606w and f814w images simultaneously .
a total of 4429 stars were ultimately detected and measured in the three wf frames . due to the degree of undersampling and relatively poor psf representation in the wf chips ,
psf - fitting of brighter stars is subject to greater uncertainties in the absence of crowding than is pure synthetic - aperture photometry .
however , crowding and poor signal - to - noise ratio become ever increasing problems as one considers fainter and fainter stars , and the allframe - derived magnitudes yield a tighter main sequence than is obtainable using aperture photometry . in order to retain the small uncertainties associated with aperture photometry of bright , uncrowded giants and to simultaneously extend the main sequence as far as possible
, the color - magnitude diagram was constructed using both aperture and psf - fitted magnitudes .
the aperture magnitudes were used whenever the predicted total light within an aperture of radius 2 pixels from all other resolved stars ( based on the known characteristics of the wfpc2 psf ) within 2@xmath10 was less than 1% of the light from the star centered in the aperture .
corrections to a 05 aperture were made in each case , and the synthetic zero points of holtzman ( 1995 ) were used to compute wfpc2-system magnitudes .
completeness tests were carried out by adding @xmath11 stars at a time to each of the f606w and f814w images .
stars were added in integer
f814w steps , and with f606w - f814w @xmath12 mag .
the frames were then processed using daophot ii and allframe in a manner identical to that applied to the original data .
the results of these tests are shown in figure [ fig : completeness ] , which shows that the 50% completeness level occurs at f814w @xmath13 .
the faintest stars we can reliably measure therefore have masses of @xmath14 .
the tests also show that the photometric uncertainties are @xmath15 mag rms at f814w = 23 , and @xmath16 mag rms at f814w = 25 .
the f814w vs f606w
f814w color - magnitude ( cm ) diagram , constructed using both aperture photometry for uncrowded stars ( 269 ) and allframe fits for crowded stars ( 4156 ) is shown in figure [ fig : cm ] .
an electronic version of the photometry table is available on request from cjg .
the morphology in figure [ fig : cm ] above the turnoff is very similar to that found by stetson ( 1985 ) and cs86 .
owing to the superior resolution afforded by the _ hst _
data , we are able to extend the main sequence 2 magnitudes fainter than these ground - based studies . the turnoff region is well resolved and reveals that the bluest main - sequence stars have = 0.36 .
there are evidently fewer blue stragglers than in the cm diagram of cs86 , consistent with a ratio of @xmath17 in field of view .
unlike cs86 , we see no evidence for distinct multiple turnoffs ; most of the stars within the wfpc2 field of view appear to be approximately coeval .
we caution however that , while our photometry may be more accurate , we have sampled only one fifth as many stars in the turnoff region as have cs86 , and it may be premature to rule out the multiple turnoffs in draco .
additional wfpc2 fields of draco will be required before we can seriously address this issue . in figure
[ fig : cmtracks ] we compare the draco cm diagram with fiducial sequences for the galactic globular clusters m68 ( walker 1994 ) and m92 ( heasley & janes 1996 ) , which have [ fe / h ] = -2.09 and -2.24 dex , respectively ( djorgovski 1993 ) .
the draco magnitudes have been dereddened assuming e(_b - v _ ) = 0.03 ( stetson 1979a ) using the absorptions tabulated by holtzman ( 1995 ) for stars of k5 spectral type .
the globular sequences were dereddened and translated using the color excesses and distances tabulated by peterson ( 1993 ) .
a mean abundance for giants in draco has been spectroscopically measured to be [ fe / h ] @xmath18 dex ( lehnert 1992 ) . adopting a mean horizontal branch magnitude @xmath19 ( stetson 1979a ) , e@xmath20 = 0.03 ( stetson 1979b ) , and @xmath21 = 0.15 [ fe / h ] + 0.82 ( carney 1992 )
, we use @xmath22 the isochrones shown in figure [ fig : cmtracks ] are from vandenberg & bell ( 1985 ) , with colors from the @xmath23 vs @xmath24 relations for dwarfs of johnson ( 1966 ) , and for giants of ridgway ( 1980 ) .
@xmath23 colors have been transformed to @xmath25 empirically and in the correct metallicity regime via the preliminary multimetallicity color - color diagrams of worthey & fisher ( 1996 ) , and @xmath25 have in turn been converted to f606w f814w using the transformation equations in holtzman ( 1995 ) .
a sequence of normal points representing the color - magnitude sequence was generated by defining an inclusion envelope around the main sequence , sub - giant , and giant branches , and computing mean magnitudes and colors in 0.4 magnitude - wide bands . due to the scarcity of giant branch stars and the associated large uncertainties in their mean colors ,
the locus of the giant branch above f814w = 20.5 was estimated by eye .
this sequence , resampled using spline fits , was then compared using a maximum - likelihood technique with isochrones ranging in age from 12 to 18 gyrs , and in [ fe / h ] ranging from -1.2 to -2.2 .
the normal points were uniformly weighted , and the best - fitting isochrone had an age of 16 gyrs and [ fe / h ] = -2.2 . given the weighting of the data , a formal @xmath26 estimate of the uncertainty was not straightforward , and the fitting uncertainty was instead estimated by generating one hundred realizations of the sequence of normal points ( using gaussian deviates based on the dispersion in the mean colors in each magnitude bin ) and comparing them to each of the isochrones .
this yielded an age uncertainty , based purely on random and fitting errors , of 0.5 gyr . shown in figure [ fig : cmtracks ]
are isochrones corresponding to an age of 16 gyr and [ fe / h ] ranging from -1.2 to -2.2 dex .
an age of 16 gyrs is not atypical of ages found for several metal - poor galactic globular clusters using this set of isochrones .
however , despite the small uncertainty due to random effects , it is apparent from figure [ fig : cmtracks ] that there are clear , systematic differences between the isochrones and the data .
whereas the fiducial sequences of m68 and m92 agree reasonably well with the cm distribution of draco stars , the isochrones generally appear too red , particularly along the giant branch .
this is undoubtedly due in large part to the roundabout technique used above to transform values of @xmath24 to f606w
f814w colors .
given the numerous calibrations involved and the systematic discrepancies which can arise , it is clearly dangerous to place much faith in the age we derive using this method . a _ calibration - independent _ method for measuring relative ages ) , then the _ absolute _ ages of the globular clusters ( and draco ) are reduced by @xmath0 gyrs .
] has been described by vandenberg , bolte , & stetson ( 1990 ) .
this involves measuring the difference in color between the turnoff and the giant branch , which is relatively insensitive to metallicity but is a monotonic function of age . in figure
[ fig : cmaligned ] we show the fiducial sequence for draco ( generated as described above ) compared with similar sequences for the galactic globular clusters m68 and m92 . also shown
are isochrones having [ fe / h ] = -2.0 and ages ranging from 13 to 18 gyr .
the globular cluster sequences and the isochrones have been shifted horizontally to match the draco sequence at the bluest point of the turnoff region , and vertically to agree with one another at a point 0.05 magnitudes redward of the turnoff on the main sequence .
the draco sequence evidently agrees very well with the globular cluster sequences , though there are small differences in the shape of the turnoff and the slope of the giant branch .
the relative ages between sequences are best measured in this technique by comparing the colors of the giant branch , where the color separation is largly independent of magnitude .
looking at the isochrone spacing , it is evident that the color difference between giant branches is not uniform , but depends on absolute age .
if we calibrate age differences using the oldest ( leftmost ) two isochrones , then the mean color difference measured in the region 20 @xmath27 f814w @xmath27 21 ( below the hb and in a region where the giant branch is relatively well sampled ) yields a relative age difference between draco and either of the globular clusters of considerably less than one gyr .
note that using increasingly younger isochrones to establish the color difference age calibration would reduce the inferred age differences even more .
given the photometric scatter apparent on the main sequence , the presence of blue stragglers or large numbers of binary stars could conceivably affect the determination of the mean colors of the turnoff and the main sequence . to estimate our sensitivity to these effects ,
we have recomputed the draco normal points using an inclusion envelope alternately more exclusive of stars on the blue side of the turnoff and on the red side of the main sequence . in the latter case , we have excluded approximately three times as many outliers to the red of the main sequence as to the blue .
using these computed normal sequences ( in addition to a number of hand - sketched realizations ) and bringing all other sequences into alignment as before , we estimate our measuring uncertainty to be @xmath28 gyr .
the distribution of relative ages measured in this way is slightly asymmetric in the sense that an older draco is preferred over a younger .
additional uncertainty arises from possible errors in the color terms used to transform between photometric systems .
two such color terms are used here : the term used to transform the ground - based data into v - i , and that used to transform v - i to f606w - f814w . the close similarity in age
we infer for m68 and m92 agrees with the results of vandenberg et al .
( 1990 ) based on independent b , v data sets and suggests that errors in the ground - based color terms are likely to be small .
errors in the transformation from v - i to f606w - f814w arise from inaccuracies in the fitting formula used on the synthetic photometry results and from the error made by applying a transformation based on stars of solar metallicity to a metal - poor population .
we estimate that fitting errors would introduce an error of order 0.02 magnitudes in the difference in color between the turnoff and the giant branch .
this translates to an age uncertainty of about 2.3 gyrs .
applying transformation coefficients based on solar - metallicity stars to the more metal - poor stars of m68 and m92 has the effect of making the globular cluster giant branches in figure [ fig : cmaligned ] appear somewhat bluer relative to draco than they should . using kurucz model atmospheres , we find that stars at the turnoff and on the giant branch are fainter in f606w by 0.014 and 0.024 magnitudes , respectively , than what we compute using the holtzman ( 1995 ) coefficients .
the f814w transformation for these stars is much less sensitive to metallicity , and we estimate that the differences in f606w - f814w color between the turnoffs and the giant branches of m68 and m92 should be about 0.014 magnitudes larger than shown . correcting our measured result accordingly and combining the two major sources of error above , we conclude that draco is @xmath1 gyrs older than either m68 or m92 . given the very different horizontal branch morphologies of draco and the two globular clusters , what can we say about the 2nd parameter problem ?
in figure [ fig : ldz ] we partially reproduce the [ fe / h ] vs. hb morphology diagram of lee ( 1994 ) .
we have determined the morphological type quantity ( b - r)/(b+v+r ) for draco using the color - magnitude diagrams of both stetson ( 1979a ) and cs86 .
counting stars on the blue and red sides of the instability strip as well as likely rr lyrae variables , we find ( b - r)/(b+v+r ) = -.77 and -0.70 for the two data sets , respectively .
we plot the mean of these two values in figure [ fig : ldz ] along with similar quantities tabulated by lee ( 1994 ) for m68 and m92 . for both uniform mass - loss models and models
in which the mass - loss rate increases with age ( reimers 1977 ) , the draco point appears to fall about 2 gyr below the isochrone which best matches m68 and m92 .
as this is only 1.4@xmath29 our estimated age uncertainty , we can not completely rule out consistency between draco and an age - dependent hb morphology of the type modeled by lee .
wfpc2 photometry of old globular clusters , sufficient to reduce the relative age uncertainty to @xmath30 gyr , will be required before we can put firmer constraints on the models .
zinn ( 1978 ) and lehnert ( 1992 ) have demonstrated an abundance spread of approximately 1 dex in draco . aside from this chemical inhomogeneity ,
how well does the draco dwarf spheroidal galaxy fit the _ simple stellar population _ paradigm of renzini & buzzoni ( 1986 ) ?
we have seen that the turnoff cm diagram morphology of draco is very similar to that of m92 ; globular clusters are the prototype simple stellar populations , coeval systems of initially homogeneous composition .
clearly , draco is predominantly a simple stellar population originating as early as anything in the galaxy .
close examination of the cm diagram shows a significant population of blue stragglers ( see also cs86 ) .
such stars are often seen in simple stellar populations , including both ancient globular and younger open star clusters , but they can also result from an @xmath31 star formation history .
we can count them in color profiles ( figure [ fig : color ] ) formed out of half - magnitude strips from the turnoff region .
the surface density of foreground stars in this region of the cm diagram ( specifically a region 0.3 magnitudes wide in color and 2 magnitudes high in f606w ) is predicted by the galactic model of bahcall and soneira ( 1980 ) to be 17 per square degree .
thus , there is only a 2% probability of finding a single foreground star among the @xmath32 stars we identify as blue stragglers ( roughly speaking , those stars with f814w @xmath33 and @xmath34 ) .
there are between 60 and 70 stars in the same region ( @xmath35 and @xmath36 ) of the cm diagram of cs86 .
given that the area surveyed by cs86 is about 5 times that of the area covered by our wfpc2 field , a ratio of @xmath37 in the number of blue stragglers indicates reasonable agreement .
simulations of the turnoff region using the yale isochrones ( green , demarque & king 1984 ) require 6% of the mass to be contained in a second population , half the age of the primary population , in order to fully account for the blue stragglers in figure [ fig : cm ] .
we note that 6% is an upper limit on the size of a second younger component of draco , since blue stragglers can be an intrinsic part of a simple stellar population .
a putative extended star formation history strengthens the link between draco and more fully developed examples of this phenomenon , such as carina ( smecker - hane 1994 ; 1996 ) .
the difference between draco and carina and galaxies such as lgs 3 ( mould 1997 ) which are still forming stars today ( at 10% of the original rate in the latter case ) is then simply one of degree .
draco simply had a shorter tail to its initial burst of star formation .
there is considerable evidence that wind - induced mass - loss can not account for the stellar populations seen in several local group dwarfs , and that stripping or capture of gas must have occurred ( see gallagher & wyse 1994 for a review ) .
draco is currently among the nearest of the dwarf galaxies and , if its orbit is largely contained within its present distance , the probability of capturing significant amounts of rejuvenating circumgalactic gas may be substantially smaller than for dwarfs in the outskirts of the local group . alternatively , a high ultraviolet flux from the precursors to the galaxy and m31 may have removed substantially more gas from the nearby dwarfs ( van den bergh 1994 ) .
none of these remarks bear at all on the better - known curious phenomenon of draco s red horizontal branch .
the finding that draco is as old or older than m92 requires us to focus on the problem of understanding draco s rhb .
higher s / n ( and preferably f555w
f814w ) photometry performed differentially with respect to [ fe / h ] = 2 clusters have the potential to sharpen our conclusions further .
if the relative age uncertainty could be reduced to 0.5 gyr , one would have to turn to parameters such as cluster density ( e.g. fusi pecci 1996 ) as the second parameter in the globular cluster system .
the age spread of the halo , which is the defining feature of the searle & zinn ( 1978 ) galaxy formation scenario ( as opposed to eggen , lynden - bell , & sandage 1962 ) would then need to be defined by the ages of the low density element of the halo , namely draco , its sibling dwarf spheroidals , and the galactic bulge .
figure [ fig : lf ] shows the differential luminosity function ( lf ) for the stars in our sample .
also shown is a lf computed from walker s ( 1994 ) photometry of m68 .
no completeness corrections have been applied to the m68 data , and the counts have been arbitrarily normalized to our data at f814w = 23 .
note that the number of measured stars in the m68 sample is about @xmath38 larger than in the draco sample .
the draco lf appears to be somewhat steeper on the giant branch , and somewhat shallower below the turnoff .
uncertainties introduced by differences in noise characteristics , resolution , and detection efficiency make it difficult to draw firm conclusions but , over the ( admittedly narrow ) mass range considered here , there does not appear to be an obvious bias towards lower - mass stars which might be held to account for the very high estimates of draco s mass - to - light ratio .
based on the present data set , all we can say is that if the bulk of the matter is in the form of stars , such stars must have masses less than @xmath39 .
the observed lf can be used to place limits on the initial mass function ( imf ) in draco .
the current observations are of stars with masses between 0.6 and 0.9 .
we parameterize the imf in this mass range by a simple power law , with @xmath40 .
even stars of this low mass evolve significantly over a hubble time , so the inferred imf slope depends on the age of the stellar population in draco .
we construct model lfs using a variety of imfs and compare each model distribution with the observed one to see if it is plausible that they are drawn from the same parent population .
the construction of model lfs allows us to incorporate photometric errors ( both random and systematic ) and incompleteness into the analysis ; the incompleteness and errors which we apply to the model lfs are those determined from our artificial star experiments .
we can also include the effect of a binary star contribution if we make an assumption about the frequency and distribution of the relative masses of stars in binary pairs .
for draco , we assume a metallicity of [ fe / h]=-2 , a distance and extinction as quoted above , and use vandenberg ( 1983 ) isochrones ( from which worthey s models are derived for low mass stars ) along with kurucz model atmospheres to convert from masses to assumed luminosities . we define as unacceptable imf slopes for which the model lf and the observed lf are inconsistent with each other at the 95% confidence level as determined by a kolmogorov - smirnoff test . with this definition ,
we constrain the imf slope to fall within the range @xmath41 for an assumed age of 18 gyr , @xmath42 for an assumed age of 15 gyr , @xmath43 for an age of 12 gyr , and @xmath44 for an age of 10 gyr when we fit the f606w lf over the range @xmath45 with stellar models of a single metallicity , [ fe / h]=-2 .
the dependence on age is easy to understand : as a population evolves , the lf around the turnoff steepens , leading to a flatter inferred imf for an older population .
these results were derived assuming a binary fraction of 0.5 with uncorrelated masses of the binary stars ; with no binaries , we get slopes which are slightly shallower .
similarly , small variations on the above allowed ranges are derived with different assumptions about the metallicity , distance , and/or a different adopted magnitude range over which the comparison with the observed lf is made .
if we fit the f814w lf , we infer a range of possible imf slopes which are slightly steeper than those inferred from the f606w lf . for no binaries , the inferred slope range overlaps between the two bandpasses ; for a binary fraction of 0.5 , all of the acceptable imf slopes for f814w are steeper than those allowed by the f606w lf . for comparison ,
the imf in this mass range in the solar neighborhood is proportional to @xmath46 ( kroupa , tout , & gilmore 1993 ) .
the allowed slopes for draco are remarkably similar to this value .
based on photometry of wfpc2 images of the draco dwarf spheroidal , we conclude the following : * draco is predominantly an old system , being @xmath47 gyrs older than either m68 or m92 . *
we see no evidence for multiple turnoffs which would be indicative of episodic star formation . *
the presence of significant numbers of blue stragglers _ may _ indicate an extended period of star formation , though the numbers of younger stars @xmath48% of the first generation . *
the luminosity function looks in all respects nearly identical to that of the old globular cluster m68 . *
the slope of the initial mass function appears to be very similar to that of the imf determined for the solar neighborhood .
future work should be aimed at refining the relative ages of draco and blue horizontal branch clusters like m92 .
obtaining both higher signal - to - noise ratio data as well as wfpc color - magnitude diagrams of metal poor globular clusters should allow us to reduce the relative age uncertainty to @xmath49 gyr .
such data would considerably strengthen constraints on the process and mechanisms which led to the formation of our galaxy .
we are grateful to gary da costa for a critical reading of an earlier version of this manuscript .
this research was conducted by the wfpc2 investigation definition team , supported in part by nasa grant no .
nas7 - 1260 .
fusi pecci , f. , bellazzini , m. , ferraro , f. r. , buonanno , r. , & corsi , c. e. 1996 , in `` formation of the galactic halo ... inside and out '' , eds .
h. morrison & a. sarjedini , a.s.p .
conference series 92 , p. 221 .
smecker - hane , t. a. , stetson , p. b. , hesser , j. e. , vandenberg , d. a. 1996 , in `` from stars to galaxies : the impact of stellar physics on galaxy evolution '' , eds . c. leitherer , u. fritze - van alvensleben , and j. huchra , ( asp : san francisco ) , asp conference series . | we present an f606w - f814w color - magnitude diagram for the draco dwarf spheroidal galaxy based on _ hubble space telescope _ wfpc2 images .
the luminosity function is well - sampled to @xmath0 magnitudes below the turn - off .
we see no evidence for multiple turnoffs and conclude that , at least over the field of the view of the wfpc2 , star formation was primarily single - epoch . if the observed number of blue stragglers is due to extended star formation , then roughly 6% ( upper limit ) of the stars could be half as old as the bulk of the galaxy .
the color difference between the red giant branch and the turnoff is consistent with an old population and is very similar to that observed in the old , metal - poor galactic globular clusters m68 and m92 . despite its red horizontal branch
, draco appears to be older than m68 and m92 by @xmath1 gyrs , lending support to the argument that the `` second parameter '' which governs horizontal branch morphology must be something other than age .
draco s observed luminosity function is very similar to that of m68 , and the derived initial mass function is consistent with that of the solar neighborhood
. epsf |
the * traveling salesman problem ( tsp ) * is one of the best known and most fundamental problems in combinatorial optimization .
determining how well it can be approximated in polynomial time is therefore a major open problem , albeit one for which the solution still seems elusive . on the algorithmic side ,
the best known efficient approximation algorithm for the symmetric case is still a 35-year old algorithm due to christofides @xcite which achieves an approximation ratio of @xmath4 . however , recently there has been a string of improved results for the interesting special case of graphic tsp , improving the ratio to @xmath5 @xcite . for the asymmetric case ( atsp ) , it is not yet known if a constant - factor approximation is even possible , with the best known algorithm achieving a ratio of @xmath6 @xcite .
unfortunately , there is still a huge gap between the algorithmic results mentioned above and the best currently known hardness of approximation results for tsp and atsp . for both problems ,
the known inapproximability thresholds are small constants ( @xmath0 and @xmath1 ( cf .
@xcite ) , respectively ) . in this paper
, we try to improve this situation somehow by giving modular hardness reductions that slightly improve the hardness bounds for both problems to @xmath2 and @xmath3 , respectively .
the latter bound is the first , for more than a decade now , improvement of papadimitriou and vempala bound @xcite for the atsp .
the method of our solution differs essentially from that of @xcite and uses some new paradigms of the bounded occurrence optimization which could be also of independent interest in other applications .
similarly to @xcite , the hope is that the modularity of our construction , which goes through an intermediate stage of a bounded - occurrence constraint satisfaction problem ( csp ) , will allow an easier analysis and simplify future improvements .
indeed , one of the main new ideas we rely on is a certain new variation of the wheel amplifiers first defined by berman and karpinski @xcite to establish inapproximability for 3-regular csps .
this construction , which may be of independent interest , allows us to establish inapproximability for a 3-regular csp with a special structure .
this special structure then makes it possible to simulate many of the constraints in the produced graph essentially `` for free '' , without using gadgets to represent them .
thus , even though for the remaining constraints we mostly reuse gadgets which have already appeared in the literature , we are still able to obtain improved bounds .
let us now recall some of the previous work on the hardness of approximation of tsp and atsp .
papadimitriou and yannakakis @xcite were the first to construct a reduction that , combined with the pcp theorem @xcite , gave a constant inapproximability threshold , though the constant was not more than @xmath7 for the tsp with distances either one or two .
engebretsen @xcite gave the first explicit approximation lower bound of @xmath8 for the problem .
the inapproximability factor was improved to @xmath9 by bckenhauer and seibert @xcite , who studied the restricted version of the tsp with distances one , two and three .
papadimitriou and vempala @xcite proved that it is -hard to approximate the tsp with a factor better than @xmath10 .
presently , the best known approximation lower bound is @xmath0 due to lampis @xcite . the important restriction of the tsp , in which we consider instances with distances between cities being values in @xmath11 , is often referred to as the @xmath12-tsp .
the best known efficient approximation algorithm for the @xmath13-tsp has an approximation ratio @xmath14 and is due to berman and karpinski @xcite . as for lower bounds , engebretsen and karpinski @xcite gave inapproximability thresholds for the @xmath12-tsp problem of @xmath15 for @xmath16 and @xmath17 for @xmath18 .
more recently , karpinski and schmied @xcite obtained improved inapproximability factors for the @xmath13-tsp and the @xmath19-tsp of @xmath20 and @xmath21 , respectively . for atsp
the currently best known approximation lower bound was @xmath1 due to papadimitriou and vempala @xcite .
when we restrict the problem to distances with values in @xmath11 , there is a simple approximation algorithm with approximation ratio @xmath22 that constructs an arbitrary tour as solution .
blser @xcite gave an efficient approximation algorithm for the @xmath23-atsp with approximation ratio @xmath24 .
karpinski and schmied @xcite proved that it is -hard to approximate the @xmath25-atsp and the @xmath19-atsp within any factor less than @xmath26 and @xmath27 , respectively .
for the case @xmath18 , engebretsen and karpinski @xcite gave an inapproximability threshold of @xmath28 .
* overview : * in this paper we give a hardness proof which proceeds in two steps .
first , we start from the max - e3-lin2 problem , in which we are given a system of linear equations mod @xmath29 with exactly three variables in each equation and we want to find an assignment such as to maximize the number of satisfied equations .
optimal inapproximability results for this problem were shown by hstad @xcite .
we reduce this problem to a special case where variables appear exactly @xmath30 times and the linear equations have a particular structure .
the main tool here is a new variant of the wheel amplifier graphs of berman and karpinski @xcite . in the second step ,
we reduce this @xmath30-regular csp to tsp and atsp .
the general construction is similar in both cases , though of course we use different gadgets for the two problems .
the gadgets we use are mostly variations of gadgets which have already appeared in previous reductions .
nevertheless , we manage to obtain an improvement by exploiting the special properties of the @xmath30-regular csp . in particular , we show that it is only necessary to construct gadgets for roughly one third of the constraints of the csp instance , while the remaining constraints are simulated without additional cost using the consistency properties of our gadgets
. this idea may be useful in improving the efficiency of approximation - hardness reductions for other problems .
thus , overall we follow an approach unlike that of @xcite , where the reduction is performed in one step , and closer to @xcite .
the improvement over @xcite comes mainly from the idea mentioned above , which is made possible using the new wheel amplifiers , as well as several other tweaks .
the end result is a more economical reduction which improves the bounds for both tsp and atsp .
an interesting question may be whether our techniques can also be used to derive improved inapproximability results for variants of the atsp and tsp ( cf .
@xcite,@xcite and @xcite ) or other graph problems , such as the steiner tree problem .
in the following , we give some definitions concerning directed ( multi-)graphs and omit the corresponding definitions for undirected ( multi-)graphs if they follow from the directed case . given a directed graph @xmath31 and @xmath32 , for @xmath33 , we define @xmath34 and @xmath35 . for convenience , we abbreviate a sequence of edges @xmath36 , @xmath37 by @xmath38 . in the undirected case ,
we use sometimes @xmath39 instead of @xmath40 , @xmath41 . given a directed ( multi-)graph @xmath42 , an _
cycle in @xmath42 is a directed cycle that traverses all edges of @xmath42 exactly once .
we refer to @xmath42 as _ eulerian _ , if there exists an eulerian cycle in @xmath42 . for a multiset @xmath43 of directed edges and @xmath44
, we define the outdegree ( indegree ) of @xmath45 with respect to @xmath43 , denoted by @xmath46 ( @xmath47 ) , to be the number of edges in @xmath43 that are outgoing of ( incoming to ) @xmath45 .
the _ balance _ of a vertex @xmath45 with respect to @xmath43 is defined as @xmath48 . in the case of a multiset @xmath43 of undirected edges ,
we define the balance @xmath49 of a vertex @xmath44 to be one if the number of incident edges in @xmath43 is odd and zero otherwise .
we refer to vertices @xmath44 with @xmath50 as _ balanced _ with respect to @xmath43 .
it is well known that a ( directed ) ( multi-)graph @xmath31 is eulerian if and only if all edges are in the same ( weakly ) connected component and all vertices @xmath51 are balanced with respect to @xmath52 . given a multiset of edges @xmath43 , we denote by @xmath53 the number of ( weakly ) connected components in the graph induced by @xmath43 .
quasi - tour _ @xmath43 in a ( directed ) graph @xmath42 is a multiset of edges from @xmath52 such that all vertices are balanced with respect to @xmath43 and @xmath54 .
we refer to a quasi - tour @xmath43 in @xmath42 as a _ tour _ if @xmath55 .
given a cost function @xmath56 , the cost of a quasi - tour @xmath43 in @xmath42 is defined by @xmath57 .
in the asymmetric traveling salesman problem ( atsp ) , we are given a directed graph @xmath31 with positive weights on edges and we want to find an ordering @xmath58 of the vertices such as to minimize @xmath59}d_g(v_i , v_{i+1})$ ] , where @xmath60 denotes the shortest path distance in @xmath42 . in this paper , we will use the following equivalent reformulation of the atsp : given a directed graph @xmath42 with weights on edges , we want to find a tour @xmath43 in @xmath42 , that is , a spanning connected multi - set of edges that balances all vertices , with minimum cost .
the metric traveling salesman problem ( tsp ) is the special case of the atsp , in which instances are undirected graphs with positive weights on edges .
in this section , we define the bi - wheel amplifier graphs which will be our main tool for proving hardness of approximation for a bounded occurrence csp with some special properties .
bi - wheel amplifiers are a simple variation of the wheel amplifier graphs given in @xcite .
let us first recall some definitions ( see also @xcite ) .
if @xmath42 is an undirected graph and @xmath61 a set of vertices , we say that @xmath42 is a _ @xmath62-regular amplifier _ for @xmath63 if the following conditions hold : * all vertices of @xmath63 have degree @xmath64 and all vertices of @xmath65 have degree @xmath62 . * for every non - empty subset @xmath66
, we have the condition that @xmath67 , where @xmath68 is the set of edges with exactly one endpoint in @xmath69 .
we refer to the set @xmath63 as the set of _ contact _ vertices and to @xmath65 as the set of _ checker _ vertices .
amplifier graphs are useful in proving inapproximability for csps , in which every variable appears a bounded number of times . here
, we will rely on @xmath30-regular amplifiers . a probabilistic argument for the existence of such graphs
was given in @xcite , with the definition of wheel amplifiers .
a wheel amplifier with @xmath70 contact vertices is constructed as follows : first construct a cycle on @xmath71 vertices .
number the vertices @xmath72 and select uniformly at random a perfect matching of the vertices whose number is not a multiple of 7 .
the matched vertices will be our checker vertices , and the rest our contacts .
it is easy to see that the degree requirements are satisfied .
berman and karpinski @xcite gave a probabilistic argument to prove that with high probability the above construction indeed produces an amplifier graph , that is , all partitions of the sets of vertices give large cuts .
here , we will use a slight variation of this construction , called a bi - wheel .
a * bi - wheel amplifier * with @xmath70 contact vertices is constructed as follows : first construct two disjoint cycles , each on @xmath73 vertices and number the vertices of each @xmath74 .
the contacts will again be the vertices whose number is a multiple of 7 , while the remaining vertices will be checkers . to complete the construction , select uniformly at random a perfect matching from the checkers of one cycle to the checkers of the other . intuitively , the reason that amplifiers are a suitable tool here is that , given a csp instance , we can use a wheel amplifier to replace a variable that appears @xmath70 times with @xmath71 new variables ( one for each wheel vertex ) each of which appears 3 times .
each appearance of the original variable is represented by a contact vertex and for each edge of the wheel we add an equality constraint between the corresponding variables .
we can then use the property that all partitions give large cuts to argue that in an optimal assignment all the new vertices take the same value .
we use the bi - wheel amplifier in our construction in a similar way .
the main difference is that while cycle edges will correspond to equality constraints , matching edges will correspond to inequality constraints .
the contacts of one cycle will represent the positive appearances of the original variable , and the contacts of the other the negative ones .
the reason we do this is that we can encode inequality constraints more efficiently than equality with a tsp gadget , while the equality constraints that arise from the cycles will be encoded in our construction `` for free '' using the consistency of the inequality gadgets .
before we apply the construction however , we have to prove that the bi - wheel amplifiers still have the desired amplification properties .
[ thm : biwheel ] with high probability , bi - wheels are 3-regular amplifiers .
exploiting the similarity between bi - wheels and the standard wheel amplifiers of @xcite , we will essentially reuse the proof given there .
first , some definitions : we say that @xmath69 is a _
bad set _ if the size of its cut is too small , violating the second property of amplifiers .
we say that it is a _ minimal bad set _
if @xmath69 is bad but removing any vertex from @xmath69 gives a set that is not bad .
recall the strategy of the proof from @xcite : for each partition of the vertices into @xmath69 and @xmath75 , they calculate the probability ( over the random matchings ) that this partition gives a minimal bad set .
then , they take the sum of these probabilities over all potentially minimal bad sets and prove that the sum is at most @xmath76 for some constant @xmath77 .
it follows by union bound that with high probability , no set is a minimal bad set and therefore , the graph is a proper amplifier .
our first observation is the following : consider a wheel amplifier on @xmath71 vertices where , rather than selecting uniformly at random a perfect matching among the checkers , we select uniformly at random a perfect matching from checkers with labels in the set @xmath78 to checkers with labels in the set @xmath79 .
this graph is almost isomorphic to a bi - wheel .
more specifically , for each bi - wheel , we can obtain a graph of this form by rewiring two edges , and vice - versa .
it easily follows that properties that hold for this graph , asymptotically with high probability also hold for the bi - wheel .
thus , we just need to prove that a wheel amplifier still has the amplification property if , rather than selecting a random perfect matching , we select a random matching from one half of the checker vertices to the other .
we will show this by proving that , for each set of vertices @xmath80 , the probability that @xmath80 is a minimal bad set is roughly the same in both cases .
after establishing this fact , we can simply rely on the proof of @xcite .
recall that the wheel has @xmath81 checker vertices . given a set @xmath80 with @xmath82 ,
what is the probability that exactly @xmath83 edges have exactly one endpoint in @xmath80 ? in a standard wheel amplifier the probability is @xmath84 where we denote by @xmath85 the product of all odd natural numbers less than or equal to @xmath86 , and we assume without loss of generality that @xmath87 is even . let us explain this : the probability that exactly @xmath83 edges cross the cut in this graph is equal to the number of ways we can choose their endpoints in @xmath80 and in its complement , times the number of ways we can match the endpoints , times the number of matchings of the remaining vertices , divided by the number of matchings overall .
how much does this probability change if we only allow matchings from one half of the checkers to the other ? intuitively , we need to consider two possibilities : one is that @xmath80 is a balanced set , containing an equal number of checkers from each side , while the other is that @xmath80 is unbalanced .
it is not hard to see that if @xmath80 is unbalanced , then , we can easily establish that the cut must be large .
thus , the main interesting case is the balanced one ( and we will establish this fact more formally ) .
suppose that @xmath82 and @xmath80 contains exactly @xmath88 checkers from each side .
then the probability that there are exactly @xmath83 edges crossing the cut is @xmath89 let us explain this . if @xmath80 is balanced and there are @xmath83 matching edges with exactly one endpoint in @xmath80 , then , exactly @xmath90 of them must be incident on a vertex of @xmath80 on each side , since the remaining vertices of @xmath80 must have a perfect matching .
again , we pick the endpoints on each side , and on the complement of @xmath80 , select a way to match them , select matchings on the remaining vertices and divide by the number of possible perfect matchings . using stirling formulas ,
it is not hard to see that @xmath91 .
also @xmath92 .
it follows that @xmath93 is roughly the same as @xmath94 in this case , modulo some polynomial factors which are not significant since the probabilities we are calculating are exponentially small .
let us now also show that if @xmath80 is unbalanced , the probability that it is a minimal bad set is even smaller .
first , observe that if @xmath80 is a minimal bad set whose cut has @xmath83 edges , we have @xmath95 .
the reason for this is that since @xmath80 is bad , then , @xmath83 is smaller than the number of contacts in @xmath80 minus the number of cycle edges cut .
it is not hard to see that , in each fragment , that is , each subset of @xmath80 made up of a contiguous part of the cycle , two cycle edges are cut .
thus , the extra edges we need for the contacts the fragment contains are at most @xmath96 of its checkers .
suppose now that @xmath80 contains @xmath97 checkers on one side and @xmath98 checkers on the other .
the probability that @xmath83 matching edges have one endpoint in @xmath80 is @xmath99 the reasoning is the same as before , except we observe that we need to select more endpoints on the side where @xmath80 is larger , since after we remove checkers matched to outside vertices @xmath80 must have a perfect matching .
observe that for @xmath100 this gives @xmath93 .
we will show that for the range of values we care about @xmath101 achieves a maximum for @xmath100 , and can thus be upper - bounded by ( essentially ) @xmath94 , which is the probability that a set is bad in the standard amplifier .
the rest of the proof follows from the argument given in @xcite . in particular
, we can assume that @xmath102 , since @xmath103 edges are cut with probability 1 . to show that the maximum is achieved for @xmath100
, we look at @xmath104 .
we will show that this is less than 1 . using the identity @xmath105
, we get @xmath106 using the fact that @xmath107 , we end up needing to prove the following . @xmath108
combining that without loss of generality @xmath109 holds with the bounds of @xmath83 and @xmath110 we have already mentioned , the inequality @xmath111 is straightforward to establish .
by using the bi - wheel amplifier from the previous section , we are going to prove hardness of approximation for a bounded occurrence csp with very special properties .
this particular csp will be well - suited for constructing a reduction to the tsp given in the next section .
as the starting point of our reduction , we make use of the inapproximability result due to hstad @xcite for the , which is defined as follows : given a system @xmath112 of linear equations mod 2 , in which each equation is of the form @xmath113 with @xmath114 , we want to find an assignment to the variables of @xmath112 such as to maximize the number of satisfied equations .
let @xmath112 be an instance of the and @xmath115 the set of variables , that appear in @xmath112 .
we denote by @xmath116 the number of appearances of @xmath117 in @xmath112 . for every @xmath118 , there exists a constant @xmath119 such that given an instance @xmath112 of the with @xmath120 equations and @xmath121 } d(i)\leq b_{\epsilon}$ ] , it is -hard to decide whether there is an assignment that leaves at most @xmath122 equations unsatisfied , or all assignment leave at least @xmath123 equations unsatisfied .
similarly to the work by berman and karpinski @xcite ( see also @xcite and @xcite ) , we will reduce the number of occurrences of each variable to @xmath30 .
for this , we will use our amplifier construction to create special instances of the hybrid problem , which is defined as follows : given a system @xmath124 of linear equations mod @xmath29 with either three or two variables in each equation , we want to find an assignment such as to maximize the number of satisfied equations .
in particular , we are going to prove the following theorem .
[ thm : hybrid ] for every constant @xmath125 and @xmath126 , there exist instances of the hybrid problem with @xmath127 equations such that : @xmath128 each variable occurs exactly three times .
@xmath130 equations are of the form @xmath131 , @xmath132 equations are of the form @xmath133 and @xmath120 equations are of the form @xmath134 .
@xmath135 it is -hard to decide whether there is an assignment to the variables that leaves at most @xmath136 equations unsatisfied , or every assignment to the variables leaves at least @xmath137 equations unsatisfied .
let @xmath118 be a constant and @xmath112 an instance of the with @xmath121 } d(i ) \leq b_{\epsilon}$ ] .
for a fixed @xmath126 , we can flip some of the literals such that all equations in the instance @xmath112 are of the form @xmath138 , where @xmath139 are variables or negations . by constructing three more copies of each equation , in which all possible pairs of literals appear negated
, we may assume that each variable occurs the same number of times negated as unnegated .
let us fix a variable @xmath117 in @xmath112 .
then , we create @xmath140 new variables @xmath141 .
in addition , we construct a bi - wheel amplifier @xmath142 on @xmath143 vertices ( that is , a bi - wheel with @xmath116 contact vertices ) with the properties described in theorem [ thm : biwheel ] . since @xmath144 is a constant , this can be accomplished in constant time . in the remainder , we refer to _ contact _ and _ checker _ variables as the elements in @xmath145 , whose corresponding index is a contact and checker vertex in @xmath142 , respectively .
we denote by @xmath146 the associated perfect matching on the set of checker vertices of @xmath142 .
in addition , we denote by @xmath147 and @xmath148 the set of edges contained in the first and second cycle of @xmath142 , respectively .
let us now define the equations of the corresponding instance of the hybrid problem .
for each edge @xmath149 , we create the equation @xmath150 and refer to equations of this form as _ matching equations_. on the other hand , for each edge @xmath151 in the cycle @xmath152 with @xmath153 , we introduce the equation @xmath154 . equations of this form will be called _
cycle equations_. finally , we replace the @xmath155-th unnegated appearance of @xmath117 in @xmath112 by the contact variable @xmath156 with @xmath157 , whereas the @xmath155-th negated appearance is replaced by @xmath158 .
the former construction yields @xmath120 equations with three variables in the instance of the hybrid problem , which we will denote by @xmath124 . notice that each variable appears in exactly @xmath30 equations in @xmath124 . clearly , we have @xmath159 equations , thereof @xmath132 matching equations , @xmath130 cycle equations and @xmath120 equations of the form @xmath160 .
a consistent assignment to @xmath145 is an assignment with @xmath161 and @xmath162 for all @xmath163 $ ] , where @xmath126 .
a consistent assignment to the variables of @xmath124 is an assignment that is consistent to @xmath145 for all @xmath164 $ ] . by standard arguments using the amplifier constructed in theorem [ thm : biwheel ] ,
it is possible to convert an assignment to a consistent assignment without decreasing the number of satisfied equations and the proof of theorem [ thm : hybrid ] follows .
this section is devoted to the proof of the following theorem .
[ thm : tspmain ] it is -hard to approximate the tsp to within any constant approximation ratio less than @xmath2 .
let us first sketch the high - level idea of the construction .
starting with an instance of the hybrid problem , we will construct a graph , where gadgets represent the equations .
we will design gadgets for equations of size three ( figure [ fig : eqn3 ] ) and for equations of size two corresponding to _ matching _ edges of the bi - wheel ( figure [ fig : eqn2 ] ) .
we will not construct gadgets for the cycle edges of the bi - wheel ; instead , the connections between the matching edge gadgets will be sufficient to encode these extra constraints .
this may seem counter - intuitive at first , but the idea here is that if the gadgets for the matching edges are used in a consistent way ( that is , the tour enters and exits in the intended way ) then it follows that the tour is using all edges corresponding to one wheel and none from the other .
thus , if we prove consistency for the matching edge gadgets , we implicitly get the cycle edges `` for free '' .
this observation , along with an improved gadget for size - three equations and the elimination of the variable part of the graph , are the main sources of improvement over the construction of @xcite .
we are going to describe the construction that encodes an instance @xmath124 of the hybrid problem into an instance of the tsp problem .
due to theorem [ thm : hybrid ] , we may assume that the equations with three variables in @xmath124 are all of the form @xmath165 . in order to ensure that some edges are to be used at least once in any valid tour ,
we apply the following simple trick that was already used in the work by lampis @xcite : let @xmath166 be an edge with weight @xmath167 that we want to be traversed by every tour .
we remove @xmath166 and replace it with a path of @xmath168 edges and @xmath169 newly created vertices each of degree two , where we think of @xmath168 as a large constant .
each of the @xmath168 edges has weight @xmath170 and any tour that fails to traverse at least two newly created edges is not connected . on the other hand , a tour that traverses
all but one of those edges can be extended by adding two copies of the unused edge increasing the cost of the underlying tour by a negligible value . in summary
, we may assume that our construction contains _ forced _ edges that need to be traversed at least once by any tour . if @xmath171 and @xmath172 are vertices , which are connected by a forced edge @xmath166 , we write @xmath173 or simply @xmath174 . in the following ,
we refer to unforced edges @xmath166 with @xmath175 as _ simple_. all unforced edges in our construction will be simple .
let us start with the description of the corresponding graph @xmath176 : for each bi - wheel @xmath177 , we will construct the subgraph @xmath178 of @xmath176 . for each vertex of the bi - wheel
, we create a vertex in the graph and for each cycle equation @xmath179 , we create a simple edge @xmath180 . given a matching equation between two checkers @xmath181 , we connect the vertices @xmath182 and @xmath183 with two forced edges @xmath184 and @xmath185 .
we have @xmath186 for each @xmath187 .
additionally , we create a central vertex @xmath188 that is connected to gadgets simulating equations with three variables .
let @xmath189 be the @xmath155-th equation with three variables in @xmath124 .
we now create the graph @xmath190 displayed in figure [ fig : eqn3 ] @xmath191 , where the ( contact ) vertices for @xmath192 have already been constructed in the cycles . the edges
@xmath193 with @xmath194 and @xmath195 are all forced edges with @xmath196 .
furthermore , we have @xmath197 for all @xmath194 .
@xmath198 and @xmath199 are both forced edges , whereas all remaining edges of @xmath190 are simple .
this is the whole description of @xmath176 .
c|c & + @xmath191 & @xmath200 given an instance @xmath124 of the hybrid problem and an assignment @xmath201 to the variables in @xmath124 , we are going to construct a tour in @xmath176 according to @xmath201 and give the proof of one direction of the reduction .
in particular , we are going to prove the following lemma .
[ lem : tsp1 ] if there is an assignment to the variables of a given instance @xmath124 of the hybrid problem with @xmath127 equations and @xmath202 bi - wheels , that leaves @xmath110 equations unsatisfied , then , there exists a tour in @xmath176 with cost at most @xmath203 .
before we proceed , let us give a useful definition .
let @xmath42 be an edge - weighted graph and @xmath43 a multi - set of edges of @xmath52 that defines a quasi - tour .
consider a set @xmath204 .
the local edge cost of the set @xmath205 is then defined as @xmath206 in words , for each vertex in @xmath205 , we count half the total weight of its incident edges used in the quasi - tour ( including multiplicities ) . observe that this sum contains half the weight of edges with one endpoint in @xmath205 but the full weight for edges with both endpoints in @xmath205 ( since we count both endpoints in the sum ) . also note that for two sets @xmath207 , we have @xmath208 ( with equality for disjoint sets ) and that @xmath209 . first , note that it is sufficient to prove that we can construct a quasi - tour of the promised cost which uses all forced edges exactly once .
since all unforced edges have cost 1 , if we are given a quasi - tour we can connect two disconnected components by using an unforced edge that connects them twice ( this is always possible since the underlying graph we constructed is connected ) .
this does not increase the cost , since we added two unit - weight edges and decreased the number of components . repeating this results in a connected tour .
let @xmath210 be the associated set of bi - wheels of @xmath124 .
for a fixed bi - wheel @xmath177 , let @xmath211 be its associated set of variables .
due to the construction of instances of the hybrid problem in section [ sec : hybrid ] , we may assume that all equations with two variables are satisfied by the given assignment .
thus , we have @xmath212 , @xmath213 and @xmath214 for all @xmath215 $ ] .
assuming @xmath216 for some @xmath217 , we use once all simple edges @xmath218 with @xmath219 $ ] and the edge @xmath220 .
we also use all forced edges corresponding to matching equations once . in other words , for each biwheel we select the cycle that corresponds to the assignment 1 and use all the simple edges from that cycle .
this creates a component that contains all checker vertices from both cycles and all contacts from one cycle .
c|c & + @xmath191 & @xmath200 as for the next step , we are going to describe the tour traversing @xmath190 with @xmath221 $ ] given an assignment to contact variables .
let us assume that @xmath190 simulates @xmath222 . according to the assignment to @xmath171 , @xmath172 and @xmath223
, we will traverse @xmath190 as follows : in all cases , we will use all forced edges once . + * case ( @xmath224 ) : * then , we use @xmath225 for all @xmath226 and @xmath227 with @xmath228 . for @xmath229 with @xmath230
, we use @xmath231 for all @xmath226 . + * case ( @xmath232 with @xmath126 ) : * in both cases , we traverse @xmath233 for all @xmath234 and @xmath235 . + * case ( @xmath236 ) : * we use @xmath237 with @xmath238 .
furthermore , we include @xmath239 for both @xmath226 .
let us now analyze the cost of the edges of our quasi - tour given an assignment .
for each matching edge @xmath240 consider the set of vertices made up of its endpoints .
its local cost is 5 : we pay 4 for the forced edges and there are two used simple edges with one endpoint in the set .
let us also consider the local cost for a size - three equation gadget , where we consider the set to contain the contact vertices @xmath241 as well the other 8 vertices of the gadget . the local cost here is 9.5 for the forced edges .
we also pay 6 more ( for a total of 15.5 ) when the assignment satisfies the equation or 7 more when it does not .
thus , we have given a covering of the vertices of the graph by @xmath132 sets of size two , @xmath120 sets of size 11 and @xmath242 .
the total edge cost is thus at most @xmath243 . to obtain an upper bound on the cost of the quasi - tour , we observe that the tour has at most @xmath244 components ( one for each bi - wheel and one containing @xmath188 ) . the lemma follows . in this section
, we are going prove the other direction of our reduction . given a tour in @xmath176
, we are going to define an assignment to the variables of the associated instance of the hybrid problem and give the proof of the following lemma .
[ lem : tsp2 ] if there is a tour in @xmath176 with cost @xmath245 , then , there is an assignment to the variables of the corresponding instance of the hybrid problem that leaves at most @xmath110 equations unsatisfied .
again , let us give a useful definition .
consider a quasi - tour @xmath43 and a set @xmath204 .
let @xmath246 be the number of connected components induced by @xmath43 which are fully contained in @xmath205 .
then , the full local cost of the set @xmath205 is defined as @xmath247 . by the definition , the full local cost of @xmath248 is equal to the cost of the quasi - tour ( plus 2 ) .
intuitively , @xmath249 captures the cost of the quasi - tour restricted to @xmath205 : it includes the cost of edges and the cost of added connected components .
note that now for two disjoint sets @xmath207 we have @xmath250 since @xmath251 could contain more connected components than @xmath207 together .
if we know that the total cost of the quasi - tour is small , then @xmath252 is small ( less than @xmath253 ) .
we can use this to infer that the sum of the local full costs of all gadgets is small .
the high - level idea of the proof is the following : we will use roughly the same partition of @xmath248 into sets as in the proof of lemma [ lem : tsp1 ] .
for each set , we will give a lower bound on its full local cost for any quasi - tour , which will be equal to what the tour we constructed in lemma [ lem : tsp1 ] pays .
if a given quasi - tour behaves differently its local cost will be higher .
the difference between the actual local cost and the lower bound is called the credit of that part of the graph .
we construct an assignment for @xmath124 and show that the total sum of credits is higher that the number of unsatisfied equations . but using the reasoning of the previous paragraph , the total sum of credits will be at most @xmath110 .
we are going to prove a slightly stronger statement and show that if there exists a quasi - tour in @xmath176 with cost @xmath245 , then , there exists an assignment leaving at most @xmath110 equations unsatisfied .
recall that the existence of a tour in @xmath176 with cost @xmath254 implies the existence of a quasi - tour in @xmath176 with cost at most @xmath254 .
we may assume that simple edges are contained only once in @xmath43 due to the following preprocessing step : if @xmath43 contains two copies of the same simple edge , we remove them without increasing the cost , since the number of components can only increase by one . in the following ,
given a quasi - tour @xmath43 in @xmath176 , we are going to define an assignment @xmath255 and analyze the number of satisfied equations by @xmath255 compared to the cost of the quasi - tour .
the general idea is that each vertex of @xmath176 that corresponds to a variable of @xmath124 has exactly two forced and exactly two simple edges incident to it .
if the forced edges are used once each , the variable is called honest .
we set it to 1 if the simple edges are both used once and to 0 otherwise .
it is not hard to see that , because simple cycle edges connect vertices that represent the variables , this procedure will satisfy all cycle equations involving honest variables .
we then argue that if other equations are unsatisfied the tour is also paying extra , and the same happens if a variable is dishonest .
let us give more details .
first , we concentrate on the assignment for checker variables .
let us consider the following equations with two variables @xmath256 , @xmath257 , @xmath258 , @xmath259 and @xmath260 .
we are going to analyze the cost of a quasi - tour traversing the gadget displayed in figure [ fig : eqn2 ] @xmath191 and define an assignment according to @xmath43 .
let us first assume that our quasi - tour is honest , that is , the underlying quasi - tour traverses forced edges only once .
* honest tours : * for @xmath261 , we set @xmath262 if the quasi - tour traverses both simple edges incident on @xmath171 and @xmath263 , otherwise .
since we removed all copies of the same simple edge , we may assume that cycle equations are always satisfied .
if the tour uses @xmath264 , we get @xmath265 , @xmath266 and @xmath267 satisfied equations . given @xmath268
, we obtain @xmath267 satisfied equations as well .
let us define @xmath269 .
notice that in both cases , we have local cost @xmath270 .
we claim that @xmath271 for a valid quasi - tour . in order to obtain a valid quasi - tour ,
we need to traverse both forced edges in @xmath272 and use at least two simple edges , as otherwise , it implies @xmath273 .
given a quasi - tour @xmath43 , we introduce a local credit function defined by @xmath274 . if @xmath275 forms a connected component , we get @xmath276 satisfied equations and @xmath277 , which is sufficient to pay for the unsatisfied equation @xmath278 . on the other hand , assuming @xmath265 and @xmath279 , we get @xmath280 and @xmath281 unsatisfied equation . +
* dishonest tours : * we are going to analyze quasi - tours , which are using one of the forced edges twice . by setting @xmath282
, we are able to find an assignment that always satisfies @xmath278 and two other equations out of the five that involve these dishonest variables .
the local cost in this case is at least @xmath283 .
hence , the credit @xmath284 is sufficient to pay for the two unsatisfied equations .
again , we will distinguish between honest tours ( which use forced edges exactly once ) and dishonest tours .
this time we are interested in seven equations : the size - three equation @xmath285 and the six cycle equations containing the three contacts .
observe that the local cost of @xmath286 is at least 15.5 .
the local edge cost of any quasi - tour is 9.5 for the forced edges . for each component
@xmath287 with @xmath288 , we need to pay at least 2 more because there are two vertices with odd degree ( @xmath289 ) and we also need to connect the component to the rest of the graph ( otherwise the component already costs 2 more ) .
let us define the credit of @xmath290 with respect to @xmath43 by @xmath291 .
+ + * honest tours : * for each @xmath234 , we set @xmath228 if the tour uses both simple edges incident on @xmath292 and @xmath293 , otherwise . notice that in the case @xmath294 with @xmath295 , this satisfies all seven equations and the tour has local cost at least @xmath296 .
case @xmath297 the assignment now failed to satisfy the size - three equation , so we need to prove that the quasi - tour has local cost at least 16.5 .
since all vertices are balanced with respect to @xmath43 , the quasi - tour has to use at least one edge incident on @xmath298 and @xmath299 besides @xmath300 and @xmath301 .
if the quasi - tour takes @xmath302 for a @xmath303 and all @xmath304 , since all simple edges incident on @xmath192 are used , we get at total cost of at least 16.5 , which gives a credit of 1 . case @xmath305 without loss of generality , we assume that @xmath306 holds . again , only the size - three equation is unsatisfied , so we must show that the local cost is at least @xmath307 .
we will discuss two subcases .
@xmath128 there is a connected component @xmath308 for some @xmath309 .
we obtain that @xmath310 and therefore , a lower bound on the total cost of @xmath307 .
@xmath129 since we may assume that @xmath311 , @xmath312 , @xmath313 and @xmath314 are balanced with respect to @xmath43 , we have that @xmath315 for all @xmath316 and @xmath317 . because @xmath318 are also balanced , we obtain @xmath319 for all @xmath316 , which implies a total cost of @xmath307
. + * dishonest tours : * let us assume that the quasi - tour uses both of the forced edges @xmath320 and @xmath321 for some @xmath322 twice . we delete both copies and add @xmath323 instead which reduces the cost of the quasi - tour .
hence , we may assume that only one of the two incident forced edges is used twice .
first , observe that if all forced edges were used once , then there would be eight vertices in the gadget with odd degree : @xmath324 .
if exactly one forced edge is used twice , then seven of these vertices have odd degree .
thus , it is impossible for the tour to make the degrees of all seven even using only the simple edges that connect them .
we can therefore assume that if a forced edge is used twice , there exists another forced edge used twice .
we will now take cases , depending on how many of the vertices @xmath192 are incident on forced edges used twice .
note that if one of the forced edges incident on @xmath171 is used twice , then exactly one of the simple edges incident on @xmath171 is used once .
so , first suppose all three of @xmath192 have forced edges used twice .
the local cost from forced edges is at least @xmath325 .
furthermore , there are three vertices of the form @xmath326 , for @xmath288 and @xmath327 with odd degree .
these have no simple edges connecting them , thus the quasi - tour will use three simple edges to balance their degrees .
finally , the used simple edges incident on @xmath192 each contribute @xmath328 to the local cost .
thus , the total local cost is at least @xmath329 , giving us a credit of @xmath30 .
it is not hard to see that there is always an assignment satisfying four out of the seven affected equations , so this case is done .
second , suppose exactly two of @xmath192 have incident forced edges used twice , say , @xmath330 . for @xmath223 ,
we select the honest assignment ( 1 if the incident simple edges are used , 0 otherwise ) and this satisfies the cycle equations for this variable . we can select assignments for @xmath330 that satisfy three of the remaining five equations , so we need to show that the cost in this case is at least @xmath331 .
the cost of forced edges is at least @xmath332 , and the cost of simple edges incident on @xmath330 adds @xmath281 to the local cost .
one of the vertices @xmath333 and one of @xmath334 have odd degree , therefore the cost uses two simple edges to balance them .
finally , the vertices @xmath335 have odd degree .
if two simple edges incident to them are used , we have a total local cost of 17.5 .
if the edge connecting them is used , then the two simple edges incident on @xmath223 must be used , again pushing the local cost to 17.5 .
finally , suppose only @xmath171 has an incident forced edge used twice . by the parity argument given above , this means that one of the forced edges incident on @xmath188 is used twice .
we can satisfy the cycle equations for @xmath336 by giving them their honest assignment , and out of the three remaining equations some assignment to @xmath171 satisfies two .
therefore , we need to show that the cost is at least 16.5 .
the local cost from forced edges is @xmath337 and the simple edge incident on @xmath171 contributes @xmath328 .
also , at least one simple edge incident on @xmath312 or @xmath311 is used , since one of them has odd degree . for @xmath334 ,
either two simple edges are used , or if the edge connecting them is used the simple edges incident on @xmath172 contribute @xmath281 more . with similar reasoning for @xmath335 , we get that the total local cost is at least @xmath338 .
+ let us now conclude our analysis .
consider the following partition of @xmath339 : we have a singleton set @xmath242 , @xmath132 sets of size @xmath29 containing the matching edge gadgets and @xmath120 sets of size @xmath340 containing the gadgets for size - three equations ( except @xmath188 ) .
the sum of their local costs is at most @xmath341 .
but the sum of their local costs is ( using the preceding analysis ) equal to @xmath342 .
thus , the sum of all credits is at most @xmath110 .
since we have already argued that the sum of all credits is enough to cover all equations unsatisfied by our assignment , this concludes the proof .
we are ready to give the proof of theorem [ thm : tspmain ] .
we are given an instance @xmath112 of the with @xmath202 variables and @xmath120 equations .
for all @xmath343 , there exists a @xmath110 such that if we repeat each equation @xmath110 time we get an instance @xmath344 with @xmath345 equations and @xmath202 variables such that @xmath346 .
then , from @xmath344 , we generate an instance @xmath124 of the hybrid problem and the corresponding graph @xmath176 . due to lemmata [ lem : tsp1 ] ,
[ lem : tsp2 ] and theorem [ thm : hybrid ] , we know that for all @xmath347 , it is -hard to tell whether there is a tour with cost at most @xmath348 or all tours have cost at least @xmath349 .
the ratio between these two cases can get arbitrarily close to @xmath2 by appropriate choices for @xmath350 .
in this section , we prove the following theorem . [ thm : atspmain ] it is -hard to approximate the atsp to within any constant approximation ratio less than @xmath3 .
let us describe the construction that encodes an instance @xmath124 of the hybrid problem into an instance of the atsp .
again , it will be useful to have the ability to force some edges to be used , that is , we would like to have bidirected forced edges .
a bidirected forced edge of weight @xmath167 between two vertices @xmath171 and @xmath172 will be created in a similar way as undirected forced edges in the previous section : construct @xmath169 new vertices and connect @xmath171 to @xmath172 through these new vertices , making a bidirected path with all edges having weight @xmath170 .
it is not hard to see that without loss of generality we may assume that all edges of the path are used in at least one direction , though we should note that the direction is not prescribed . in the remainder
, we denote a directed forced edge consisting of vertices @xmath171 and @xmath172 by @xmath351 , or @xmath352 . let @xmath124 consist of the collection @xmath353 of bi - wheels .
recall that the bi - wheel consists of two cycles and a perfect matching between their checkers .
let @xmath354 be the associated set of variables of @xmath177 .
we write @xmath355 to denote the function which , given the index of a checker variable @xmath182 returns the index @xmath155 of the checker variable @xmath183 to which it is matched ( that is , the function @xmath356 is a permutation function encoding the matching ) .
we write @xmath357 to denote the inverse function @xmath358 . now , for each bi - wheel @xmath177 , we are going to construct the corresponding directed graph @xmath359 as follows .
first , construct a vertex for each checker variable of the wheel . for each matching equation @xmath360 , we create a bidirected forced edge @xmath361 with @xmath362 . for each contact variable @xmath363 , we create two corresponding vertices @xmath364 and @xmath365 , which are joined by the bidirected forced edge @xmath366 with @xmath367 .
next , we will construct two directed cycles @xmath368 and @xmath369 .
note that we are doing arithmetic on the cycle indices here , so the index @xmath370 should be read as equal to @xmath281 . for @xmath368 , for any two consecutive checker vertices @xmath371 on the un - negated side of the bi - wheel , we add a simple directed edge @xmath372 . if the checker @xmath182 is followed by a contact @xmath373 in the cycle , then we add two simple directed edges @xmath374 and @xmath375 .
observe that by traversing the simple edges we have just added , the forced matching edges in the direction @xmath376 and the forced contact edges for the un - negated part in the direction @xmath377 we obtain a cycle that covers all checkers and all the contacts of the un - negated part .
we now add simple edges to create a second cycle @xmath369 .
this cycle will require using the forced matching edges in the opposite direction and , thus , truth assignments will be encoded by the direction of traversal of these edges .
first , for any two consecutive checker vertices @xmath378 on the un - negated side of the bi - wheel , we add the simple directed edge @xmath379 .
then , if the checker @xmath380 is followed by a contact @xmath381 in the cycle then we add the simple directed edges @xmath382 and @xmath383 .
now by traversing the edges we have just added , the forced matching edges in the direction @xmath384 and the forced contact edges for the negated part in the direction @xmath385 , we obtain a cycle that covers all checkers and all the contacts of the negated part , that is , a cycle of direction opposite to @xmath368 .
what is left is to encode the equations of size three .
again , we have a central vertex @xmath188 that is connected to gadgets simulating equations with three variables . for every equation with three variables ,
we create the gadget displayed in figure [ fig : eqn3 ] @xmath200 , which is a variant of the gadget used by papadimitriou and vempala @xcite .
let us assume that the @xmath155-th equation with three variables in @xmath386 is of the form @xmath387 .
this equation is simulated by @xmath388 .
the vertices used are the contact vertices @xmath389 , which we have already introduced , as well as the vertices @xmath390 \}$ ] . for notational simplicity , we define @xmath391 , \gamma\in \{x , y , z\ } , { \alpha}\in \{r , l\ } \big\}$ ] .
all directed non - forced edges are simple .
the vertices @xmath392 and @xmath393 are connected to @xmath188 by forced edges with @xmath394 , where @xmath395 is a small fixed constant . to simplify things
, we also force them to be used in the displayed direction by deleting the edges that make up the path of the opposite direction .
this is the whole description of the graph @xmath396 .
we are going to construct a tour in @xmath396 given an assignment to the variables of @xmath124 and prove the following lemma .
[ lem : atsp1 ] given an instance @xmath124 of the hybrid problem with @xmath202 bi - wheels and an assignment that leaves @xmath110 equations in @xmath124 unsatisfied , then , there exists a tour in @xmath396 with cost at most @xmath397 .
before we proceed , let us again give a definition for a local edge cost function .
let @xmath42 be an edge - weighted digraph and @xmath43 a multi - set of edges of @xmath52 that defines a tour .
consider a set @xmath204 .
the local edge cost of the set @xmath205 is then defined as @xmath398 in words , for each vertex in @xmath205 we count the total weight of its outgoing edges used in the quasi - tour ( including multiplicities ) .
thus , that this sum contains the full weight for edges with their source in @xmath205 , regardless of where their other endpoint is .
also note that again for two sets @xmath207 we have @xmath208 ( with equality for disjoint sets ) and that @xmath209 .
let @xmath177 be a bi - wheel with variables @xmath399 .
given an assignment to the variables of @xmath124 , due to theorem [ thm : hybrid ] , we may assume that either @xmath400 for all @xmath215 $ ] or @xmath401 for all @xmath215 $ ] .
we traverse the cycle @xmath368 if @xmath402 and the cycle @xmath369 otherwise .
this creates @xmath202 strongly connected components .
each contains all the checkers of a bi - wheel and the contacts from one side . for each matching edge gadget ,
the local edge cost is @xmath30 .
we pay two for the forced edge and @xmath281 for the outgoing simple edge .
we will account for the cost of edges incident on contacts when we analyze the size - three equation gadget below .
let us describe the part of the tour traversing the graph @xmath388 , which simulates @xmath403 . recall that if @xmath171 is set to true in the assignment we have traversed the bi - wheel gadgets in such a way that the forced edge @xmath404 is used , and the simple edge coming out of @xmath312 is used . according to the assignment to @xmath171 , @xmath172 and @xmath223 , we traverse @xmath388 as follows : * case @xmath405 : *
let us assume that @xmath406 holds .
then , we use @xmath407 .
the cost is @xmath408 for the forced edges , @xmath409 for the simple edges inside the gadget , plus @xmath281 for the simple edge going out of @xmath312 .
total local edge cost cost : @xmath410 .
* case @xmath411 : * then , we use @xmath412 .
again we pay @xmath408 for the forced edges , @xmath276 for the simple edges inside the gadget and @xmath30 for the outgoing edges incident on @xmath413 .
total local edge cost : @xmath414 .
* case @xmath415 : * let us assume that @xmath416 holds .
then , we use @xmath417 with total local edge cost @xmath418 . * case @xmath419 : * we use @xmath420 with @xmath418 .
the total edge cost of the quasi - tour we constructed is @xmath421 .
we have at most @xmath422 strongly connected components : one for each bi - wheel and one containing @xmath188 . a component representing a bi - wheel
can be connected to @xmath188 as follows : let @xmath333 be two contact vertices in the component .
add one copy of each edge from the cycle @xmath423 .
this increases the cost by @xmath424 but decreases the number of components by one . in this section
, we are going to prove the other direction of the reduction .
[ lem : atsp2 ] if there is a tour with cost @xmath425 , then , there is an assignment that leaves at most @xmath110 equations unsatisfied .
given a tour @xmath43 in @xmath396 , we are going to define an assignment to checker and contact variables . as in lemma
[ lem : tsp2 ] , we will show that any tour must locally spend on each gadget at least the same amount as the tour we constructed in lemma [ lem : atsp1 ] .
if the tour spends more , we use that credit to satisfy possible unsatisfied equations .
let us consider the following equations with two variables @xmath426 , @xmath427 , @xmath428 , @xmath429 , @xmath430 and the corresponding situation displayed in figure [ fig : eqn2 ] @xmath200 . since @xmath43 is a valid tour in @xmath396 , we know that @xmath431 is traversed and due to the degree condition , for each @xmath261 , the tour uses another incident edge @xmath166 on @xmath171 with @xmath432 . therefore , we have that @xmath433 .
the credit assigned to a gadget is defined as @xmath434 .
let us define the assignment for @xmath182 and @xmath183 .
a variable @xmath182 is honestly traversed if either both the simple edge going into @xmath182 is used and the simple edge coming out of @xmath183 is used , or neither of these two edges is used . in the first case , we set @xmath182 to @xmath281 , otherwise to @xmath293 .
similarly , @xmath183 is honest if both the edge going into @xmath183 and the edge out of @xmath182 are used , and we set it to @xmath281 in the first case and @xmath293 otherwise .
* honest tours : * first , suppose that both @xmath182 and @xmath183 are honest .
we need to show that the credit is at least as high as the number of unsatisfied equations out of the five equations that contain them .
it is not hard to see that if we have set @xmath282 all equations are satisfied .
if we have set both to @xmath281 , then the forced edge must be used twice , making the local edge cost at least @xmath409 , giving a credit of @xmath30 , which is more than sufficient . * dishonest tours : * if both @xmath182 and @xmath183 are dishonest the tour must be using the forced edge in both directions .
thus , the local cost is @xmath267 or more , giving a credit of @xmath29 .
there is always an assignment that satisfies three out of the five equations , so this case is done .
if one of them is dishonest , the other must be set to @xmath281 to ensure strong connectivity .
thus , there are two simple edges used leaving the gadget , making the local cost @xmath276 ( perhaps the same edge is used twice ) .
we can set the honest variable to @xmath281 ( satisfying its two cycle equations ) , and the other to @xmath293 , leaving at most one equation unsatisfied .
first , we note that for any valid tour , we have @xmath435 .
this is because the two forced edges of weight @xmath436 must be used , and there exist @xmath437 vertices in the gadget for which all outgoing edges have weight @xmath281 .
let us define the credit @xmath438 .
* honest traversals : * we assume that the underlying tour is honest , that is , forced edges are traversed only in one direction .
we set @xmath171 to @xmath281 if the forced edge is used in the direction @xmath404 and @xmath293 otherwise . in the first case
we know that the simple edges going into @xmath311 and out of @xmath312 are used . in the second , the edges @xmath439 and @xmath440 are used .
we do similarly for @xmath336 .
we are interested in the equation @xmath441 and the six cycle equations involving @xmath192 .
the assignment we pick for honest variables satisfies the cycle equations , so if it also satisfies the size - three equation we are done .
if not , we have to prove that the tour pays at least @xmath442 .
* case @xmath443 : * due to our assumption , we know that @xmath444 is a part of the tour . since @xmath43 is a tour , there exists a vertex in @xmath445 that is visited twice and we get @xmath446 .
thus , we can spend the credit @xmath447 on the unsatisfied equation @xmath441 .
* case @xmath415 : * without loss of generality , let us assume that @xmath416 holds .
then , we know that @xmath448 is a part of the tour .
but , this implies that there is a vertex in @xmath449 that is visited twice .
hence , we have that @xmath450
. * dishonest traversals : * consider the situation , in which some forced edges @xmath451 are traversed in both directions for some variables @xmath227 . for the honest variables
, we set them to the appropriate value as before , and this satisfies their cycle equations .
observe now that if a forced edge @xmath452 is also used in the opposite direction , then there must be another edge used to leave the set @xmath225 .
thus the local edge cost of this set is at least @xmath30 .
it follows that the credit we have for the gadget is at least as large as the number of dishonest variables .
we can give appropriate values to them so each satisfies one cycle equation and the size - three equation is satisfied .
thus , the number of unsatisfied equations is not larger than our credit .
in summary , for every tour @xmath43 in @xmath396 , we can find an assignment to the variables of @xmath124 such that all unsatisfied equations are paid by the credit induced by @xmath43 .
we are ready to give the proof of theorem [ thm : atspmain ] .
we are again given an instance @xmath112 of the with @xmath202 variables and @xmath120 equations .
for all @xmath343 , there exists a @xmath110 such that if we repeat each equation @xmath110 time we get an instance @xmath344 with @xmath345 equations and @xmath202 variables such that @xmath453 .
then , from @xmath344 , we generate an instance @xmath124 of the hybrid problem and the corresponding directed graph @xmath396 . due to lemmata [ lem : atsp1 ] ,
[ lem : atsp2 ] and theorem [ thm : hybrid ] , we know that for all @xmath347 , it is -hard to tell whether there is a tour with cost at most @xmath454 or all tours have cost at least @xmath455 , for some @xmath456 depending only on @xmath457 .
the ratio between these two cases can get arbitrarily close to @xmath3 by appropriate choices for @xmath458 .
in this paper , we proved that it is hard to approximate the atsp and the tsp within any constant factor less than @xmath3 and @xmath2 , respectively .
the proof method required essentially new ideas and constructions from the ones used before in that context . since the best known upper bound on the approximability is @xmath459 for atsp and @xmath4 for tsp , there is certainly room for improvements .
especially , in the asymmetric version of the tsp , there is a large gap between the approximation lower and upper bound , and it remains a major open problem on the existence of an efficient constant factor approximation algorithm for that problem .
furthermore , it would be nice to investigate if some of the ideas of this paper , and in particular the bi - wheel amplifiers , can be used to offer improved hardness results for other optimization problems , such as the steiner tree problem .
c. papadimitriou and s. vempala , _ on the approximability of the traveling salesman problem _ , in proc .
32nd acm stoc ( 2000 ) , pp .
126133 , 2000 ; see also a corrected version in combinatorica 26 , pp . 101120 , 2006 . | in this paper , we study the approximability of the metric traveling salesman problem ( tsp ) and prove new explicit inapproximability bounds for that problem .
the best up to now known hardness of approximation bounds were @xmath0 for the symmetric case ( due to lampis ) and @xmath1 for the asymmetric case ( due to papadimitriou and vempala ) .
we construct here two new bounded occurrence csp reductions which improve these bounds to @xmath2 and @xmath3 , respectively .
the latter bound is the first improvement in more than a decade for the case of the asymmetric tsp .
one of our main tools , which may be of independent interest , is a new construction of a bounded degree wheel amplifier used in the proof of our results . |
polyelectrolytes ( pe s ) are polymer chains that carry ionizable groups along their backbones . upon solution in aqueous solvents ,
these groups dissociate , leaving behind a charged chain and the corresponding counterions in the solution .
chemically anchoring @xmath0 such pe - chains on a colloidal particle of radius @xmath1 gives rise to a spherical polyelectrolyte brush ( spb ) . in the limit in which the height @xmath2 of the brush greatly exceeds @xmath1 , one talks instead about polyelectrolyte stars . a great deal of theoretical@xcite and
experimental@xcite effort has been devoted in the recent past with the goal of understanding the conformations and the interactions of spb s and pe - stars .
the reasons are many - fold .
the grafted pe - chains can provide an electrosteric barrier against flocculation of the colloidal particles on which the chains are grafted , rendering the systems very interesting from the point of view of colloidal stabilization .
moreover , they can act as control agents for gelation , lubrication and flow behavior . from the point of view of fundamental research , pe - stars are a novel system of colloidal particles that combine aspects from two different parts of colloid science : polymer theory and the theory of charged suspensions . accordingly , the derivation of the effective interaction between pe - stars has led to their description as ultrasoft colloids and to theoretical predictions on their structural and phase behavior@xcite that have received already partial experimental confirmation.@xcite moreover , certain similarities in the structural and phase behavior of pe - stars and ionic microgels have been established , demonstrating the close relationship between the two systems.@xcite by and large , the theoretical investigations involving pe - stars have been limited to the study of either single pe - stars or bulk solutions of the same . however , a new field of promising applications and intriguing physics is arising when pe - stars or spb s are mixed with hard colloids or brought in contact with planar walls .
indeed , pe - stars can be used to model cell adhesion and are also efficient drug - delivery and protein - encapsulation and immobilization agents.@xcite on the other hand , hydrogels , which are physically similar to pe - stars , adsorbed on planar walls , form arrays of dynamically tunable , photoswitchable or bioresponsive microlenses.@xcite the encapsulation properties as well as the characteristics of the microlenses depend sensitively on the interactions between the pe - stars and the colloidal particles or the wall , respectively .
therefore , there exists a need to undertake a systematic effort in trying to understand these interactions physically and make quantitative predictions about the ways to influence them externally . in this paper , we take a first step in this direction by considering a pe - star in the neighborhood of a planar , uncharged and purely repulsive wall .
we analyze the mechanisms that give rise to an effective star - wall repulsion and identify the counterion entropy and the chain compression against the planar wall as the major factors contributing to this force .
our findings are corroborated by comparisons with molecular dynamics simulation results for a wide range of parameters that structurally characterize the stars .
this work provides the foundation for examining effects of wall curvature and charge .
our paper is organized as follows : sec .
[ sec : md ] is devoted to the description of the system we are investigating , the simulation model , and the simulation techniques used .
moreover , we specify the physical quantities of interest . our theoretical approach is addressed in detail in sec . [
sec : theory ] . in sec .
[ sec : comparison ] , we quantitatively compare and discuss the respective results . finally , we conclude and present an outlook to possible future work in sec.[sec : conclusion ] .
we start with a definition of the system under consideration including its relevant parameters and a description of the simulation model used .
we study a dilute , salt - free solution of pe - stars confined between two hard walls parallel to the @xmath3-@xmath4-plane at positions @xmath5 , resulting in an overall wall - to - wall separation @xmath6 .
we assume a good solvent that is only implicitly taken into account via its relative dielectric permittivity @xmath7 , i.e. , we are dealing with an aqueous solution . to avoid the appearance of image charges,@xcite we assume the dielectric constants to be the same on both sides of the respective confining walls .
it turns out , however , that this is not a severe assumption , because the effective charge of a pe - star is drastically reduced compared to its bare value , due to the strong absorption of neutralizing counterions.@xcite hence , the influence of image charges can be expected to be of minor importance .
the pe - stars themselves consist of @xmath0 pe - chains , all attached to a common colloidal core of radius @xmath1 , whose size equals the monomer size and is therefore much smaller than the typical center - to - end length of the chains for all parameter combinations .
the introduction of such a core particle is necessary to place the arms in the vicinity of the center , where the monomer density can take very high values .
the theoretical approach pertains to the limit of vanishingly small core size . in order to remove effects arising from the small ( but finite ) value of the core in the simulation model and to provide a comparison with theory
, we will henceforth employ a consistent small shift of the simulation data by @xmath1 .
this standard approach has been applied already , e.g. , to the measurement of the interactions between neutral @xcite and charged stars @xcite , in order to isolate the direct core - core interaction effects from the scaling laws valid at small interstar separations .
the pe - chains are modeled as bead - spring chains of @xmath8 lennard - jones ( lj ) particles .
this approach was first used in investigations of neutral polymer chains and stars @xcite and turned out to be reasonable .
the method was also already successfully applied in the case of polymer - colloid mixtures@xcite or polyelectrolyte systems.@xcite to mimic the above mentioned good solvent conditions , a shifted and truncated lj potential is introduced to depict the purely repulsive excluded volume interaction between the monomers : @xmath9&r\leq 2^{1/6}\sigma_{\rm lj}\\ 0&r>2^{1/6}\sigma_{\rm lj}. \end{cases}\ ] ] here , @xmath10 is the spatial distance of two interacting particles .
the quantities @xmath11 and @xmath12 set the basic length and energy scales for the system , respectively . in
what follows , we fix the system s temperature to the value @xmath13 , where @xmath14 denotes boltzmann s constant .
the pe - chain connectivity is modeled by employing a standard finite extension nonlinear elastic ( fene ) potential:@xcite @xmath15&r\leq r_0\\ \infty&r > r_0 , \end{cases}\ ] ] with a spring constant @xmath16 .
the divergence length @xmath17 limits the maximum relative displacement of two neighboring monomers and is set to @xmath18 in the scope of this work .
the described set of parameters determines the equilibrium bond length , in our case resulting in a value @xmath19 . when modeling the interactions between the monomers and a star s colloidal core , the finite radius @xmath1 of the latter has to be taken into account .
all monomers experience a repulsive interaction with the central particle , in analogy to eq .
( [ eq : pot : lj ] ) reading as @xmath20 in addition , there is an attraction between the innermost monomers in the arms chain sequence and the core which is of fene type and can be written as ( cf .
( [ eq : pot : fene ] ) ) : @xmath21 the chains are charged in a periodic fashion by a fraction @xmath22 in such a way that every @xmath23-th monomer carries a monovalent charge .
consequently , there is a total number of @xmath24 monomer ions per pe - star . to ensure electroneutrality of the system as a whole
, we include the same amount of oppositely charged , monovalent counterions in our considerations .
since the latter are able to freely move , they have to be simulated explicitly .
furthermore , they are of particular importance because they are expected to crucially affect the physics of the system .
one example is the aforementioned fact that they induce a reduction of the stars bare total charges .
two charged beads with spatial distance @xmath25 interact by a full coulomb potential , i.e. , the electrostatic interaction energy is @xmath26 thereby , @xmath27 are the valencies of monomer ions and counterions , respectively , @xmath28 denotes the absolute value of the unit charge , and @xmath29 is the inverse temperature . in the above equation ,
the so - called bjerrum length @xmath30 was introduced .
it is defined as the distance at which the electrostatic energy equals the thermal energy .
thus , it characterizes the interaction strength of the coulomb coupling . in the case of water at room temperature ,
one obtains @xmath31 . in our simulations
, we fix the bjerrum length to @xmath32 , thus corresponding to an experimental particle diameter @xmath33 .
this is a realistic value for typical polyelectrolytes.@xcite for the purpose of completing the set of interaction potentials needed to describe the system at hand , we have to define particle - wall interactions . for technical reasons ( see below ) , we do not regard the walls as true hard walls , as common in monte - carlo ( mc ) simulation studies.@xcite following the course of our above modeling and based on eq .
( [ eq : pot : lj ] ) , we assume them to be of truncated - and - shifted - lj type instead , leading to the following monomer - wall interaction : @xmath34 whereas @xmath35 refers to the @xmath35-component of the position vector of the particular bead .
likewise , the potential function for a star s core interacting with the confining walls is yielded by combining eqs .
( [ eq : pot : c ] ) and ( [ eq : pot : w ] ) : @xmath36 to finalize this section , we present a short summary of the simulation techniques used .
we perform monomer - resolved molecular dynamics ( md ) simulations in the canonical ensemble , employing a rectangular simulation box of total volume @xmath37 that contains a single pe - star .
we apply periodic boundary conditions in the @xmath3- and @xmath4-directions , while the box is confined with respect to the @xmath35-direction . here , we always fix @xmath38 .
this provides a sufficiently large simulation box to suppress any undesirable side - effects _ a priori _ and emulates a dilute solution of pe - stars ( cf .
above description of the physical problem under investigation ) .
for the numerical integration of the equations of motion , we adopt a so - called verlet algorithm in its velocity form.@xcite in order to stabilize the system s temperature , we make use of a langevin thermostat@xcite that introduces additional friction and random forces with appropriately balanced , temperature - dependent amplitudes . due to the periodic boundary conditions and the long - range character of the coulombic forces , a straight - forward calculation of the latter pursuant to eq .
( [ eq : pot : coul ] ) is not feasible .
therefore , we have to evaluate the forces using lekner s well - established summation method@xcite in its version for quasi two - dimensional geometries .
thereby , the convergence properties of the sums occurring during the computation are enhanced by a mathematically accurate rewriting , allowing a proper cut - off .
for performance reasons , the forces have to be tabulated .
the typical time step is @xmath39 , with @xmath40 being the associated time unit and @xmath41 the monomer mass . in our case ,
the counterions are taken to have the same mass and size as the ( charged ) monomers . in our simulations
, we measure the effective core - wall forces as a function of the center - to - wall distance @xmath42 .
furthermore , we examine various other static quantities and their @xmath42-dependence , namely the center - to - end distance @xmath2 , the radius of gyration @xmath43 , the number of condensed counterions and the density profiles of all particle species involved . for this purpose , for every particular value of @xmath42 the system is equilibrated for about @xmath44 time steps . after this equilibration period ,
we perform production runs lasting between @xmath45 time steps .
we carry out the above described measurements for a variety of arm numbers @xmath0 ( @xmath4610 , 18 , 30 ) and charging fractions @xmath22 ( @xmath471/5 , 1/4 , 1/3 ) in order to make systematic predictions for the @xmath0- and @xmath22-dependencies of all theoretical parameters . in all cases ,
we fix the degree of polymerization , i.e. , the number @xmath8 of monomers per arm , to a value @xmath48 .
a pe - star in the neighborhood of a planar , impenetrable wall , undergoes conformational changes that modify its ( free ) energy in comparison to the value it has when the wall is absent or far away from the star .
this separation - dependent difference between the free energies is precisely the effective interaction between the pe - star and the wall .
there are three distinct mechanisms that give rise to an effective interaction in our case : the change in the electrostatic energy of the star , the change in the counterion entropy arising from the presence of the geometric confinement as well as contributions from compression of the stiff pe - chains against the flat wall .
the former two are intricately related to each other , as the counterion distribution is dictated by the strong coulomb interactions , hence they will be examined jointly , the latter is a distinct phenomenon arising in the presence of impenetrable walls . to obtain theoretical predictions for the electrostatic - entropic contribution to the free energy for a pe - star close to a flat , hard wall , we employ a mean - field approach , inspired by and akin to that developed in refs . , and .
let @xmath49 be the density of the solution of pe - stars , where @xmath50 denotes the total number of pe - stars in the macroscopic volume @xmath51 .
in general , a single pe - star of total charge @xmath52 is envisioned as a spherical object of radius @xmath2 , embedded in a spherical wigner - seitz cell of radius @xmath53 .
additionally , the latter contains @xmath54 counterions , restricted to move in the wigner - seitz cell only and forming an oppositely charged background of total charge @xmath55 .
the cell s radius is connected to the star density via @xmath56 . in the case of center - to - wall separations
smaller than @xmath53 and/or @xmath2 , we simply cut the counterion cloud and/or the star itself at the confining wall and treat them as chopped spheres instead of full spheres .
obviously , the compound system composed of a pe - star and its counterions is electroneutral as a whole . since we are interested in the case of dilute pe - solutions only , we can limit ourselves to the consideration of a single pe - star .
clearly , for center - to - wall distances @xmath57 there is no interaction between a pe - star and the wall within the framework of our theory .
accordingly , we will deal with the case @xmath58 only .
[ fig : sketch ] sketches the physical situation and visualizes the decisive length scales of the problem .
a mechanism crucially influencing the physics of pe - systems is the so - called manning condensation of counterions.@xcite a parameter indicating whether or not such a condensation effect will occur is the dimensionless fraction @xmath59 .
if it exceeds unity , counterions will condense on the arms of the stars.@xcite in our case , this condition is true for all parameter combinations examined ( see below ) .
hence , the condensation effect has to be taken into account in our theoretical modeling .
for that reason , we partition the counterions in three different states , an ansatz already put forward in refs . , and : @xmath60 of the @xmath54 counterion are in the condensed state , i.e. , they are confined in imaginary hollow tubes of outer radius @xmath61 and inner radius @xmath11 around the arms of the star . the total volume with respect to the value given in the main text
is neglected .
anyway , this simplification does not considerably affect the theoretical results . ]
accessible for the counterions in this state is @xmath62 .
a number of @xmath63 counterions is considered to be trapped in the interior of the star .
these ions are allowed to explore the overall volume @xmath64 of the ( possibly chopped ) star except the tubes introduced above , i.e. , @xmath65 .
thereby , the volume @xmath66 of a chopped sphere can be derived by straightforward calculations . with @xmath67 ,
one obtains : @xmath68 note that the total volume of the star is big compared to the volume of the tubes surrounding the arms , @xmath69 .
thus , we will make use of the approximation @xmath70 in all steps to come .
the remaining @xmath71 counterions are able to freely move within the outer shell of volume @xmath72 surrounding the star , where @xmath73 is the total volume of the wigner - seitz cell .
the subdivision of counterions in the three different states is also depicted in fig .
[ fig : sketch ] , showing a star with 5 arms that are assumed to be fully stretched for demonstration .
the electrostatic and entropic part @xmath74 of the effective interaction @xmath75 between the pe - star and the wall , kept at center - to - wall distance @xmath42 , results after taking a canonical trace over all but the star center degree of freedom . with @xmath76 being the variational helmholtz free energy for a system where a single star faces the confining wall , and which contains just the electrostatic and entropic parts
, it is defined as:@xcite @xmath77 in principle , the equilibrium values of @xmath2 and @xmath78 are determined by the above minimization .
we will discuss this point in more detail shortly .
note that we neglected a second , @xmath42-independent term on the right - hand side of eq .
( [ eq : potential ] ) representing the contribution to the free energy for an infinitely large center - to - wall separation .
since it makes up a constant energy shift only , it does not influence the effective forces between the star and the wall we are mainly interested in .
the electrostatic - entropic force contributions are obtained by differentiating with respect to @xmath42 : @xmath79 now , we derive concrete expressions for the terms of which @xmath76 is built up , whereby we want to include both the counterions entropic contributions and the electrostatic energy . to keep our theory as simple as possible
, we will omit other thinkable contributions , like elastic energies@xcite of the chains or flory - type terms arising through self - avoidance.@xcite . as we will see in sec .
[ sec : comparison ] , our theoretical results in combination with chain - compression terms to be deduced in what follows , are capable of producing very good agreement between the theory and the corresponding simulational data ; these simplifications are therefore _ a posteriori _ justified .
consequently , the free energy @xmath76 reads as : @xmath80 the electrostatic mean - field energy @xmath81 is assumed to be given by a so - called hartree - type expression .
let @xmath82 and @xmath83 denote the number densities of the monomers and the three different counterion species , respectively , measured with respect to the star s geometrical center .
these density profiles then determine the overall local charge density @xmath84 of our model system .
therewith , we have : @xmath85 it is convenient to separate the total charge density @xmath84 into two contributions : @xmath86 in the interior of the star , i.e. , the volume @xmath87 , and @xmath88 in the outer region , i.e. , the volume @xmath89 .
let @xmath90 and @xmath91 be the contribution to the electrostatic potential at an arbitrary point @xmath92 in space caused by the respective charge density . using this definitions ,
we can rewrite eq .
( [ eq : electrostatics ] ) and get @xmath93 \varrho_{\rm in}({\bf r})\right.\nonumber\\ & + \left.\int_{\omega_{\rm out}}{\rm d}^3r\ ; \left[\phi_{\rm in}({\bf r})+\phi_{\rm out}({\bf r})\right ] \varrho_{\rm out}({\bf r } ) \right\}.\end{aligned}\ ] ] on purely dimensional grounds , we expect a result having the general form @xmath94,\end{aligned}\ ] ] where we introduced dimensionless functions @xmath95 arising from the integrations of the products @xmath96 in eq .
( [ eq : electrostatics:2 ] ) . here , one should remember that @xmath97 is the number of uncompensated charges of the star and characterizes its effective valency .
the specific shape of the @xmath98-functions depends on the underlying charge distributions @xmath86 and @xmath88 alone .
the terms @xmath99 represent the ideal entropic free energy contributions of the different counterion species .
they take the form @xmath100 \nonumber\\ & + 3n_i\log\left(\frac{\lambda}{\sigma_{\rm lj}}\right),\end{aligned}\ ] ] where @xmath101 is the thermal de - broglie wavelength . in writing the sum in eq .
( [ eq : helmholtz ] ) , the respective last terms in eq .
( [ eq : entropy ] ) yield an trivial additive constant only , namely @xmath102 , which will be left out in what follows .
now , we have to quantitatively address the electrostatic and entropic terms pursuant to eqs .
( [ eq : electrostatics ] ) to ( [ eq : electrostatics:3 ] ) and ( [ eq : entropy ] ) . for that purpose ,
we first of all need to specify the above introduced number densities , @xmath82 and @xmath83 . here , we model the arms of the pe - star to be fully stretched , or to put it in other words , the monomer density profile inside the star to fall of as@xcite @xmath103 .
this is a good approximation , as measurements yield a scaling behavior with an only somewhat smaller exponent @xmath104 , indicating an almost complete stretching of the chains.@xcite since the monomer ions are placed on the backbone of the chains in a periodical manner ( cf .
[ sec : md ] ) , their density within the interior of the star must obviously show an identical @xmath10-dependence .
moreover , the profile for the trapped counterions exhibits the same scaling , due to the system s tendency to achieve local electroneutrality.@xcite therefore , it seems to be a good choice to use an ansatz @xmath105 for the trapped counterions and @xmath106 for the overall charge density in the inner region .
clearly , these distributions have to be normalized with respect to the total number of trapped counterions @xmath63 and the effective charge of the star @xmath107 by integrating over the related volumes , @xmath108 and @xmath87 . in doing so ,
we obtain @xmath109 and @xmath110 , with @xmath111.\ ] ] we presume the condensed counterions to be uniformly distributed within the tubes surrounding the pe - chains , i.e. , we use @xmath112 .
this approach is supported by simulation results on single pe - chains@xcite and was successfully put forward in previous studies of pe - star systems.@xcite in a similar fashion , we assume an also uniform distribution of the free counterions within the outer shell @xmath113 , i.e. , @xmath114 , implying an associated charge density @xmath115 . on this basis , we are able to explicitly calculate the variational free energy in virtue of eqs .
( [ eq : helmholtz ] ) to ( [ eq : entropy ] ) .
as far as the entropic contributions are concerned , an analytical computation is feasible in a straightforward manner . for reasons of clarity
, we leave out any intermediate steps and present our final results only : @xmath116 - 1,\\ \frac{\beta s_2}{n_2}=&\log\left(\frac{n_2}{2\pi c}\right ) + \frac{d}{rc}\log^2\left(\frac{d}{r}\right ) \nonumber\\ \label{eq : s2 } & -3\log\left(\frac{r}{\sigma_{\rm lj}}\right)+1,\\ \frac{\beta s_3}{n_3}=&\log\left\{\frac{n_3 } { \frac{2\pi}{3}\left[\left(\frac{r_{\rm w}}{r}\right)^3 - 1\right ] + \frac{\pi d}{r}\left [ \left(\frac{r_{\rm w}}{r}\right)^2 + 1\right]}\right\ } \nonumber\\ \label{eq : s3 } & -3\log\left(\frac{r}{\sigma_{\rm lj}}\right)-1.\end{aligned}\ ] ] now , we want to investigate the electrostatic term @xmath81 in more detail . to begin with , we have to derive expressions for the potential functions @xmath90 and @xmath91 .
this computations are rather technical and we will outline the course of action only . in general , the electrostatic potential @xmath117 due to a charge density @xmath84 in a dielectric medium is given by @xmath118 in order to calculate the integral above for a chopped sphere , yielding the function @xmath90 , we decompose it in infinitesimally thin discs which are oriented perpendicular to the @xmath35-axis and cover the whole sphere .
the radii of these discs obviously depend on their position with respect to the sphere s center . afterwards , each disc is object to further decomposition into concentric rings . provided that the charge density is spherical symmetric , as is in the case under consideration , the charge carried by each ring can be easily calculated .
the electrostatic potential of a charged ring is known from literature,@xcite therefore the potential of a disc can be derived by integrating over all corresponding rings .
an analytical determination of this integral is possible .
the potential of the chopped sphere itself can then be obtained by another integration over all discs .
the latter can not be performed analytically , thus one has to resort to a simple , one - dimensional numerical integration . for a complete and explicit description of the procedure
, we refer to ref .
, appendix a , where an almost identical derivation was put forward .
so far , the potential @xmath91 caused by the hollow chopped sphere of volume @xmath89 containing the free counterions remains to be acquired .
thereto , we employ the superposition principle .
the hollow region of uniform charge density @xmath88 can be apprehended as a superposition of two chopped spheres with radii @xmath2 and @xmath53 , and charge densities @xmath119 and @xmath88 , respectively . in doing so
, the problem is reduced to the calculation of the electrostatic potential of a chopped sphere with uniform charge density , which can be computed following the method described above for the inner sphere of volume @xmath87 ( see also ref . , appendix b ) .
clearly , with knowledge of both @xmath90 and @xmath91 , the dimensionless functions @xmath95 and thus the electrostatic energy @xmath81 can be obtained according to eqs .
( [ eq : electrostatics:2 ] ) and ( [ eq : electrostatics:3 ] ) using numerical techniques . in principle , the electrostatic - entropic interaction potential @xmath74 is obtained by adding up the entropic and electrostatic contributions following eq .
( [ eq : helmholtz ] ) and minimizing the helmholtz free energy @xmath76 with respect to @xmath2 and @xmath78 , cf .
( [ eq : potential ] ) . however , the star radii are in good approximation unaffected by the center - to - wall separation @xmath42 , as confirmed by our simulations .
the physical reason lies in the already almost complete stretching of the chains due to their charging . in the simulation the radius @xmath2 of the star
was measured according to @xmath120 where @xmath121 stands for the position vector of the last monomer of the @xmath122-th arm of the star and @xmath123 the core position .
[ fig : r ] illustrates the weak @xmath2 vs. @xmath42-dependence for an exemplarily chosen parameter combination .
therefore , it is convenient not to determine @xmath2 through the variational calculation , but to use average values @xmath124 as obtained from md simulations instead .
the latter are comparable to the corresponding radii for isolated pe - stars according to ref . , see table [ table : radii ] .
the amount of condensed counterions @xmath60 was measured by counting the number of such particles separated from the monomer ions by a distance smaller than @xmath61 and performing a statistical average .
the total number of captured counterions , @xmath125 , was measured by counting all counterions within a sphere having the instantaneous , arm - averaged center - to - end distance and again taking a time average .
since the @xmath78 are related through the equation @xmath126 , only two independent variational parameters remain , say @xmath60 and @xmath63 . in our simulations
, we have found that the number of condensed counterions , @xmath60 , is also approximately constant with respect to @xmath42 , see fig .
[ fig : n - species ] .
hence , we will treat @xmath60 as a fit parameter , held constant for all @xmath42 and chosen in such a way to achieve optimal agreement with simulation results .
therefore , @xmath76 is minimized with respect to the number of trapped counterions , @xmath63 , only , reflecting the possibility of redistribution of counterions between inside / outside the stars as the distance to the wall is varied .
[ fig : f - contrib ] shows a comparison of the entropic and the electrostatic contributions @xmath127 and @xmath81 to the electrostatic - entropic effective potential , @xmath128 .
as one can see from the exemplary plot , the total entropic term @xmath127 is the major contribution to @xmath128 and therefore determines the functional form of the latter , while the electrostatic term is of minor importance . at first glance
, this may seem to be counterintuitive in a pe - system , but electrostatics do in fact indirectly affect the interaction potential : due to the presence of charges , the conformations and thus the radii of pe - stars are strongly changed compared to the neutral star case , leading to an increased range of the interaction . in the inset of fig .
[ fig : f - contrib ] it can be seen that total entropy of the system at hand is mainly determined by the trapped counterions contribution @xmath129 .
as one can see , @xmath130 is only weakly influenced by the star - to - wall separation @xmath42 , and @xmath131 is even completely independent of @xmath42 , as was evident from eq .
( [ eq : s1 ] ) .
we have to emphasize , however , that even though @xmath131 contributes a constant value only and therefore does not influence the effective force @xmath132 at all , the number of condensed counterions @xmath60 nevertheless plays an fundamental role in our problem . since @xmath133 set the overall scale for of the term @xmath129 , @xmath60 becomes relevant in renormalizing the effective interaction .
the treatment of @xmath60 and the net charge , consequently , as a fit parameter is a common approach for charged systems , known as charge renormalization.@xcite in anticipation of sec .
[ sec : comparison ] , we want to point out that significant deviations arise between the effective star - wall forces as obtained by our computer simulations and their theoretical counterparts @xmath134 calculated using the formalism presented in sec .
[ subsec : ee ] for intermediate center - to - surface distances @xmath135 .
now , we want to elucidate the physical mechanism leading to these deviations and derive simple expressions for additional contributions to both the total effective forces and potentials , @xmath136 and @xmath137 .
therewith , we will conclude the development of our theoretical approach .
the need to introduce these supplementary contributions is due to so far unconsidered conformational changes enforced by the presence of the confining wall . during the construction of the mean - field part of our theory , we neglected such changes by assuming density profiles of monomers and counterions which are undisturbed compared to the case of isolated stars in full 3d geometries .
the influence of the wall became manifest in a truncation of the spheres representing star and associated counterion cloud , only .
but in reality , for @xmath42 in the order of the typical length of an arm of the star , the star will undergo strong configurational variations to avoid the wall since the monomers are not able to interpenetrate it .
thereby , due to the presence of neighboring arms , it can be energetically favorable for chains directing towards the wall to compress instead of bending away from the surface , although this compression leads to an extra cost in electrostatic energy .
the latter obviously originates in a decrease of the ion - ion distances along the backbones of the chains .
[ fig : arm - comp](a ) and [ fig : arm - comp](b ) depict the situation .
once a critical value of the chain length is reached , a further shortening of the chains under consideration becomes disadvantageous .
electrostatic repulsions and excluded volume interaction increase more and more strongly , the chains will preferably curve and start to relax in length , see fig . [
fig : arm - comp](c ) .
the occurrence of such compression - decompression processes is proved by our simulation runs . in what follows
, we will use a simplified picture and model the affected arms as rigid , uniformly charged rods of common length @xmath138 and diameter @xmath11 for both the compressed and the bent regime . in the former case
, we assume the rods to have orientations perpendicular to the wall surface . in the latter case
, we will account for the bending only via the re - lengthening of the rods and their change in orientation with respect to the @xmath35-axis .
[ fig : arm - comp](c ) visualizes this approximation , whereby the shaded straight chain represents the imaginary rod that replaces the bent arm within the framework of our modeling . the total charge of one such rod is just @xmath139 , leading to an @xmath138-dependent linear charge density , @xmath140 , given by @xmath141 .
we consider then the self - energy of one rod , as a function of its respective length @xmath138 , to be made up of a purely repulsive contribution which is electrostatic in nature and arises due to like - charge repulsions , and a second harmonic term with an effective spring constant @xmath142 describing the binding of the chain monomers in a coarse - grained fashion .
thus , we have:@xcite @xmath143 from eq .
( [ eq : correction:1 ] ) , we then obtain the corresponding force @xmath144 that acts on the ends of the rod parallel to its direction ( a negative / positive sign of @xmath145 corresponds to a force that tends to compress / stretch the rod . ) the competition of a repulsive and an attractive part results in a finite equilibrium length @xmath146 of the rod , i.e. , @xmath147 vanishes for @xmath148 .
clearly , @xmath146 is related to the above introduced spring constant @xmath142 by the following condition : @xmath149 now , let @xmath150 and @xmath151 determine the range of center - to - surface distances for which the above described processes take place .
moreover , @xmath152 denotes the critical length for which the transition from the compressing to the bending regime is observed . according to our md results ,
we will fix @xmath153 , @xmath154 , and @xmath155 .
in addition , we require @xmath156 in what follows . we know from our simulation
runs that it is a reasonable first - order approximation to assume the length of the affected chains as a function of @xmath42 to be : @xmath157\\ d_{\rm max } & d\in[d_{\rm max},\infty[. \end{cases}\ ] ] by using this empirical fact , we implicitly include effects due to chain bending and entropic repulsions of neighboring chains , even if we did not consider corresponding energy contributions explicitly in eq .
( [ eq : correction:1 ] ) .
therewith , a promising estimate for the chain compression contribution to the total effective force is : @xmath158\\ 0 & { \rm otherwise}. \end{cases } \label{eq : correction:2}\ ] ] here , @xmath159 is the total number of affected chains . assuming that the chains are regularly attached to the colloidal core
, we expect a linear relation between @xmath159 and @xmath0 , namely @xmath160 .
simulation data indicate @xmath161 to be a good choice for all parameter combinations under investigation .
note that the pre - factor @xmath162 results from geometrical considerations , as can be seen from fig .
[ fig : arm - comp](c ) . based on eq .
( [ eq : correction:2 ] ) , we obtain the corresponding energy term @xmath137 by a simple integration : @xmath163 finally , the total effective forces and interaction potentials are obtained as the sum of the electrostatic - entropic and compression terms : @xmath164 and @xmath165 , respectively .
[ fig : f - correct ] shows a concluding comparison of the full effective potential @xmath75 and the same without the inclusion of the compression term @xmath166 .
according to the plot , the theoretically predicted total potential is purely repulsive and ultra - soft . as we will see in sec .
[ sec : comparison ] , the total effective forces show a striking step - like shape for intermediate distances .
this is in contrast to the well - known case of star - star interactions.@xcite
in this section we test our theoretical model against corresponding simulation results to confirm its validity .
using standard md techniques , a straightforward measurement of ( effective ) interaction potentials is not possible .
thereto , one would have to apply more sophisticated methods.@xcite but since the mean force acting on the center of a pe - star can be easily received from computer experiments,@xcite we will focus on a comparison of such effective forces and corresponding theoretical predictions . to be more precise
: when choosing the colloidal center of a star at a fixed position @xmath167 as effective coordinate in our simulations , the effective force can be measured as the time average over all instantaneous forces @xmath168 acting on the core ( cf . also the simulation model in sec .
[ sec : md ] ) by means of the expression : @xmath169 in all considerations to come , we will deal with absolute values of the forces only .
it is obvious from the chosen geometry that the effective forces are in average directed perpendicular to the confining wall , i.e. , parallel to the @xmath35-axis .
thus , the mean values of the @xmath3- and @xmath4-components vanish , leading to the relation @xmath170 . for symmetry reasons ,
the effective forces and potentials depend on the @xmath35-component of @xmath167 alone , whereas @xmath171 .
consequently , the forces are connected to the potential by the following simple equation : @xmath172 therewith , predictions for the effective forces can be computed starting from theoretical results for the corresponding potential ( see sec . [
sec : theory ] , fig .
[ fig : f - correct ] ) .
this allows the desired comparison to md data .
[ fig : md - snapshot ] shows typical simulation snapshots for two different values of the center - to - surface separation @xmath42 , illustrating the conformational changes a star undergoes due to the presence of the confining wall . in part ( b ) of the figure ,
one may particularly note the pe - chains directed perpendicular to the wall .
these arms are object to the compression - decompression mechanism described in detail in sec .
[ subsec : cc ] and sketched in fig .
[ fig : arm - comp ] .
moreover , the snapshots qualitatively visualize our quantitative finding that the majority of counterions is captured within the interior of a pe - star , cf .
[ fig : n - species ] .
we want to point out that , compared to the bulk case,@xcite the counterion behavior in this regard is not significantly changed by introducing ( hard ) walls . in this context
, we also refer to table [ table : radii ] , where the number of condensed counterions @xmath60 as obtained from our md runs is shown for different parameter combinations together with corresponding values for isolated pe - stars in unconfined geometries .
[ fig : f - results ] is the core piece of this work and finally presents both our simulational and theoretical results for the effective star - wall forces .
as can be seen from the plots , we find a very good agreement for all combinations of arm numbers and charging fractions .
the theory lines exhibit a remarkable step - like shape for intermediate values of the center - to - surface distance @xmath42 arising due to the introduction of the chain compression term @xmath136 .
the latter had to be added to the electrostatic - entropic forces @xmath134 to account for conformational changes induced by the spatial proximity of the star to the planar wall .
in doing so , the functional form of the theoretically predicted forces accurately renders the results found by our computer experiments .
when using the respective star s radius as basic length scale for the plots , the position of the step is independent of the parameters considered . in this sense
, the effect is universal and there is no influence by specific details of the system .
note that the distinct kink is an artifact resulting from the special construction of @xmath136 , it does not have any physical meaning .
again , it should be emphasized that the described behavior is in contrast to the well - known star - star case.@xcite the difference between both systems mainly lies in the fact that a star is strictly forbidden to access the region of the wall , while it is allowed to interpenetrate another star - branched macromolecule to a certain degree . to put it in other words :
a wall is impenetrable , while a second star is a diffuse , soft object . here , we also want to emphasize that the step is no hysterisis / metastability effect , as in our md simulations the center - to - surface separation @xmath42 is varied by pulling the stars away from the wall instead of pushing them towards it .
this procedure prevents any coincidental trapping of individual pe - chains .
the particular values of the fit parameter @xmath60 used to calculate the theoretical curves in fig .
[ fig : f - results ] are given in the corresponding legend boxes . obviously , @xmath60 significantly grows with both the charging fraction @xmath22 and the arm number @xmath0 , as is physically reasonable .
table [ table : radii ] compares the fit values used to their counterparts obtained by our simulations .
thereby , one recognizes systematic deviations , i.e. , the fit values are always slightly off from the numbers computed using the md method . both quantities
are of the order of 50 - 60% of the total counterion number , for all functionalities @xmath0 and charging fractions @xmath22 .
the discrepancies emerge due to the different meanings of the quantities star radius @xmath2 and number of condensed counterions @xmath60 . within the scope of the simplified theoretical model ,
a pe - star is a spherical object of well - defined radius , while we observe permanent conformational fluctuations of the simulated stars .
thus , the md radius does not identify a sharp boundary , but determines a typical length scale only . in this sense
, the star can be viewed as a fluffy object .
most of the time , there are monomers , monomer ions and condensed counterions located at positions outside the imaginary sphere of radius @xmath2 , and such counterions are not counted when measuring @xmath60 , thus the theory value typically exceeds the simulation one . with increasing arm number @xmath0 and charging fraction @xmath22 , there is less variation in length of different arms of a star , resulting in smaller deviations between the theoretically predicted @xmath60 and its md equivalent .
table [ table : radii ] confirms this trend .
a last remark pertains to the behavior for small center - to - surface distances @xmath42 .
since our simulation model includes a colloidal core of finite radius @xmath1 , the forces obtained from md runs diverge for vanishing wall separations due to the core - wall lj repulsion which mimics the excluded volume interaction .
in contrast , our theoretical model has no core , for this reason the predicted forces remain finite for the whole range of distances . according to this ,
our theoretical approach is in principle not capable of reproducing a divergence of the forces , even in case of close proximity to the confining wall .
we have measured by means of molecular dynamics simulations and analyzed theoretically the effective forces emerging when multiarm polyelectrolyte stars approach neutral , impenetrable walls .
the forces have the typical range of the star radius , since osmotic pe - stars reabsorb the vast majority of the counterions and are thereby almost electroneutral ; longer - range forces that could arise due to the deformation of the diffuse counterion layer outside the corona radius are very weak , due to the small population of the free counterion species .
the dominant mechanism that gives rise to the soft repulsion is the entropy of the absorbed counterions and the reduction of the space available to them due to the presence of the impenetrable wall . at the same time , a novel , additional mechanism is at work , which stems from the compression of a fraction of the star chains against the hard wall . for star - wall separations that are not too different from the equilibrium star radius ,
the chain compression gives rise to an additional repulsive contribution to the force . at close star - wall approaches ,
the compressed chains ` slip away ' to the side , orient themselves parallel to or away from the wall and thus the compression process ceases and the additional contribution to the force vanishes .
the compression force could play a decisive role in influencing the cross - interaction between pe - stars and spherical , hard colloids of larger diameter .
indeed , for this case , the cross - interaction could be calculated by using the results of the present work as a starting point and performing a derjaguin approximation
. it will be interesting to see what effects the cross - interaction has in the structural and phase behavior of such pe - star
colloidal mixtures .
this issue , along with the study of the effects of charged walls and wall - particle dispersion forces , are the subject of ongoing work . .
[ table : radii ] conformational properties as obtained from md simulations and corresponding fit parameters used in our theoretical approach for different arm numbers and charging fractions .
in addition , results from ref . are presented for comparison .
the chain length is fixed to @xmath48 . for our data ,
the wigner - seitz radius is @xmath173 .
note that there are in part insignificant discrepancies between the parameters used here and in ref . .
[ cols="^,^,^,^,^,^,^,^",options="header " , ] fig . 1 : sketch visualizing the physical situation at hand , namely showing a pe - star of typical spatial extent @xmath2 , and its counterion background of radius @xmath174 .
the three possible counterion states are illustrated : condensed , trapped , and free .
for further explanations , see main text .
fig . 2 : star radii vs. the star - wall separation @xmath42 as obtained by md simulations , referring to parameters @xmath175 , @xmath176 .
the mean value , averaged with respect to @xmath42 , is shown as a horizontal line for comparison .
3 : fractions of condensed counterions @xmath177 and captured counterions @xmath178 for pe - stars with @xmath175 arms and various values of @xmath22 .
as one can see from the plots , both quantities are in very good approximation @xmath42-independent and the majority of counterions is captured within the interior of the star .
fig . 4 : comparison of the entropic and the electrostatic contributions @xmath127 and @xmath81 to the electrostatic - entropic part of the effective potential , @xmath128 , exemplarily shown for @xmath175 , @xmath176 , and @xmath179 .
obviously , the entropic contribution dominantly determines the @xmath42-dependence of @xmath128 . in the inset
, @xmath127 is decomposed into the different counterion species contributions @xmath99 , illustrating that the term @xmath129 governs the functional form of the total entropy .
note that the different contributions in both the main plot and the inset were shifted by constant values to enhance the readability of the plot .
fig . 5 : schematic illustrating the physical mechanism leading to the necessity to introduce the energy contribution @xmath137 .
hollow circles denote neutral monomers , gray balls are monomer ions .
the bigger black ball is the colloidal core of the star .
counterions are omitted in the depiction for reasons of clarity . in part ( c ) of the figure , the dashed rod replacing the chain bent to the left
is shown mirror - reflected to the right to avoid crowding . for a more detailed discussion of the compression effect ,
see main text . fig . 6 : comparison of the total and the electrostatic - entropic contributions to the effective potential , @xmath75 and @xmath74 , for @xmath175 , @xmath176 , and @xmath179 .
the inset shows the functional form of the compression term @xmath166 .
the main text provides an explanation for the need of the latter contribution and contains an explicit derivation .
fig . 7 : simulation snapshots of a pe - star with functionality @xmath180 and charging fraction @xmath176 . the center - to - wall distances are @xmath181 ( a ) and @xmath182 ( b ) .
bright gray balls are neutral monomers , while the dark spheres along the chains indicate monomer ions .
the counterions are the small , dark gray spheres surrounding the star .
fig . 8 : effective star - wall forces with the colloidal core of the respective stars taken as effective coordinate , obtained from both computer simulations and our theoretical approach .
the fit parameters @xmath60 used are specified in the legend boxes . here
, we show results for stars with functionalities @xmath175 ( first row ) , @xmath180 ( second row ) , and @xmath183 ( third row ) . in all cases
, we are considering charging fractions @xmath184 ( first column ) , @xmath185 ( second column ) , and @xmath176 ( third column ) . in the rightmost picture of the first row , we additionally included a theoretic line calculated without taking the compression term @xmath137 into account to illustrate the need of such a contribution .
since the theoretical model has , in contrast to the simulation model , no ( hard ) colloidal core , all simulation data have to be displaced by the core radius @xmath1 . | we employ monomer - resolved molecular dynamics simulations and theoretical considerations to analyze the conformations of multiarm polyelectrolyte stars close to planar , uncharged walls .
we identify three mechanisms that contribute to the emergence of a repulsive star - wall force , namely : the confinement of the counterions that are trapped in the star interior , the increase in electrostatic energy due to confinement as well as a novel mechanism arising from the compression of the stiff polyelectrolyte rods approaching the wall .
the latter is not present in the case of interaction between two polyelectrolyte stars and is a direct consequence of the impenetrable character of the planar wall . |
owing to the impressive progress in experimental techniques , ultracold neutral atoms trapped in optical lattices are nowadays widely recognized as a versatile toolbox bringing into reality ideal models of condensed matter physics @xcite .
a prominent example in this respect is no doubt the bose - hubbard model , originally introduced to sketch the physics of superfluid helium in porous media @xcite and subsequently shown to be realizable in terms of optically trapped ultracold bosonic atoms @xcite .
this model , describing interacting bosonic particles hopping across the @xmath0 sites of a discrete structure , is characterized by a hamiltonian of the form @xmath2 where @xmath3 , @xmath4 and @xmath5 are on - site bosonic operators , @xmath6 measures the strength of the ( on - site ) boson - boson interaction , @xmath7 is the hopping amplitude across neighboring sites and @xmath8 is the so - called adjacency matrix , describing the lattice topology .
its generic entry , @xmath9 , equals 1 if the sites @xmath10 and @xmath11 are adjacent , and 0 otherwise .
parameters @xmath6 and @xmath7 are directly related to well defined experimental quantities , i.e. the scattering length of the bosonic atoms and the intensity of the laser beams giving rise to the optical lattice , respectively @xcite . the possibility of tuning the lattice strength over a wide range of values played a fundamental role in the experimental observation @xcite of the superfluid - insulator quantum phase transition predicted for hamiltonian ( [ e : bhh ] ) in the case of repulsive interaction @xcite .
further aspects of versatility of optically trapped ultracold atoms lie in the possibility of tuning the atomic scattering length , and hence the boson - boson interaction , via feshbach resonances @xcite and realizing a great variety of different lattices , by means of a suitable laser beam setup @xcite , possibly involving holographic elements @xcite . in this paper
we are concerned with the case of attractive bosons in a 1d periodic lattice , whose experimental realizability is made concrete by a recently proposed optical trapping scheme @xcite and by the control over the sign and intensity of the boson - boson interaction granted by feshbach resonances @xcite .
this symmetric , _ necklace - like _ geometry is particularly interesting since , in the strong interaction limit , the system ground state is a quantum superposition of macroscopic spatially localized states , i.e. a so - called _ schrdinger - cat _
state @xcite .
although the state - of - art experimental setups ensure a remarkable control and insulation from the environment , the realization of such deeply quantum states presents several difficulties , which will be briefly discussed in the final part of this work .
our main aim here is a study of the ground state properties of quantum hamiltonian ( [ e : bhh ] ) , which we compare to the corresponding state of a suitable semiclassical approximation to the bose - hubbard model . before going into details , we mention that the attractive bose - hubbard model has been addressed by two recent works @xcite . in ref .
@xcite a further nonlinear ( cubic ) repulsive term is introduced in hamiltonian ( [ e : bhh ] ) in order to prevent excessive localization for large values of the interaction .
such a term is ascribed to three body interactions coming into play at elevate atomic densities . however , experimental observations suggest that these interactions produce different phenomena resulting into a leak of condensate atoms rather than into a repulsive term @xcite
. hence , here we do not consider such a repulsive term , and assume that the total boson population is sufficiently low to rule out the above phenomena . in some respects
our analysis is similar to that of ref .
@xcite , whose results we compare to ours .
other recent results related to ours are found in refs .
@xcite , which also focus on annular geometries , yet in the absence of a lattice potential . finally , we mention that the attractive bose - hubbard model ( [ e : bhh ] ) has received considerable attention as the quantum counterpart of the discrete self - trapping ( dst ) or quantum discrete non - linear schrdinger ( dnls ) equations @xcite , particularly in relation to the issues of discrete quantum solitons @xcite and breathers @xcite .
the layout of this paper is as follows . in section [ s : quantum ]
we rapidly review the key features of the bose - hubbard model ( [ e : bhh ] ) . focusing on the special case of a periodic one - dimensional lattice comprising @xmath0 sites , we recall the main properties arising from translational invariance . in section [ s : semicl ] we describe a semiclassical variational approach to hamiltonian ( [ e : bhh ] ) based on a number - conserving trial state .
this approach results into a set of dnls equations featuring a correction factor that makes them effective also in the deeply quantum regime characterized by large attractive interactions and small boson populations .
furthermore , we describe the three regimes that can be recognized for the ground state of the one - dimensional periodic lattice under investigation . a _ schrdinger - cat regime _ for small values of the parameter @xmath12 , where @xmath13 is the total number of bosons in the system ; a _ soliton regime _ , for intermediate values of @xmath14 ; a _ superfluid regime _ for large values of @xmath14 .
we show that the crossover between the schrdinger - cat and soliton regimes occurs around a size - independent threshold @xmath15 , while the critical point marking the boundary between the soliton and superfluid regimes , @xmath16 , exhibits a clear dependence on the lattice size @xmath0 , which becomes linear for large sizes . in section [ s : results ]
we illustrate the effectiveness of the semiclassical description introduced in section [ s : semicl ] . in particular ,
we show that in the schrdinge - cat regime , the delocalized ground state of the quantum model is very well described by a coherent superposition of the degenerate localized ground states of the semiclassical picture .
furthermore , we discuss the relevance of the semiclassical thresholds @xmath15 and @xmath16 in the quantum picture .
in particular we show that the local observables considered in ref .
@xcite are not sensitive to size dependent threshold @xmath16 , which is related to the finiteness of the system , but rather to the size - independent threshold @xmath15 , which is related to the discrete nature of the system .
we also show that @xmath16 is signaled by nonlocal observables . in appendix
[ s : sum ] we show that the variational trial state introduced in section [ s : semicl ] is actually a su@xmath1 coherent state , which is a further justification ( ? ) for the effectiveness of the semiclassical picture . in appendix
[ s : pert ] we show that in the schrdinger - cat regime the localized ground state of the semiclassical picture coincides at the first perturbative order in @xmath14 with a coherent superposition of the almost degenerate @xmath0 lowest states of hamiltonian ( [ e : bhh ] ) .
let us now summarize the main properties of hamiltonian ( [ e : bhh ] ) .
first of all , we recall that the relevant hilbert space is in principle infinite , and it is spanned by the fock vectors generated by the action of the positive powers of the creation operators @xmath17 onto the vacuum state @xmath18 , i.e. the state such that @xmath19 for any @xmath20
. however , hamiltonian ( [ e : bhh ] ) commutes with the total number operator , @xmath21 , and hence its study can be restricted to the finite size eigenspaces of @xmath22 .
henceforth , we assume to be in the eigenspace relevant to the total population @xmath13 , whose size is @xmath23 .
as we mention , we are interested in the case of attractive bosons ( @xmath24 ) on a 1d homogeneous periodic lattice .
hence we particularize the adjacency matrix in eq .
( [ e : bhh ] ) to @xmath25 where , on a periodic lattice comprising @xmath0 sites , the site labels @xmath26 and @xmath27 correspond to sites @xmath0 and @xmath28 , respectively . with this choice hamiltonian ( [ e : bhh ] ) is clearly translationally invariant .
hence all of the relevant observables can not break this symmetry , i.e. exhibit a dependence on the site label . owing to the translational invariance
the eigenstates of @xmath29 can be labeled by a further quantum number .
this is made quite evident by switching to the reciprocal space , i.e. by introducing the transformed operators @xmath30 making the kinetic term of hamiltonian ( [ e : bhh ] ) diagonal .
of course , the reciprocal space has the same size @xmath31 as the direct space , and it is spanned by the reciprocal fock vectors , generated by the action of the powers of the @xmath32 s on @xmath18 .
note indeed that @xmath33 and that @xmath34 for any @xmath35 . in this representation
the displacement operator changing @xmath3 into @xmath36 is simply @xmath37 since @xmath38=0 $ ] , the hamiltonian can be studied separately in the @xmath0 eigenspaces of @xmath39 .
note indeed , that @xmath40 , which entails that the eigenvalues of the displacement operator are the @xmath41 roots of unity . after noticing that the reciprocal fock vectors are clearly eigenstates of @xmath39 , the block diagonal structure of @xmath29 can be made evident by simply arranging these states according to the relevant eigenvalue , which we denote by @xmath42 . a perhaps more convenient choice for the displacement quantum number is the integer @xmath43 $ ] such that @xmath44 .
in particular , the ground state of the system is the lowest eigenstate of the block relevant to @xmath45 or , equivalently , @xmath46 .
introducing a further integer label @xmath47 ranking the eigenstates of of hamiltonian ( [ e : bhh ] ) according to their energy , we set @xmath48 .
in particular , @xmath49 and @xmath50 denote the ground state of @xmath29 and its energy .
we recall that this state is translationally invariant , while , more in general , all of the eigenstates of @xmath29 are delocalized .
that is to say , these eigenvalues can not give rise to site - dependent expectation values of local operators , such as e.g. @xmath51 .
this means that , at least in principle , the states @xmath52 exhibit the features of the celebrated _ schrdinger cat _ state .
indeed , as it is well known , for @xmath53 the ground states of the @xmath0 blocks of @xmath29 are almost degenerate coherent superpositions of macroscopically localized states of the form @xmath54 ) perturbatively @xcite . in sec .
[ s : results ] we show that the _ schrdinger cat _ structure of eq .
( [ e : loc ] ) can be extended to a whole interval of @xmath55 ratios , by replacing the fock vectors @xmath56 with suitable su@xmath1 coherent states which are strictly related to the localized ground state of the semiclassical approximation to hamiltonian ( [ e : bhh ] ) .
as we recall above , the attractive bose - hubbard model ( [ e : bhh ] ) has received significant attention also as the quantum counterpart @xcite of the so - called discrete self - trapping ( dst ) or discrete non - linear schrdinger ( dnls ) equations @xcite .
these semiclassical equations are equivalently obtained by discretizing the gross - pitaevskii equations @xcite or applying to hamiltonian ( [ e : bhh ] ) a time dependent variational principle ( tdvp ) based on glauber s coherent states @xcite .
here we adopt the same approach as ref .
@xcite , using a macroscopic trial state of the form @xmath57 where @xmath58 appearing in eq .
( [ e : twf ] ) is a dynamical variable describing the bosons at lattice site @xmath10 through a macroscopic local phase and population , identified with @xmath59 and @xmath60 .
the request that the trial state satisfies the schrdinger equation on the average @xcite , @xmath61 , leads to the identification of @xmath62 with the effective lagrangian for the dynamical variables @xmath63 , and @xmath64 with the corresponding effective hamiltonian @xcite .
minimizing @xmath65 with respect to the @xmath63 s provides a semiclassical variational approximation to the ground state of the bose - hubbard hamiltonian . introducing the chemical potential @xmath66 , i.e. a lagrange multiplier allowing for the normalization constraint in eq .
( [ e : twf ] ) , the ensuing equations are @xmath67 we recall that eqs .
( [ e : fpe ] ) are actually the fixed point equations corresponding to the equations of motion for the @xmath63 s @xmath68 obtained via the canonical poisson brackets @xmath69 . here by fixed points we mean the `` normal modes '' whose time evolution is entirely contained in a simple phase factor @xmath70 . equations ( [ e : dst ] ) clearly belong to the dnls class and , as we show in the following section , they are able to capture many of the features of the quantum ground state also in the extreme quantum regime of large interactions and small populations .
this is due to the extra @xmath71 factor appearing in their interaction term , which is absent in the dnls equations standardly adopted in the semiclassical approach to the bose - hubbard model .
the presence of this extra factor is a direct consequence of the macroscopic trial state , eq .
( [ e : twf ] ) , being an eigenstate of the total number operator @xmath22 .
this is not true of the glauber coherent state product adopted as a trial function in standard tdvp , which conserves the total number of bosons only on the average @xcite .
hence the resulting dynamical equations , which are perfectly equivalent to standard dnls equations , do not feature the extra @xmath72 factor .
the effectiveness of the description provided by the macroscopic state ( [ e : twf ] ) can be explained by noticing that it is actually a coherent state itself , of su@xmath1 .
note that @xmath13 is the algebra representation index of the present two - boson realization of su@xmath1 as confirmed by the fact that @xmath22 commutes with the algebra generators @xmath73 for any @xmath74 .
hence the conservation of the boson population is already incorporated in this coherent state picture .
this is illustrated in appendix [ s : sum ] as it is well known , the nonlinear terms can break the translational symmetry characterizing hamiltonian ( [ e : scbh ] ) with the choice in eq .
( [ e : homj ] ) .
indeed , if the interaction @xmath6 is sufficiently strong , the ground state of the system is arbitrarily localized at one of the @xmath0 equivalent lattice sites .
this ground state is a fixed point of eqs .
( [ e : dst ] ) , and can be easily determined numerically by treating eqs .
( [ e : fpe ] ) as a set of self - consistent linear equations @xcite .
more in general , depending on the value of the @xmath55 ratio , it is possible to recognize three regimes for the semiclassical ground state . for large ratios ,
the ground state is superfluid , the boson population being evenly spread over the entire lattice . for sufficiently low values of the @xmath55 ratio
this delocalized superfluid solution becomes unstable , and the ground state of the system becomes soliton - like , featuring a localization peak breaking the translational symmetry of hamiltonian ( [ e : scbh ] ) .
the critical point marking the boundary between the _ superfluid _ and _ soliton regime _ is @xmath75^{-1}$ ] , where we recall that we define @xmath76 @xcite .
as @xmath14 is further decreased , the localization peak which initially involves a significant number of lattice sites becomes sharper and sharper , until the entire boson population is basically concentrated at a single lattice site .
this situation is the is the semiclassical , symmetry - breaking counterpart of the schrdinger - cat state , where the boson localization is symmetrically distributed over all of the lattice sites . as we discuss in the sequel , it is possible to recognize a threshold , or rather a crossover , between this _ schrdinger - cat regime _ and the soliton regime where the localization peak involves a large portion of the lattice .
note indeed that in the latter situation the periodic boundary conditions and the discretization of the lattice produce a small effect , and the ground state is well described by the solution to the nonlinear schrdinger equations on a 1d infinite domain , @xmath77 , where @xmath78 is the characteristic width of the population peak .
the inset of fig .
[ f : regimes ] shows that this is actually a satisfactory approximation .
hence , the crossover between the schrdinger cat and soliton regime can be estimated by requiring that this characteristic width equals the lattice spacing , i.e. @xmath79 , which results into @xmath80 .
note that this is equivalently obtained by equalling the energies of the system in the two extreme situations @xmath81 and @xmath82 . indeed in the former case the ground state is completely localized and its energy is @xmath83 , while in the latter case the ground state is delocalized and its energy is @xmath84 .
we emphasize that exactly the same estimate is obtained considering the energy of the quantum ground state .
conversely , the delocalization transition at @xmath85 , marking the boundary between the soliton and superfluid regime , corresponds to the peak width being of the same order of the lattice size , @xmath86 .
note indeed that @xmath16 is linear in @xmath0 for sufficiently large lattices @xcite .
the regimes discussed above are quite evidently signalled by the behavior of the ( normalized ) interaction energy @xmath87 as a function of @xmath14 , where @xmath88 denotes expectation value on the trial state ( [ e : twf ] ) equipped with the ground state solution to the fixed point equations ( [ e : fpe ] ) .
( color online ) normalized interaction energy , eq .
( [ e : int_en ] ) , for two different lattice sizes , @xmath89 ( light solid line ) and @xmath90 ( dark solid line ) .
the delocalization transition clearly occurs at @xmath91 .
furthermore , the plotted quantity exhibits a size independent behavior in the small @xmath14 s regime .
the crossover at @xmath92 corresponds to a change in the convexity plotted function .
this is related to the ultraviolet cutoff due to the discreteness of the lattice .
note indeed that in the continuum limit the corresponding quantity would diverge as @xmath93 for @xmath94 .
the inset shows a comparison between the localized ground state of @xmath65 ( bullets ) on a lattice of size @xmath90 and the corresponding solution in the continuum limit ( solid line).,width=321 ] this is shown in fig .
[ f : regimes ] , where two lattice size are considered , @xmath89 and @xmath90 . recalling that @xmath95 , it is easy to see that the quantity in eq .
( [ e : int_en ] ) is also related to the fluctuation of the local density of bosons .
in this section we compare the features of the semiclassical picture introduced in the preceding section to the corresponding features of the quantum model , eq .
( [ e : bhh ] ) .
first , we direct our attention to the energy and structure of the ground state . then , we discuss the relevance of the semiclassical thresholds in the quantum picture .
we consider lattice comprising up to @xmath89 sites , and populations consisting of up to @xmath96 bosons .
all of the quantum observables are obtained evaluating the eigenstates of hamiltonian ( [ e : bhh ] ) and the relevant energies by means of lanczos - like algorithms .
the translational symmetry of hamiltonian ( [ e : bhh ] ) is conveniently taken into account in order to reduce significantly the computational demand . as it can be seen in fig .
[ f : deltae ] , the ground state energy of the quantum model , eq .
( [ e : bhh ] ) is in general remarkably well described by the relevant semiclassical result .
( color online ) relative energy difference between the exact quantum ground state and the relevant semiclassical approximation .
the curves refer to boson population @xmath96 and lattice sizes @xmath97 .
the inset shows the energy range spanned by the ground states @xmath98 ( shaded region ) and the first excited states @xmath99 ( region between the dotted lines ) of the blocks relevant to different values of the quantum number @xmath100 associated to the translational invariance of @xmath29 [ see eq .
( [ e : hspec ] ) ] .
these result refer to @xmath101 and @xmath102 .
the vertical dashed lines signal the crossover at @xmath92 . ,
width=321 ] indeed the maximum in the relative error settles at about 2% as the lattice size increases .
furthermore , the agreement between the quantum and the classical result is extremely satisfactory both in the large @xmath14 and in the small @xmath14 region .
the agreement in the former region is expected , since in the noninteracting limit the quantum ground state of the system is exactly given by eq .
( [ e : twf ] ) with @xmath103 , which is nothing but the superfluid semiclassical ground state taking over after the delocalization transition at @xmath16 .
the agreement in the strong interaction limit is more surprising , and it requires further discussion .
indeed , in this region the semiclassical and the quantum ground states are significantly different , the former being strongly localized at one lattice site , while the latter being always completely delocalized . as we recall in sec .
[ s : quantum ] , the eigenvalues of the quantum hamiltonian ( [ e : bhh ] ) are labeled by a quantum number @xmath100 ensuing from translational invariance , eq .
( [ e : hspec ] ) . the inset of fig .
[ f : deltae ] clearly shows that , in the small @xmath14 region the actual ground state , @xmath104 , is almost degenerate with the ground states of all of the remaining hamiltonian blocks , @xmath105 .
hence , in this region , the actual ground state of the system is satisfactorily described by any superposition of the states @xmath105 . of course this is not an eigenstate of @xmath29 , but the relevant dephasing time is extremely long , being inversely proportional to the energy spreading of the involved eigenstates .
in particular , one can choose a superposition of the form @xcite latexmath:[\[\label{e : coh_sup } which features a population peak localized at lattice site @xmath47 .
we emphasize that the phase factors @xmath107 are not arbitrary , but have to be chosen to make sure that the @xmath105 s are phase coherent .
recall indeed that a generic eigenstate is defined up to an arbitrary phase factor .
the superposition in eq .
( [ e : coh_sup ] ) features the desired localization properties only if the @xmath107 s compensate for possible phase differences among the @xmath105 s @xcite . as it is shown in fig .
[ f : loc_st ] , the boson density @xmath108 is in remarkable agreement with the corresponding semiclassical quantity , @xmath60 for sufficiently small values of @xmath14 . at larger values
the agreement is lost since the energy spreading of the eigenstates involved in eq .
( [ e : coh_sup ] ) becomes significant , and the superposition can not be confused with the ground state any more . furthermore , as it is shown in the inset of fig .
[ f : deltae ] , the energies of the @xmath109 s become comparable to those of the @xmath105 s .
( color online ) normalized boson density @xmath108 as provided by the quantum localized state ( solid lines ) , compared to the corresponding semiclassical quantity @xmath60 ( bullets ) for @xmath102 bosons on @xmath110 lattice sites . , width=321 ] by resorting to perturbation theory , it is possible to show that the agreement between the trial state ( [ e : twf ] ) and the localized state @xmath111 , eq .
( [ e : coh_sup ] ) , goes beyond the local density of bosons . indeed , at the first order in the perturbative parameter @xmath14 , they are actually the same state .
this is discussed in appendix [ s : pert ] .
notice that the above discussion entails that , in the strong coupling , schroedinger - cat regime , the actual ground state of the quantum hamiltonian ( [ e : bhh ] ) is a coherent superposition of localized states of the form ( [ e : twf ] ) , where the @xmath63 s are obtained solving the corresponding semiclassical problem ( [ e : fpe ] ) .
once again , we emphasize that the extra @xmath71 in the dnls equations ( [ e : dst ] ) , stemming from the choice of a number - conserving trial state , eq .
( [ e : twf ] ) , plays a fundamental role in ensuring the remarkable agreement between the quantum and the semiclassical picture , eqs .
( [ e : bhh ] ) and ( [ e : scbh ] ) , respectively .
let us now discuss the relevance of the semiclassical thresholds @xmath15 and @xmath16 in the quantum picture of the system .
this issue has been considered to some extent in ref .
@xcite , where only the delocalization transition occurring at @xmath91 is taken into account .
the authors report that this critical point is signalled by the behavior of some quantum observables , such as the one - particle and @xmath112-particle correlation functions , @xmath113 where @xmath88 denotes expectation value over the ground state .
more precisely , they observe that , with increasing interaction , @xmath114 features a sudden drop around the semiclassical delocalization threshold @xmath16 , while @xmath115 features a surge roughly at the same point .
furthermore they report that this threshold is also signalled by the number distribution , i.e. the probability @xmath116 that the boson population at a given lattice site equals @xmath117 .
more precisely , they report that this distribution becomes double peaked for @xmath118 . here
we show that the above quantum observables do not signal the delocalization threshold , but rather the crossover between the soliton and schrdinger - cat regime occurring around @xmath92 .
first of all , we observe that the variation of these observables is not so sudden to identify a precise transition point .
they rather signal a crossover interval .
furthermore , in the case considered in ref .
@xcite , @xmath119 , @xmath102 , the two semiclassical thresholds @xmath120 and @xmath121 are relatively close , and both compatible with such crossover interval . however , as we illustrate in section [ s : semicl ] , @xmath122 is substantially size - independent , while @xmath16 clearly depends on the lattice size .
a study of the scaling properties of the above quantum observables shows that the features considered in ref .
@xcite do not depend on the lattice size , and occur in the region around @xmath15 @xcite .
( color online ) top : contour plot of the number density @xmath123 for @xmath124 bosons on a lattice of size @xmath110 . the darker the shade , the lower the density .
the white bullets signal the value @xmath125 at which the density is minimum .
the vertical lines signal the semiclassical thresholds .
bottom : plot of @xmath125 for several choices of @xmath13 and @xmath0 .
, width=321 ] let us first of all consider the number distribution @xmath126 , i.e. the quantity such that @xmath127 and , in general , @xmath128 .
it is clear that in the strong interaction limit this distribution features two peaks at @xmath129 and @xmath130 , while it is substantially vanishing for all of the remaining numbers @xcite .
recall indeed that in this limit , , i.e. deep in the schrdinger - cat regime , the ground state has the form in eq .
( [ e : loc ] ) @xcite . in this regime
the number fluctuations are very large .
actually , as we discuss in sec .
[ s : semicl ] , this situation is the quantum counterpart of the semiclassical situation where the boson population is almost entirely concentrated at one of the lattice sites .
hence , the ground state can not be influenced by the boundary conditions or by the lattice size , and this regime is expectedly entered at @xmath131 .
this is actually confirmed by fig .
[ f : rho_n ] , which illustrates some key features of the number distribution .
as it can be seen in the contour plot of in the top panel , for small @xmath14 the number distribution @xmath123 features two maxima at the extreme values @xmath132 and @xmath133 , while its minimum is at @xmath134 .
as soon as @xmath14 becomes larger than @xmath15 the relative maximum at @xmath133 disappears , and the minimum suddenly moves to this extreme value .
the location of such minimum , @xmath135 , is signalled by the white bullets on the contour plot .
note that no special feature can be recognized in correspondence of the delocalization threshold , signalled by the vertical dashed line .
the bottom panel of fig .
[ f : rho_n ] reports the ( normalized ) location @xmath136 of the minimum of @xmath123 for several choices of the lattice size @xmath0 and of the boson population @xmath13 . in all cases
the sudden displacement of the minimum signals that the number distribution becomes single peaked at the semiclassical crossover @xmath15 , i.e. at the crossover between the schrdinger - cat and soliton regime .
behavior of the @xmath112-particle correlation function @xmath115 ( top ) and of its first derivative ( bottom ) with respect to @xmath14 .
the dotted vertical lines signal the semiclassical crossover at @xmath15 .
the dashed vertical lines signal the semiclassical delocalization thresholds relevant to the lattice sites considered @xmath137 .
, width=321 ] basically the same discussion as above applies to the @xmath112-particle correlation function @xmath115 defined in eq .
( [ e : corfun ] ) .
it is indeed clear that it must exhibit a significant surge when the schrdinger - cat regime is entered and the number distribution becomes double peaked , i.e. around @xmath138 .
indeed , for this correlation to be non - vanishing it is necessary that the probability of finding all of the @xmath13 bosons at the same lattice site is significant .
this is confirmed by fig .
[ f : cn ] , reporting the behavior of @xmath139 and of its derivative with respect to @xmath14 for several choices of the boson population and of the lattice size .
no particular feature occurs at the @xmath0 dependent delocalization thresholds , signalled by the vertical dashed lines . actually , a dependence on the lattice size @xmath0 is hardly recognizable , the curves referring to the same number of bosons @xmath112 being almost perfectly coincident .
conversely , the location where the plotted quantity becomes significant and its concavity changes are satisfactorily estimated by the semiclassical size - independent crossover at @xmath92 .
behavior of the single - particle correlation function @xmath114 ( top ) and of its first derivative ( bottom ) with respect to @xmath14 .
the dotted vertical lines signal the semiclassical crossover at @xmath15 .
the dashed vertical lines signal the semiclassical delocalization thresholds relevant to the lattice sites considered @xmath137 .
, width=321 ] figure [ f : c1 ] shows that also the substantial drop in the single - particle correlation function @xmath114 considered in ref .
@xcite actually occurs around @xmath138 .
this is confirmed by the behavior of the derivative of this quantity , shown in the bottom panel of the same figure .
again , the plotted quantity is not sensitive to the size - dependent delocalization threshold @xmath16 , essentially due to the fact that it is a local quantity , as it can be seen from its definition , eq .
( [ e : corfun ] ) .
the size - dependent delocalization threshold between the soliton and superfluid regime is signalled by a typically nonlocal quantity such as the occupation number @xmath140 of the lowest reciprocal mode , where @xmath141 is defined in eq .
( [ e : bk ] ) .
this is clearly evident in fig .
[ f : nk0 ] , where @xmath142 is plotted with respect to both @xmath14 ( top panel ) and @xmath143 ( bottom panel ) , for several lattice sizes and boson populations . unlike the quantities in figs .
rho_n]-[f : c1 ] , @xmath142 exhibits a clear dependence on the lattice size , when plotted against @xmath14 .
conversely , when the abscissae of the plot are rescaled so that all of the size - dependent delocalization thresholds occur at the same point ( signalled by the vertical dashed line ) , a very satisfactory data collapse becomes evident .
this means that the drop in the occupation of the lowest reciprocal mode is related to the delocalization transition observed in the semiclassical model .
behavior of the occupation of the lowest reciprocal mode @xmath140 with respect to @xmath14 ( top ) and with respect to @xmath143 .
the dashed vertical lines signal the semiclassical delocalization thresholds relevant to the lattice sites considered @xmath137 .
the data collapse in the bottom panel shows that @xmath140 actually signals the semiclassical delocalization threshold at @xmath16 .
, width=321 ]
in this paper we consider the bose - hubbard hamiltonian ( [ e : bhh ] ) describing ultracold attractive bosons trapped into the _ necklace - like _ lattice whose realization in terms of optical potentials is described in ref .
in particular , we study the properties of the ground - state of such extremely symmetric quantum hamiltonian , which in the strong interaction limit is known to exhibit the features of a _ schrdinger - cat _ state .
we introduce a variational approximation to this ground state based on a su@xmath1 coherent state , eq .
( [ e : twf ] ) .
the equations governing the dynamics of the labels of this coherent state belong to the class of dnls equations , yet they feature an unusual correction term ensuing from the total number conservation . owing to this correction term
, the ground - state of this semiclassical approximation provides a remarkably satisfactory description of its quantum counterpart not only as expected in the small - interaction superfluid regime , but also in the deeply quantum schredinger - cat regime , i.e. for strong attractive interactions .
in particular we show that in this regime the delocalized quantum ground state of the system can be seen as a coherent superposition of the localized ground states of the variational approach .
furthermore , we show that the three regimes clearly recognizable in the semiclassical picture are signalled by suitably chosen quantum observables .
the latter result is compared with similar findings reported in ref .
@xcite . in the intermediate regime
the description provided by the above semiclassical picture is not as satisfactory as in the two outermost regimes .
the structure of the quantum ground state in this parameter range is different , and probably much more complex , than that of the variational trial state ( [ e : twf ] ) , and requires further investigations . before concluding this paper we briefly consider some experimental issues .
first of all , we observe that attaining the deeply quantum schrdinger - cat regime might prove very challenging from the experimental point of view . indeed , this would be most likely obtained by starting from the ground state of the ( almost ) noninteracting superfluid regime , which is relatively easy to realize . by making use of feshbach resonances and/or by increasing the strength of the optical lattice , the ratio @xmath144 can be subsequently decreased down to suitably small values , so that the system is driven into the schrdinger - cat regime .
note that , independent of the value of such ratio , the ground state of the system has always the same symmetry , i.e. it always belongs to the same eigenspace of the displacement operator defined in eq .
( [ e : disp ] ) .
this means that , at least in principle , only the relevant block of hamiltonian ( [ e : bhh ] ) is involved in the process .
actually , this is true provided that the system is perfectly isolated from the environment . under this rather ideal assumption ,
the achievement of the schrdinger - cat ground state basically depends on the adiabaticity of the parameter variation .
however , the interaction with the environment much likely introduces a coupling between the ideally uncoupled blocks of hamiltonian ( [ e : bhh ] ) , which has dramatic effects in the strong interaction regime .
indeed in this situation the actual ground state is almost degenerate with the lowest states of the blocks relevant to @xmath145 , and an arbitrarily small perturbation can induce a transition to one of these excited states . another issue worth mentioning is the actual possibility of making measurements on a system such as the one considered here . as we recall in section [ s : intro ] , there is an upper bound to the number of attractive bosons @xmath13 that can be loaded into the optical lattice .
indeed , the maximum attainable boson density @xmath146 must be sufficiently low to rule out the complex phenomena described in ref . @xcite . on a one - dimensional system like the one considered here
, the maximum boson population compatible with such constraint can be too small to allow measurements , , such as e.g. absorption imaging .
this difficulty can be possibly circumvented by considering a stack of coaxial _ necklace _ optical lattices isolated from each other , in place of a single necklace @xcite .
assuming that these lattices can be loaded with the same number of bosons , they are expected to behave like copies of the system discussed in this paper . on the other hand ,
the larger overall boson population should allow for effective measurements .
the authors wish to thank f. minardi , p. maioli and p.g .
kevrekidis for stimulating discussions .
one of the authors ( p.b . ) acknowledges a grant from the _ lagrange project _ - crt foundation .
in this section we give a sketch of the proof that the variational trial state introduced in eq .
( [ e : twf ] ) is actually a su@xmath1 coherent state , where @xmath0 is the number of sites composing the lattice . according to the standard theory @xcite , the coherent states of a given group @xmath147
are represented , up to a phase factor , as @xmath148 , where @xmath149 and @xmath150 is a so - called _ extremal vector_. the latter is identified by determining the maximal isotropy sub algebra @xmath151 , where @xmath152 is the complexified version of the algebra @xmath153 associated to the group @xmath147 . in our particular case , @xmath153 is the su@xmath1 algebra represented in terms of the @xmath0 lattice boson operators @xmath3 .
hence @xmath154 and a generic element of the su@xmath1 group is given by @xmath155 } \ ] ] where the cartan operators @xmath156 are suitable linear combinations of the on - site number operators @xmath157 , @xmath158 and @xmath159 . the relevant maximal isotropy sub algebra @xmath160 is formed by ( complex valued ) linear combinations of all of the generators of @xmath153 , except the @xmath161 operators of the form @xmath162 , where @xmath163 corresponds to one of the lattice sites , and @xmath164 .
the extremal state @xmath165 , is defined as the state such that @xmath166 for any @xmath167 , with @xmath168 @xcite .
this yields @xmath169 , where @xmath170 is defined in eq .
( [ e : loc ] ) . of course
, all of the @xmath0 possible choices for @xmath163 , and hence for the extremal state , are equivalent . for definiteness , here we set @xmath171 .
it is possible to show that for any @xmath149 there exists an unique decomposition @xmath172 , where @xmath167 and the so - called _ displacement operator _ has the form @xmath173 .
\label{e : cs}\ ] ] hence , up to an irrelevant phase factor arising from @xmath174 , one gets @xmath175 on the other hand , the variational trial state introduced in eq .
( [ e : twf ] ) can be rewritten as @xmath176 observing that @xmath177 , and hence that @xmath178 , where @xmath179 one gets @xmath180 where we made use of the fact that @xmath181 .
after some further algebraic manipulations , state ( [ e : twf3 ] ) reduces to ( up to an irrelevant phase factor ) @xmath182 where @xmath183 . the comparison of the latter expression with eq .
( [ e : cs ] ) fully discloses the su(@xmath0)-coherent - state structure of the variational trial state @xmath184 defined in eq .
( [ e : twf ] ) .
it is important to notice that , while @xmath184 defined as in eq .
( [ e : twf2 ] ) is operationally useful to derive the semiclassical picture relevant to eqs .
( [ e : scbh ] ) and ( [ e : dst ] ) , the hidden , equivalent form of @xmath184 as a su@xmath1 coherent - state ( cs ) indeed is responsible for such a semiclassical scenario . in general ,
cs are known to provide a formulation of quantum - mechanical problems @xcite particularly close to the corresponding classical hamiltonian picture .
let us consider a classical hamiltonian @xmath185 for @xmath186 degrees of freedom and the corresponding quantum model @xmath187 obtained via canonical quantization .
the dynamics of the quantum model can be investigated enacting a time - dependent variational principle ( tdvp ) based on a coherent state @xmath188 , with @xmath189 .
this allows one to recast the schrdinger equation for @xmath187 into a classical problem whose hamiltonian is @xmath190 .
the corresponding hamiltonian equations describe the evolution of the cs labels @xmath191 , which thus play the role of dynamical parameters .
simple analytic formulas show that @xmath191 s can be rewritten in terms of the canonical variables appearing in @xmath185 ( see , e. g. , ref .
@xcite ) . more in general , the fact that @xmath188 is a coherent state ensures that the domain where @xmath192 ranges ( cs manifold ) coincides with the phase space where the dynamics of @xmath185 is represented .
the cs variational method exhibits two further noticeable features .
first , the classical poisson brackets characterizing the @xmath192 dynamics preserve the algebraic structure of the operators constituting @xmath187 and generating the group @xmath147 .
also , the coherent - state labels @xmath191 , beyond their canonical role in the @xmath193 dynamics , are the expectation values of hamiltonian operators that embody information on microscopic interaction processes . for example , in the specific case under investigation , @xmath194 , @xmath195 in hamiltonian ( [ e : scbh ] ) are manifestly related to @xmath196 and @xmath51 in hamiltonian ( [ e : bhh ] ) .
notice that hamiltonian ( [ e : scbh ] ) matches exactly the classical version of eq .
( [ e : bhh ] ) as soon as @xmath197
. an interesting results in the present work is the profound similarity between the ( purely quantum ) properties exhibited by the ground state of hamiltonian ( [ e : bhh ] ) and the properties of its semiclassical minimum - energy configuration derived from state ( [ e : twf2 ] ) in the schrdinger - cat regime .
in this section we show that in the strong interaction limit the superposition defined in eq .
( [ e : coh_sup ] ) coincides with the coherent state ( [ e : twf ] ) where the @xmath63 s correspond to the semiclassical solution to eqs .
( [ e : fpe ] ) localized at lattice site @xmath47 . we do this resorting to perturbative theory , i.e. by assuming that @xmath198 and @xmath199 are the unperturbed and perturbative terms of the quantum hamiltonian ( [ e : bhh ] ) , respectively . at the @xmath200 order in @xmath201 the quantum superposition @xmath111 in eq .
( [ e : coh_sup ] ) coincides with the fock state @xmath202 defined in eq .
( [ e : loc ] ) .
the latter belongs to the @xmath0-fold degenerate lowest eigenspace of @xmath203 .
indeed , @xmath204 as to the first - order perturbative correction to @xmath111 , it is standardly obtained as @xmath205 .
after some simple algebraic manipulations we get latexmath:[\[\label{e : firstb } on the other hand it is easy to show that , to the first perturbative order in @xmath14 , the solution to the semiclassical equations ( [ e : fpe ] ) localized at lattice site @xmath47 is given by @xmath208 plugging this expression into eq .
( [ e : twf ] ) gives @xmath209 ) .
hence , to the first perturbative order , the quantum superposition of the @xmath0 lowest eigenstates of @xmath29 defined in eq .
( [ e : coh_sup ] ) coincides with the state obtained plugging the suitable localized solution to the semiclassical equations ( [ e : fpe ] ) into the coherent state ( [ e : twf ] ) .
we emphasize that in the strong interaction limit the zero temperature properties of @xmath29 can be investigated based on the ground state of the corresponding semiclassical model , eq .
( [ e : scbh ] ) , which is evaluated at a considerably lower computational cost . e. schrdinger , naturwissenschaften * 23 * , 807 ( 1935 ) .
more recent references focusing on system similar to the one considered here are j.i .
cirac , m. lewenstein , k. mlmer , and p. zoller , phys .
a * 57 * 1208 ( 1998 ) ; d.a.r dalvit , j. dziarmaga and w.h .
zurek , phys .
a * 62 * , 013607 ( 2000 ) ; a. polkovnikov , _ ibid . _ * 68 * , 033609 ( 2003 ) ; k.w .
mahmud , a.m. leung and w.p .
reinhardt , e - print cond - mat/0403002 .
see e.g c. a. sackett , j. m. gerton , m. welling and r. g. hulet , phys .
lett . * 82 * , 876 ( 1999 ) ; j. m. gerton , d. strekalov , i. prodan and r. g. hulet , nature 408 , * 692 * ( 2000 ) ; e. a. donley , n. r. claussen , s. l. cornish , j. l. roberts , e. a. cornell , c. e. wieman , nature * 412 * , 295 ( 2001 ) . l.d .
carr , c.w .
clark and w. p. reinhardt , phys .
a 62 , 063611 ( 2000 ) ; g.m .
kavoulakis , _ ibid . _ * 67 * , 011601r ( 2003 ) ; r. kanamoto , h. saito , and m. ueda , _ ibid . _ * 67 * , 013608 2003 ; * 68 * , 043619 2003 ; phys
lett . * 94 * , 090404 2004 .
very recently , results similar to those described in sec .
iii of the present paper albeit in the @xmath210 limit have been obtained for a toroidal bose - einstein condensate .
see a. montina and f.t .
arecchi , phys .
a * 71 * 063615 ( 2005 ) .
more precise estimates can be obtained considering the analytical solution to the nlse on a ring domain @xcite .
note that in this continuous model an `` infrared '' threshold related to the finite size of the system is naturally present .
conversely , the `` ultraviolet '' crossover related to the discreteness of the lattice has to be estimated requiring that @xmath212 .
this is obtained setting @xmath213 where @xmath214 is a state with a non vanishing projection on all of the @xmath105 .
a suitable choice is @xmath215 .
note that @xmath216 is perfectly equivalent to @xmath105 , as an eigenstate of @xmath29 .
however , the @xmath217 s are in some sense phase coherent , in that their phases are all aligned to that of the reference state @xmath214 , i.e. @xmath218 . in ref .
@xcite all of the quantum observables are plotted vs. @xmath219^{-1}$ ] . here
we prefer to adopt @xmath14 , which makes easier to put into evidence the scaling properties of the plotted quantities . also note that our definition of @xmath114 differs from that in ref .
@xcite by an irrelevant yet graphically convenient normalization factor . | we study the ground state properties of the bose - hubbard model with attractive interactions on a @xmath0-site one - dimensional periodic necklace - like lattice , whose experimental realization in terms of ultracold atoms is promised by a recently proposed optical trapping scheme , as well as by the control over the atomic interactions and tunneling amplitudes granted by well - established optical techniques .
we compare the properties of the quantum model to a semiclassical picture based on a number - conserving su@xmath1 coherent state , which results into a set of modified discrete nonlinear schrdinger equations .
we show that , owing to the presence of a correction factor ensuing from number conservation , the ground - state solution to these equations provides a remarkably satisfactory description of its quantum counterpart not only as expected in the weak - interaction , _ superfluid regime _ , but even in the deeply quantum regime of large interactions and possibly small populations . in particular ,
we show that in this regime , the delocalized , schrdinger - cat - like quantum ground state can be seen as a coherent quantum superposition of the localized , symmetry - breaking ground - state of the variational approach .
we also show that , depending on the hopping to interaction ratio , three regimes can be recognized both in the semiclassical and quantum picture of the system . |
rate coefficients are necessary and sufficient information in the rate laws of kinetic phenomena .
to predict the average behavior of populations with these laws , it is essential to reliably estimate the rate coefficient .
when fluctuations away from the average are significant in irreversible decay , predictions depend on a distribution of rate coefficients or a time varying rate coefficient @xcite .
the terms static and dynamic disorder describe these theoretical constructs and the underlying kinetic process @xcite .
examples of disordered kinetics include enzyme catalyzed reactions @xcite , the dissociation of single molecules during pulling experiments @xcite , and hydrogen bond breaking @xcite . however , static and dynamic disorder are not mutually exclusive classifications of a process @xcite or straightforward to assign _ a priori _
@xcite , which raises fundamental and practical questions .
how can we measure disorder from kinetic data when microscopic heterogeneity manifests in the rate coefficient ? or
, when is the rate coefficient well defined ?
when is traditional kinetics valid ? answering these questions requires the ability to measure disorder .
an example of recent progress is an information theoretic measure of dynamic disorder @xcite .
such measures could enable new methods for minimizing the statistical description of rate coefficients and maximizing their predictive fidelity for data collected from simulations and experiments .
in this work we demonstrate a theoretical framework for this purpose that applies to irreversible decay .
central to the theory is a measure of temporal and spatial fluctuations in the rate coefficient , when the disorder is static , dynamic , or both .
our main result is an inequality that only reduces to an equality in the absence of disorder .
this result is a necessary condition for the traditional kinetics of irreversible first - order decay to be valid and the associated rate coefficients to be unique .
consider a population of species irreversibly decaying over time .
the survival probability , @xmath0 , is the probability the initial population survives up to a time @xmath1 .
this probability defines the observed or effective rate coefficient , @xmath2 , through the differential rate law @xmath3 characterizing the first - order decay of the entire population .
if individual members of the population are experimentally indistinguishable and the overall decay of @xmath0 is non - exponential @xcite , a disordered model for @xmath2 may be necessary .
non - exponential decay can imply an underlying mechanism where members decay in different structural or energetic environments , or a local environment that fluctuates in time @xcite . for irreversible decay , the survival probability is the ratio of the number of members in the population at a time @xmath1 , @xmath4 , and the number initially , @xmath5 . in terms of the effective rate coefficient
it is @xmath6 survival probabilities are the input to the theory here and a common observable ; they are measurable from computer simulations and experiments .
while measurements of @xmath0 may have non - negligible statistical fluctuations , we use non - fluctuating survival probabilities to simplify the presentation of our theoretical results .
one purpose of the theory here is to measure the fluctuations in the statistical parameter @xmath2 for statically and dynamically disordered kinetics . for activated escape on an energy surface , measures of the fluctuations in @xmath2
characterize the mechanism of decay over a distribution of barriers or a single barrier whose height varies in time .
our approach is to use the fisher information , which is a natural measure of the ability to estimate statistical parameters from the probability distribution of a fluctuating observable .
colloquially , the fisher information is the amount of `` disorder '' in a system or phenomenon @xcite , but is it the amount of static and dynamic disorder in a kinetic process ? to answer this question
, a relationship between the fisher information and rate coefficients , the parameters in kinetic phenomena , is necessary .
we relate @xmath2 and a modified form of the fisher information , expressed in terms of the survival probability of a decaying population through equation [ eqn : key ] . assuming an exact , non - fluctuating survival probability varying smoothly over time describes the evolution of the population , a reasonable form of the fisher information at time @xmath1 is @xmath7 which shows the direct relation to the effective rate coefficient .
a more general form of the fisher information may be necessary when there are statistical fluctuations in @xmath0 @xcite .
rearranging equation [ eqn : fisherinfo ] gives @xmath8 the remainder of our results stem from this connection between the fisher information and the statistical parameter , @xmath2 , of interest in disordered kinetics .
a particularly useful aspect of the fisher information , in general , is that it naturally defines a `` statistical distance '' between probability distributions .
though the notion of statistical distance was originally applied to quantum states @xcite , it is applicable to any two probability distributions . with the definition of the fisher information above
, we can measure the statistical distance between survival probabilities and relate it to the observed rate coefficient . if we take a microscopic , stochastic perspective , the rate coefficient of decay , @xmath2 , is the transition probability per unit time . in a given time interval between @xmath1 and @xmath9 ,
it is the ratio of the number of reacted molecules @xmath10 $ ] and the number of unreacted molecules @xmath11 @xmath12 from this discrete form of equation [ eqn : key ] and the discrete form of the relation between the time - dependent rate coefficient and the fisher information in equation [ eqn : fisherinrate ] , we define a dimensionless statistical distance @xmath13 ^ 2 = \left[\delta \ln s\right]^2\ ] ] between the logarithm of the survival probabilities at @xmath1 and @xmath14 .
if the fractional change in the population from one time to the next , @xmath15 , is a fixed value , the rate coefficient is independent of time and the distance between @xmath16 and @xmath17 is constant . we can interpret the statistical distance as a criterion for the distinguishability of @xmath16 at two times , since it is zero if the survival probability is time independent , or as the square of the transition probability @xmath18 during @xmath19 . during an irreversible decay process from an initial time @xmath20 to a final time @xmath21 , and letting @xmath22 , we can integrate the arc length of the logarithmic survival curve @xmath16 @xmath23 to get the statistical length , @xmath24 , measuring the cumulative rate coefficient .
the statistical length is infinite for an infinite time interval @xmath25 .
another useful quantity , related to the statistical length , is the fisher divergence , @xmath26 .
the fisher divergence we define as @xmath27 the time integrated square of the rate coefficient ( times the magnitude of the time interval ) . both the statistical length and the fisher divergence are cumulative properties of the rate coefficient , identified through equation [ eqn : fisherinrate ] for irreversibly decaying populations .
these properties for the history of @xmath2 satisfy an inequality @xmath28 which is our main result .
an analogous inequality in finite - time thermodynamics is a bound on the dissipation in an irreversible process @xcite .
we will show the inequality above measures disorder in kinetic phenomena and is an equality in the absence of both static or dynamic disorder .
some evidence for this comes from the inequality @xmath29 in terms of the observed rate coefficient @xmath30 ^ 2 & \geq & 0.\end{aligned}\ ] ] for decay processes with a time independent rate coefficient , there is no disorder , and the bound holds @xmath31 . while this finding is suggestive , we can show more concretely that the inequality measures the variation of the rate coefficient in irreversible decay kinetics .
we turn to this now and show the inequality measures the amount of disorder during a given time interval . as a proof - of - principle
we apply the theory to widely used kinetic models for populations decaying non - exponentially .
note , however , that the theory is model independent .
to demonstrate the inequality measures dynamic disorder , we consider the kohlrausch - williams - watts ( kww ) stretched exponential survival function .
this empirical model is a widely used phenomenological description of relaxation in complex media .
it has been applied to the discharge of capacitors @xcite , the dielectric spectra of polymers @xcite , and more recently the fluorescence of single molecules @xcite .
the survival probability , @xmath32 $ ] , has two adjustable parameters @xmath33 - a characteristic , time - independent rate coefficient or inverse time scale - and @xmath34 measuring the `` cooperativity '' of decay events - being nearer or equal to @xmath35 for independent decay events and nearer @xmath36 for coupled decay events ( figure [ fig : kww ] ) . fitting data to this function when the decay is non - exponential implicitly assumes the kinetics are dynamically disordered .
figure [ fig : kww](a ) shows analytical results for how the natural logarithm of the survival function versus time depends on the value of @xmath37 .
there is a linear dependence on time only when @xmath38 , corresponding to exponential kinetics and a time independent rate coefficient @xmath39 .
for all other @xmath37 values the survival function decays faster than exponential before , and slower than exponential after , @xmath40 .
these stretched exponential decays have a time - dependent rate coefficient @xmath41 when @xmath42 from figure [ fig : kww](b ) .
the effective rate coefficient is the ( negative of the ) slope of the graph of @xmath16 versus @xmath1 . through the rate law ,
the decrease in @xmath2 over time indicates a decrease in the rate of decay ; this also means the number of species decaying in a unit time interval is decreasing .
furthermore , if we interpret the decay as an activated process over an energy barrier @xcite , a decrease in @xmath2 implies an increase in barrier height and slowing rate of escape . in the limit of long times , the rate coefficient @xmath2 is effectively constant , regardless of @xmath37 .
the value of @xmath37 determines the rate at which @xmath2 reaches this limit : smaller @xmath37 values lead to a more rapid decline of @xmath2 .
( color online)(a ) logarithmic kohlrausch - williams - watts survival curves @xmath16 for stretched exponential decay with @xmath43 and @xmath44 , @xmath45 , @xmath46 , and exponential decay @xmath47 .
( b ) the rate of decay of the observed rate coefficient @xmath2 over time depends on @xmath37 , @xmath48 .
( c ) @xmath49 over the range of @xmath37 values for the final times of @xmath50 , @xmath51 , @xmath52 , @xmath53 and an initial time of @xmath54 , in arbitrary time units .
the difference is scaled by an arbitrary factor of @xmath55 , @xmath56 , @xmath57 , and @xmath35 for a direct comparison .
( d ) @xmath58 versus the disorder parameter @xmath37 on a logarithmic scale .
@xmath59 for all @xmath37 and any time interval when there is no dynamic disorder
when @xmath47 and @xmath60 . ] from the effective rate coefficient , we find the statistical length @xmath61 and , for @xmath62 , the divergence @xmath63 for @xmath64 , the divergence is @xmath65 . from these results for stretched exponential kinetics of the kww type , we find @xmath66 ( dropping the subscript kww for clarity ) and can show the inequality measures the amount of dynamic disorder associated with the empirical parameter @xmath37 .
the temporal variation of @xmath2 on the observational time scale dictates the magnitude of @xmath67 .
since @xmath37 determines how strongly @xmath2 varies in time , we show @xmath68 over the range of @xmath37 for select final times in figure [ fig : kww](c ) and ( d ) .
the plots are at four different final times @xmath69 , @xmath70 , @xmath71 , @xmath72 and an initial time @xmath54 in arbitrary units .
most striking in figure [ fig : kww](c ) and ( d ) , is the maximum in @xmath67 .
we see the maximum of the inequality is at a @xmath37 greater than or equal to @xmath73 for all non - zero time intervals and all @xmath21 . at final times less than @xmath74 ,
the difference is symmetric with a maximum at @xmath75 . with increasing @xmath21 ,
the curve becomes asymmetric and the maximum shifts to higher @xmath37 values . in part ,
the maximum comes from @xmath67 being zero at the limiting values of @xmath37 where the effective rate coefficient is time - independent . when @xmath38 , the decay is exponential , @xmath76 , and @xmath59 for all final and initial times @xmath77 .
the effective rate coefficient is also zero when @xmath78 , though this limit is typically excluded from the kww model ; however , when the rate coefficient is zero the equality again holds . since @xmath67 must be zero at the limits of the @xmath37 range , and
is non - zero between these limits , there must be a maximum at intermediate @xmath37 values . in this case , and the model we discuss next , there is a single maximum in @xmath68 versus @xmath37 . a single maximum results when the decay is monotonic , and non - exponential , with a time - varying rate coefficient between two limits where the rate coefficient is constant .
however , if the decay is more complex , or if the kinetics are driven , more maxima are possible .
the maximum in @xmath67 shown in figure [ fig : kww](c ) and ( d ) is also the result of how the history of @xmath2 depends on @xmath37 in this model .
panel ( b ) shows that decreasing @xmath37 also increases the difference between @xmath79 and @xmath80 , which would imply an increase in the difference between the barrier heights at @xmath21 and @xmath20 for an activated process . as @xmath37 decreases from one to zero , this increase in the variation of the rate coefficient initially leads to a greater inequality between @xmath26 and @xmath81 .
however , as @xmath37 continues to decrease , the effective rate coefficient also becomes a more rapidly decreasing function of time ( figure [ fig : kww](b ) ) .
consequently , significant changes in the magnitude of @xmath2 can become relatively small portions of the observational time window .
for example , at @xmath82 the effective rate coefficient varies little over the majority of the observation time for @xmath83 , and @xmath84 .
but , for @xmath85 , @xmath2 is more strongly dependent on time over the interval @xmath86 and @xmath87 . overall , the magnitude of the @xmath68 reflects the both the variation of @xmath2 in time and the fraction of the observation time where those changes occur .
the maximum in the inequality versus @xmath37 reflects the competition of these factors .
the difference between @xmath26 and @xmath88 also measures the amount of static disorder in irreversible decay kinetics over a time interval . as a tractable example , consider the model with two experimentally indistinguishable states , @xmath89 and @xmath90 @xcite .
these states might be the internal configurations of a reactive molecule that can interconvert through isomerization or two energy levels leading to unimolecular dissociation in the gas phase .
the states are in equilibrium and so a population occupying these states will undergo transitions between them .
a population occupying the @xmath89 and @xmath90 states can also decay irreversibly to @xmath91 with rates characterized by two time - independent rate coefficients , @xmath33 and @xmath92 .
if the individual decay out of @xmath89 and @xmath90 have different characteristic rates , it may be necessary to characterize the collective decay from the @xmath89 states with both rate coefficients .
since only two rate coefficients @xmath33 and @xmath92 represent the `` fluctuations '' in the rate coefficient , this is the simplest case of potentially statically disordered kinetics ( figure [ fig : plonka ] ) , and is readily generalizable to a continuum of @xmath89 states .
the phenomenological rate equations are @xmath93\nonumber { \frac{\displaystyle d n_{a'}}{\displaystyle d t } } & = & -(\omega ' + a\lambda)n_{a ' } + \lambda n_a\end{aligned}\ ] ] for the number of species in state @xmath89 , @xmath94 , and @xmath90 , @xmath95 .
these numbers are exact and not fluctuating in time . whether the decay from @xmath89 and @xmath90 is exponential or non - exponential depends on the relative rates of interconversion and reaction .
two limiting solutions to these rate equations are of particular interest @xcite .
each limit affects the analytical form of the survival probability for the population of the @xmath89 and @xmath90 states @xmath96 ( color online)(a ) logarithmic survival plot for the kinetic scheme in the text for @xmath97 and @xmath98 .
the decay kinetics are statically disordered when @xmath99 , leading to the non - exponential relaxation of @xmath0 and the non - linearity of @xmath16 in time .
the kinetics are exponential when @xmath100 and @xmath16 versus @xmath1 is linear ( black line ) .
there is no disorder when @xmath99 at long times when the slower channel dominates decay .
( b ) the graph of @xmath67 reflects these facts , showing this difference measures the amount of static disorder over a range of relative decay rates @xmath101 ( @xmath98 ) and time intervals @xmath102 , @xmath103 , and @xmath104 on a logarithmic scale .
the inset is on a linear scale for @xmath105 .
@xmath67 is zero if decay is exponential ( when @xmath100 or when @xmath106 ) , but nonzero otherwise . ]
first , in the limit where interconversion between @xmath89 and @xmath90 is fast compared to the reaction , @xmath107 , the survival probability is @xmath108}.\ ] ] and the population decays exponentially . from equation [ eqn : key ] , the observed rate coefficient @xmath2 is time independent and a weighted - average of @xmath33 and @xmath92 , @xmath109 . from this limiting survival probability
it follows that the divergence and length squared are equal , @xmath110 .
this result illustrates that to satisfy the bound , the @xmath89 states must be kinetically indistinguishable , which happens here because the population collectively decays to products from one effective state .
second , in the limit @xmath111 , the transitions between the @xmath89 states are slow relative to those of reaction .
when the decay from @xmath89 and @xmath90 is independent , the net decay is bi - exponential @xmath112 the logarithmic dependence of this survival function versus time in figure [ fig : plonka](a ) is linear , and the decay of @xmath0 is effectively exponential , if the decay from the @xmath89 and @xmath90 states have the same rate coefficients @xmath100 . in this case
, we again find @xmath113 .
however , this condition breaks when the kinetics are statically disordered @xmath99 and decay from @xmath89 and @xmath90 competitively depletes the population , leading to the non - exponential decay in figure [ fig : plonka](a ) .
the @xmath0 in this regime causes the observed rate coefficient @xmath2 to vary in time .
only when there is static disorder in this model does @xmath81 not equal @xmath26 .
the relative magnitude of @xmath33 and @xmath92 , which we measure with the ratio @xmath101 , and the time interval , @xmath114 , control the degree of static disorder .
the difference @xmath58 is sensitive to both of these factors .
figure [ fig : plonka](b ) illustrates the dependence of the inequality on @xmath101 and assumes the survival function is valid , @xmath115 , over the range of interest .
varying @xmath101 shows the difference between the divergence and length squared is positive , has a maximum that depends on the time interval , and tends to zero with increasing @xmath101 . in the figure ,
the difference is shown for three final times @xmath116 , @xmath117 , and @xmath70 and @xmath54 , in arbitrary units .
a logarithmic scale allows direct comparison of these curves ( the vertical axis in the inset has a linear scale ) .
an inequality between @xmath81 and @xmath26 gives quantitative insight into the mechanism of kinetic processes , when only the survival of the entire population is available . from figure
[ fig : plonka](b ) the inequality grows but as the disparity in the rate coefficients of each decay path increases further , there becomes a separation of timescales , one path decays on times that are increasingly short in comparison to the time interval of interest .
just as for the kww model , there is a maximum in the inequality between the divergence and length squared . in the limit @xmath106 , @xmath118 , indicating exponential decay and a single characteristic rate coefficient .
the maximum is understandable from the survival function : with increasing @xmath33 or @xmath1 , the term @xmath119 decreases until , ultimately , the decay channel contributes negligibly to the overall decay of the population
the kinetics are exponential with the characteristic rate of the slower decay channel , @xmath120 .
the survival function in equation [ eqn : s ] becomes @xmath121 .
together these findings demonstrate the inequality is a statement about whether a single rate coefficient is sufficient for irreversible rate processes .
the rate coefficient is a unique constant only when @xmath118 , and the decaying states are kinetically indistinguishable . otherwise , the inequality between @xmath26 and @xmath81 measures the dispersion in the rate coefficient though it does not distinguish between statically and dynamically disordered kinetics .
the inequality both results from , and measures , the heterogeneity of microscopic environments that ultimately produce disordered kinetics . for activated processes over an energy barrier , heterogeneous environments can cause either fluctuating barrier heights or a distribution of barriers through steric / energetic effects . consequently
, the inequality could be used to investigate the local structure around reaction sites in disordered media and the dependence of rate processes on energetic or structural effects of the environment @xcite .
the equality signifies when the interconversion between decaying states is faster than decay , when there is no microscopic heterogeneity to cause a distribution of rate coefficients , and when observations of the process are on a time scale where fast decay processes are insignificant to the longer term survival of the population . under these circumstances traditional kinetics holds
in summary , the inequality between @xmath88 and @xmath26 measures the amount of static and dynamic disorder in irreversible decay kinetics .
the inequality follows from the relation between the rate coefficient for decay and the fisher information .
it relies on two functions of the fisher information - a statistical length and the fisher divergence for the decaying population - which are properties of the history of fluctuations in the rate coefficient .
this relationship is a quantitative signature of when decay kinetics are accurately described by a fluctuating rate coefficient ( measured by the inequality ) and when a single , unique rate constant is sufficient .
traditional kinetics corresponds to the condition @xmath122 .
our results suggest minimizing the difference between @xmath88 and @xmath26 , when it is desirable to minimize the statistical description of kinetic phenomena with rate coefficients or to maximize the predictive fidelity of rate coefficients extracted from experimental or simulation data . in the future , this framework may also be useful in the analysis of , potentially complex , chemical reactions with fluctuating rates .
this work was supported as part of the non - equilibrium energy research center ( nerc ) , an energy frontier research center funded by the u.s .
department of energy , office of science , basic energy sciences under award # de - sc0000989 .
this material is based upon work supported in part by the u.s .
army research laboratory and the u.s .
army research office under grant number w911nf-14 - 1 - 0359 .
28ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _ _ ( , ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , ed .
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | rate coefficients can fluctuate in statically and dynamically disordered kinetics . here
we relate the rate coefficient for an irreversibly decaying population to the fisher information . from this relationship
we define kinetic versions of statistical - length squared and divergence that measure cumulative fluctuations in the rate coefficient .
we show the difference between these kinetic quantities measures the amount of disorder , and is zero when the rate coefficient is temporally and spatially unique . |
quite a few nearby galaxies seem to have compact radio cores in their nuclei , prominent cases are the milky way ( sgr a * ) , m31 and m81 .
those radio cores resemble the cores of radio loud quasars , showing a very high brightness temperature and a flat to inverted radio spectrum that extends up to submm wavelengths .
several models have been developed to explain those cores in the context of black hole accretion : melia ( 1992a&b ) suggested a spherical accretion model for sgr a * and , what he called , m31*. falcke et al .
( 1993 ) and falcke , mannheim , & biermann ( 1993 , hereinafter fmb93 ) proposed an alternative model , where sgr a * was explained as the core of a radio jet , fed by an underluminous , starving accretion disk ( see also falcke 1996a , for a review , and falcke & heinrich 1994 for m31 * ) .
this model evolved into the jet - disk symbiosis approach ( falcke & biermann 1995 , hereinafter fb95 ; falcke 1996b ) that unified the explanation for radio cores of quasars ( falcke , malkan , & biermann 1995 , hereinafter fmb95 ) , galactic jet sources ( falcke & biermann 1996 ) and sgr a * into a single picture .
the basic idea was to postulate that black holes , jets and disks form closely coupled systems with little variations from one system to another except for the accretion rate .
recent vlbi ( bietenholz et al . 1996 ) , multiwavelengths ( ho , filipenko & sargent 1996 ) and submm observations ( reuter & duschl 1996 ) have now shown that the radio core in m81 which in analogy to sgr a * and m31 * , we will label m81 * hereafter is very similar to sgr a*. this wealth of data now makes m81 * an excellent laboratory to test the jet - disk symbiosis in detail .
one of the interesting findings of the vlbi observations ( de bruyn et al . 1976 ; bartel et al .
1982 ; bietenholz et al .
1996 ) is that the radio core of m81 * is elongated and its size is intrinsically frequency dependent with @xmath0 .
a frequency dependent size was one of the basic predictions of the jet model and excludes any homogenous , optically thin models ( e.g. as in duschl & lesch 1994 ) .
however , the measured size - frequency relation is slightly shallower than in the blandford & knigl ( 1979 , hereinafter bk79 ) free jet model adopted by fb95 , and the radio spectrum of sgr a * and m81 * are both slightly inverted rather than flat as in the simple model . here
we show , that a self - consistent treatment of the bk79 model implies a weak acceleration of the bulk jet flow due to its longitudinal pressure gradient which can naturally explain the observed size- and spectral indices in compact radio cores . as in fb95
, we shall describe the jet core within the framework of relativistic gasdynamics of a relativistic gas with adiabatic index @xmath1 .
we consider only the supersonic regime and impose as boundary condition , that the jet expands freely with its initial soundspeed @xmath2 without any lateral gradients behind the jet nozzle .
this leads to the familiar conical jet with @xmath3 ( bk79 ) , where adiabatic losses due to lateral expansion need not be considered . for a confined jet scenario ( e.g. sanders 1983 ) at larger scales the adiabatic losses could
, however , be quite severe with energy densities scaling as @xmath4 .
nonetheless , even in a simple , free jet at least some work due to expansion will be done , because there is always a longitudinal pressure gradient , which will accelerate the jet along its axis .
this acceleration is described by the z - component of the modified , relativistic euler equation ( e.g. pomraning 1973 , eq .
9.171 ) in cylindrical coordinates where we set @xmath5 and @xmath6 @xmath7 here , @xmath8 ist the enthalpy density of the jet , @xmath9 is the internal energy density , @xmath10 is the particle density , and @xmath11 is the pressure in the jet ( all in the local rest frame ) . for
the ` maximal jets ' the radiatively most efficient type of jet ( fb95 ) to be discussed here , we demand the equivalence of internal energy and kinetic ( or better rest mass ) energy @xmath12 , hence @xmath13 and @xmath14const at the sonic point @xmath15 . in the free jet with conical shape
the energy density evolves as @xmath16 and we do not consider any loss mechanisms other than adiabatic losses due to the longitudinal expansion . using the relations mentioned above , the euler equation becomes @xmath17 with @xmath18 . for @xmath19 this is analogous to the euler equation for the well known isothermal , pressure driven winds , with the wind speed replaced by the proper jet speed and the sound speed is constant and fixed by the required equivalence between internal and rest mass energy density at a value @xmath20 . the asymptotic solution of eq .
[ euler2 ] for @xmath21const and @xmath22 is @xmath23 .
if we ignore terms of the order @xmath24 we could approximate the solution by @xmath25 .
for the following calculations we will , however , use the exact , numerical solution to eq . [ euler2 ] ( @xmath26const ) , but the deviation from the approximate , asymptotic solution is rather small .
the basic ideas how to derive the synchrotron emissivity and basic properties of a jet in a coupled jet - disk system have been described extensively in fmb93 and fb95 .
the power @xmath27 of _ one _ jet cone is a fixed fraction of the disk luminosity , relativistic particles and magnetic field @xmath28 are in equipartition within a factor @xmath29 which we hereafter set to one , and the energy fluxes are conserved along the conical jet ( for a moment we will forget the adiabatic losses ) .
mass conservation requires that the mass loss in the jet @xmath30 is smaller than the mass accretion rate in the disk @xmath31 , thus @xmath32 , while mass and energy of the jet are coupled by the relativistic bernoulli equation @xmath33 ( fmb93 ) . here
we will make use of the same logic and notation but with three changes with respect to fb95 , namely ( 1 ) applying the velocity law eq .
[ v ] , ( 2 ) neglecting the energy contents in turbulence and ( 3 ) considering only a quasi monoenergetic energy distribution for the electrons at a lorentz factor @xmath34 ( usually @xmath35 ) .
the latter is indicated by the steep submm - ir cut - off of sgr a * ( zylka et al .
1995 ) and probably also m81 * ( reuter & lesch 1996 ) which is different from typical blazar spectra and precludes an initial electron powerlaw distribution ( as in fmb93 ) .
the semi - opening angle of the jet is @xmath36 , and the magnetic field in the comoving frame of a maximal jet with the given sound speed , @xmath37 erg / sec , @xmath38 , and @xmath39 cm becomes ( see eq . 19 in fb95 ) @xmath40 the number of relativistic electrons that are to be in equipartition with the magnetic field are a fraction @xmath41 of the total particle number density , and the energy density ratio between relativistic protons and electrons is @xmath42 . from the energy equation @xmath43 we find that the characteristic lorentz factor and electron density required to achieve equipartition are @xmath44 finally , to incorporate adiabatic losses we now have to make the following transitions with respect to the equations in fb94 @xmath45 the local spectrum of the jet will be @xmath46 between the synchrotron self - absorption frequency @xmath47 and the characteristic frequency @xmath48 = @xmath49 ( here we will use an average pitch angle @xmath50 ) . for the maximal jet using equations [ b ] , [ gamma ] , [ n ] and [ adloss ] , we have in the observers frame @xmath51 with the doppler factor @xmath52 for an inclination @xmath53 of the jet axis to the line of sight . to find the local synchrotron self - absorption frequency @xmath54 we have to solve the equation @xmath55 and transform into the observers frame .
the jet radius is @xmath56 and the synchrotron self - absorption cross section for a monoenergetic electron distribution is given by @xmath57 @xmath58 @xmath59 , yielding @xmath60 if we now use the velocity profile eq .
[ v ] we can invert eq . [ nuc ] to find the size of the jet as a function of observed frequency .
for the asymptotic regime @xmath61 , @xmath62 is approximated for a fixed inclination angle to within a few per - cent by a powerlaw in @xmath63 , such that @xmath64 where the parameters are @xmath65 au , @xmath66 ghz , @xmath67 and @xmath68 au for @xmath69 .
the total spectrum of the jet is obtained by integrating the synchrotron emissivity @xmath70 along the jet : @xmath71 , which can again be approximated using powerlaws : @xmath72 for @xmath73 mpc , and @xmath74 ) respectively .
this means that the resulting spectrum is no longer flat but tends to be inverted with @xmath75 for @xmath76 .
the spectral index , like the size index , is a function of the inclination angle of the jet and tends to become flatter for smaller @xmath53 .
we can now apply this formalism to the situation in m81 * , where we have five observed quantities , namely the time averaged spectral index @xmath77 ( reuter & lesch 1996 ) , the flux @xmath78 mjy , the size @xmath79 au , the frequency dependence of the size @xmath80 with @xmath81 ( bietenholz et al . 1996 ) , and the bolometric luminosity of the nuclear source ( ho , filipenko , & sargent 1996 ) which is @xmath82 erg / sec . most of the model parameters are already well constrained . as the cosmic - ray ratio @xmath83 and the fractional number of electrons
@xmath84 are almost interchangeable we keep the latter fixed at @xmath85 .
the jet - disk ratio @xmath86 was determined in fmb95 to be @xmath87 for a flat spectrum . to bring this model with an inverted spectrum and a peak at submm wavelengths to the same fluxscale one
has to increase this value to @xmath88 where disk luminosity and jet have equal powers ; here we do not intent to change this value .
the size of the nozzle @xmath89 enters only weakly into the equations ; if we assume that the high - frequency cut - off which is somewhere around 1000 ghz correspond to the size scale of the nozzle ( in analogy to sgr a * in falcke 1996 ) we find by extrapolation that @xmath90 au as an order of magnitude estimate . for a nozzle size of 10 - 100 gravitational radii
this correponds to black hole masses of @xmath91 which is consistent with dynamical estimates ( ho et al .
1996 ) .
this leaves us with two main input parameters : @xmath53 and @xmath92 .
the cosmic - ray ratio determines the electron lorentz factor and is well constrained ; by definition @xmath93 and from the condition @xmath94 we here have @xmath95 ( eqs .
[ nuc]&[nus ] ) .
figure 1 shows the predicted sizes and fluxes of the model for various @xmath92 and @xmath53 , indicating that in m81 * @xmath96 and @xmath97 ( fmb95 used @xmath98 for quasars ) .
the predicted spectral index then is @xmath99 and the predicted size index is @xmath100 . considering the simplicity of the model , those values are reasonably close to the observed ones .
the jet - disk symbiosis emission model by fmb93 and fb95 which was initially developed to explain the radio core of sgr a * and radio loud quasars can also explain the nuclear radio source in m81 in detail . simply by scaling the accretion disk luminosity by several powers of ten to the observed value properly predicts flux , size , spectral index , and size index of the vlbi radio source . if one inserts m81 * into the universal @xmath101radio correlation presented in falcke & biermann ( 1996 ) it falls right onto the line that connects sgr a * and radio loud quasars .
this suggests that all these sources are powered by a very similar ( jet / disk ) engine . in this context
it is quite interesting to note that m81 * and sgr a * are in spiral galaxies , yet appear radio loud in these diagrams .
m81 also has recently shown double - peaked broad emission lines , a feature usually seen only in radio - loud galaxies ( bower et al .
1996 ) .
like in quasars and sgr a * , one has to choose the most efficient , equipartition jet model to explain m81 * , with high internal energy and a jet power that equals the disk luminosity .
this leaves one with no free parameters for m81 * other than the jet inclination angle which according to the current model is around @xmath102 .
some caution is , however , neccessary , as the pa of the vlbi component seems to change with frequency , something we have ignored here .
it is not clear whether this indicates bending , helical motion , or an extrinsic effect and thus is difficult to interpret .
some of these effects could change the results slightly but not to an order of magnitude . in this _
letter _ we have considered for the first time the effects of the longitudinal pressure gradient in the bk79 and fb95 jet model . without any sophisticated mechanisms , this gradient alone will already lead to a moderate acceleration of the jet to bulk lorentz factors of 2 - 3 .
consequently , the lorentz factor will increase towards lower frequencies and for a fixed viewing angle outside the boosting cone the flux will become doppler - dimmed with respect to higher frequencies .
this can provide a natural explanation for the inverted spectrum seen in compact radio cores in general and for the size / frequency relation of m81*. such a mildly accelerating jet may also be of interest for quasar and bl lac radio cores or galactic jet sources .
the biggest advantage , however , is that the jet velocity , which was previously a free parameter , now is fixed by a simple physical model .
another interesting point of the presented model is the usage of an initial , quasi monoenergetic electron distribution for which we find @xmath103 .
this not only reduces another free parameter ( the electron distribution index ) but also can explain the submm - ir cut - offs seen in sgr a * and m81*. if the jet in m81 * does not have strong shocks to re - accelerate those electrons into the ususal power - law , the jet would have size - dependent high - frequency cut - offs and thus explain the apparent absence of extended vlbi components in m81*. this research was supported by nasa under grants nagw-3268 and nag8 - 1027 .
i thank an anonymous referee for helpful comments , and p. biermann , w. duschl , h. lesch , p. mezger and a. wilson for numerous discussions .
xx bartel , n. , shapiro , i.i . , corey , b.e . , et al .
1982 , apj 262 , 556 bietenholz , m.f .
, bartel , n. , rupen , m.p . ,
conway , j.e . ,
beasley , a.j .
1996 , apj 604 , 28 blandford , r.d . , &
knigl , a. 1979 , apj 232 , 34 bower , g.a . ,
wilson , a.s . , heckman , t.m . ,
richstone , d.o .
1996 , aj , in press de bruyn , a.g . , crane , p.c . , price , r.m .
, & carlson , j.b .
1976 , a&a 46 , 243 falcke , h. 1994 , phd thesis , rfw universitt bonn falcke , h. 1996 , in `` unsolved problems of the milky way '' , iau symp .
169 , l. blitz & teuben p.j .
( eds . ) , kluwer , dordrecht , p. 163
falcke h. 1996 , to appear in : `` jets from stars and galactic nuclei '' , ed .
w. kundt , springer lecture notes ho l.c . ,
filipenko , a.v . , & sargent , w.l.w .
1996 , apj , in press melia , f. 1992a , apj 387 , l25 melia , f. 1992b , apj 398 , l95 pomraning g.c . 1973 ,
`` the equations of radiation hydrodynamics '' , pergamon press , oxford reuter , h .-
p . , & lesch , h. 1996 , a&a subm .
sanders , r.h .
1983 , apj 266 , 73 | in this paper we apply the jet - disk symbiosis model developed for sgr a * to m81 * the nucleus of the nearby galaxy m81 .
the model accurately predicts radio flux and size of m81 * for the observed bolometric luminosity of the nuclear source with no major free parameter except for the inclination angle .
we point out that the usually applied free , conical jet emission model implies a longitudinal pressure gradient that must lead to a moderate acceleration of the jet along its flow direction .
this , usually neglected , gradual acceleration naturally accounts for the inverted spectrum and the size / frequency relation of m81 * and may be a general feature of radio cores .
m81 * is so far the best case for a radio - loud jet nature of the compact radio core in the nucleus of a nearby spiral galaxy .
the fact that one can account for sgr a * and m81 * with the same model by simply changing the accretion rate , strongly supports the jet - disk symbiosis model as an explanation for the compact radio cores of galaxies in general . |
the non - thermal x - ray emission from large scale jets in nearby radio galaxies of the fanaroff - riley class i is expected to be common feature of such type of objects starting from its discovery in two close galaxies cen a or m 87 ( feigelson et al . 1981 , biretta et al .
this x - ray emission smoothly extends up to the radio energy range supporting the idea on its common origin in the synchrotron process of tev electrons ( hardcastle et al .
tev electrons are expected to meet strong soft radiation field , produced either in the synchrotron process or by the stars in the galactic bulge and also the microwave background radiation ( mbr ) , fro its comptonization to tev energies ( e.g. stawarz et al .
2003 , hardcastle & croston 2011 ) .
observations of persistent tev @xmath0-ray emission from such sources should provide important constraints on the high energy particles in the kpc scale jets .
however , since such electrons has to lose energy on a short time scale , they require an _ in situ _ acceleration process .
therefore , efficient acceleration of electrons should occur not only within the inner jet but also in the kpc scale jets .
in fact , the observed x - ray knots in the intermediate scale jets seems to provide conditions for electron acceleration to tev energies .
these might correspond to shocks formed in collisions of winds from compact objects with the jet plasma ( e.g. hardcastle et al .
2003 ) .
a jet , launched from the vicinity of the central super - massive black hole ( smbh ) , has to pass through the stellar bulge and globular cluster ( gc ) halo around central engine . since the number of stars in the central stellar cluster is expected to be huge , many of them have to be immersed in the jet plasma .
in fact , it has been argued that as a result of the star - jet collisions , stellar winds can provide large amount of barion matter into the jet ( e.g. komissarov 1994 , bowman et al .
1996 , laing & bridle 2002 , hubbard & blackman 2006 ) .
shocks formed in collisions of stellar winds with the plasma in the inner jet can be also responsible for the acceleration of electrons which , comptonizing stellar radiation , might be responsible for the production of @xmath0-ray flares ( e.g. bednarek & protheroe 1997 , barkov et al .
2010 , bosch - ramon et al .
2012 , araudo et al .
2013 , wykes et al .
also collisions of relativistic jet plasma with dense clouds can turn to the @xmath0-ray production ( e.g. blandford & knigl 1979 , dar & laor 1997 , beall & bednarek 1999 , purmohammad & samimi 2001 , araudo et al .
2010 , barkov et al .
these clouds might be fragments of the close to the jet supernova explosion blandford & knigl ( 1979 ) .
large number of objects , moving in the galactic bulge , are expected to be immersed within the kpc scale jets .
galaxies of different types are surrounded by spherical halos of gcs with the radii of several kpcs .
gcs are expected to produce winds which are the products of the mixture of the relativistic winds from a large population of millisecond pulsars ( msps ) within the gc and slow barionic winds from the red giant stellar population within gcs ( bednarek & sobczak 2014 ) . the mixed millisecond pulsar and stellar winds , expelled from the gc , interact with a relatively rare relativistic intermediate scale jet . as a result , a double shock structure forms around gc .
we argue that particles can be accelerated on such shocks to multi - tev energies .
they produce non - thermal radiation by interacting with the magnetic field in the jet ( the synchrotron radiation ) and with the soft radiation from the galactic bulge at the sub - kpc distance scale and also with the mbr at the a kpc distance scale via inverse compton process . in this paper
we explore the hypothesis that relativistic electrons appear in the intermediate scale jet as a result of their acceleration on the shocks formed in collisions of many compact objects ( red giants , massive stars or gcs ) with the jet .
it is assumed that at the kpc scale distances jets are already non - relativistic and the doppler boosting effects are small . under this assumption
, we calculate the synchrotron and inverse compton @xmath0-ray emission produced in the intermediate jet for different parameters of such scenario . as an example , we show that the high energy tev @xmath0-ray component in the cen a spectrum , observed by the hess collaboration ( aharonian et al . 2009 ,
sahakyan et al .
2013 ) , might originate within the intermediate scale jet as a result of acceleration of electrons in collisions of many compact objects with the jet plasma .
active galaxies are complicated systems surrounded at larger scale by a few important elements . the central smbh , with the mass in the range @xmath1 m@xmath2 , is surrounded by the bulge , which has quasi - spherical shape , containing late type stars within the radius of the order of @xmath3kpc .
the mass of the bulge is related to the mass of the smbh and its luminosity is of the order of a few tens percent of the whole galaxy .
many smbhs are surrounded by central clusters of luminous and massive stars .
such clusters have usually extended shape with the main axis in the galactic plane .
its mass and dimension is supposed to be determined by the destruction process of another galaxy by the massive galaxy harboring smbh . finally , active galaxy is surrounded by a halo of gcs within the distance scale of several kpc .
the number of gcs is related to the mass of the parent galaxy and the mass of the smbh .
the parameters of these stellar clusters around nucleus of active galaxy are discussed below .
nuclear clusters of massive stars are observed not only around smbhs in massive active galaxies but also in the center of our galaxy . in the case of our galaxy
such cluster is relatively small .
it has a half mass radius of @xmath3 pc and the mass of @xmath4 m@xmath2 ( e.g. schdel et al .
2014 ) in the case of the nearby active radio galaxy cen a , the massive stellar cluster has been formed around 50 - 100 myrs ago as a result of a merger process with a gas reach galaxy .
the observed star formation region is centered on the cen a nucleus .
it has the visual dimensions of @xmath53 kpc@xmath6 ( neff et al .
the current star formation rate has been estimated on 2 m@xmath2yr@xmath7 which will result in production of @xmath8 m@xmath2 of young stars ( wykes et al .
2014 ) . assuming the standard initial mass function of stars
, we estimate the total number of stars with masses above 20 m@xmath2 in the nuclear cluster in cen a on @xmath9 .
a small part of these stars can pass through the jet expelled from the super - massive black hole .
the rough estimate of the number of massive o and wr type stars , contained within the 1.5 kpc from the base of the jet with the opening angle of the order of 0.1 rad for the cluster with dimension mentioned above , is of the order of a hundred .
this simple estimate shows that many massive stars can interact with the jet plasma .
central regions of galaxies are also surrounded by the quasi - spherical concentrations of late type stars ( galactic bulges ) with typical dimensions of the order of @xmath31 kpc and luminosity of several percent of the total luminosity of the parent galaxy .
interestingly , the masses of bulges are related to masses of the central black holes ( for review on scaling formulas see e.g. czerny & nikoajuk 2010 ) . in the case of cen
a , the number of bulge stars within the jet has been estimated on @xmath10 ( wykes et al .
if only a small number of these stars are in the red giant phase , of the order of @xmath310@xmath11 , then the total number of red giants in cen a jet can be as large as @xmath12 .
as we mentioned in the introduction also gcs are expected to enter the jet from time to time at the kpc distances from the base of the jet .
in fact the number of observed ( or estimated ) gcs around galaxies seems large enough to guarantee encounters with the jet .
for example , in the case of our galaxy more than 150 has been discovered ( harris 2003 ) .
larger galaxies are expected to have more gcs .
the andromeda galaxy is expected to be surrounded by a halo of @xmath3500 gcs ( barmby & huchra 2001 ) but the giant radio galaxy , m87 , contains already @xmath313000 gcs ( mclaughlin et al .
it is shown that in active galaxies , containing super - massive black holes ( smbhs ) , the number of gcs in the galactic halo is correlated with the mass of the central black hole ( e.g. spitler & forbes 2009 , burkert & tremaine 2010 , harris et al .
for example , in the radio galaxy cen a the number of gcs is estimated on @xmath13 ( gltekin et al .
2009 ) .
we consider active galaxy which has a jet propagating from the smbh .
the jet has a conical structure on the intermediate scale .
the jet opening angle at a kpc distance scale is expected to be of the order of @xmath14 rad , see e.g. observations of the nearby active galaxy cen a ( e.g. clarke et al .
1992 , hardcastle & croston 2011 ) .
then , the solid angle subtended by the jet is , @xmath15 . since the number of compact objects surrounding central engine can be very large ( see above ) , many of the late type stars , massive stars , and gcs can pass through the jet ( see fig.1 on the left ) . as we mentioned above the number of these compact objects
is shown to be related to the mass of the central black hole .
therefore , it is expected that very massive black holes , which are on average more luminous , are also surrounded by a larger number of stars and stellar clusters .
for example , we estimate the number of gcs within the jets of two well known nearby active galaxies , cen a and m87 , on @xmath16 and @xmath370 gcs for the jet opening angle equal to 0.15 rad .
the power of jets in the nearby radio galaxies are typically of the order of @xmath310@xmath17 erg s@xmath7 ( e.g. in cen a , wykes et al .
2013 ) . the ram pressure of the jet plasma can be estimated from ( e.g. bednarek & protheroe 1997 ) , @xmath18 where the power of the jet is @xmath19 erg s@xmath7 , the distance from the base of the jet is @xmath20 kpc and @xmath21 is the velocity of light .
compact objects , mentioned above , are also surrounded by winds which pressure can be balanced by the pressure of the jet plasma . as a result , a double shock structure is expected to appear around the compact objects within the jet as already discussed in the case of different types of stars . in the case of gcs ,
hadronic winds from significant population of red giant stars mix efficiently with the winds from many millisecond pulsars ( msps ) within the gc . as a result , the wind emanates from the gc with the power determined by the energy output from the msp population within the gc .
this mixture of red giant / msp winds collide with the jet s plasma as shown schematically in fig . 1 ( on the right ) .
we argue that particles can be accelerated at the shocks from the jet site to multi - tev energies .
they produce non - thermal radiation by interacting with the magnetic field in the jet ( the synchrotron radiation ) and with the soft radiation from the galactic bulge at the kpc distance scale and also with the microwave background radiation ( mbr ) at the distance scale up to a few tens of kpcs ( the inverse compton radiation ) .
below we consider separately the case of collision of a star with the jet and a gc with the jet since these types of objects differ significantly in the parameters describing their winds .
as we argue above two different classes of stars are expected to surround the central engine of active galaxy at distances up to the order of a kpc .
they are massive stars from the nuclear stellar cluster and late type stars ( red giants ) from the galactic bulge .
in fact , collisions of the winds from massive stars with the inner jet plasma in active galaxies , as a possible scenario for acceleration of particles and their radiation , have been already discussed in bednarek & protheroe ( 1997 ) .
more recently , distraction process of red giant stars as a result of collisions with the inner jets has been also considered as the mechanism for @xmath0-ray flares
( e.g. barkov et al . 2010 and
bosch - ramon et al . 2012 ) .
such destruction process is possible only in the case of very powerful jets and at a few pc scale distances from the base of the jet . at large distances , where the pressure of the jet plasma is weak , the red giant star winds are strong enough to stop the jet plasma above the red giant surface . as a result ,
the scenario for collisions of red giant stars becomes similar to the scenario of collisions of massive stars with jets on the intermediate scale distances from the base of the jet .
recently , collision of different types of stars with the intermediate scale jets , as the mechanism for the non - thermal synchrotron emission from jets , has been discussed by wykes et al .
( 2014 ) with the application to cen a. first , we consider collisions of stars with parameters of the red giant stars .
the scenario for collisions of massive stars will be in principle the same .
we will comment on possible differences at the end of this section .
we assume that a small part of the bulge stars , of the order of @xmath22 , are in the phase of the red giant . as an example , we consider red giants with following typical parameters , the radius @xmath23 cm , the surface temperature @xmath24 k , the luminosity @xmath25 l@xmath2 , the mass loss rate @xmath26 m@xmath2 yr@xmath7 , and the wind velocity @xmath27 cm s@xmath7 . as discussed in sect . 2 , a large number of red giant stars , of the order of @xmath28 , is expected to be immersed in the intermediate jet of active galaxy on the distance scale of @xmath29 kpc (
see estimates for the numbers of bulge stars within the jet in wykes et al .
2014 ) . the maximum power , that can be extracted from the jet by such a shock structure around a single star , is estimated on , @xmath36 note that this power depends only on the parameters of the red giants star . since the total number of red giants in the jet can be quite large
, a significant energy of the jet can be extracted and eventually transferred to relativistic particles in the acceleration mechanism operating at the collision region of the jet and stellar winds .
the parameters of the massive stars , which can also enter the jet , are clearly much more extreme than expected for the red giants .
for example , the terminal velocities of the winds of massive o type stars can be @xmath37 km s@xmath7 and the mass loss rates @xmath38 m@xmath2 yr@xmath7 and of the wr type stars @xmath39 km s@xmath7 and the mass loss rates @xmath40 m@xmath2 yr@xmath7 , respectively ( lang 1992 ) .
stars in different phases of evolution can have very large mass loss rates of the order of @xmath41 m@xmath2 yr@xmath7 , e.g. asymptotic giant branch stars or luminous blue variable stars ( see table 1 in wykes et al .
these massive stars , although more rare than red giants , will produce shocks in jets which are able to extract @xmath33 - 4 orders of magnitude more energy from the jet plasma than a single red giant . as described in sect . 2 ,
gcs are expected to create a halo on several kpc distance scale around the central engine of active galaxy .
a substantial number should be submerged in the jet with the opening angle of the order of @xmath30.1 - 0.15 rad .
gcs are composed from old low mass stars which significant number , on average about @xmath3100 , is expected to be in the red giant phase .
these red giants are expected to be the main contributors of the mass into the gc through their stellar winds .
in fact , the mass loss rate of isolated red giant is expected to be in the range between @xmath42 m@xmath2 yr@xmath7 to @xmath43 m@xmath2 yr@xmath7 ( boyer et al .
2008 , meszaros et al .
the mass loss rate in the range of @xmath44 m@xmath2 yr@xmath7 to @xmath45 m@xmath2 yr@xmath7 is estimated from 100 red giants in gc .
it is expected that the single solar mass star should lose about @xmath46 m@xmath2 during its evolution path to the white dwarf ( e.g. see discussion in heyl et al .
simple estimates of the mass loss rates in a single gc , with the total number of stars equal to @xmath47 , gives us the average mass loss rate during 10 gyr equal to @xmath48 m@xmath2 yr@xmath7 , consistently with the above estimated value . in our example calculations , we use the mass loss rates of stars in gcs close to this value . on the other hand ,
gcs contain many millisecond pulsars which has been accelerated to short periods due to the transfer of momentum from the stellar companions .
these msps have typical surface magnetic fields of the order of @xmath310@xmath49 g and rotational periods of the order of @xmath33 ms .
the number of such pulsars in a single gc can be of the order of @xmath3100 , as estimated from their observed @xmath0-ray luminosity ( abdo et al .
these msps lose rotational energy in the form of the relativistic pulsar winds that should mix efficiently with slowly moving winds from the red giants ( bednarek & sobczak 2014 ) . as a result
, a barion loaded pulsar wind slows down to the sub - relativistic velocities .
the velocity of the mixed pulsar - stellar wind is estimated to be , @xmath50 where the mass loss rate of stars within the gc scales with @xmath51 m@xmath2 yr@xmath7 and @xmath52 erg s@xmath7 is the rotational energy loss rate of all millisecond pulsars within the gc . the pressure of the gc wind is estimated from ( bednarek & sobczak 2014 ) , @xmath53 the location of this shock in the jet around the gc , @xmath54 , can be estimated by comparing the pressure of the wind from the gc with the ram pressure of the jet plasma , @xmath55 the radius of the shock should be larger than the core radius of the gc .
note that , most of gcs ( @xmath56 ) in our galaxy have the core radius below 1 pc ( harris 1996 ) .
therefore , for the parameters considered in this paper we expect that this condition is usually fulfilled for the gcs at the distance larger than @xmath31 kpc from the central engine .
the shock structure , formed in collision of the jet with gc wind , is expected to have a double structure with different conditions on both sites ( see fig . 1 on the right ) .
the shock from the jet site is semi - relativistic ( the velocity of plasma after the strong shock drops by a factor of three ) and has relatively strong magnetic field . on the other hand ,
the shock in the gc wind is non - relativistic with weak magnetic field .
we expect that only shock in the jet plasma can accelerate efficiently particles to large energies transferring significant part of the jet power to relativistic electrons .
the power which can be extracted by the shock from the jet plasma can be estimated on , @xmath57 note that the power estimated above for the single collision of the gc with the jet is clearly larger than in the case of a single collision of the red giant star with the jet but it becomes of the order of the power expected in the collision of a single massive star with the jet .
therefore , collisions of gcs with jets at the distance scale of a few kpcs from the central engine can be also responsible for the formation of powerful shocks in jets at such large distances . in the next section we discuss possible acceleration of electrons on the shocks formed in collisions of compact objects with jet plasma in a relatively nearby active galaxies .
in order to estimate the parameters of particles which could be eventually accelerated in such scenario , we consider the physical conditions within the jet . the upper limit on the magnetic field at the base of the jet
can be estimated by assuming that the poynting flux ( @xmath58 ) through the jet , is a part ( @xmath59 ) of the total jet power ( @xmath60 ) , i.e. @xmath61 where the inner radius of the jet can be related to the smbh mass @xmath62 cm , the mass of the bh is @xmath63 , the energy density of the magnetic field is @xmath64 , and @xmath65 is the lorentz factor of the jet .
the jet can be either poynting flux dominated ( then @xmath66 ) or matter dominated ( then @xmath67 ) . the magnetic field at the base of the jet can be estimated by reversing eq .
( [ eq10 ] ) , @xmath68 the perpendicular component of the magnetic field in the jet ( in respect to the jet axis ) at the distance , @xmath69 , from its base can be approximated by , @xmath70 for large distances from the base of the jet . in this case , @xmath71 does not depend on the mass of the black hole .
we assume that electrons are accelerated at the shock from the site of the jet plasma to energies limited by their synchrotron energy losses .
the energy gain rate by electrons is assumed to be determined by the so called acceleration coefficient , @xmath72 , @xmath73 where @xmath74 tev is the energy of electrons and @xmath75 is the larmor radius of electrons .
this acceleration process can be saturated by the energy losses of electrons .
the synchrotron energy loss rate is , @xmath76 where the energy density of magnetic field is , @xmath77 @xmath78 is the thomson cross section and @xmath79 is the electron rest mass .
then , the maximum energy of electrons , limited by the synchrotron energy losses , are @xmath80 electrons with these maximum energies can not have the larmor radii larger than the dimension of the shock around the compact objects .
the conditions , @xmath81 and @xmath82 , introduce another absolute limit on the maximum electron energy , equal to @xmath83 tev and @xmath84 tev .
the limits due to the larmor radius are clearly above the maximum energies of accelerated electrons .
however , another limit on the maximum energies of electrons is introduced by the time scale spend by electrons at the shock region .
this limit differ for stars and gc .
therefore , we consider it separately . the maximum energies of electrons can be also limited by their advection time scale along the shock with the flowing plasma .
this time scale can be estimated from , @xmath85 where @xmath86 is the velocity of the jet plasma in units of the velocity of light .
the comparison of the acceleration time scale ( @xmath87 , see eq .
[ eq13 ] ) with the advection time scale , @xmath88 , gives us the limit on the maximum energies of electrons accelerated in the stellar shock , @xmath89 note that advection limit on the energy of electrons mainly depends on the parameters of the stellar wind . on the other hand ,
the synchrotron limit mainly depends on the parameters of the jet . in general
, the advection limit becomes important in respect to the synchrotron limit for the red giant stars in contrast to the case of the massive stars .
electrons accelerated at the shock around the star can also lose efficiently energy on the interaction with the stellar radiation . in order to evaluate importance of this process
we calculate the energy density of stellar photons at the shock region , @xmath90 $ ] ev cm@xmath11 , where the surface temperature of the red giant type star is @xmath92 k and its radius @xmath93 cm , and @xmath94 is the stefan - boltzmann constant . on the other hand , energy density of the magnetic field at the shock from the jet site
is given by eq .
the relative energy losses of electrons on these two processes ( in the thomson regime , i.e. for @xmath95 gev ) depends only on the ratio of the energy densities of the radiation and the magnetic field .
therefore , electrons with energies clearly below @xmath96 gev will lose energy more efficiently on the ic process than on the synchrotron process .
however , the energy losses of electrons becomes comparable for tev energies to which these particles are expected to be accelerated at the shocks provided that the jet is poynting flux dominated ( @xmath66 ) .
since the winds of massive stars are clearly more extreme than those of red giants , a relatively stronger ic @xmath0-ray emission ( in respect to the synchrotron emission ) is expected in the case of massive stars within the jets .
the power in relativistic electrons extracted from the shock , created in collision of the jet with stellar winds , is estimated on , @xmath97 where @xmath98 is the energy conversion efficiency from the shock to relativistic electrons , and @xmath99 is the number of stars within the jet . the energy conversion efficiency , @xmath100 , is usually expected to be in the range @xmath30.1 - 0.2 .
the number of massive stars within the jet is expected to be a few orders of magnitudes lower than the number of red giants ( see sect . 2 ) . however
the product of the mass loss rate and the wind velocity of a massive star is typically 3 - 4 orders of magnitudes larger ( see sect .
3.1 ) . as a result
, the power in relativistic electrons accelerated on the shocks formed by the whole population of red giants and massive stars is expected to be on the similar level .
as above , we constrain the maximum energies of electrons due to their advection along the shock around the gc within the jet .
we compare the acceleration time scale , @xmath101 , with @xmath102 .
this condition limits energies of electrons to , @xmath103 in general , this limit is less restrictive than the advection limit for stars within the jet , since the shock dimensions around gcs are expected to be clearly larger for typical parameters of the scenario . the synchrotron limit on the acceleration of electrons dominates over the advection limit for the following condition on the least constrained parameter of the model , i.e. the acceleration coefficient should be , @xmath104 .
if the above condition is fulfilled , then we can estimate the characteristic energies of synchrotron photons produced by electrons on @xmath105 kev , where @xmath106 gs is the critical magnetic field strength . in the case of cen
a , the x - ray emission up to @xmath16 kev is observed ( hardcastle et al
. 2006 ) .
therefore , the acceleration coefficient is expected to be at least @xmath107 .
electrons with such energies should produce synchrotron emission extending to x - ray energies but also tev @xmath0-rays by scattering optical radiation produced within the galactic bulge and the mbr .
let us check whether electrons , accelerated at the shock around gc , will manage to cool efficiently on the synchrotron or the inverse compton ( ic ) processes .
the electron cooling mean free path on the synchrotron process in the magnetic field of the jet is , @xmath108 the synchrotron cooling process of electrons will occur locally in the jet , provided that the synchrotron mean free path becomes shorter than the characteristic distance scale which we identify with the distance of the injection place of electrons from the base of the jet , i.e. @xmath109 . the above condition is fulfilled for electrons with energies @xmath110 tev .
the ic energy losses of electrons can in principle be determined by the soft radiation produced by the gc itself , the galactic bulge thermal radiation and the mbr .
the bulge radiation might play an important role in the inner part of the intermediate scale jet ( i.e @xmath3kpc scale ) .
the mbr can start to be important in the outer parts of the large scale jet when the energy density of the magnetic field drops significantly .
the energy density of radiation from the gc at the location of the shock can be estimated on @xmath111 ev cm@xmath11 , where the optical luminosity of the gc is @xmath112 erg s@xmath7 and @xmath113 is the luminosity of the sun .
the energy density of the bulge radiation is estimated on @xmath114 ev cm@xmath11 , where the bulge luminosity is @xmath115 l@xmath2 , and the bulge radius is @xmath116 kpc .
the energy density of the mbr is @xmath117 ev cm@xmath11 .
the mean free path for electron energy losses on ic process in the thomson regime in these radiation fields can be estimated on , @xmath118 where @xmath119 ev cm@xmath11 .
we expect that multi - tev electrons will mainly scatter stellar radiation in the klein - nishina regime but the estimate based on the thomson cross section allows us to have impression about importance of the ic energy losses .
the ic losses becomes important when the mean free path of electrons is comparable to the characteristic distance scale on which they propagate . in the case of the radiation field from the gc
, @xmath120 is always clearly lower than @xmath121 , for reasonable parameters of the scenario .
therefore , energy losses of electrons in the gc radiation can be safely neglected . in the case of the bulge radiation ,
the mean free path of the ic energy losses can become comparable to the characteristic size of the bulge ( of the order of kpc ) and also to the mean free path of electrons on the synchrotron process .
therefore , the ic energy losses in the bulge radiation should be taken into account on the kpc distance scale from the base of the jet .
the mean free path for the ic energy losses of electrons in the mbr is of the order of @xmath3kpcs , for electrons with energies of a hundreds tev .
therefore , ic losses of electrons are expected to become important in the outer parts of the jets at the distance scale of several kpc from the base of the jet .
note that the gc halo still extends to such distances .
we conclude that in fact , electrons can be accelerated to large energies at the shocks formed in collisions of gcs with jet plasma .
they should produce synchrotron radiation extending up to x - ray energy range and also multi - tev @xmath0-rays by scattering radiation from the galactic bulge and the mbr .
the total power in relativistic electrons , accelerated at the shocks around gcs , can be estimated on , @xmath122 where the shock power from a single gc is given by eq .
[ eq9 ] , and @xmath123 is the number of gcs within the jet .
the maximum energies of electrons are limited in this case by their synchrotron energy losses to values given by eq .
( 16 ) . for the expected values of the parameters describing considered scenario , the power in electrons , accelerated at the shocks around gcs ,
is estimated to be about an order of magnitude lower than in the case of the shocks around stars .
however , electrons , accelerated at the shocks around gcs , are injected into the jet at much larger distances from its base ( i.e. above @xmath3kpc ) . at these distances
the jet magnetic field strength is already weak .
therefore , electrons lose relatively more efficiently energy on the ic process ( producing @xmath0-rays ) than on the synchrotron process ( producing radiation below x - ray energies ) . in the next section
we perform detailed calculations of the multi - wavelength spectra expected from the intermediate scale jet assuming that different types of compact objects ( red giants , massive stars , gcs ) form multiple shocks in the jet plasma . as an example , the parameters of the nearby radio galaxy , cen a , are applied .
we calculate expected synchrotron and ic spectra produced by electrons accelerated on the fronts of multiple shocks within the jet for different parameters describing compact objects and the jet content .
it is assumed that electrons are accelerated with the differential power law spectra above some minimum energy @xmath124 . the spectral index equal to @xmath125 is selected for the example calculations as expected in the shock acceleration scenario .
the maximum energies of electrons , @xmath126 , are determined either by the synchrotron energy losses ( eq . [ eq16 ] ) or by the escape from the acceleration region due to the advection of electrons with the jet plasma ( eq . [ eq18 ] or [ eq20 ] ) .
these electron spectra have been normalized to the power transferred to electrons from the shocks in the jet ( see eqs .
[ eq19 ] and [ eq23 ] ) . in order to obtain the synchrotron and the ic @xmath0-ray spectra
, we inject electrons at different distances from the base of the jet and simulate their propagation in the jet magnetic field and the radiation fields , discussed above , by applying the monte carlo method .
the klein - nishina effect has been taken into account when calculating the ic @xmath0-ray spectra . in these example calculations ,
we assume that the jet is semi - relativistic , with the apparent velocity equal to @xmath127 and moves at relatively large angle to the observer s line of sight estimated on @xmath128 in the case of cen a. for the above parameters the velocity of the jet is estimated on @xmath129 and its lorentz factor on @xmath130 .
the doppler factor of the jet in cen a at the distance of a hundred pc from its base is then estimated on @xmath131\approx 1.282 $ ] .
note that at the kpc distance scales the jet may significantly decelerate .
therefore , doppler effects can be negligable in the example case of cen a. such assumption clearly simplifies the calculations of the non - thermal radiation presented in this paper .
the jet has a constant opening angle equal to @xmath132 rad and the jet power is @xmath133 erg s@xmath7 . for these basic parameters ,
we investigate the spectra as a function of two other parameters determining the relativistic electrons in the jet , the jet magnetization parameter @xmath59 and the acceleration efficiency of electrons in the jet @xmath134 .
these two parameters determine the relative part of energy of relativistic electrons lost on the synchrotron and on the ic process .
9.5truecm the calculations of the ic @xmath0-ray spectra in the case of collisions of stars with the inner jet ( i.e. at distances @xmath135 pc ) have been already considered by e.g. bednarek & protheroe ( 1997 ) . in such case ,
the shock is located relatively close to the massive companion star and relativistic electrons initiate cascade in the dense stellar radiation field .
the calculations of the synchrotron emission from electrons accelerated on the multiple shock structures around the stars in the intermediate scale jets has been recently considered by wykes et al .
( 2014 ) . but
no ic @xmath0-ray emission has been calculated in this model .
we consider the star - jet interaction process on a large distance scale from the base of the jet .
in such case , the cascade process initiated by electrons accelerated around massive stars is inefficient since the shock is too far from the star .
electrons escape with significant energies from the shock and propagate mainly in the jet volume . at this stage , electrons lose energy on the synchrotron process in the magnetic field of the jet and on the ic process by scattering different radiation fields such as radiation of compact star , diluted radiation from the galactic bulge and the mbr .
electrons propagating outside the bulge , with typical dimension of @xmath136 = 1 kpc , will see the radiation field diluted by the factor @xmath137 . in these example calculations
we assume that electrons are accelerated on shocks around stars entering uniformly the jet between 10 pc up to 1 kpc .
note that in such case maximum energies of electrons injected from the shocks into the jet depend on the distance from the base of the jet as given by eq .
[ eq16 ] or eq .
the parameters of stars have been fixed on @xmath138 m@xmath2 yr@xmath7 and @xmath139 km s@xmath7 ( for the red giant stars ) and on @xmath140 m@xmath2 yr@xmath7 and @xmath141 km s@xmath7 ( for the massive stars ) .
results of these example calculations of the synchrotron and ic @xmath0-ray spectra are shown in fig .
we investigate the spectra for the poynting flux dominated jets ( the magnetization parameter of the jet @xmath142 ) and for the matter dominated jets ( described by @xmath143 and @xmath144 ) .
we also show the spectra obtained for different values of the acceleration efficiency of electrons @xmath145 and @xmath146 . in all considered cases ,
the synchrotron spectra dominate over the ic spectra .
as expected , the ic @xmath0-ray spectra are stronger for less magnetized jets . in the case of red giants , the cut - off in synchrotron spectrum
is determined by the advection time scale of electrons along the shock structure .
in contrast , the synchrotron spectra cuts - off independently on the jet magnetization for shocks around massive stars .
collisions of gcs with jets in active galaxies have not been considered up to now as a source of disturbances in the jet plasma ( shocks ) which might be responsible for acceleration of particles . however , a large number of gcs around active galaxies , and evidences that their numbers are correlated with the mass of the central black hole ( harris et al .
2014 ) , suggest that significant number of gcs is immersed in the jet .
we showed above that shocks around gcs provide conditions for acceleration of electrons to hundreds of tev . applying the numerical code for the propagation of electrons in the jet and their radiation described above , we calculate the synchrotron and ic spectra .
since such calculations have not been considered before , we also show the results of calculations of the spectra produced at different distances from the base of the jet .
the spectra are investigated as a function of parameters describing the acceleration process of electrons ( i.e. the magnetization parameter of the jet @xmath59 and the acceleration efficiency of electrons @xmath134 ) .
electrons are injected at fixed distance from the base of the jet and , after leaving the shock , they are frozen in the jet plasma . in such case ,
the synchrotron energy losses of electron are determined by the profile of the magnetic field along the jet ( see eq .
[ eq12 ] ) . on the other hand
, these electrons also comptonize radiation from the galactic bulge ( with photon density dropping with the distance along the jet ) and the mbr .
note that electrons are accelerated to larger energies farther along the jet .
as in the case of collisions of stars with the jet , we assume that electrons reach the power law spectrum extending up to the maximum energies described above . in the case of gcs , the minimum energy of electrons is fixed on 1 tev which corresponds to the characteristic energy of leptons injected by the millisecond pulsars into the mixed pulsar / stellar wind from the gc .
8.5truecm in the example calculations of the synchrotron and the inverse compton spectra , we apply the parameters derived for the nearby radio galaxy cen a : the mass of the black hole @xmath147 , the jet power @xmath148 , the jet opening angle @xmath149 , the jet velocity @xmath150 ( @xmath151 ) , and the parameters of typical for gcs : gc stellar luminosity @xmath152 , the power supplied by msps @xmath153 , the mass loss rate of red giant stars @xmath154 , and the parameters of the host galaxy bulge , the bulge stellar luminosity @xmath155 , the bulge radius @xmath156 .
the acceleration coefficient is taken to be @xmath157 in order to be consistent with the observations of the non - thermal x - ray emission from the jet .
the radius of the shock around gc is then @xmath158 pc , the magnetic field strength along the jet is @xmath159 gs , the maximum energies of electrons due to synchrotron energy losses are @xmath160 tev , the maximum energies of electrons due to advection from the shock are @xmath161 tev .
the comparison of these two last limits gives us the range of distances for which synchrotron limit is more restrictive , i.e. @xmath162 .
we conclude that the maximum electron energies are always limited by synchrotron losses in the case of poynting dominated jet .
they are limited by the advection process in the case of the matter dominated jet with @xmath163 . for @xmath143 the acceleration process of electrons
is limited by the synchrotron energy losses in the inner part of the jet , i.e. below @xmath164 , but in the outer parts of the jet by the advection process .
the mean free paths of electrons are @xmath165 pc on the synchrotron process , @xmath166 kpc on the ic scattering of bulge radiation , and @xmath167 kpc on the ic scattering of the mbr .
the characteristic distance scales for ic energy losses of electrons on the border between the thomson ( t ) and the klein - nishina ( kn ) regimes in the bulge radiation are comparable to the bulge dimension .
these distance scales for ic energy losses of electrons with the tev energies in the mbr are comparable to the dimensions of jets in agns . in fig . 3
, we show the synchrotron and the inverse compton spectra for different distances from the base of the jet , @xmath168 kpc , 3 kpc , and 10 kpc .
the poynting flux dominated jets ( @xmath142 ) and the matter dominated jets ( @xmath143 and 0.01 ) are considered . in the case of the poynting dominated jets , the ic spectra are on a much lower level than the synchrotron spectra especially at lower distances from the base of the jet . on the other hand ,
the ic spectra start to dominate over synchrotron spectra at larger distances from the base of the jet in the case of matter dominated jets defined by the magnetization parameter @xmath169 .
therefore , most of the energy of relativistic electrons can be transferred to the tev @xmath0-rays in jets at distances of a few kpcs from its base .
the synchrotron spectra , produced in the jet by these electrons , clearly extend up to the x - ray energy range , even for the advection dominated acceleration process of electrons , for the considered range of parameters , @xmath170 and @xmath171 .
the synchrotron and ic spectra , produced by electrons accelerated at shocks formed by gcs entering the jet at the range of distances from its base between 1 kpc and 20 kpc , are shown in fig .
it is assumed that gcs enter the jet homogeneously over such range of distances and ejects electrons with the spectra and powers as described above .
the results are shown for the parameters as in previous figures .
it is clear that in the case of the matter dominated jets the level of the tev @xmath0-ray emission is comparable to the level of the synchrotron x - ray emission . in this case , the @xmath0-ray spectra peak at multi - tev energies and the synchrotron spectra extends above @xmath310 kev .
however , in the case of poynting dominated jets , the synchrotron emission clearly dominates over the tev @xmath0-ray emission .
the @xmath0-ray luminosity is expected to be about an order of magnitude below the x - ray luminosity and the @xmath0-ray spectra peak at gev energies .
we show that in the case of matter dominated jet models , the collisions of gcs with the kpc scale jets can be responsible for the detectable synchrotron x - ray emission and also for the tev @xmath0-ray emission .
the relative levels of the emission in these two energy ranges should provide important constrains on the content of jets at kpc scale distances .
note that in the matter dominated jets the maximum energies of accelerated electrons are determined by their escape from the shock region due to the advection process .
these electrons are not completely cooled when propagating in the jet on the distance scale of 20 kpc considered in the paper .
therefore , the ic spectra are so hard below the peak at tev energies .
the lower energy electrons are cooled on much longer distance scale .
they should produce steeper spectra on similar angular scales in the case of distant agns for which the contribution to the ic spectrum comes from more extended part of the jet .
therefore , such hard spectra are only expected from the intermediate scale jets in the nearby agns such as cen a for which the distance of 20 kpc corresponds to angular size of the source several arc min .
we considered only semi - relativistic jets as observed from nearby agns . in the case of relativistic jets , expected in the distant agns ,
these maximum energies of accelerated electrons are expected to be lower ( see eq .
therefore , the spectra may not extend to @xmath310 tev .
these spectra are not expected to be hard enough to provide interesting constraints on the extra - galactic background light .
centaurus a is the closest radio galaxy ( distance @xmath172 mpc ) associated with the elliptical galaxy ngc 5128 ( harris et al .
the smbh in cen a has the mass estimated on @xmath173 m@xmath2 ( cappellari et al .
two jets , clearly asymmetric on kpc scale , are observed in cen a. the viewing angle of the jet has been estimated on @xmath350@xmath174 ( tingay et al .
1998 , hardcastle et al . 2003 ) . the projected speed of the jet at hundred pc distance scale has been measured on 0.5c ( hardcastle et al .
the power of the jet has been estimated on @xmath175 erg s@xmath7 ( wykes et al .
2013 ) .
due to its proximity , x - ray emission is clearly seen from the kpc scale jet in cen a ( feigelson et al . 1981 ) . at the distance between @xmath31 - 3 kpcs ,
the _ chandra _ observations ( 0.4 - 2.5 kev energy range ) show that the diffusive x - ray emission dominates over x - ray emission from the knots .
this x - ray emission has been interpreted as due to the synchrotron process ( hardcastle et al .
such synchrotron radiation requires the presence of the tev electrons in the kpc scale jet .
they have to be accelerated close to the radiation site due to short energy loss time scales .
recently , a point like @xmath0-ray source with the power low spectrum ( spectral index @xmath32.7 ) extending to @xmath32 - 3 gev has been detected by _
satellite from the core of cen a ( abdo et al .
2010a ) . the analysis of the 4 years of the _
fermi _ data shows that @xmath0-ray emission extends up to @xmath3 50 gev .
however , the spectrum flattens above @xmath34 gev ( sahakyan et al .
the spectral index changes from @xmath176 below 4 gev to @xmath177 at higher energies .
this flattening of the @xmath0-ray spectrum is consistent with the detection of the tev source towards cen a by the hess collaboration ( aharonian et al .
it suggests that the tev and hard gev emission might originate in this same mechanism .
the lower energy points of these tev observations ( @xmath178 gev ) links nicely to the hard high energy component detected by _
fermi _ but the spectral index of the tev emission is again better described by a steep spectrum ( spectral index @xmath179 , see aharonian et al .
2009 ) . such curious spectral behavior , lack of detected clear variability , and pure angular resolution of @xmath0-ray observations suggest involvement of at least two radiation processes or emission regions ( sahakyan et al .
2013 ) . due to the large angular extend of the cen a jets on the sky ( of the order of @xmath310@xmath174 ) ,
the _ fermi _ observatory was also able to discover gev @xmath0-ray emission from the regions of the giant radio lobes formed in collisions of jets with intergalactic medium ( abdo et al .
all these high energy observations show that the high energy processes are characteristic not only for the direct vicinity of the smbh but also for the large scale jets .
we wonder whether the highest energy component ( at @xmath31 tev ) , observed from cen a by the hess collaboration , can originate in the kpc scale jet due to the acceleration of electrons on the shocks in the jet which are produced in the collisions of compact objects ( stars , gcs ) with the jet plasma . in order to test this hypothesis , we performed calculations of the synchrotron and the ic spectra expected in terms of the above jet / compact objects collision model , applying the known parameters of the jet and the surrounding medium of cen a. the results of calculations are compared with the observed high energy spectrum from cen a ( fig .
5 ) . in this figure
the x - ray spectrum represents emission from the kpc scale jet ( hardcastle et al .
2006 ) . the _ fermi _ spectrum comes from the central part of the galaxy harboring cen a ( sahakyan et al .
the origin and location of this steady spectral component is at present unknown due to the pure angular resolution of the _ fermi_-lat telescope .
finally , the data points show the measurements of the tev @xmath0-ray emission from cen a inner / intermediate scale jet .
we assumed that the red giant stars and massive stars enter the jet uniformly at distances between 10 pc and 1 kpc from the base of the jet .
electrons are accelerated at the shocks with the energy conversion efficiency of @xmath180 .
the spectra are calculated for the presence of @xmath47 red giants ( dotted curves ) and 200 massive stars ( dashed curves ) within the jet .
the example parameters of these stars are described in sect .
5.1 . for the considered parameters of the acceleration model ( @xmath163 , @xmath181 ) , the emission from electrons accelerated on the shock around red giants
is not expected to contribute significantly to the high energy spectrum observed from cen a. on the other hand , about a hundred of massive stars in the jet can produce synchrotron emission on the level observed from the jet of cen a. the accompanying @xmath0-ray emission , from the componization of background radiation by these same electrons , is on the level of the tev @xmath0-ray emission reported by hess from cen a. we also show the synchrotron and ic spectra produced in collisions of 20 gcs with the jet at larger distance scale ( dot - dashed curves ) for these same parameters of the jet . the synchrotron emission is not expected to contribute significantly to the observed x - ray spectrum . however , the ic @xmath0-ray emission peaks at the tev energies .
the cumulative @xmath0-ray spectrum from collisions of massive stars and gcs with the jet is clearly consistent with the level of the tev @xmath0-ray emission observed by hess .
therefore , such model might be responsible for the highest energy @xmath0-ray component observed from cen a. if the jet in cen a still moves with the velocity of the order of @xmath182 over kpc distances , then the calculated spectra ( in fig .
5 ) should be enhanced by a factor of @xmath183 . as a result
, our estimates of the number of compact objects within the jet should be reduced by this factor .
we conclude that the steady tev @xmath0-ray emission , observed from the central part of the radio galaxy cen a , can originate in the kpc scale jet . this emission ( or at least a part of this emission ) should be steady independently on the activity state of the inner jet .
since the expected emission is above the level of sensitivity of the future cherenkov telescope array ( cta ) , the question on the existence of the low level persistent emission from the intermediate scale jet in cen a should be definitively answered in the coming years .
we propose that the compact object / jet collision model can be responsible for the non - thermal emission from the intermediate scale jets observed in close radio galaxies .
we consider collisions of different type of objects , starting from the red giant stars from the galactic bulge , through the massive stars from the nuclear central cluster and finishing on the globular clusters in the galactic halo .
collisions produce multiple shocks in the jet on which leptons can be accelerated to tev energies .
these leptons are responsible for the diffusive synchrotron and inverse compton radiation from the kpc scale jets .
contrary to the inner jet collision models , leptons accelerated on the shocks do not lose energy close to the shock region ( near the stars ) but they are injected into the jet .
they are transported with the jet plasma to farther distances from the base of the jet losing energy on the mentioned above radiation processes .
therefore , this high energy emission is expected to be steady during the time scales of the jet dimension ( i.e. thousands of years ) , in contrast to the emission expected from collisions of stars with the inner jets ( below @xmath31 pc , e.g. bednarek & protheroe 1997 ) . moreover , due to the saturation of the acceleration process of leptons by the synchrotron or the advection time scales , the larger energy leptons are expected to be accelerated at farther along the jet . therefore , non - thermal emission from jets should show clear dependence on the energy with the lower energy @xmath0-rays produced closer to the base of the jet .
we predict that the multi - tev @xmath0-rays should have tendency to be produced farther from the base of the jet .
in contrast , such feature may not be observed in the case of the synchrotron x - ray emission if the acceleration of leptons is saturated by synchrotron energy losses .
only when the maximum electron energies are constrained by their escape ( advection ) from the acceleration region , the maximum energies of synchrotron radiation should show energy dependence on the distance from the base of the jet .
we compared the example spectra , obtained in terms of such a model , with the observations of the intermediate scale jet in cen a ( fig .
it is found that in general the observed level of the x - ray emission can be explained as a result of the synchrotron radiation produced by electrons accelerated on the shocks around massive stars . on the other hand ,
the tev @xmath0-ray emission is expected to be produced by leptons accelerated on the shocks around massive stars ( below @xmath31 tev ) and by electrons accelerated on the shocks around gcs ( above @xmath31 tev ) .
this hypothesis could be tested in the near future with the operation of the next generation @xmath0-ray telescopes such as cta .
cta is expected to reach @xmath31 arc min resolution at 10 tev and the integral sensitivity of @xmath184 tev @xmath185 s@xmath7 at a few tev ( acharya et al .
2013 ) . with such angular resolution the tev @xmath0-ray emission from the kpc scale jet in cen a could be resolved and the prediction on the energy dependent emission with the distance from the base of the jet could be confirmed or disproved .
we applied the collision model for the radio galaxies whose intermediate jets are expected to be rather slow and observed at a relatively large angle to the line of sight . if the jets still moves relativistically at the kpc distance scale , then the non - thermal radiation can be significantly doppler boosted ( see e.g. recent paper by bosch - ramon 2015 ) .
in such case the collisions of different types of compact objects ( massive stars , red giants , globular clusters ) with the relativistic jets could provide natural explanation for the powerful non - thermal emission from blazars observed at small angle to then line of sight . | massive black holes in active galaxies are immersed in huge concentrations of late type stars in the galactic bulges and also early type massive stars in the nuclear stellar clusters which are additionally surrounded by quasi - spherical several kpc scale halos containing from a few hundred up to several thousand globular clusters ( gcs ) .
it is expected that significant numbers of red giant stars , massive stars and also gcs can move through the jet expelled from the central engine of active galaxy .
we consider collisions of stars from the galactic bulge , nuclear cluster and globular clusters with the jet plasma . as a result of such collisions ,
multiple shocks are expected to appear in the jet around these compact objects .
therefore , the plasma in the kpc scale jet can be significantly disturbed .
we show that particles can be accelerated on these shocks up to the multi - tev energies .
tev leptons emit synchrotron radiation , extending up to the x - ray energies , and also comptonize radiation produced in a stellar cluster and also the microwave background radiation to tev @xmath0-ray energies .
we show that such non - thermal radiation is likely to be detectable from the intermediate scale jets of the nearby active galaxies for reasonable number of stars and gcs immersed within the jet . as an example , we calculate the expected non - thermal emission in the x - ray and gamma - ray energies from the nearby radio galaxy cen a from which a steady gamma - ray emission with the complex spectrum has been recently reported by the _
fermi _ and the hess observatories . |
higher twist terms contribute to the nucleon structure functions at lower scales @xmath4 .
the range in which these terms may be safely neglected against the leading twist contributions , partly depends on the size of the experimental errors in the respective measurement .
highly precise data at low values of @xmath4 allow to access these contributions , the detailed physical understanding of which is presently still in an early stage .
it has been outlined in refs .
@xcite how the higher twist contributions can be extracted in a phenomenological way in case of the structure functions @xmath5 and @xmath6 in the valence quark region . in this note
we report on recent results of an improved analysis .
another interesting question concerns the structure function @xmath3 in the polarized case , which has been measured to a higher precision during the last years @xcite . here
we try to extract first information on the twist-3 contributions to @xmath3 .
we have carried out a qcd analysis in the valence region including more recent data from jlab following earlier work @xcite . in the present analysis tails from sea - quarks and the gluon in the valence region were dealt with based on the abkm distributions @xcite .
both the valence quark distributions @xmath7 and @xmath8 at @xmath9 are effected only very little .
the values of @xmath10 change marginally w.r.t .
the earlier analysis @xcite .
we obtain : @xmath11 . here
, the n@xmath1lo@xmath12-analysis accounts for the three - loop wilson coefficients and a pad - model for the non - singlet four - loop anomalous dimension , to which we attached a @xmath13 uncertainty , cf .
@xcite for details .
furthermore , we found that the response of the individual deep - inelastic data sets in the valence region respond stable values , which are in accordance with the central value obtained moving from nlo to n@xmath1lo@xmath12 .
the present result agrees very well with determinations of @xmath10 in refs .
@xcite , see also @xcite .
a survey on the current status of @xmath10 based on precision measurements in different reactions has been given in @xcite .
in the present analysis we obtain a lower value of @xmath14 than the world average , cf .
@xcite , and values being obtained in @xcite at nnlo .
reasons for the difference to the values given in @xcite have been discussed in refs .
@xcite in detail .
in particular , the partial response of @xmath14 in case of the bcdms and slac data in @xcite turns out to be partly different comparing to the results in @xcite .
there are also differences between the analyses @xcite and @xcite w.r.t .
several data sets contributing . the higher twist contributions can be determined by extrapolating the fit - results at leading twist for @xmath15 to the region @xmath16 , cf .
@xcite . below the valence region have been determined . ]
the results for the coefficients @xmath17 @xmath18}{f_2(x , q^2 ) } + \frac{c_{\rm ht}(x)}{q^2[\gev^2 ] } \right]\end{aligned}\ ] ] are shown in figure 1 , where we averaged over the respective range in @xmath4 .
we applied the target mass corrections @xcite to the leading twist contributions . and @xmath19 including the jlab data , has been performed in @xcite recently . ]
the result for the higher twist coefficients for proton and deuteron targets depends on the order to which the leading twist distribution is described .
the higher twist terms become smaller moving from nlo to n@xmath1lo@xmath12 . within the present theoretical and experimental accuracy
the curves stabilize for @xmath20 , while at larger values there are still differences .
higher twist contributions to the polarized structure function @xmath22 have been studied in refs .
@xcite in phenomenological approaches aiming on the twist-4 contributions . however , the structure function @xmath3 , together with other polarized electro - weak structure functions @xcite , receives also twist-3 contributions .
@xmath3 obeys the burkhardt - cottingham relation @xcite @xmath23 since the wandzura - wilczek relation @xcite implies , that the first moment of the twist-2 part vanishes separately also @xmath24 holds .
the errors on the present world data from e143 , e155 , hermes and nmc @xcite on @xmath3 are still large but yet one may try the fit of a profile in @xmath2 . in ref .
@xcite the parameterization @xmath25 + ( 1-x)^3\left[b - c(1-x ) + d(1-x)^2\right ] \label{sha}\end{aligned}\ ] ] has been proposed .
since the data points are measured at different values of @xmath4 an evolution has to be performed to a common scale .
furthermore , the target mass corrections @xcite have to be taken into account . in figure
2 the results of the fit to @xmath21 are presented for @xmath26 .
we limited the analysis to the region @xmath27 .
the present errors are still large and the data of e155 dominate in the fit .
we may compare with a theoretical prediction given in @xcite .
indeed both results are quite similar .
the twist-3 contribution to the structure function @xmath22 can be obtained from that to @xmath3 by the integral - relation @xcite @xmath28~,\end{aligned}\ ] ] cf . @xcite . due to the large errors of the data the present results are of more qualitative character . to study the twist-3 contributions both to the structure functions @xmath3 and @xmath22 in detail , a high luminosity machine , like the planned eic @xcite ,
is needed .
we performed a re - analysis of the present deep - inelastic world data on proton and deuteron targets for the structure function @xmath5 in the valence region @xmath29 accounting for remaining non - valence tails , which were calculated using the abkm09 distributions @xcite .
we obtain a slightly lower value of @xmath10 than in our previous analysis @xcite at n@xmath1lo@xmath12 , however , far within the @xmath30 error range .
very stable predictions are obtained going from nlo to n@xmath1lo@xmath12 , both for the valence distribution functions and @xmath10 .
the values being obtained for the different sub - sets of experimental data in the present fit are well in accordance with our global result .
we do not confirm the significant differences reported by mstw between the slac @xmath31 and @xmath32 data at nnlo @xcite .
we also disagree with the large value of nnpdf @xcite for the bcdms data at nlo , which also contradicts the corresponding result by mstw @xcite .
our results are in agreement with those of the gjr collaboration @xcite and the singlet analyses @xcite .
we obtained an update of the dynamical higher twist contributions to @xmath0 in the valence region , which depends on the order to which the leading twist contributions were calculated .
the effect stabilizes including corrections up to n@xmath1lo@xmath12 in the range @xmath33 . at larger values of @xmath2
still higher order corrections may be needed . a first estimate on the quarkonic twist-3 contributions to the polarized structure function @xmath3
is given in a fit to the available world data on @xmath3 .
the contributions to @xmath22 are obtained by an integral relation , cf . ref .
for discussions we would like to thank s. alekhin . this work has been supported in part by dfg sonderforschungsbereich transregio 9 , computergesttzte theoretische teilchenphysik , and eu network lhcphenonet pitn - ga-2010 - 264564 .
99 j. blmlein , h. bttcher and a. guffanti , nucl .
b * 774 * ( 2007 ) 182 [ hep - ph/0607200 ] ; nucl .
* 135 * ( 2004 ) 152 [ hep - ph/0407089 ] .
j. blmlein and h. bttcher , phys .
b * 662 * ( 2008 ) 336 [ arxiv:0802.0408 [ hep - ph ] ] .
e143 coll .
, k. abe et al . , phys . rev .
d 58 , 112003 ( 1998 ) ; + e155 coll . , p. l. anthony et al .
b 553 , 18 ( 2003 ) ; + a. airapetian , n. akopov , z. akopov , e. c. aschenauer , w. augustyniak , r. avakian , a. avetissian and e. avetisyan _ et al . _ ,
j. c * 72 * ( 2012 ) 1921 [ arxiv:1112.5584 [ hep - ex ] ] .
s. alekhin , j. blmlein , s. klein and s. moch , phys .
d * 81 * ( 2010 ) 014032 [ arxiv:0908.2766 [ hep - ph ] ] .
m. glck , e. reya and c. schuck , nucl .
* b754 * ( 2006 ) 178 [ arxiv : hep - ph/0604116 ] ; + p. jimenez - delgado and e. reya , phys .
* d79 * ( 2009 ) 074023 [ arxiv:0810.4274 [ hep - ph ] ] . s. alekhin , j. blmlein and s. moch , arxiv:1202.2281 [ hep - ph ] . s. alekhin , j. blmlein and s. moch , eur .
j. c * 71 * ( 2011 ) 1723 [ arxiv:1101.5261 [ hep - ph ] ] .
s. bethke _ et al .
_ , proceedings of the workshop on precision measurements of @xmath14 , arxiv:1110.0016 [ hep - ph ] .
a. d. martin , w. j. stirling , r. s. thorne , g. watt , eur .
j. * c64 * ( 2009 ) 653 [ arxiv:0905.3531 [ hep - ph ] ] .
s. lionetti _ et al .
_ , phys .
b * 701 * ( 2011 ) 346 [ arxiv:1103.2369 [ hep - ph ] ] .
j. blmlein and h. bttcher , in preparation . h. georgi and h. d. politzer , phys .
d * 14 * ( 1976 ) 1829 .
m. e. christy , j. blmlein and h. bttcher , arxiv:1201.0576 [ hep - ph ] .
e. leader , a. v. sidorov and d. b. stamenov , phys .
d * 80 * ( 2009 ) 054026 [ arxiv:0908.2390 [ hep - ph ] ] .
j. blmlein and h. bttcher , nucl .
b * 841 * ( 2010 ) 205 [ arxiv:1005.3113 [ hep - ph ] ] .
j. blmlein and n. kochelev , phys .
b * 381 * ( 1996 ) 296 [ hep - ph/9603397 ] .
j. blmlein and n. kochelev , nucl .
b * 498 * ( 1997 ) 285 [ hep - ph/9612318 ] .
j. blmlein and a. tkabladze , nucl .
b * 553 * ( 1999 ) 427 [ hep - ph/9812478 ] .
h. burkhardt and w. n. cottingham , annals phys .
* 56 * ( 1970 ) 453 .
s. wandzura and f. wilczek , phys . lett .
b * 72 * ( 1977 ) 195 .
v. m. braun , t. lautenschlager , a. n. manashov and b. pirnay , phys .
d * 83 * ( 2011 ) 094023 [ arxiv:1103.1269 [ hep - ph ] ] .
j. blmlein , h. bttcher , and a. de freitas , in preparation .
d. boer , m. diehl , r. milner , r. venugopalan , w. vogelsang , d. kaplan , h. montgomery and s. vigdor _ et al .
_ , arxiv:1108.1713 [ nucl - th ] . | we report on recent results on higher twist contributions to the unpolarized structure functions @xmath0 at n@xmath1lo in the large @xmath2 region and constraints on the twist3 contribution to polarized structure function @xmath3 . |
the aim of this communication is to present and describe a computational package to handle first - principles density functional ( dft ) based studies of electronic structure of systems without long - ranged lattice translational symmetry .
bulk disordered alloys and surfaces and interfaces which are either flat , corrugated or rough fall under this category which our formalism should be able to take care of .
the aim is also to go beyond the usual mean - field approaches like the coherent potential approximation ( cpa ) and be able to take into account configuration fluctuations of the local environment .
lack of lattice translational symmetry means that the standard reciprocal space techniques based on the powerful bloch theorem can no longer be applicable and we shall depend on alternative techniques based purely on real space approaches .
our formalism will be a marriage of three distinct methods which have been individually applied extensively : namely , the recursion method ( rm ) of haydock @xcite-@xcite , the augmented space method proposed by one of us @xcite-@xcite ( asr ) and the tight - binding , linear muffin - tin orbitals method ( tb - lmto ) @xcite .
the last mentioned provides us with a dft self - consistent sparse representation of the hamiltonian in a real - space minimal basis @xmath0 which spans the hilbert space @xmath1 . here
@xmath2 labels the sites where the ion - cores sit and @xmath3 are the angular momentum indexes . for a disordered system the matrix elements of the hamiltonian representation in this basis are random .
this representation is then taken over by the asr to generate a modified hamiltonian representation in the outer product space of @xmath1 and the space @xmath4 of configuration fluctuations of the random parameters .
this modified hamiltonian in the augmented space @xmath5 represents a collection of all possible hamiltonians for all possible configurations of hamiltonian representations .
once this is done the rm allows us to obtain the matrix elements of the green functions related to the kohn - sham equation for the electronic states .
the augmented space theorem ( ast ) @xcite relates a specific matrix element of the green function in the augmented space @xmath6 to the configuration averaged green functions .
the rm is a purely real - space based technique , and therefore as applicable to a bulk system as one with surfaces , interfaces or extended defects . in the following
we shall describe each of the points raised above in some detail .
the tb - lmto has been described in great detail earlier @xcite,@xcite-@xcite .
we shall only quote here the main results which will be relevant for setting up the asr programme .
the starting point is the kohn - sham equation with the muffin - tin effective crystal potential .
the basis chosen for representation of the wave - function are the muffin - tin orbitals @xmath0 : if we expand the wave - function in terms of a linear combination of these muffin - tin orbitals and substitute the expansion in the kohn - sham equation we obtain the korringa - kohn - rostocker ( kkr ) secular equation : @xmath7 energy linearization gives the eigen - type lmto secular equation : e__n_m_ll - h^(2)__nl,_ml= 0 where , the ` second order ' hamiltonian is given by : ^(2)__n,_m = _ _ _ n_m+__n,_m - _ _
n,_k _ _ k _ _ k,_m [ ham ] note that each element is a matrix in the @xmath8 space . in the screened representation
the structure matrix @xmath9 is short - ranged and the hamiltonian is sparse : @xmath10 the expansion energies @xmath11 are chosen suitably by us at the center of the energy window of our interest and the potential parameters @xmath12 and @xmath13 are diagonal in @xmath8 space and are self - consistently generated .
the structure matrix @xmath14 is obtained from the geometry of the lattice .
note that the tb - lmto basis is minimal and hence usually it is enough to take @xmath15 and in a majority of cases one does not have to go beyond @xmath16 at this point we should comment on several possible generalizations : we have energy linearized the secular equation . in case
we do not wish to do so , we can still deal with a energy dependent hamiltonian " or secular matrix " @xmath17 .
this is the tb - kkr .
the subsequent recursion becomes energy - dependent , however each recursion at each energy point can be parallely carried out for efficiency .
such energy dependent recursion has been carried out by us earlier @xcite .
the assumption ( a posteriori shown after calculations ) was that the recursion coefficients are weakly energy dependent , therefore recursion is carried out an equi - spaced
seed " points and the intermediate points found by interpolation .
further , if we allow for third - nearest neighbour sparsity in the hamiltonian we could start with the real - space full - potential lmto ( rs - fplmto ) of eriksson @xcite .
whereas , the above formulation is general enough , we should stress on the fact that the hamiltonian representation is an infinite matrix .
lack of lattice translation symmetry means that we are unable to symmetry reduce the dimensionality of the matrix to essentially @xmath18 .
we therefore have to resort to techniques which allow us to deal with infinite matrices .
the recursion method was exactly such a technique introduced by haydock @xcite-@xcite to obtain the green functions associated with the secular equation .
this would give us the density of states from which we may obtain properties like the fermi energy , the charge and magnetization densities , the magnetic moment and the band energy .
a recent generalization also gives us correlation functions associated with response functions @xcite-@xcite .
the rm begins by recursively changing the basis through a three term recurrence relation : @xmath19 mutual orthogonality of this new basis gives : @xmath20 to understand how the above equations are efficiently operationally coded on the computer , let us first describe the hamiltonian as an operator @xmath21 . from eqn.([ham ] ) @xmath22 of these operators four are ` diagonal ' : * e*@xmath23 , * c * , @xmath24 and * o*. their structure are all of the form : @xmath25 . while the structure matrix is off - diagonal : @xmath26 .
if we represent a hilbert space ` vector ' @xmath27 by a matrix as shown below : @xmath28 the action of the diagonal operators * d * is rather simple : 1 . one by one choose non - zero rows @xmath29 of @xmath27 2 .
multiply @xmath30 and add this to the @xmath31-th row of a new @xmath32 .
3 . repeat this for all non - zero rows of @xmath27 .
finally @xmath33 the action of the off - diagonal operator * s * is more complicated as it has the information of the underlying lattice or network embedded in it .
the first step would be to prepare the neighbour map " ( nm ) .
the nm is a matrix @xmath34 where @xmath35 is the @xmath36-th neighbour of the @xmath31-th site on the lattice or network .
we have first to number the lattice / network points by integers .
the near neighbour vectors point in different directions , while numbering this directionality has to be carefully recorded , since the structure matrix between @xmath2 and @xmath37 will depend upon the direction of @xmath38 . in figure [ map ]
we show the geometry of a bulk square lattice , a square lattice with a plane surface ( 100 ) , one with a corrugated surface ( 110 ) and a network with four local bonds .
since the integer numbering increases always clockwise from a site , the directionality is preserved .
the neighbour maps , also preserving directionality , for the first three examples are : @xmath39 the first four columns give us the neighbours in these four directions and the last column gives us the number of neighbours .
for the fourth example of a fourfold coordinated planar random network , the neighbour map is the same for the square lattice , however the * s * is different .
the action of * s * on a vector @xmath27 is prompted by the relevant neighbour map .
the operations proceed as follows : 1 .
one by one choose rows @xmath29 of @xmath27 2 . from the neighbour map
choose @xmath40 ( which is the @xmath41-th neighbour of @xmath31 ) 3 .
multiply @xmath42 and add this to the @xmath43-th row of a new @xmath32 4 .
repeat this for all the non - zero rows of @xmath27 .
finally @xmath44 . the the recursive operations described in eqns . ( [ rec ] ) can be encoded into the following modules : + a. ( @xmath45 ) , which describes the action of the hamiltonian on a ` vector ' .
+ b. the ( @xmath46 ) which calculates the coefficients @xmath47 recursively : + the two vectors @xmath48 and @xmath49 are input / output vectors of the module hop , while the three vectors @xmath50 and @xmath51 are dummies which are needed within the module . the space for these dummies may be dynamically allocated during this procedure and released after the operation is over .
the same is true for the vector @xmath48 in the module rec .
( [ albe ] ) indicates that in the new basis the hamiltonian representation is tri - diagonal .
it follows immediately that the green function for the system is given by a continued fraction : @xmath52 the asymptotic part of the green function @xmath53 is called the _
terminator_. many terminators are available in literature .
the suitable one depends upon the way in which the coefficients @xmath54 behave as @xmath55 .
the most commonly used is the square root terminator which is suitable when @xmath56 .
luchini and nex @xcite have suggested a modification when calculations up to a large @xmath57 is not possible .
if we carry out the terminator approximation after @xmath57 steps the first @xmath58 moments of the density of states are _ exact _ and it has been shown that the asymptotic moments are also accurately reproduced .
beer and pettifor @xcite have suggested an alternative way of obtaining the terminator .
they note that @xmath59 so that the beer - pettifor prescription is a closed algorithm .
it wins over the square - root terminator described above when the convergence of the coefficients is either oscillatory or slow .
viswanathan and muller @xcite have suggested several other terminators like the terminator with an exponential tail , suitable when @xmath60 and the terminator with a gaussian tail , suitable when @xmath61 .
we shall have options for several terminating procedures so that the user may choose according to his need . c. the ( @xmath62 ) which calculates the green function .
+ the partial density of states , projected onto a particular @xmath63 , is : n__il(e ) = - _ 0 m g__il,_il(e+i ) it is clear from this discussion that the recursion procedure is general enough to deal with lattices of any complexity , surfaces and interfaces and disordered networks .
finally we come to the main part of the package : that which deals with disorder
. the augmented space formalism @xcite allows us to directly compute the configuration average of the green function by constructing a hamiltonian in the augmented space of configurations of the random parameters .
the formalism is based on ideas prevalent in quantum measurement theory : we associate with each random parameter @xmath64 an operator @xmath65 whose spectrum @xmath66 are the measured values of the parameter .
the configuration state of @xmath64 are the eigenkets @xmath67 of @xmath65 , which , therefore is an operator in the space @xmath68 spanned by the different configuration states . given this ,
the probability density of the parameter is @xmath69 where @xmath70 the augmented space theorem @xcite then gives us the configuration average of _ any _ function of the random variables : ( \{n_i})= ( \{n_i } ) here , 1 .
it is a member of the basis set @xmath72 which spans the configuration space @xmath73 of the set of random parameters @xmath74 .
@xmath75 is an operator in the configuration space @xmath76 .
it is the _ same _ operator functional of @xmath77 as @xmath78 was a function of @xmath74 .
the expression is exact .
the configuration average is done exactly first , then the approximations may be carried out maintaining various physical constraints .
the philosophy is very different from the mean - field approximations like the cpa .
if we apply this to the configuration averaged green function of a random hamiltonian @xmath79 : @xmath80 let us take as an example the case of a disordered binary alloy : the hamiltonian has the same form as eqn .
( [ ham ] ) but : @xmath81 the random occupation " variables @xmath64 take the values 0 and 1 with probabilities proportional of the concentrations x and y of the constituents a and b. @xmath65 has a @xmath82 representation : @xmath83 .
if we denote the basis of the above representation by @xmath84 , then @xmath85 and each of the potential parameters become operators in the configuration space , for example : @xmath86 where @xmath87 and @xmath88
. also , since the structure matrix is not random : = _ ij _ _
i_j i_ij [ str ] @xmath89 if we now compare eqns .
( [ avgreen ] ) and ( [ ham - as ] ) with eqns .
( [ ham2])-([gf ] ) it becomes clear that once we have defined the hamiltonian in the space of configurations , the recursion method may be directly used to obtain the configuration averaged green function .
the approximation involved will then only be the termination " approximation .
heine s black - body theorem " @xcite indicates that most electronic structure energetics is dominated by the immediate environment of a site .
this is a major justification of the termination approximation . unlike the cpa which gives only the first eight moments of the density of states exactly and
the asymptotic moments accurately , the augmented space recursion gives @xmath58 moments exactly ( where the termination is done after @xmath57 steps ) and the asymptotic moments also accurately . in most of our calculations
we can take @xmath57 about 8 - 9 steps .
the full package is divided into several modules .
the first is the preparation module .
this module has two branches , one each for each constituent of the alloy .
each branch runs parallely in two different slave processors .
the structure matrix is prepared in the master processor .
this module prepares the following : 1 .
it prepares the two control files for the different alloy constituents .
this part is interactive with the user .
2 . it carries out the simple hartree calculation and prepares the atomic radii of the constituents and the initial charge density .
3 . it calculates the overlap between atomic spheres and inputs empty spheres maintaining the symmetries .
4 . it calculates the structure matrix from the inputs of the control files .
the preparation module is followed by the main asr - module .
this module is called by the routine _ lmasr_.
this main module is divided into five smaller modules : * : this module reads the data generated by the preparation module .
it checks the inputs for consistency .
the routines in this module are again those in the stuttgart lmto47 , but modified to read the inputs for _ both _ the two constituents of the alloy .
0.4 cm * : at the start of the dft self - consistency loop , this module takes the overlap of a simple hartree atomic density calculation in the preparation module and generates the hartree and exchange - correlation potentials , spheridizes them and inputs them to the atomic module that follows . in later steps of the self - consistency loop ,
this module first mixes the charge densities of the earlier steps and prepares the hartree and exchange - correlation potentials , spheridizes them for input into the atomic module .
0.2 cm + - at this point there is a choice of using either the standard dft exchange - correlation potentials or , alternatively , there is a branch module _
harbola - sahni _ which sets up the harbola - sahni potential for the study of excited states @xcite .
0.4 cm * : this is the atomic module .
it takes the spheridized kohn - sham potential generated in module b and solves the radial kohn - sham equation numerically .
the kohn - sham orbitals and energies then lead to the potential parameters for each constituent .
those for the two constituents are calculated on different processors .
the parameters are first calculated in the orthogonal representation .
a new routine _ gtoa _ , not present in the stuttgart lmto47 package , then transforms them to the most screened tight - binding representation . 0.4 cm * : this is the main asr module .
the routines herein have been fully developed by us and form the main backbone of the package .
the input are the tight - binding hamiltonian parameters from the atomic module .
first they are combined with the alloy composition to prepare the augmented space hamiltonian . the nearest neighbour map in augmented space is then generated .
next , the recursion is carried out for each @xmath8 value , terminators generated and the @xmath8 projected density of states are calculated .
each different recursion for each @xmath8 value is carried out on a different processor , thus vastly accelerating the calculations .
+ we then proceed to calculate the total density of states and the fermi energy .
again , branching out into different processors , we calculate the @xmath8-dependent moments and magnetic moments . of all the modules , this is the one amenable to maximum parallelization .
0.2 cm + - at this point we have the possibility of introducing short ranged order.the asr for short - ranged order has been described in some detail earlier @xcite .
the branching for this choice occurs just before we set up the augmented space hamiltonian .
the extra input is the warren - cowley short ranged order parameter .
0.2 cm + - also at this point we have the option to introducing disorder in the structure matrix because of size mismatch between the two constituents of the alloy .
the branching now takes place earlier in the preparation module where we generate not one , but three different structure matrices : @xmath90 and @xmath91 . in the ` end point ' approximation
@xcite the eqn .
( [ str ] ) is now replaced by : + = _ ij \ { + ' '' '' + _
ij + ^(1)_ij ( n_i+n_j ) + ^(2)_ij n_i_j } _ ij + where @xmath92,@xmath93 and @xmath94 .
this modification will be available in this module . 0.2 cm +
it is in these last two options that the asr really scores over the cpa , which can not really deal with either short - ranged order or off - diagonal disorder as both involve more than one site .
the competing methodology is the special quasi - random structures ( sqs ) @xcite .
however , if we are dealing with materials with many atoms per unit cell and non - stoichiometric compositions , the sqs required will be rather large and will involve use of _ huge _ unit cells . here
the tb - lmto - asr with the use of much smaller unit cells will score .
we have shown earlier @xcite that both these two methods give virtually the same results for the density of states .
0.4 cm * : in this module the @xmath8-dependent moments are used to obtain the charge density .
we also calculate the total energy , including the madelung term . for the disordered alloy
, the madelung term is obtained from the procedure suggested by skriver and ruban @xcite . finally the old and new moments are mixed , the mixed charge density thus obtained is input back in module b. this is iterated till convergence in both energy and charge density is achieved .
fe@xmath95 for different compositions : ( left ) cr@xmath96fe@xmath97 ; ( middle ) cr@xmath98fe@xmath98 ; ( right ) cr@xmath97fe@xmath96 ; ( bottom ) dos for cu@xmath99zn@xmath95 for different compositions : ( left ) cu@xmath100zn@xmath101 ; ( middle ) cu@xmath102zn@xmath102 ; ( right ) cu@xmath103zn@xmath104 .
energy(ryd ) is plotted along @xmath105-axis ; @xmath106-axis is dos ( states / ryd ) ] for the dft part , our code depends heavily on stuttgart - lmto routines developed by anderson and co - workers @xcite two independent dft codes run parallely to produce the potential parameters of the two constituent atoms of the binary alloy .
these potential parameters are used at the input point of our asr routines in module d. our present dft modules deal only with collinear magnetism . for
spin dependent calculations the hamiltonian is separable in spin space .
thus the spin is merged with the label @xmath8 which is now @xmath107 and we have just to carry out twice the number of recursions : for @xmath108 and @xmath109 . in case
we wish to introduce spin - orbit coupling and possibility of non - collinear magnetism , we have to replace the dft module with one dealing with density matrices , rather than densities @xcite and the asr module with one applying generalized or vector recursion @xcite-@xcite which was designed to deal with hamiltonians whose representation in spin - space or ` spinor ' bases is not diagonal : @xmath110 this new module is under preparation @xcite and will be incorporated once the basic checks are carried out . here
we mention this in order to bring out the different possibilities and versatility of the package .
we have described the contents and commented upon the efficiency and versatility of the tb - lmto - asr package .
we shall conclude by describing two different applications : namely , the alloys @xmath111 where both the constituents are non - magnetic in the bulk and the @xmath112-band centres of cu and zn are well separated ; and fe@xmath99cr@xmath95 where both constituents are magnetic and their @xmath112-bands overlap .
lcc + & old tb - lmto - asr & new tb - lmto - asr + wall time & 699 sec & 225 sec + efficiency@xmath113 & 0.714 & 0.97 + [ comp ] the alloys have been studied earlier by us using the old version of the tb - lmto - asr @xcite,@xcite-@xcite and tb - lmto - ccpa @xcite , both introduced by us , and the kkr - icpa @xcite which is also based on the augmented space formalism , as well as through the kkr - cpa @xcite , paw - sqs@xcite and kkr - nl - cpa @xcite-@xcite .
they are therefore ideal system for benchmarking the new tb - lmto - asr package .
comparison of the densities of states shown in fig .
[ dos ] with the results shown in the above references will convince us that for fe@xmath99cr@xmath95 there is hardly anything to choose between the various techniques and the packages based on them .
however , for the split band alloy cu@xmath99zn@xmath95 the tb - lmto - asr , tb - lmto - ccpa and kkr - nl - cpa scores over the cpa versions , as expected from earlier analysis .
one of our our main point of interest is the relative runtime and efficiency of this new version of the tb - lmto - asr , as compared to the several earlier versions proposed by us . to benchmark these characteristics
we take a specific calculation on fe@xmath99cr@xmath95 .
each calculation is done on a 73109 site augmented space map , with n recursion steps with @xmath114 and @xmath115 , and @xmath116 , followed by a beer - pettifor termination scheme .
the old and new versions of the tb - lmto - asr are compared in the table [ comp ] . here
wall time is the same as run time and efficiency is the ratio between the wall time and cpu time .
gnu profiles ( gprof ) for the new tb - lmto - asr output and the old serial version are given in the table ( [ gpr ] ) . .gprof
data for the optimized and parallelized asr code and ( in brakets ) the old serial code .
the discontinuity between the cumulative time of ` spectral ` and ` splint ` is due to the machine routine which are not included here . [
cols="^,^,^,^,^,^,^ " , ]
we have presented here a computational package that combines three different techniques to allow us to study the electronic properties of disordered systems .
the package can deal with bulk disordered alloys as well as surfaces and interfaces , both smooth , stepped and rough and also structurally distorted lattices .
the package can be generalized at many stages which have been clearly commented upon and the generalizations are being carried out step by step .
99 r. haydock , v. heine , and m. kelly , 5 ( 1972 ) 2845 r. haydock , v. heine , and m. kelly . ,
8 ( 1975 ) 2591 a. mookerjee in _ electronic structure of alloys , surfaces and clusters _ ( taylor - francis , uk ) ed . d.d .
sarma and a. mookerjee ( 2003 ) .
a. chakrabarti and a. mookerjee,_e .
j. _ b44 ( 2005 ) 21 andersen o k , kasowski r v , b 4 ( 1971 ) 1064 a. mookerjee , ( 1973 ) 1340 r.v .
kasowski , o.k .
andersen o k , _ solid state commun .
_ 11 ( 1972 ) 1064 o.k .
andersen , r.g.woolley , _
_ 28 ( 1973 ) 905 o.k .
andersen , b 12 ( 1975 ) 3060 s. ghosh , n. das and a. mookerjee , _ int j. mod .
. _ b13 ( 1999 ) 723 j.m .
wills , o. eriksson and m. aloouani , in _ structure and physical properties of solids _ ed .
h. dreyse ( berlin , springer - verlag ) ( 2000 ) 148 t. j. godin , r. haydock , , ( 1988 ) 5237 t. j. godin , r. haydock , _ comp .
_ 64 ( 1991 ) 123 n.u .
luchini and c.m.m .
nex , _ j. phys .
c ( solid state ) _ 20 ( 1987 ) 3125 n. beer and d. g. pettifor , in _ the electronic structure of complex systems _ , ed .
p. phariseau and w. m. temmerman , ( plenum , new york , 1982 ) , 113 769 v.s .
viswanath and g. mller , in the user friendly recursion method " , ( troisieme cycle de la physique , en suisse romande ) ( 1993 ) v. heine , in solid state theory vol 35 ( academic press , new york ) ( 1980 ) o. k. andersen , phys .
b 12 ( 1975 ) 3060 m. rahaman , s. ganguly , p. samal , t. saha - dasgupta , a. mookerjee and m.k .
harbola , _ physica _ b : condens matter 404 ( 2009 ) 1137 a. mookerjee a. and r. prasad r. , ( 1993 ) 17724 t. saha and a. mookerjee , 8 ( 1996 ) 2915 a. zunger , s.h.- wei , l.g .
ferreira and j. bernard , ( 1990 ) 353 k. tarafder , s. ghosh .
b. sanyal , o. eriksson , a. mookerjee and a.chakrabarti , 20 ( 2008 ) 445201 a.v .
ruban and h.l .
skriver , ( 2003 ) 024201 o.k .
andersen and o. jepsen , ( 1984 ) 2571 r. lizrraga , s. ronnenteg , r .
berger , a. bergman , p. mohn , o. erikkson and l. nordstrm , b 70 ( 2004 ) 024407 s. ganguli , a. bergman and a. mookerjee , ( under preparation ) ( 2009 ) t. saha , i. dasgupta and a. mookerjee , b 50 ( 1994 ) 13267 t. saha , i. dasgupta and a. mookerjee , 8 ( 1996 ) 1979 a. chakrabarti and a. mookerjee , _
b 44 ( 2005 ) 21 m. rahaman and a. mookerjee , b 79 ( 2009 ) 054201 s. ghosh , p.l . leath and m.h .
cohen b66 , ( 2003 ) 214206 a. chakrabarti and a. mookerjee ,
_ e. phys .
b 44 ( 2005 ) 21 p.e.a .
turchi , l. reinhard and g.m .
stocks , b 50 , ( 1994 ) 15542 m. jarrell and h.r .
krishnamurthy , ( 2001 ) 125102 d.a .
rowlands , j.b .
staunton , b.l .
gyorffy , e. bruno and b. ginatempo , ( 2005 ) 045101 | we present here an optimized and parallelized version of the _ augmented space recursion code _ for the calculation of the electronic and magnetic properties of bulk disordered alloys , surfaces and interfaces , either flat , corrugated or rough , and random networks .
applications have been made to bulk disordered alloys to benchmark our code .
* h * # 1 # 1 : |
circular polarization ( cp ) observations of radio - loud active galactic nuclei ( agn ) provide a unique opportunity , especially in conjunction with linear polarization observations , for constraining the magnetic field geometry , particle energy distribution , and even particle composition ( @xmath3 vs. @xmath4 ) of their jets .
a number of possible mechanisms have been proposed for the production of circular polarization in agn ( see @xcite and references therein ) .
the two most likely are the _ intrinsic _ circular polarization of synchrotron radiation and faraday _ conversion _ of linear to circular polarization , a propagation effect , e.g. @xcite .
@xcite catalog a large number of circular polarization observations of agn , mainly from the 1970s , and find typical levels of @xmath5% . after the survey by @xcite
, there have been no major cp observations published until recently . in @xcite , we reported the detection of circular polarization in the parsec - scale jet of 3c279 .
detailed analysis suggested the jet is predominately a pair plasma .
circular polarization has since been detected in the galactic center @xcite , the galactic x - ray binary ss433 @xcite , the intra - day variable source pks 1519 - 273 @xcite , and a sample of more than twenty agn by @xcite , hereafter rn&s . in @xcite ,
hereafter h&w , we presented multi - epoch , parsec - scale cp detections of four agn , made with the very long baseline array ( vlba ) at 15 ghz . here
we present results from the first large scale circular polarization survey made at very long baseline interferometry ( vlbi ) resolution . in december 1996
, we observed a sample of 40 agn at 5 ghz ( @xmath6 6 cm ) with the vlba .
these observations were originally made as part of a continuing study of the parsec - scale linear polarization properties of blazars ( @xcite ; attridge et al .
, in prep . ) .
they do not comprise a complete sample .
we report the results of a re - analysis of these data for circular polarization .
the observations were scheduled in a continuous 48 hour block , with source changes after every 5.5 minute scan .
sources were highly interleaved in the schedule to maximize ( u , v)-coverage .
each source was observed for at least ten scans , five on each day , giving a total integration time of nearly an hour .
the data were recorded at each station using 1-bit sampling and were correlated at the vlba correlator in socorro , nm .
the correlated data contained all four cross - correlations ( rr , rl , lr , ll ) , each with four intermediate frequencies ( ifs ) . the data were loaded into nrao s astronomical imaging processing system ( aips ) @xcite and calibrated using standard techniques for vlbi polarization observations , e.g. , @xcite .
the vlba station in hancock , nh was found to have a low flux calibration and significantly higher noise than the other antennas and was dropped from the observations . for a specific description of the original calibration steps
see @xcite .
the reader is referred to h&w for a detailed discussion of how to detect small levels of circular polarization with the vlba . there
we describe three techniques for robustly detecting circular polarization with circularly polarized antenna feeds . in our re - analysis of the 5 ghz survey for circular polarization
, we applied the _ gain transfer _ technique which provides a direct measure of the circular polarization in a source .
we were unable to apply the other techniques described in h&w due to both a lack of strong circular polarization signals in the sample ( most are a few mjy or less ) and the lack of sufficiently bright extended structure .
the application of the gain transfer technique is straight forward , requiring that we simply make no assumption about the circular polarization of a source during self - calibration , i.e. we only assume @xmath7 .
this will leave uncalibrated the @xmath8 complex gain ratio at each antenna .
for the source model , we used the best model obtained from the original analysis , and self - calibrated the data in amplitude and phase .
after self - calibration , we removed the effects of antenna feed leakage determined by applying the aips task pcal to the unpolarized source oq208 .
this step is crucial as uncorrected feed leakage terms may induce non - closing errors in the parallel hand data .
after feed leakage correction , we self - calibrated the data again in amplitude and phase using the same model as before .
the data were then carefully edited to remove clearly discrepant polarization data from stokes q , u , and v in an antenna based manner .
these `` final '' data , could then be used to either ( 1 ) solve for the @xmath8 complex gain ratios by self - calibration assuming no circular polarization in that source ( @xmath9 , @xmath10 ) , or ( 2 ) look for circular polarization on a source by applying a smoothed set of @xmath8 complex gains from a sub - set of sources .
the key to gain transfer calibration is determining a smoothed set of @xmath8 complex antenna gains . as suggested in h&w , we initially used nearly the entire 40 source sample to produce a smoothed set of gains .
the large sample size was a great asset , and by smoothing the gains over several hours and averaging over many sources , we were insensitive to the particular characteristics ( such as strong circular polarization ) of any one source .
we then applied this smoothed set of gains back to the sources , imaged in circular polarization , and removed the most strongly polarized sources from contributing to the smoothed gains .
we repeated this process through a few iterations until there were no sources with apparent cp @xmath11% that contributed to the smoothed gains .
mjy ) which would not produce at least @xmath12 mjy of cp at @xmath13% .
] this cutoff gave thirteen sources contributing to the smoothed gains with a boxcar averaging window of six hours .
our main results are insensitive to both the details of which sources contribute to the gains , and our choice of smoothing function and time - scale .
the reader is referred to h&w for a detailed discussion of the uncertainties associated with circular polarization observations with circularly polarized feeds . here
we would like to briefly remind the reader of two points .
( 1 ) while it is important to remove feed leakage terms from the parallel hand data , contributions from the small uncertainties in the determination of the feed leakage terms are insignificant relative to the gain uncertainties .
( 2 ) beam squint , resulting from pointing errors , shows up as a pure amplitude gain effect and is therefore naturally included in our estimate of gain uncertainties . in the appendix
we have refined our estimate of the gain calibration uncertainties as presented in h&w . as a final check on our results
, we divided the single 48-hour data set into two sub - sets and imaged each separately in cp .
this provided a direct test of our understanding of the uncertainties from all but the longest time - scale effects ( @xmath14h ) .
we found no significant differences between the two parts of the 48-hour observation .
table [ t : results ] lists our results , and figure [ f:0607 ] presents an image of the circular polarization distribution of a typical source ( pks 0607 - 157 ) . the sources in this sample tend to be strongly core dominated , and the cp we observe is coincident with the bright vlbi cores .
any small displacements from the vlbi cores do not appear significant within our positional uncertainties which are approximately a beam - width divided by the snr of the cp measurement .
for this reason , we simply report the measured cp as a fraction of the peak total intensity .
the uncertainties and limits listed in table [ t : results ] are dominated by gain calibration uncertainties which are estimated in the appendix .
the total quoted uncertainty , @xmath15 , includes both the gain uncertainty and the rms noise in the stokes v map , added in quadrature .
the limits in table [ t : results ] are the largest of @xmath12 mjy , @xmath16 , or @xmath17 .
figure [ f : detect ] compares our detection rates for circular polarization for bl lacs and quasars with the sample of rn&s ( also at 5 ghz ) . for this comparison
we have divided by the total vlbi flux ( rather than the peak vlbi flux ) to better compare to the australian telescope compact array ( atca ) integrated measurements .
our detection rates are very similar to those of rn&s .
we can do a similar comparison to the 15 ghz vlbi sample of h&w .
here we compare the peak values for fractional circular polarization , @xmath18 , in our 5 ghz measurements to @xmath18 of the integrated core at 15 ghz , including model components out to 1.5 mas .
this comparison essentially eliminates resolution based differences between the samples . while we find only 2/36 quasars or bl lacs with @xmath19% in the 5 ghz measurements presented here , the 15 ghz sample of h&w has a significantly higher detection rate for strong cp .
there are five closely spaced ( @xmath20 2 month intervals ) epochs in h&w which include the same 11 quasars and bl lacs , three of which show repeated detections of cp .
averaging over the epochs in h&w , we have a mean of 2.4/9.6 detections of @xmath19% from the quasars and bl lacs strong enough in total intensity in those epochs to have detectable cp at this level .
although the total number of sources in h&w is small , this detection rate is nearly five times that seen here at 5 ghz , and we note that most of the detections in h&w are @xmath21% cp , and we observe no source with cp that strong in this 5 ghz sample . the higher incidence of strong circular polarization detections in h&w may be due to one of three possible causes : ( 1 ) random fluctuations in small samples , ( 2 ) a broad frequency dependent effect , or ( 3 ) source selection criteria .
we can evaluate option ( 1 ) by assuming simple binomial statistics based on our 5 ghz observations ( i.e. assume the chance of observing @xmath19% is @xmath22 ) .
we find we would need approximately a 2.5 sigma random variation to explain the 15 ghz results of h&w .
any broad frequency dependent effects , option ( 2 ) , depend on the generation mechanism of the circular polarization .
while both intrinsic cp and faraday conversion have specific , frequency dependent signatures , the inhomogeneous , self - similar nature of vlbi radio cores complicates the analysis considerably ( e.g. @xcite ) and may lead to a flat @xmath18 spectrum for both mechanisms @xcite . with
no clear frequency dependent effect , we suspect that source selection criteria , option ( 3 ) above , may explain the higher levels of circular polarization observed in h&w .
the sample of h&w was part of multi - epoch monitoring program of the strongest , most currently active blazars ( @xcite ; ojha et al . in prep . ) , and the three quasars with high levels of cp were undergoing violent core outbursts during the observations .
this is consistent with the report by rn&s of a significant correlation between overall source variability and fractional cp . for rn&s , who have only integrated observations , the correlation with source variability
may be confused by a correlation with overall source `` compactness '' , as extended structure will decrease both the fractional cp of a source and its fractional variability . for the vlbi observations of h&w compactness
is not an issue , and the higher levels of fractional cp appear correlated with the selection criteria of currently active sources for that sample .
it is unclear how current core outburst events may lead to increased fractional circular polarization via the intrinsic mechanism .
outbursts are typically associated with shocks that tend to order the magnetic field and increase particle densities . increased particle densities should not enhance the _
fractional _ contribution of intrinsic circular polarization , and it is difficult to imagine how the _ uni - directional _ field necessary for intrinsic cp should be preferentially enhanced by a shock .
if the uni - directional field is along the jet axis and the shocks are transverse , just the opposite should happen , and the relative contribution from intrinsic circular polarization should decrease . for faraday conversion ,
ordered magnetic fields are necessary for both the generation of linear polarization and its conversion to circular polarization .
shocks in the core may very well increase the contribution of faraday conversion to the circular polarization produced .
rn&s found no significant difference between the distribution of circular polarization in the quasars and bl lacs in their sample , although they did find that bl lacs and quasars had significantly higher levels of cp than radio galaxies .
their detailed analysis was made possible by the great precision ( @xmath23% ) with which atca can measure integrated cp .
due to our higher uncertainty ( @xmath24% ) , which induces a dependence on source strength , we can not do a similar detailed analysis .
the bl lacs in our sample tend to be weaker than the quasars and many have upper limits higher than the typical detection level for quasars .
figure [ f : mc_ml ] is a plot of fractional core circular polarization , @xmath25 , versus fractional core linear polarization , @xmath26 .
beyond the fact that linear polarization is almost always stronger than circular polarization , there does not appear to be any correlation between linear and circular polarization of the core .
this was also noted by rn&s .
one might expect that fractional linear polarization , a measure of field order , should correlate with fractional cp , as disordered magnetic fields will not generate significant _ net _ cp by either the intrinsic or conversion mechanism .
the issue is complicated , however , by the possibility of faraday depolarization in the cores of these objects .
large rotation measures have been observed in the cores of quasars by @xcite , and if the faraday effect is occurring in an external screen , the cp will remain largely unmodified ( cgs ) , e.g. @xcite ] , although the linear polarization may be significantly reduced .
the scintillation mechanism described by @xcite for the production of cp has no correlation with linear polarization , but is expected to have rapid changes ( @xmath20 minutes hours ) in sign which are not observed in agn ( see [ s : consist ] ) .
we find no correlation for fractional cp with either total intensity or redshift in this sample .
we have a single source in common with the observations of rn&s : pks 0454 - 234 .
they found @xmath27% integrated circular polarization in 1997.17 at 5 ghz , compared to our @xmath28 result of @xmath29% core circular polarization measured 11 weeks earlier ( also at 5 ghz ) .
for the source pks 0607 - 157 we detect circular polarization of approximately @xmath30% in january of 1998 at 8 ghz ( homan et al .
, in prep . )
, a signal which is stronger , but of the same sign as we report here at 5 ghz in observations made a full year earlier .
this kind of short - term sign consistency in cp observations , as well as the possibility for longer - term consistency , is discussed in [ s : consist ] . in 3c279
we observe strong circular polarization for five epochs in 1996 at 15 ghz ( h&w ) ; however , in the 5 ghz sample presented here , we do not detect significant cp for 3c279 in december of 1996 a little more than two months after the final 15 ghz cp observations by h&w .
in @xcite we argue that the cp in 3c279 is associated with the core - west ( cw ) component which appears to be part of a core outburst event in 3c279 during 1996 .
this component would have been strongly self - absorbed at 5 ghz in late 1996 , so we should not expect to see any cp from it .
@xcite noted that circularly polarized sources tended to have a _ preferred handedness _ or _
sign _ which they maintained over the course of a few years at 5 ghz . in h&w
, we observed the same effect at 15 ghz over five epochs , taken at two month intervals during 1996 .
here we present evidence that this sign consistency may persist for decades . between our sample and those of rn&s and h&w , there are six sources with circular polarization detected both recently and @xmath1 years ago @xcite : 3c84 , pks 0537 - 441 , 3c273 , 3c279 , pks 1921 - 293 , and pks 1934 - 638 .
of those six sources , five have the same sign of circular polarization today as they preferred two decades ago measurements , cta 102 and 3c454.3 , were also detected @xmath1 years ago with the same sign @xcite . ] .
though the statistics are small , this result is potentially very interesting .
it suggests a long - lived property of the jets , stable on the time - scale of multiple outbursts , which sets the preferred sign of circular polarization @xcite . a natural possibility for this property is the polarity of a net magnetic flux in the jet , reflecting the magnetic flux at the central engine
this is a fundamental parameter in electro - magnetic models of jet production ( e.g. @xcite ) .
a net flux can generate either intrinsic circular polarization or may drive the faraday conversion process with a small amount of faraday rotation , e.g. @xcite .
another possible long - term property is the handedness of a helical field which may produce significant quantities of faraday conversion , e.g. @xcite .
we have presented circular polarization results from a survey of the parsec - scale polarization polarization of 40 agn at 5 ghz .
we found 11 circular polarization detections at the @xmath0 level or higher .
this nearly quadruples the number of sources detected in cp at vlbi resolutions .
we observed only two sources with high levels of cp , @xmath31% , and we believe that current outbursts in the sources observed by h&w could be related to the higher levels of cp detected there .
we found no correlation between fractional linear and circular polarization across our sample .
this may be due to faraday depolarization in the cores of agn which will leave circular polarization unaffected . in comparing these and other recent results to cp observations made @xmath1 years ago , we found that , in five of six cases , sources have the same sign of cp today as they did @xmath2 years ago .
this suggests the presence of a long term property of the jets , such as the polarity of a net magnetic flux , that is stable on the time - scale of multiple outbursts .
this work has been supported by nsf grants ast 95 - 29228 and ast 98 - 02708 .
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 .
this research has also made use of nasa s astrophysics data system abstract service .
the calibration of circular antenna feeds for circular polarization is described in detail in h&w .
the _ gain transfer _ procedure allows calibration of the @xmath8 complex antenna gain ratio .
] at each antenna by self - calibration assuming zero circular polarization ( @xmath32 , @xmath33 ) on a subset of sources , smoothing the derived antenna gains , and applying the smoothed gains to all sources .
figure [ f : brgains ] shows these gains for a typical antenna in our observations . in h&w
, we showed that there were two main contributions to the complex antenna gain ratios : ( 1 ) a long - term offset and ( 2 ) a short - term rapid variation .
both contributions were of the order of a percent , and while the long - term offset could be easily corrected , the short - term rapid variation went essentially uncorrected .
we noted that the short - term rapid variations seemed uncorrelated between antenna , if , and scan , and used this to crudely estimate our uncertainties from gain transfer calibration . in this appendix
we provide a more detailed analysis of the uncertainties associated with this calibration technique . for a given antenna ,
assume we have corrected the complex antenna gain ratio , leaving some residual error @xmath34 in amplitude and @xmath35 in phase . under these circumstances ,
the residual , uncorrected gain ratio can be written for antenna @xmath36 as @xmath37 where both @xmath34 and @xmath35 are on the order of 0.005 with the largest contributions from rapid , uncorrelated variations . in the stokes
@xmath38 visibilities , these errors show up , to first order , multiplying the total intensity : @xmath39\ ] ] here we consider amplitude errors , as only they show up on the phase center and may be confused for a real circularly polarized signal . for point - like sources , phase errors will not contribute to first order , and for sources with significant extended structure , phase errors will appear in an anti - symmetric fashion about the phase center .
the total uncertainty from gain transfer , @xmath40 , in the fractional circular polarization includes contributions to @xmath34 from three sources : ( 1 ) uncertainty in the determination of the smoothed antenna gains , ( 2 ) uncertainty in the true circular polarization of the calibrator sources , and ( 3 ) uncorrected rapid variations in the antenna gain ratios .
is estimated by examining the scan to scan variations of the thirteen sources used to determine the average gains .
we computed the rms variation in the individual gains from the six hour boxcar average determined on these sources . before computing the rms , we took a direct average of the variations across all ifs for each scan . this is to avoid having to make any assumptions about the degree of correlation in the variations between ifs higher than simply assuming the four ifs were completely uncorrelated in their variations . ] .
we found an rms variation for the calibrators of @xmath41 .
the uncertainty in any given six hour average is given by @xmath42 , where @xmath43 is the number of scans in that averaging period , and the contribution to the circular polarization of a target source is @xmath44 for these observations the number of antennas , @xmath45 , was nine .
the number of independent averages , @xmath46 , which apply to a given source was taken to be two as each source was observed over a period of five to six hours on each of the two days of the experiment .
the average numbers of calibrator scans , @xmath43 , contributing to the gains applied to each source were computed and are listed in table [ t : errors ] .
is easily estimated by finding the rms apparent cp of the calibrators and dividing by the square root of the number of calibrators contributing to a given gain average : @xmath47 for these observations , the rms peak circular polarization of the calibrators was @xmath48 .
as each source ( targets and calibrators alike ) was observed over a five to six hour period on each day , and the gain averages are done over a six hour interval , each source overlaps with many calibrator sources . unfortunately , the contributions of each of these calibrator sources were not equal , and we needed to use a weighted average computed from the experiment schedule and listed in table [ t : errors ] .
these weighted averages were determined in a conservative fashion : the calibrator source with the largest contribution ( number of calibrator scans that influenced each target scan ) to a given source was treated as a 1.0 contribution , and the other calibrators had their _ relative _ contribution added to this .
can be directly computed for each source : @xmath49 the rms variation for each source was computed relative to a twelve - hour average on that source , averaging the variations across if before computing the rms so that we made no assumption about the extent of correlated variation between if smaller if we assumed the ifs were uncorrelated in their variations . ] .
ccccccccc 0215@xmath51015 & pks & @xmath52 & bl & @xmath53 & @xmath54 & & @xmath55 & + 0219@xmath56164 & pks & @xmath57 & bl & @xmath58 & @xmath59 & & @xmath60 & + 0219@xmath51428 & 3c066a & @xmath61 & bl & @xmath62 & @xmath63 & & @xmath64 & + * 0336@xmath56019 * & * cta 26 * & @xmath65 & * q * & @xmath66 & @xmath67 & @xmath68 & @xmath69 & @xmath70 + 0403@xmath56132 & pks & @xmath71 & q & @xmath72 & @xmath73 & & @xmath74 & + 0420@xmath56014 & pks & @xmath75 & q & @xmath76 & @xmath77 & & @xmath78 & + 0422@xmath56004 & pks & @xmath79 & bl & @xmath80 & @xmath81 & @xmath82 & @xmath83 & + 0454@xmath56234 & pks & @xmath84 & q & @xmath85 & @xmath86 & @xmath87 & @xmath88 & @xmath89 + 0605@xmath56085 & pks & @xmath90 & q & @xmath91 & @xmath73 & & @xmath92 & + * 0607@xmath56157 * & * pks * & @xmath93 & * q * & @xmath94 & @xmath95 & @xmath96 & @xmath97 & @xmath98 + * 0743@xmath56006 * & * pks * & @xmath99 & * q * & @xmath100 & @xmath73 & @xmath101 & @xmath102 & @xmath103 + 0846@xmath51513 & & @xmath104 & bl & @xmath105 & @xmath70 & & @xmath106 & + 0851@xmath51202 & oj287 & @xmath107 & bl & @xmath108 & @xmath109 & & @xmath92 & + 0855@xmath51143 & 3c212 & @xmath110 & q & @xmath111 & @xmath112 & & @xmath113 & + 0906@xmath51015 & da 263 & @xmath114 & q & @xmath115 & @xmath116 & & @xmath117 & + * 0906@xmath51430 * & * 3c216 * & @xmath118 & * q * & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath70 + 1055@xmath51018 & pks & @xmath123 & q & @xmath124 & @xmath125 & @xmath126 & @xmath64 & + * 1150@xmath51497 * & * 4c49.22 * & @xmath127 & * q * & @xmath128 & @xmath129 & @xmath130 & @xmath131 & @xmath132 + 1156@xmath51295 & 4c29.45 & @xmath133 & q & @xmath134 & @xmath135 & & @xmath78 & + 1253@xmath56055 & 3c279 & @xmath136 & q & @xmath137 & @xmath138 & @xmath139 & @xmath78 & + 1334@xmath56127 & pks & @xmath140 & q & @xmath141 & @xmath142 & @xmath143 & @xmath144 & @xmath145 + 1404@xmath51286 & oq208 & @xmath146 & syi & @xmath147 & @xmath148 & & @xmath92 & + 1413@xmath51135 & oq122 & @xmath149 & bl & @xmath150 & @xmath148 & @xmath151 & @xmath152 & + 1502@xmath51106 & or103 & @xmath153 & q & @xmath154 & @xmath155 & & @xmath156 & + 1504@xmath56166 & pks & @xmath157 & q & @xmath158 & @xmath159 & @xmath160 & @xmath161 & + 1510@xmath56089 & pks & @xmath162 & q & @xmath163 & @xmath164 & @xmath165 & @xmath166 & + * 1546@xmath51027 * & * pks * & @xmath167 & * q * & @xmath168 & @xmath169 & @xmath170 & @xmath171 & @xmath172 + 1656@xmath51053 & pks & @xmath173 & q & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath135 + * 1842@xmath51681 * & & @xmath178 & * q * & @xmath179 & @xmath59 & @xmath180 & @xmath181 & @xmath182 + * 1921@xmath56293 * & * pks * & @xmath183 & * q * & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath125 + 1958@xmath56179 & pks & @xmath188 & q & @xmath189 & @xmath190 & & @xmath92 & + 2032@xmath51107 & pks & @xmath191 & bl & @xmath192 & @xmath169 & & @xmath161 & + * 2201@xmath51171 * & * pks * & @xmath193 & * q * & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 + * 2223@xmath56052 * & * 3c446 * & @xmath199 & * q * & @xmath200 & @xmath195 & @xmath201 & @xmath202 & @xmath203 + 2227@xmath56088 & pks & @xmath204 & q & @xmath205 & @xmath206 & & @xmath156 & + 2230@xmath51114 & cta 102 & @xmath207 & q & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath135 + 2234@xmath51282 & & @xmath212 & q & @xmath213 & @xmath214 & @xmath215 & @xmath216 & + * 2243@xmath56123 * & *
pks * & @xmath217 & * q * & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath182 + 2251@xmath51158 & 3c454.3 & @xmath222 & q & @xmath223 & @xmath73 & @xmath224 & @xmath225 & @xmath59 + 2345@xmath56167 & pks & @xmath226 & q & @xmath227 & @xmath228 & @xmath229 & @xmath78 & + ccccccc 0215@xmath51015 & @xmath230 & @xmath67 & @xmath231 & @xmath232 & @xmath233 & @xmath234 + 0219@xmath56164 & @xmath235 & @xmath236 & @xmath237 & @xmath232 & @xmath233 & @xmath238 + 0219@xmath51428 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 + 0336@xmath56019 & @xmath245 & @xmath155 & @xmath246 & @xmath247 & @xmath248 & @xmath249 + 0403@xmath56132 & @xmath250 & @xmath63 & @xmath251 & @xmath252 & @xmath246 & @xmath253 + 0420@xmath56014 & @xmath254 & @xmath89 & @xmath255 & @xmath252 & @xmath251 & @xmath256 + 0422@xmath56004 & @xmath254 & @xmath89 & @xmath257 & @xmath252 & @xmath251 & @xmath234 + 0454@xmath56234 & @xmath258 & @xmath77 & @xmath233 & @xmath259 & @xmath260 & @xmath261 + 0605@xmath56085 & @xmath262 & @xmath155 & @xmath263 & @xmath252 & @xmath264 & @xmath265 + 0607@xmath56157 & @xmath266 & @xmath89 & @xmath267 & @xmath259 & @xmath251 & @xmath268 + 0743@xmath56006 & @xmath269 & @xmath270 & @xmath271 & @xmath252 & @xmath272 & @xmath273 + 0846@xmath51513 & @xmath274 & @xmath270 & @xmath275 & @xmath232 & @xmath272 & @xmath276 + 0851@xmath51202 & @xmath277 & @xmath67 & @xmath243 & @xmath232 & @xmath233 & @xmath273 + 0855@xmath51143 & @xmath278 & @xmath164 & @xmath279 & @xmath232 & @xmath280 & @xmath281 + 0906@xmath51015 & @xmath282 & @xmath283 & @xmath253 & @xmath232 & @xmath280 & @xmath284 + 0906@xmath51430 & @xmath285 & @xmath286 & @xmath246 & @xmath232 & @xmath280 & @xmath287 + 1055@xmath51018 & @xmath288 & @xmath198 & @xmath271 & @xmath289 & @xmath290 & @xmath291 + 1150@xmath51497 & @xmath292 & @xmath203 & @xmath291 & @xmath242 & @xmath272 & @xmath253 + 1156@xmath51295 & @xmath293 & @xmath98 & @xmath290 & @xmath289 & @xmath241 & @xmath256 + 1253@xmath56055 & @xmath292 & @xmath294 & @xmath267 & @xmath242 & @xmath264 & @xmath273 + 1334@xmath56127 & @xmath295 & @xmath77 & @xmath255 & @xmath242 & @xmath260 & @xmath256 + 1404@xmath51286 & @xmath296 & @xmath182 & @xmath297 & @xmath242 & @xmath264 & @xmath244 + 1413@xmath51135 & @xmath298 & @xmath145 & @xmath241 & @xmath247 & @xmath299 & @xmath300 + 1502@xmath51106 & @xmath301 & @xmath159 & @xmath280 & @xmath247 & @xmath302 & @xmath261 + 1504@xmath56166 & @xmath303 & @xmath145 & @xmath280 & @xmath247 & @xmath299 & @xmath261 + 1510@xmath56089 & @xmath250 & @xmath159 & @xmath238 & @xmath252 & @xmath302 & @xmath146 + 1546@xmath51027 & @xmath269 & @xmath109 & @xmath267 & @xmath252 & @xmath260 & @xmath256 + 1656@xmath51053 & @xmath304 & @xmath98 & @xmath248 & @xmath252 & @xmath241 & @xmath305 + 1842@xmath51681 & @xmath296 & @xmath286 & @xmath306 & @xmath242 & @xmath280 & @xmath261 + 1921@xmath56293 & @xmath274 & @xmath125 & @xmath263 & @xmath232 & @xmath307 & @xmath273 + 1958@xmath56179 & @xmath308 & @xmath270 & @xmath309 & @xmath289 & @xmath272 & @xmath291 + 2032@xmath51107 & @xmath310 & @xmath283 & @xmath272 & @xmath289 & @xmath280 & @xmath244 + 2201@xmath51171 & @xmath311 & @xmath116 & @xmath291 & @xmath312 & @xmath313 & @xmath261 + 2223@xmath56052 & @xmath314 & @xmath315 & @xmath271 & @xmath312 & @xmath271 & @xmath302 + 2227@xmath56088 & @xmath316 & @xmath116 & @xmath241 & @xmath312 & @xmath313 & @xmath273 + 2230@xmath51114 & @xmath317 & @xmath318 & @xmath271 & @xmath312 & @xmath313 & @xmath299 + 2234@xmath51282 & @xmath319 & @xmath318 & @xmath241 & @xmath312 & @xmath313 & @xmath273 + 2243@xmath56123 & @xmath320 & @xmath321 & @xmath313 & @xmath322 & @xmath271 & @xmath299 + 2251@xmath51158 & @xmath314 & @xmath116 & @xmath259 & @xmath312 & @xmath313 & @xmath246 + 2345@xmath56167 & @xmath323 & @xmath142 & @xmath271 & @xmath312 & @xmath313 & @xmath299 + | we present circular polarization results from a 5 ghz survey of the parsec - scale polarization properties of 40 agn made with the vlba . we find 11 circular polarization detections at the @xmath0 level or higher .
this nearly quadruples the number of sources detected in circular polarization at vlbi resolution .
we find no correlation between fractional linear and circular polarization across our sample .
a likely explanation is external faraday depolarization in the cores of agn which reduces linear polarization but leaves circular polarization unchanged . in comparing ours and other recent results to observations made @xmath1 years ago , we find that , in five of six cases , sources have the same sign of circular polarization today as they did @xmath2 years ago .
this suggests the presence of a long term property of the jets , such as the polarity of a net magnetic flux , which is stable on time - scales much longer than those of individual outbursts . |
recently , georgi @xcite pointed out an interesting possibility for the existence of a scale - invariant sector with a continuous mass distribution .
this scale invariant stuff was coined the term `` unparticle '' to describe a possible scale - invariant hidden sector sitting at an infrared fixed point at a high scale @xmath6 .
the scale - invariant sector may be weakly or strongly interacting but its effects on the standard model ( sm ) is assumed to be weakly interacting . in georgi
s scheme @xcite , the hidden sector communicates with the sm content via a messenger sector characterized by a high mass scale @xmath7 . at energy below @xmath7 , one can integrate out the messenger sector and ends up with effective operators suppressed by inverse powers of @xmath7 in the following form @xmath8 where @xmath9 and @xmath10 represent local operators of the sm and hidden sector with scaling dimensions @xmath11 and @xmath12 , respectively .
as one scales down the theory from @xmath7 , the hidden sector may flow to an infrared fixed point at the scale @xmath6 , which is generated by quantum effects via dimensional transmutation for example .
at the fixed point the hidden sector becomes scale invariant , the above operator eq .
( [ effectiveop ] ) has to be replaced by a new set of operators of similar form @xmath13 where @xmath14 is the unparticle operator with a scaling dimension @xmath15 and @xmath16 is the unknown coefficient . due to
the underlying theory is a scale invariant interacting theory , the scaling dimension @xmath15 needs not having the canonical values of integer or half - integer like the free boson or free fermion cases .
besides its scaling dimension , the unparticle operator @xmath14 can be characterized by scalar , vector , tensor , or spinor etc according to its lorentz group representation . despite the scale invariant sector remains unspecified , the 2-point function @xcite and the feynman propagator @xcite of the unparticle field operator @xmath17 can be determined by scale invariance .
the normalization of the 2-point function of unparticle operator of scaling dimension @xmath15 was fixed by georgi @xcite to be the same as the phase space of @xmath15 massless particles . on the other hand ,
the most peculiar feature of the unparticle propagator is a phase factor @xmath2 associated only with time - like momenta .
this cp - conserving phase has been shown to have interesting interference effects at high energy experiments @xcite and other phenomenology . in this work ,
we consider photon - photon scattering via unparticle exchanges .
the sm contribution to photon - photon scattering can only arise from loop diagrams with all charged particle running around the loop and thus is highly suppressed .
it is anticipated that the cross section due to unparticle exchange can easily surpass the sm cross section at high enough energies , because exchanges of unparticle are at the tree - level . moreover , photon scatters via unparticle exchanges in all @xmath3- , @xmath4- , and @xmath5-channels .
the peculiar phase @xmath2 associated with the @xmath3-channel exchange gives rise to interesting interference with the @xmath4- and @xmath5-channel amplitudes .
similar effects had been studied in the model of large extra dimensions @xcite .
note that similar ideas for the spin-0 unparticle have been pursued recently in refs .
@xcite . however , our analytic results disagree with ref .
we suspect that the phase factor @xmath2 associated with the @xmath3-channel unparticle propagator was not taken care of properly .
our results are consistent with ref .
@xcite where we overlap .
in addition , we extend these previous calculations to the spin-2 unparticle exchange , which is highly nontrivial . it has been pointed out recently by grinstein _
et al _ @xcite that the vector and tensor unparticle propagators for a conformal invariant hidden sector differ from a scale invariant ones .
unitarity constraints @xcite on the scaling dimensions of the unparticle operators with conformal symmetry are also emphasized in their work @xcite . in this work ,
we follow the original georgi s scheme by assuming just scale invariance in the derivation of the unparticle propagators .
integrating out the heavy messenger sector can also lead to contact interactions among sm fields of the form @xmath18 and they can compete with the effects from unparticle exchanges @xcite .
for example , the following two dimension 8 operators @xmath19 and @xmath20 can be induced and they can also contribute to the elastic photon - photon scattering .
we assume the coefficients of these operators are minuscule and ignore them in our analysis .
a complete analysis for the photon - photon scattering including all the interference effects among the sm contribution , unparticle exchanges as well as these contact interactions is interesting but beyond the scope of this work .
the organization of the paper is as follows . in the next section ,
we give in details the scattering amplitudes for @xmath21 via spin-0 as well as spin-2 unparticle exchange . in sec .
iii , we compare the unparticle contribution with the sm contribution in the angular distribution and in the total cross section .
we also look into the nontrivial effects of the phase @xmath2 of the @xmath3-channel propagator .
we conclude in sec . iv .
the interaction of spin-0 unparticle @xmath22 with the photon can be parameterized by @xcite @xmath23 where @xmath24 is an unknown coefficient of order @xmath25 , and @xmath26 is the field strength of the photon field .
the unparticle propagator is @xcite @xmath27 where @xmath28 is given by @xmath29 the peculiar phase associated with the propagator arises from the negative sign in front of @xmath30 in eq.([unpropagator ] ) : @xmath31 therefore , the @xmath3-channel propagator has the nontrivial phase @xmath2 while the @xmath4- and @xmath5-channel propagators do not .
there are three feynman diagrams contributing to @xmath32 with the unparticle exchanges in @xmath3- , @xmath4- , and @xmath5-channels .
the sum of amplitudes for these three diagrams is given by @xmath33 where @xmath34 in the above amplitude , we can write the mandelstam variables as @xmath35 such that the phase @xmath36 associates manifestly with the @xmath3-channel only .
it is obvious that each channel is separately gauge invariant .
the square of the amplitude averaged over initial polarizations is given by @xmath37 + ( |t| |u|)^{d_{{\cal u } } } \biggr\ } \;.\ ] ] if the phase factor @xmath38 were removed , the amplitude squared would have been symmetric in @xmath39 .
note that we have written the mandelstam variables as @xmath40 , where @xmath41 and @xmath42 and @xmath43 is the central scattering angle .
the angular distribution is given by @xmath44 + \left ( \frac{1 -\cos^2\theta}{4 } \right ) ^{d_{{\cal u } } } \biggr \ } \;,\end{aligned}\ ] ] where the range of integration for @xmath45 is from @xmath46 to 1 .
the total cross section can be obtained analytically in closed form by integrating eq.([dsigmadcos ] ) over @xmath45 , viz . , @xmath47 the effective interaction of spin-2 unparticle with the photon is given by @xcite @xmath48 where @xmath49 is an unknown effective coupling constant , of order @xmath25 .
using the feynman rules and the propagator derived in @xcite , the matrix element squared for elastic photon - photon scattering via spin-2 unparticle exchange is found to be @xmath50 + 2 \left ( t u \right)^{d_{{\cal u}}- 2 } s^4 \biggr\ } \ ; .\end{aligned}\ ] ] as @xmath51 , the above expression is proportional to @xmath52 which is the familiar result @xcite for the spin-2 kaluza - klein graviton exchange in the large extra dimensions model .
however , the @xmath3-channel unparticle exchange contains a cp - conserving phase factor @xmath2 that does not share with the @xmath4- and @xmath5-channels .
therefore , the expressions of eq.([msqspin0 ] ) and eq.([msqspin2 ] ) for the matrix element squared contain the factor @xmath53 in the interference terms between @xmath3- and @xmath4- and between @xmath3- and @xmath5-channels .
this is a unique feature of the unparticle .
normalized angular distributions of @xmath54 via spin-0 and spin-2 unparticle exchanges for various @xmath15 at @xmath55 tev .
, title="fig:",width=480 ] normalized angular distributions of @xmath54 via spin-0 and spin-2 unparticle exchanges for various @xmath15 at @xmath55 tev . ,
title="fig:",width=480 ] in fig.[norm - angular ] , we show the normalized angular distributions @xmath56 at @xmath55 tev for spin-0 and spin-2 unparticle exchanges .
note that in the part for spin-0 the scale on the @xmath57-axis is linear while that for spin-2 the scale is logarithmic . therefore , in general the spin-2 exchange will give much larger contributions in the forward region .
another interesting feature is that for spin-0 case when @xmath15 increases from @xmath58 to @xmath59 the distribution is becoming more forward .
this is because the factors of @xmath60 and @xmath61 only appear in the numerator , and so when @xmath15 increases , more powers of @xmath60 and @xmath61 are contributing in the forward region . on the other hand , for spin-2 case more powers of @xmath60 and @xmath61
appear in the denominator as @xmath15 is closer to @xmath62 .
thus , the distribution is much more forward for small @xmath15 .
in fact , it diverges at @xmath63 for @xmath64 .
we have also verified that the term containing the factor @xmath38 is affecting the distribution .
if there were no such a factor , the distribution would have been different , especially for small @xmath15 .
this demonstrates the effect of the peculiar phase associated with the @xmath3-channel propagator only .
if the phase were associated with all @xmath65 propagators , the effect would have been canceled out when we squared the amplitude .
total cross sections for @xmath66 via ( a ) spin-0 and ( b ) spin-2 unparticle exchange versus center - of - mass energy for various @xmath15 . the sm expectation is also shown .
, title="fig:",width=307 ] total cross sections for @xmath66 via ( a ) spin-0 and ( b ) spin-2 unparticle exchange versus center - of - mass energy for various @xmath15 .
the sm expectation is also shown .
, title="fig:",width=307 ] in fig .
[ cross ] , we plot the integrated cross sections versus the center - of - mass energy @xmath67 .
we also show the expectation from the sm , using the results of ref .
@xcite with the form factors from ref .
@xcite . since the sm cross section peaks in the forward and backward directions , we impose an angular cut of @xmath68 to reduce the sm cross section .
it is easy to see that the unparticle cross sections can surpass the sm one at high enough energy depending on the spin and scaling dimension of the unparticle .
the factor containing @xmath38 also affects the total cross sections to some extent , especially for small @xmath15 .
one of the most peculiar features of unparticle is the phase factor @xmath2 associated with the @xmath3-channel propagator .
we have studied its effect in @xmath54 , which would have been symmetric in @xmath3- , @xmath4- , and @xmath5-channels without the phase factor . however , since the phase is only associated with the @xmath3-channel , the effect will show up in the interference terms between @xmath3- and @xmath4- and between @xmath3- and @xmath5-channels .
the effect of such a factor affects the angular distribution and total cross sections to some significance , especially for small @xmath15 .
we have also shown that the scattering cross sections due to unparticle exchanges easily surpass the sm contribution .
thus , the possibility of studying photon scattering in the future linear collider , using either laser backscattering technique or bremsstrahlung , is important to test the existence of any tree - level photon - photon scattering .
unparticle is a unique example that allows tree - level exchange and contains a special cp - conserving phase factor @xmath2 solely in the @xmath3-channel propagator to facilitate interesting interference effects .
this research was supported in parts by the nsc under grant number nsc 96 - 2628-m-007 - 002-my3 and the ncts of taiwan ( hsinchu ) .
h. georgi , phys .
lett . * 98 * , 221601 ( 2007 ) [ hep - ph/0703260 ] .
h. georgi , phys .
b * 650 * , 275 ( 2007 ) , arxiv:0704.2457 [ hep - ph ] .
k. cheung , w. y. keung and t. c. yuan , phys .
lett . * 99 * , 051803 ( 2007 ) , arxiv:0704.2588 [ hep - ph ] ; phys .
d * 73 * , 075015 ( 2007 ) , arxiv:0706.3155 [ hep - ph ] .
k. cheung , phys .
d * 61 * , 015005 ( 2000 ) [ arxiv : hep - ph/9904266 ] ; h. davoudiasl , phys .
d * 60 * , 084022 ( 1999 ) [ arxiv : hep - ph/9904425 ] .
o. cakir and k. o. ozansoy , arxiv:0712.3814 [ hep - ph ] . t. kikuchi , n. okada and m. takeuchi , arxiv:0801.0018 [ hep - ph ] .
b. grinstein , k. intriligator and i. z. rothstein , arxiv:0801.1140 [ hep - ph ] .
v. costantinim b. de tollis and g. pistoni , nuovo cim .
* a2 * , 733 ( 1971 ) ; g. jikia and a. tkabladze , phys . lett .
* b323 * , 453 ( 1994 ) ; g. gounaris , p. porfyriadis and f. renard , eur .
j. * c9 * , 673 ( 1999 ) . | elastic photon - photon scattering can only occur via loop diagrams in the standard model and is naturally suppressed .
unparticle can induce tree - level photon - photon scattering through the operator @xmath0 for spin-0 unparticle or @xmath1 for spin-2 unparticle . due to the peculiar cp - conserving phase @xmath2 associated with the @xmath3-channel unparticle propagator ,
its interference effects with the @xmath4- and @xmath5-channels on the total cross section and the angular distribution are found to be some significance .
in addition , we show that the cross sections via unparticle exchange can be substantially larger than the standard model contribution . |
over the last decade there has been theoretical interest in curiosities dubbed warp - drives initiated by the 1994 paper by m. alcubierre [ 1 ] .
these warp drives are constructs that allow some object ( a spacecraft ) to travel at superluminal velocities by manipulating spacetime in a way such that the spacecraft never locally exceeds the speed of light , but in a manner identical to the inflationary stage of the universe the spacecraft does have a relative speed defined as change of proper spatial distance over proper spatial time faster than the speed of light .
interest in warp drives has not been solely confined to the realm of theoretical speculation as shown by the formation of the nasa breakthrough propulsion program [ 2 ] and the british aerospace project greenglow [ 3 ] both of whose purpose has been to investigate the realization of these ideas . in the spirit of the morris ,
thorne and yurtsever paper [ 4 ] these warp drives , thought highly speculative in nature , provide an unique and inspiring opportunity to ask the question what constraints do the laws of physics place on the abilities of an arbitrarily advanced civilization. in this paper a new and innovative mechanism to generate the necessary alcubierre warp bubble is proposed .
it has been suggested in the context of brane - world models that our universe is a ( 3 + 1 ) brane residing in some higher dimensional bulk [ 5 ] .
it is known phenomenologically that supersymmetry is broken on our 3-brane , however is has been suggested that it may not be broken on the bulk [ 6 ] .
existing in higher dimensional bulk.,width=312 ] unbroken susy decrees that the components of the chiral or gauge multiplets share equal masses in the bulk and have the same interaction strength , however on the 3-brane susy breaking induces a mass square difference between them .
motivated by string theory the 3-brane has an effective thickness @xmath0 characterized by the string thickness @xmath1 . as a result
the casimir energy is non - trivial in the extra dimensional volume that encompasses the brane .
this energy has the necessary features to account for the cosmological constant
. for simplicity assume an @xmath2 manifold with extra dimensional radius a. susy breaking around the brane alters the casimir energy which leads to a mass shift of the bulk fields .
it is the aim of this paper to demonstrate that the mass shift is directly related to the radius of the extra dimension and as such a local change in the radius of the extra dimension will have the effect of altering the mass shift and thus the casimir energy which locally effects the value of the cosmological constant in the region effectively creating a bubble of inflation / contraction . a spacecraft with the ability to create such a bubble will always move inside their own local light - cone however the ship can utilize the expansion of spacetime behind the ship to move away from some object at any desired speed or equivalently to contract the spacetime in front of the ship to approach any object .
the alcubierre top - hat metric is shown beneath to demonstrate the analogous effects on spacetime.,title="fig:",width=240 ] .
the alcubierre top - hat metric is shown beneath to demonstrate the analogous effects on spacetime.,title="fig:",width=240 ]
to build the model we consider for simplicity a scalar field and its susy partner a calar field .
the scalar part of the action will be @xmath3\ ] ] where @xmath4 characterizes the mass - square shift and the location of the three - brane is at y=0 .
the shift of the casimir energy density due to susy - breaking in the three brane extra volume is [ 7 ] @xmath5 where @xmath6 is some constant .
the renormalized casimir energy density is @xmath7 after integrating over the extra dimensional space the renormalized casimir energy density is @xmath8 where @xmath0 is the width of the brane and @xmath9 is the radius of the compactified extra dimension . the same calculations can be applied to the superpartner calar field .
the only change is a sign difference .
thus the total casimir energy density contribution from a scalar / calar field induced by susy breaking is @xmath10 it is important to note that although the casimir energy lies in the extra dimension it does contribute to the overall energy density of the unverse in @xmath11 and represents the @xmath12 term in einstein s equation @xmath13 the important term in equation ( 5 ) is @xmath14 which demonstrates that the casimir energy density and hence the cosmological constant term are immutably related to the radius a of the extra dimension .
consider some arbitrarily advanced civilization with the ability to locally alter the radius of the extra dimension .
this would bring about a local shift in the cosmological constant roughly of the form @xmath15\ ] ] where @xmath16 represents the modified radius .
thus , given some mechanism for locally adjusting the radius it is feasible that one could locally change the value of the cosmological constant thus providing a mechanism for pushing / pulling a spacecraft .
this is analogous to the alcubierre bubble .
\1 . m. alcubierre 1994 class . quantum grav . 11 , l73 \7 .
casimir effect in a supersymmetry - breaking brane world as dark energy . by pisin chen ( slac ) , je - an gu ( taiwan , natl .
taiwan u. and slac ) , .
slac - pub-10772 , sep 2004 .
e - print archive : astro - ph/0409238 | within the framework of brane - world models it is possible to account for the cosmological constant by assuming supersymmetry is broken on the 3-brane but preserved in the bulk .
an effective casimir energy is induced on the brane due to the boundary conditions imposed on the compactified extra dimensions .
it will be demonstrated that modification of these boundary conditions allows a spacecraft to travel at any desired speed due to a local adjustment of the cosmological constant which effectively contracts / expands space - time in the front / rear of the ship resulting in motion potentially faster than the speed of light as seen by observers outside the disturbance . |
an integer nowhere - zero @xmath0-flow on a graph @xmath1 is an assignment of a direction and a value of @xmath2 to each edge of @xmath1 such that the kirchhoff s law is satisfied at every vertex of @xmath1 .
this is the most restrictive definition of a nowhere - zero @xmath0-flow .
but it is equivalent to more flexible definitions , see e.g. @xcite .
a cubic graph @xmath1 is bipartite if and only if it has a nowhere - zero 3-flow , and @xmath3 if and only if @xmath1 has a nowhere - zero 4-flow .
seymour @xcite proved that every bridgeless graph has a nowhere - zero 6-flow .
so far this is the best approximation to tutte s famous 5-flow conjecture , which is equivalent to its restriction to cubic graphs .
[ @xcite ] [ 5fc ] every bridgeless graph has a nowhere - zero 5-flow .
kochol @xcite proved that a minimum counterexample to the 5-flow conjecture is a cyclically 6-edge - connected cubic graph .
hence , it suffices to prove conjecture [ 5fc ] for these graphs . a classical parameter to measure
how far a cubic graph is from being 3-edge - colorable is its _
oddness_. the oddness , denoted by @xmath4 , of a bridgeless cubic graph @xmath1 is the minimum number of odd circuits in a 2-factor of @xmath1 .
since @xmath1 has an even number of vertices , @xmath4 is necessarily even .
furthermore , @xmath5 if and only if @xmath1 is @xmath6-edge - colorable .
jaeger @xcite showed that cubic graphs with oddness at most 2 have a nowhere - zero 5-flow .
furthermore , a consequence of the main result in @xcite is that cyclically @xmath7-edge connected cubic graphs with oddness at most @xmath8 have a nowhere - zero 5-flow .
the following is our main theorem , and it is a strengthening for both previous results .
[ oddness4_5flow ] let @xmath1 be a cyclically 6-edge - connected cubic graph . if @xmath9 , then @xmath1 has a nowhere - zero 5-flow
in this section , we recall the concept of flow partitions , which was introduced by the second author in @xcite .
let @xmath1 be a graph and @xmath10 .
the set of edges with precisely one end in @xmath11 is denoted by @xmath12 . an _ orientation _
@xmath13 of @xmath1 is an assignment of a direction to each edge . for @xmath10 , @xmath14 ( @xmath15 )
is the set of edges of @xmath12 whose head ( tail ) is incident to a vertex of @xmath11 . the oriented graph is denoted by @xmath16 , @xmath17 and @xmath18 denote the _ indegree _ and _ outdegree _ of vertex @xmath19 in @xmath16 , respectively .
the degree of a vertex @xmath19 in the undirected graph @xmath1 is @xmath20 , and it is denoted by @xmath21 .
let @xmath0 be a positive integer , and @xmath22 a function from the edge set of the directed graph @xmath16 into the set @xmath23 .
for @xmath10 let @xmath24 .
the function @xmath22 is a @xmath0-flow on @xmath1 if @xmath25 for every @xmath10 .
the _ support _ of @xmath22 is the set @xmath26 , and it is denoted by @xmath27 .
a @xmath0-flow @xmath22 is a nowhere - zero @xmath0-flow if @xmath28
. we will use balanced valuations of graphs , which were introduced by bondy @xcite and jaeger @xcite .
a _ balanced valuation _ of a graph @xmath1 is a function @xmath29 from the vertex set @xmath30 into the real numbers , such that @xmath31 for all @xmath32 .
jaeger proved the following fundamental theorem .
[ @xcite ] [ thm_jaeger_75 ] let @xmath1 be a graph with orientation @xmath13 and @xmath33 .
then @xmath1 has a nowhere - zero @xmath0-flow if and only if there is a balanced valuation @xmath29 of @xmath1 with @xmath34 , for all @xmath35 in particular , theorem [ thm_jaeger_75 ] says that a cubic graph @xmath1 has a nowhere - zero @xmath36-flow if and only if there is a balanced valuation of @xmath1 with values in @xmath37 .
let @xmath1 be a bridgeless cubic graph , and @xmath38 be a @xmath39-factor of @xmath1 with odd circuits @xmath40 , and even circuits @xmath41 ( @xmath42 , @xmath43 ) , and let @xmath44 be the complementary @xmath45-factor .
a _ canonical _ @xmath8-edge - coloring , denoted by @xmath46 , of @xmath1 with respect to @xmath38 colors the edges of @xmath44 with color @xmath45 , the edges of the even circuits of @xmath38 with @xmath39 and @xmath6 , alternately , and the edges of the odd circuits of @xmath38 with colors @xmath39 and @xmath6 alternately , but one edge which is colored @xmath47 . then
, there are precisely @xmath48 vertices @xmath49 where color @xmath39 is missing ( that is , no edge which is incident to @xmath50 has color @xmath39 ) . the subgraph which is induced by the edges of colors @xmath45 and @xmath39 is union of even circuits and @xmath51 paths @xmath52 of odd length and with @xmath49 as ends . without loss of generality we can assume that @xmath52 has ends @xmath53 and @xmath54 , for @xmath55 .
let @xmath56 be the graph obtained from @xmath1 by adding two edges @xmath57 and @xmath58 between @xmath53 and @xmath54 for @xmath55 .
extend the previous edge - coloring to a proper edge - coloring of @xmath56 by coloring @xmath58 with color @xmath39 and @xmath57 with color @xmath8 .
let @xmath59 be the cycles of the @xmath39-factor of @xmath56 induced by the edges of colors @xmath45 and @xmath39 ( @xmath60 ) .
in particular , @xmath61 is the even circuit obtained by adding the edge @xmath58 to the path @xmath52 , for @xmath55 . finally , for @xmath62 let @xmath63 be the @xmath39-circuit induced by the edges @xmath57 and @xmath58 .
we construct a nowhere - zero @xmath8-flow on @xmath56 as follows : * for @xmath64 let @xmath65 be a nowhere - zero flow on the directed circuit @xmath66 with @xmath67 for all @xmath68 ; * for @xmath69 let @xmath70 be a nowhere - zero flow on the directed circuit @xmath61 with @xmath71 for all @xmath72 ; * for @xmath73 let @xmath74 be a nowhere - zero flow on the directed circuit @xmath63 ( choose @xmath75 such that @xmath58 receives the same direction as in @xmath76 ) with @xmath77 for all @xmath78 .
then , @xmath79 is the desired nowhere - zero @xmath8-flow on @xmath56 .
by theorem [ thm_jaeger_75 ] , there is a balanced valuation @xmath80 of @xmath56 .
it holds that @xmath81 , and hence , @xmath82 for all vertices @xmath19 .
the vertices of @xmath56 , and therefore , of @xmath1 as well , are partitioned into two classes @xmath83 and @xmath84 .
we call the elements of @xmath85 ( @xmath86 ) the white ( black ) vertices of @xmath1 , respectively .
let @xmath1 be a bridgeless cubic graph and @xmath38 a 2-factor of @xmath1 .
a partition of @xmath30 into two classes @xmath85 and @xmath86 constructed as above with a canonical @xmath8-edge - coloring @xmath46 , the @xmath8-flow @xmath87 on @xmath56 and the induced balanced valuation @xmath88 of @xmath56 is called a * flow partition * of @xmath1 w.r.t .
the partition is denoted by @xmath89 .
[ diffends_lemma ]
let @xmath1 be a bridgeless cubic graph and @xmath90 be a flow partition of @xmath30 which is induced by a canonical nowhere - zero 4-flow with respect to an edge - coloring @xmath46 .
let @xmath91 , @xmath92 be the two vertices of an edge @xmath93 . if @xmath94 , then @xmath91 and @xmath92 belong to different classes , i.e. @xmath95 if and only if @xmath96 . from a flow partition @xmath89
we easily obtain a flow partition @xmath97 such that the colors on the vertices of @xmath52 are switched .
let @xmath98 be the nowhere - zero @xmath8-flow on @xmath56 obtained by using the same @xmath39-factor @xmath38 , the same @xmath8-edge - coloring @xmath46 of @xmath1 and the same orientations for all circuits , but for one @xmath99 use opposite orientation of @xmath100 and @xmath101 with respect to the one selected in @xmath87 .
[ switching_lemma ] let @xmath1 be a bridgeless cubic graph and @xmath90 be the flow partition which is induced by the nowhere - zero @xmath8-flow @xmath87 . if @xmath102 is the flow partition induced by the nowhere - zero @xmath8-flow @xmath98 , then @xmath103 , @xmath104 , @xmath105 and @xmath106
suppose to the contrary that the statement is not true . then there is a cyclically 6-edge - connected cubic graph @xmath1 , which has no nowhere - zero 5-flow .
let @xmath38 be a 2-factor of @xmath1 with precisely four odd circuits @xmath107 .
let @xmath46 be a canonical 4-edge coloring of @xmath1 and @xmath108 be the four vertices where color @xmath39 is missing .
let @xmath109,@xmath110,@xmath111 .
note , that in any flow partition which depends on @xmath38 and @xmath46 , the vertices @xmath112 and @xmath110 ( and @xmath113 and @xmath114 as well ) belong to different color classes . by lemma [ switching_lemma ]
there are flow partitions @xmath90 and @xmath102 of @xmath1 such that @xmath115 , and @xmath116 .
hence , @xmath117 and @xmath118 .
we will prove that @xmath88 or @xmath123 is a balanced valuation of @xmath1 , and therefore , @xmath1 has a nowhere - zero @xmath36-flow by theorem [ thm_jaeger_75 ] .
hence , there is no counterexample and theorem [ oddness4_5flow ] is proved .
we will prove some properties of the edge - cuts @xmath12 and @xmath133 .
we deduce the results for @xmath11 only .
the results for @xmath134 follow analogously . if @xmath135 , then @xmath136
. therefore , @xmath11 , @xmath134 are a proper subset of @xmath137 . if @xmath138 , then @xmath139 .
since @xmath1 is cyclically @xmath140-edge - connected , it has no non - trivial @xmath6-edge - cut and no @xmath39-edge - cut .
hence , we assume that @xmath141 in the following . _ proof . _
if @xmath0 is even , then @xmath150 and @xmath151 have the same parity , and if @xmath0 is odd , then they have different parities .
since @xmath11 is the disjoint union of @xmath152 and @xmath153 it follows that @xmath0 and @xmath154 have the same parity .
since @xmath1 is cubic it follows that @xmath148 .
@xmath155 let @xmath156 ( @xmath157 ) be the number of white ( black ) vertices of @xmath11 where color @xmath39 is missing .
let @xmath158 .
since @xmath128 has two black and two white vertices , it follows that @xmath159 . if @xmath173 , then @xmath174 , and if @xmath175 , then @xmath176 . in both cases
, it follows that @xmath171 , a contradiction .
thus , @xmath177 , and therefore , @xmath178 . if @xmath179 , then @xmath180 , a contradiction .
hence , @xmath181 . @xmath155 _ proof . _ if @xmath188 , then @xmath189 .
hence , @xmath190 and @xmath191 .
the edge of @xmath192 is contained in a circuit of @xmath38 whose edges are not in @xmath162 .
hence , @xmath193 , a contradiction .
if @xmath194 , then @xmath195 . we deduce a contradiction as in the case before .
now suppose to the contrary that @xmath196 .
since @xmath197 , @xmath198 , and @xmath199 , it follows that @xmath200 .
therefore , @xmath201 . if @xmath202 , then @xmath203 and @xmath204 , a contradiction .
if @xmath205 , then @xmath206 and @xmath207 , a contradiction to claim [ c1 ] since @xmath208 .
if @xmath209 , then @xmath210 and @xmath211 , a contradiction .
hence , @xmath182 and @xmath212 and @xmath213 . that leaves the unique possibility @xmath183 and @xmath184 .
@xmath155 _ proof .
_ if @xmath214 $ ] is not connected , then there is @xmath216 such that @xmath217 has at least two components @xmath218 and @xmath219 .
since @xmath1 does not have a 2-edge - cut or a non - trivial 3-edge - cut , it follows that @xmath220 and one of @xmath218 , @xmath219 is a single vertex . hence , @xmath171 , a contradiction . @xmath155
a. @xmath223 and @xmath224 , b. @xmath221 partitions the vertices @xmath112 , @xmath110 , @xmath113 and @xmath114 into two sets @xmath225 , @xmath226 , which are in different components of @xmath217 and @xmath227 or @xmath228 .
note that @xmath227 if and only if @xmath229 .
further , only condition @xmath230 depends on the flow partition .
condition i ) depends on the canonical 4-edge - coloring of @xmath1 which is unchanged along the proof . from the previous results we deduce : bad 6-edges - cuts are the only obstacles in @xmath1 for having a nowhere - zero 5-flow . in order to deduce the desired contradiction
we will show that all 6-edge - cuts are not bad with respect to either @xmath90 or @xmath102 .
recall that , @xmath112 and @xmath113 receive the same color in @xmath90 , and that @xmath112 and @xmath114 receive the same color in @xmath102 .
for @xmath231 , let @xmath232 and @xmath233 be the corresponding set of edge - cuts . since @xmath112 and @xmath110 have different colors in both @xmath90 and @xmath102 , all edge - cuts in @xmath234 are not bad with respect to @xmath90 and with respect to @xmath102 . for @xmath235 , by claim [ c5 ]
there is a 6-edge - cut @xmath236 which is bad . by claim [ c4 ]
, @xmath237 consists of two components with vertex sets @xmath238 and @xmath239 , i.e. @xmath240 .
analogously , @xmath241 consists of two components with vertex sets @xmath242 and @xmath243 .
let @xmath244 , @xmath245 , @xmath246 and @xmath247 .
thus , @xmath248 for @xmath249 , see figure [ badcuts ] . _ proof .
_ if @xmath252 $ ] has a circuit , then @xmath253 since @xmath1 is cyclically @xmath140-edge - connected .
if this is not the case , then @xmath252 $ ] is a forest , say with @xmath254 vertices .
hence , @xmath255 .
since @xmath256 , it follows that @xmath257 .
two edges @xmath258 and @xmath259 which are incident to @xmath50 are colored with color 0 and 3 , respectively .
hence , @xmath260 , @xmath261 , and @xmath262 .
if @xmath263 , then @xmath264 , and @xmath252 $ ] is a path with two edges , one of color @xmath47 and one of color @xmath6 .
@xmath155 _ proof .
_ since @xmath270 and @xmath271 are bad , each of them has exactly two edges of color @xmath39 and four edges of color @xmath45 .
hence , each of them intersects with at most one circuit of @xmath38 . for each @xmath272 , @xmath273 , and
hence , there are @xmath274 such that @xmath275 and @xmath276 , @xmath277 contain an odd circuit of @xmath38 .
since @xmath1 is cyclically 6-edge - connected it follows that @xmath278 and @xmath279 .
let @xmath280 such that @xmath267 and @xmath268 .
for symmetry , it suffices to prove that @xmath281 .
suppose to the contrary that @xmath282 . by claim [ small_component ] ,
@xmath283 $ ] and @xmath284 $ ] are paths of length two with edges colored @xmath47 and @xmath6 .
further , @xmath285 consists of three edges of color @xmath45 and two edges of color @xmath39 , which belong to the odd circuit @xmath286 of @xmath38 .
analogously , the two edges of color 2 of @xmath287 belong to the odd circuit @xmath288 of @xmath38 .
hence , both pairs of edges of color @xmath39 in @xmath285 and @xmath287 belong to @xmath270 and they are distinct , a contradiction since @xmath270 has only two edges of color @xmath39 .
@xmath155 _ proof . _
recall that @xmath295 .
hence , @xmath296 .
due to claim [ two_large_components ] , we can assume that @xmath297 , @xmath298 , @xmath299 and @xmath300 . by adding up , we obtain @xmath301 , where each edge of @xmath270 and @xmath271 is counted exactly twice .
hence , @xmath302 .
if @xmath303 , then exactly one edge , say @xmath93 , belongs to @xmath304 . if @xmath305 , then @xmath287 and @xmath306 are distinct sets of cardinality at least @xmath140 .
hence , @xmath307 , a contradiction .
if @xmath308 , then @xmath309 or @xmath310 has cardinality at most 2 , say , without loss of generality , @xmath309 . for the same reason , @xmath311 or @xmath312
has cardinality at most 2 .
if @xmath313 , then @xmath314 , and if @xmath315 , then @xmath316 , a contradiction ( in both cases ) .
hence , @xmath317 , and therefore , @xmath291 for @xmath269 .
let @xmath322 $ ] .
the components of @xmath323 are even circuits and the two paths @xmath324 and @xmath325 , where @xmath324 has the end vertices @xmath112 , @xmath110 , and @xmath325 has the end vertices @xmath113 , @xmath114 . the paths @xmath324 and @xmath325 intersect both @xmath326 and @xmath327 an odd number of times , since both , @xmath270 and @xmath271 , separate their ends . for symmetry , we can assume that @xmath328 and @xmath329 are even , and hence , @xmath330 and @xmath331 are odd .
furthermore , we assume that @xmath332 and @xmath333 are even , and hence , @xmath334 and @xmath335 are odd . note , that every other possible choice produces an analogous configuration .
the @xmath140-edge - cut @xmath336 contains an odd number of edges of @xmath337 . since @xmath338
, it follows that an odd number of edges of @xmath336 are not in @xmath337 , a contradiction , since all other components of @xmath323 are circuits , and they intersect every edge - cut an even number of times . | tutte s 5-flow conjecture from 1954 states that every bridgeless graph has a nowhere - zero 5-flow . in 2004
, kochol proved that the conjecture is equivalent to its restriction on cyclically 6-edge connected cubic graphs .
we prove that every cyclically 6-edge - connected cubic graph with oddness at most 4 has a nowhere - zero 5-flow . |
the study of random graphs since the seminal paper of erds and rnyi @xcite has by and large been restricted to analysing random subgraphs of the complete graph .
this is not of course completely true .
there has been a lot of research on random subgraphs of the hypercube and grids ( percolation ) .
there has been less research on random subgraphs of arbitrary graphs @xmath13 , perhaps with some simple properties . in this vain
, the recent result of krivelevich , lee and sudakov @xcite brings a refreshing new dimension .
they start with an arbitrary graph @xmath13 which they assume has minimum degree at least @xmath1 .
for @xmath14 we let @xmath15 be the random subgraph of @xmath13 obtained by independently keeping each edge of @xmath13 with probability @xmath16 .
their main result is that if @xmath17 then @xmath15 has a cycle of length @xmath18 with probability @xmath19 . here
@xmath20 is a function of @xmath1 that tends to zero as @xmath21 .
riordan @xcite gave a much simpler proof of this result .
krivelevich and samotij @xcite proved the existence of long cycles for the case where @xmath22 and @xmath13 is @xmath23-free for some fixed set of graphs @xmath23 .
frieze and krivelevich @xcite showed that @xmath15 is non - planar with probability @xmath19 when @xmath22 and @xmath13 has minimum degree at least @xmath1 .
in related works , krivelevich , lee and sudakov @xcite considered a random subgraph of a `` dirac graph '' i.e. a graph with @xmath24 vertices and minimum degree at least @xmath25 .
they showed that if @xmath26 for suffficently large @xmath24 then @xmath15 is hamiltonian with probability @xmath27 .
the results cited above can be considered to be generalisations of classical results on the random graph @xmath28 , which in the above notation would be @xmath29 . in this paper
we will consider generalising another model of a random graph that we will call @xmath30 .
this has vertex set @xmath31={\left\{1,2,\ldots , n\right\}}$ ] and each @xmath3 independently chooses @xmath1 random vertices as neighbors .
thus this graph has @xmath4 edges and average degree @xmath32 .
this model in a bipartite form where the two parts of the partition restricted their choices to the opposing half was first considered by walkup @xcite in the context of perfect matchings .
he showed that @xmath33 was sufficient for bipartite @xmath34 to contain a perfect matching .
matchings in @xmath30 were considered by shamir and upfal @xcite who showed that @xmath35 has a perfect matching w.h.p .
, i.e. with probability @xmath36 as @xmath37 .
later , frieze @xcite showed that @xmath38 has a perfect matching w.h.p .
fenner and frieze @xcite had earlier shown that @xmath30 is @xmath1-connected w.h.p . for @xmath33 .
after several weaker results , bohman and frieze @xcite proved that @xmath39 is hamiltonian w.h.p .
to generalise these results and replace @xmath40 by an arbitrary graph @xmath13 we will define @xmath41 as follows : we have a fixed graph @xmath0 and each @xmath3 independently chooses @xmath1 random neighbors , from its neighbors in @xmath13 .
it will be convenient sometimes to assume that each @xmath42 makes its choices with replacement and sometimes not .
it has no real bearing on the results obtained and we will indicate our choice here . to avoid cumbersome notation , we will from now on assume that @xmath13 has @xmath24 vertices and we will refer to @xmath41 as @xmath2 .
we implicitly consider @xmath13 to be one of a sequence of larger and larger graphs with @xmath37 .
we will say that events occur w.h.p . if their probability of non - occurrence can be bounded by a function that tends to zero as @xmath43 . for a vertex @xmath3
we let @xmath44 denotes its degree in @xmath13
. then we let @xmath45 .
we will first consider what we call strong dirac graphs ( sdg ) viz graphs with @xmath46 where @xmath47 is an arbitrary positive constant .
[ th1 ] suppose that @xmath13 is an sdg
. then w.h.p .
@xmath2 is @xmath1-connected for @xmath48 .
[ th2 ] suppose that @xmath13 is an sdg .
then there exists a constant @xmath49 such that if @xmath9 then @xmath2 is hamiltonian .
note that we need @xmath50 in order to prove these results .
consider for example the case where @xmath13 consists of two copies of @xmath51 plus a perfect matching @xmath52 between the copies . in this case
there is a probability greater than or equal to @xmath53 that no edge of @xmath52 will occur in @xmath2 .
we will next turn to graphs with large minimum degree .
[ th3 ] suppose that @xmath13 has minimum degree @xmath54 where @xmath55 with @xmath24 .
for every @xmath50 there exists a constant @xmath49 such that if @xmath9 then w.h.p .
@xmath2 contains a path of length @xmath8 . using this theorem as a basis
, we strengthen it and prove the existence of long cycles .
[ th4 ] suppose that @xmath13 has minimum degree @xmath54 where @xmath55 with @xmath24 .
for every @xmath50 there exists a constant @xmath49 such that if @xmath56 then w.h.p .
@xmath2 contains a cycle of length @xmath57 .
we finally note that in a recent paper , frieze , goyal , rademacher and vempala @xcite have shown that @xmath2 is useful in the construction of sparse subgraphs with expansion properties that mirror those of the host graph @xmath13 .
in this section we will assume that each vertex makes its choices with replacement .
let @xmath58 be an sdg .
let @xmath59 .
we need the following lemma .
[ lem1 ] let @xmath13 be an sdg and let @xmath60 .
then w.h.p .
there exists a set @xmath61 , where @xmath62 , such that each pair of vertices @xmath63 have at least @xmath64 common neighbors in @xmath65 .
define @xmath66 by including each @xmath67 in @xmath68 with probability @xmath69 .
since @xmath70 , each pair of vertices in @xmath13 has at least @xmath71 common neighbors in @xmath13 .
hence , the number of common neighbors in @xmath68 for any pair of vertices in @xmath72 is bounded from below by a @xmath73 random variable . @xmath74 since the expected size of @xmath68 is @xmath75 , there exists a set @xmath65 , @xmath62 , with the desired property .
let @xmath65 be a set as provided by the previous lemma , and let @xmath76 denote the subgraph of @xmath2 induced by @xmath77 .
[ lem2 ] let @xmath78 .
then w.h.p .
all components of @xmath76 are of size at least @xmath79 .
furthermore , removing any set of @xmath80 vertices from @xmath76 produces a graph consisting entirely of components of size at least @xmath79 , and isolated vertices .
[ lem : giant ] we first show that w.h.p .
@xmath76 contains no isolated vertex .
the probability of @xmath76 containing an isolated vertex is bounded by @xmath81^k = o(1),\ ] ] where @xmath65 and @xmath82 are as in lemma [ lem1 ] .
we now consider the existence of small non - trivial components @xmath83 after the removal of at most @xmath80 vertices @xmath84 .
then , @xmath85^{lk } \\ & \leq \sum_{l=2}^{cn } \binom{n}{l } \binom{n - l}{k-1 } \left[\frac{l + c\log n}{\frac{1}{2}n}\right]^{lk } \\ & \leq \sum_{l=2}^{cn } \left(\frac{ne}{l}\right)^ln^{k-1}\left[\frac{l + c\log n}{\frac{1}{2}n}\right]^{lk } \\ & = 2^ke\sum_{l=2}^{cn } \left[\frac{2^k e(l+c\log n)^k}{n^{k-1}l}\right]^{l-1 } \frac{(l+c\log n)^k}{l}\\ & \leq 2^ke{\left ( \sum_{l=2}^{\log^2n}{\left[\frac{\log^{3k}n}{n^{k-1}}\right]}^{l-1 } + \sum_{l=\log^2n}^{cn}(4ec^{k-1})^{l-1}n^k \right)}\\ & = o(1).\end{aligned}\ ] ] this proves that w.h.p .
@xmath76 consists of @xmath86 components @xmath87 and that removing any @xmath80 vertices will only leave isolated vertices and components of size at least @xmath79 .
, for any @xmath88 , there exist @xmath1 node - disjoint paths ( of length 2 ) from @xmath89 to @xmath90 in @xmath2 .
[ lem : connect ] let @xmath91 be the number of vertices in @xmath65 which pick at least one neighbor in @xmath92 and at least one in @xmath93 . furthermore ,
let @xmath94 be the indicator variable for @xmath95 picking @xmath96 and @xmath97 as its neighbors .
note that these variables are independent of @xmath76 .
let @xmath78 as in lemma [ lem2 ] and let @xmath60 as in lemma [ lem1 ] .
for @xmath95 we let @xmath98 these are independent random variables with values in @xmath99 .
let @xmath100 .
then , @xmath101 } & = & \sum_{u\in j_1 } \sum_{v\in j_2 } \sum_{\substack{w\in l\\w\in n(j_1)\cap n(j_2 ) } } { \textbf{e}\left[x_{uvw}\right ] } \\ & = & \sum_{u\in j_1 } \sum_{v\in j_2 } \sum_{\substack{w\in l\\w\in n(j_1)\cap n(j_2 ) } } { \left ( 1-{\left ( 1-\frac{1}{d_g(u ) } \right)}^k \right)}{\left ( 1-{\left ( 1-\frac{1}{d_g(v ) } \right)}^k \right)}\\ & \geq & \sum_{u\in j_1 } \sum_{v\in j_2 } \sum_{\substack{w\in
l\\w\in n(j_1)\cap n(j_2 ) } } \frac{1}{n^2}\\ & \geq & \frac{(cn)^2 12\log n}{n^2 } \\ & = & 12c^2 \log n.\end{aligned}\ ] ] applying hoeffdings inequality we get @xmath102}}{2}\right\ } } \leq \exp \left\{-\frac{({\textbf{e}\left[x\right]})^2}{2k^2|l|}\right\ } = o(1).\ ] ] now for @xmath103 let @xmath104 be the event that @xmath105 make a common choice . then @xmath106}=o(1).\ ] ] equations and together show that w.h.p .
, there are @xmath1 node - disjoint paths from @xmath92 to @xmath93 .
since the number of giant components is bounded by a constant , this is true for all pairs @xmath107 w.h.p .
we can complete the proof of theorem [ th1 ] .
suppose we remove @xmath108 vertices from @xmath65 and @xmath109 vertices from the remainder of @xmath13 .
we know from lemma [ lem1 ] that @xmath110 induces components @xmath111 of size at least @xmath79 .
there can not be any isolated verticesin @xmath110 as @xmath2 has minimum degree at least @xmath1 .
lemma [ lem2 ] implies that @xmath112 and that every vertex in @xmath65 is adjacent to @xmath113 .
in this section we will assume that each vertex makes its choices without replacement .
let @xmath13 be a graph with @xmath70 , and let @xmath1 be a positive integer .
let @xmath114 be the @xmath115 directed graphs obtained by letting each vertex @xmath116 of @xmath13 choose @xmath1 @xmath13-neighbors @xmath117 , and including in @xmath118 the @xmath1 arcs @xmath119 .
define @xmath120 and for @xmath121 let @xmath122 . for a digraph @xmath123 we let @xmath124 denote the graph obtained from @xmath123 by ignoring orientation and coalescing multiple edges , if necessary .
we let @xmath125 for @xmath126 .
let @xmath127 be the set of @xmath1-out graphs on @xmath13 .
below , when we say that @xmath118 is hamiltonian we actually mean that @xmath128 is hamiltonian .
( it will occasionally enable more succint statements ) . for each @xmath118 ,
let @xmath129 be the @xmath130 different edge - colorings of @xmath118 in which each vertex has @xmath80 outgoing green edges and one outgoing blue edge .
define @xmath131 to be the colored ( multi)graph obtained by ignoring the orientation of edges in @xmath132 .
let @xmath133 be the subgraph induced by green edges .
@xmath134 refers to @xmath135 when @xmath136 is chosen uniformly from @xmath137 $ ] , as it will be for @xmath2 .
let @xmath138 .
there exists an @xmath139 such that the following holds w.h.p . : for any set @xmath121 of size @xmath140 , @xmath141 .
[ lem : expand ] the claim fails if there exists an @xmath83 with @xmath140 such that there exists a @xmath142 , @xmath143 such that @xmath144 .
the probability of this is bounded from above by @xmath145 \\ &
\leq & \sum_{l=1}^{\alpha n } \left(\frac{ne}{l}\right)^l \left(\frac{ne}{3l-1}\right)^{3l-1 } \left[\frac{4le}{n/2}\right]^{kl } \\ & \leq & \sum_{l=1}^{\alpha n}\left[e^4 ( 8e)^k \left(\frac{l}{n}\right)^{k-4}\right]^l\\ & = & o(1),\end{aligned}\ ] ] for @xmath146 .
we say that a digraph @xmath118 _ expands _ if @xmath147 whenever @xmath140 , @xmath146 .
since almost all @xmath118 expand , we need only prove that an expanding @xmath118 almost always gives rise to a hamiltonian @xmath128 .
write @xmath148 for the set of expanding digraphs in @xmath149 and let @xmath150 .
let @xmath151 be any graph , and suppose @xmath152 is a longest path in @xmath151 . if @xmath153 and @xmath154 , then @xmath155 is also a longest path of @xmath151 . repeating this rotation for @xmath156 and all paths created in the process , keeping the endpoint @xmath157 fixed ,
we obtain a set @xmath158 of other endpoints . for @xmath159
we let @xmath160 .
[ psa ] for any endpoint @xmath116 of any longest path in any graph @xmath151 , @xmath161 .
[ lem : posa ] we say that a graph expands if @xmath162 whenever @xmath140 , assuming @xmath163 .
consider a green subgraph @xmath133 . w.h.p .
, there exists an @xmath139 such that for every longest path @xmath156 in @xmath133 and endpoint @xmath116 of @xmath156 , @xmath164 .
[ lem : greenexpand ] let @xmath165 .
we argue that if @xmath118 expands then so does @xmath151 .
for @xmath166 let @xmath167 . if @xmath168 , then @xmath169 , since each vertex of @xmath83 picks at most one blue edge outside of @xmath83 . thus @xmath151 expands .
in particular , lemma [ lem : expand ] implies that if @xmath170 , then @xmath141 and hence @xmath171 . by lemma [ lem : posa ]
, this implies that @xmath164 for any longest path @xmath156 and endpoint @xmath116 .
let @xmath172 be the number of expanding digraphs @xmath118 among @xmath173 for which @xmath128 is not hamiltonian .
we aim to show that @xmath174 as @xmath24 tends to infinity .
suppose @xmath175 are the expanding digraphs which are not hamiltonian .
define @xmath176 to be @xmath177 if @xmath133 contains a longest path of @xmath131 , @xmath178 , and @xmath179 otherwise . for @xmath180
, we have @xmath181 .
fix @xmath180 and a longest path @xmath156 of @xmath128 .
uniformly picking one of @xmath182 , we have @xmath183 the lemma follows from the fact that there are @xmath184 colorings of @xmath118 .
let @xmath185 be expanding and non - hamiltonian and for the purposes of exposition consider its edges to be colored green .
let @xmath186 be the random digraph obtained by letting each vertex of @xmath187 choose another edge , which will be colored blue .
let @xmath188 be the event that @xmath123 contains a path longer than the longest path of @xmath187 or if @xmath123 is hamiltonian , implying that @xmath189 .
since @xmath124 is connected w.h.p . ,
this event is the union of the events @xmath190 and @xmath191 let @xmath192 be the complement of @xmath188 and for hamiltionian @xmath187 let @xmath193 .
let @xmath194 be the number of @xmath195 such that @xmath196 .
we have @xmath197 the number of non - hamiltonian graphs is bounded by @xmath198 fix a @xmath185 .
let @xmath199 be the set of vertices which are endpoints of a longest path of @xmath187 .
for @xmath200 , say @xmath116 is of type i if @xmath116 has at least @xmath201 neighbors outside @xmath156 , and type ii otherwise .
let @xmath202 be the set of type i endpoints , and @xmath203 the set of type ii endpoints .
partition the set of expanding green graphs by @xmath204 where @xmath205 is the set of hamiltonian graphs , @xmath206 the set of non - hamiltonian graphs with @xmath207 and @xmath208 the set of non - hamiltonian graphs with @xmath209 . for @xmath210 , @xmath211 .
[ lem : n1bound ] let each @xmath212 choose a neighbor @xmath213 .
the event @xmath192 is included in the event @xmath214 .
we have @xmath215 where @xmath216 denotes the number of neighbors of @xmath116 inside @xmath156 . for @xmath217 , @xmath218 .
[ lem : n2bound ] let @xmath219 be a set of @xmath220 type ii endpoints . for each @xmath221 ,
let @xmath222 be the set of type ii vertices @xmath223 such that a path from @xmath116 to @xmath224 in @xmath187 can be obtained from @xmath156 by a sequence of rotations with @xmath116 fixed . by lemma [ lem : greenexpand ] we have @xmath225 for each @xmath116 , since @xmath226 .
let @xmath227 be a path with endpoints @xmath228 obtained from @xmath156 by rotations with @xmath116 fixed , and label the vertices on @xmath227 by @xmath229 .
suppose @xmath224 chooses some @xmath230 on the path with its blue edge .
if @xmath231 , let @xmath232 .
write @xmath233 for @xmath234 .
if @xmath235 , or if @xmath224 chooses a vertex outside @xmath156 , let @xmath236 . chooses @xmath230 .
the vertex @xmath234 is included in @xmath237 if and only if @xmath238 .
] there will be at least @xmath239 choices for @xmath136 for which @xmath240 .
let @xmath241 be the number of @xmath242 such that @xmath243 is nonempty .
this variable is bounded stochastically from below by a binomial @xmath244 variable , and by a chernoff bound we have that @xmath245 define @xmath246 .
conditional on @xmath247 for all @xmath248 , let @xmath249 be @xmath250 vertices whose choice produces a nonempty @xmath243 .
let @xmath251 , and for @xmath252 define @xmath253 to be @xmath177 if @xmath254 is distinct from @xmath255 and @xmath179 otherwise .
we have @xmath256 , and each @xmath253 is bounded from below by a bernoulli variable with parameter @xmath257 . to see this , note that @xmath258 has at least @xmath259 choices resulting in a nonempty @xmath260 since @xmath116 and @xmath258 are of type ii ,
so @xmath261 since @xmath262 , @xmath263 is bounded stochastically from below by a binomial @xmath264 variable , and so @xmath265 each @xmath116 for which @xmath266 will choose a vertex in @xmath237 with probability @xmath267 hence we have @xmath268 we can now complete the proof of theorem [ th2 ] . from lemmas [
lem : n1bound ] and [ lem : n2bound ] we have @xmath269 going back to with @xmath270 we have @xmath271^n \\ & \leq & \exp\left\{-\varepsilon\left(\frac{\alpha^2}{129 } - \frac{2}{c}\right)n\right\}\\ & = & o(1),\end{aligned}\ ] ] for @xmath272 .
let @xmath273 denote the directed graph with out - degree @xmath1 defined the vertex choices . consider a depth first search ( dfs ) of @xmath273 where we construct @xmath273 as we go . at all times
we keep a stack @xmath274 of vertices which have been visited , but for which we have chosen fewer than @xmath1 out - edges .
@xmath142 denotes the set of vertices that have not been visited by dfs .
each step of the algorithm begins with the top vertex @xmath275 of @xmath274 choosing one new out - edge .
if the other end of the edge @xmath42 lies in @xmath142 ( we call this a _ hit _ ) , we move @xmath275 from @xmath142 to the top of @xmath274 .
when dfs returns to @xmath276 and at this time @xmath42 has chosen all of its @xmath1 out - edges , we move @xmath42 from @xmath274 to @xmath83 . in this way we partition @xmath277 into * vertices that have chosen all @xmath1 of its out - edges . *
vertices that have been visited but have chosen fewer than @xmath1 edges . *
unvisited vertices .
key facts : let @xmath278 denote the number of hits at any time and let @xmath279 denote the number of times we have re - started the search i.e. selected a vertex in @xmath142 after the stack @xmath83 empties . 1 . @xmath280 increases by @xmath177 for each hit , so @xmath281 .
2 . more specifically , @xmath282 .
3 . @xmath283 contains a path which contains all of @xmath274 at all times .
the goal will be to prove that @xmath284 at some point of the search , where @xmath47 is some arbitrarily small positive constant .
after @xmath285 steps , i.e. after @xmath285 edges have been chosen in total , the number of hits @xmath286 w.h.p .
[ lem : hits ] since @xmath287 , each tree component of @xmath2 has at least @xmath1 vertices , and at least @xmath288 edges must be chosen in order to complete the search of the component . hence , after @xmath285 edges have been chosen , at most @xmath289 tree components have been found .
this means that if @xmath290 after @xmath291 edges have been sent out , then * p2 * implies that @xmath292 .
so if @xmath293 each edge chosen by the top vertex @xmath275 has probability at least @xmath294 of making a hit . hence , @xmath295 for @xmath296 , by the chernoff bounds
. we can now complete the proof of theorem [ th3 ] . by lemma [ lem : hits ] , after @xmath291 edges have been chosen we have @xmath297 w.h.p . for a vertex to be included in @xmath83
, it must have chosen all of its edges .
hence , @xmath298 , and we have @xmath284 . finally observe that @xmath274 is the set of vertices of a path of @xmath2 .
suppose now that we consider @xmath299 where each vertex makes @xmath1 choices each of the colors light red " , dark red " , light blue " and dark blue " .
@xmath300 respectively are the graphs induced by the differently colored choices .
we have by theorem [ th3 ] that w.h.p .
there is a path @xmath156 of length at least @xmath8 in the light red graph @xmath301 .
at this point we start using a modification of dfs ( denoted by @xmath302 ) and the differently colored choices to create a cycle .
we divide the steps into epochs @xmath303 , indexed by binary strings .
we stop the search immediately if there is a high chance of finding a cycle of length at least @xmath304 .
if executed , epoch @xmath305 will extend the exploration tree by at least @xmath306 vertices , unless an unlikely failure occurs .
theorem [ th3 ] provides @xmath307 . in the remainder
, we will assume @xmath308 .
epoch @xmath309 will use light red colors if @xmath136 has odd length and ends in a @xmath179 , dark red if @xmath136 has even length and ends in a @xmath179 , light blue if @xmath136 has odd length and ends in a @xmath177 , and dark blue if @xmath136 has even length and ends in a @xmath177 .
epochs @xmath310 and @xmath311 ( where @xmath312 denotes the string obtained by appending @xmath313 to the end of @xmath314 ) both start where @xmath309 ends , and this coloring ensures that every vertex discovered in an epoch will initially have no adjacent edges in the color of the epoch . during epoch @xmath309
we maintain a stack of vertices @xmath315 .
when discovered , a vertex is placed in one of the three sets @xmath316 , and simultaneously placed in @xmath315 if it is placed in @xmath317 .
once placed , the vertex remains in its designated set even if it is removed from @xmath315 .
let @xmath318 be the length of the unique path in the exploration tree @xmath142 from @xmath42 to @xmath319 .
we designate the set for @xmath42 as follows . * @xmath42 has less than @xmath320 @xmath13-neighbors in @xmath142 .
* @xmath42 has at least @xmath320 @xmath13-neighbors in @xmath142 , but less than @xmath321 @xmath13-neighbors @xmath319 such that @xmath322 .
* @xmath42 has at least @xmath320 @xmath13-neighbors in @xmath142 , and at least @xmath321 @xmath13-neighbors @xmath319 such that @xmath322 . at the initiation of epoch @xmath309
, a previous epoch will provide a set @xmath323 of @xmath324 vertices , as described below . starting with @xmath325 , each vertex of @xmath323
is placed in @xmath326 or @xmath327 according to the rules above .
let @xmath328 , ordered with the latest discovered vertex on top .
an epoch @xmath309 consists of up to @xmath333 steps , and each step begins with a @xmath334 at the top of the stack @xmath315 .
this vertex is called _
active_. if @xmath42 has chosen @xmath1 neighbors , remove @xmath42 from the stack and perform the next step .
otherwise , let @xmath42 randomly pick one neighbor @xmath319 from @xmath335 .
if @xmath336 , then @xmath319 is assigned to @xmath326 or @xmath327 as described above .
if @xmath337 , perform the next step with @xmath319 at the top of @xmath315 . if @xmath338 , perform the next step without placing @xmath319 in @xmath315 . note that unless @xmath329 or @xmath330 , we initially have @xmath339 , guaranteeing that @xmath333 steps may be executed .
epoch @xmath309 _ succeeds _ and is ended ( possibly after fewer than @xmath333 steps ) if at some point we have @xmath340 .
if all @xmath333 steps are executed and @xmath341 , the epoch fails .
an epoch fails if less than @xmath343 steps result in the active vertex choosing a neighbor outside @xmath142 . since the active vertex is always in @xmath317
, we have @xmath344 for @xmath345 , by hoeffding s inequality .
this proves the lemma .
the epoch produces a tree which is a subtree of @xmath142 .
let @xmath346 be the longest path of vertices in @xmath317 , and let @xmath347 be the set of vertices discovered during @xmath309 which are not in @xmath346 .
if the epoch succeeds , @xmath346 has length at least @xmath348 , and at most @xmath349 vertices discovered during @xmath309 are not on the path .
indeed , a vertex is outside @xmath346 if and only if it is in @xmath317 and has chosen all its @xmath1 neighbors , or if it is in @xmath350 .
thus , the number of vertices not on the path is bounded by @xmath351 if the epoch fails , the path @xmath346 may be shorter , but @xmath352 is still bounded by @xmath324 .
if @xmath309 succeeds , the epochs @xmath310 and @xmath311 will be initiated at the end of @xmath309 , by letting @xmath353 and @xmath354 be the last @xmath324 vertices discovered during @xmath309 .
if @xmath309 fails , @xmath310 and @xmath311 will not be initiated .
the exploration tree @xmath142 will resemble an unbalanced binary tree , in which each successful epoch gives rise to up to two new epochs .
epochs are ordered after their binary value , so that @xmath355 is initiated before @xmath356 if and only if @xmath357 , ordered according to the numerical value of the binary strings .
suppose that no epoch ends with @xmath329 or @xmath330 . under this assumption
, we may model the exploration as a galton - watson branching process , in which a successful @xmath309 gives rise to at least @xmath358 successful epochs , where @xmath359 with probability @xmath360 , @xmath361 with probability @xmath362 and @xmath363 with probability @xmath364 . the offspring distribution for this lower bound
has generating function @xmath365 let @xmath366 be the smallest fixed point @xmath367 .
we have @xmath368 as @xmath369 . hence , the probability that the branching process never expires is at least @xmath370 , which tends to @xmath177 .
we may now finish the proof of the theorem .
condition first on @xmath331 being stopped by an epoch @xmath309 having @xmath330 . in this case , let each @xmath371 choose @xmath1 neighbors using eges with the epoch s color .
each choice has probability at least @xmath47 of finding a cycle of length at least @xmath304 , by choosing a neighbor @xmath319 such that @xmath322 .
the probability of not finding a cycle of length at least @xmath304 is bounded by @xmath372 now condition on @xmath331 being stopped by an epoch @xmath309 having @xmath373 .
note that we must have @xmath374 for some @xmath375 .
indeed , if @xmath376 , then any @xmath42 discovered in @xmath314 must have at least @xmath377 @xmath13-neighbors at distance at least @xmath304 , at its time of discovery .
if not , and @xmath378 then it has at most @xmath379 @xmath13-neighbors outside @xmath142 , at most @xmath380 @xmath13-neighbors in @xmath381 .
there are at most @xmath382 @xmath13-neighbors in @xmath383 at distance less than @xmath382 and so there are at least @xmath377 @xmath13-neighbors in @xmath142 at distance at least @xmath382 from @xmath42 , which implies that @xmath384 , contradiction . note also that @xmath385 for any @xmath386 .
this can be seen as follows : for any @xmath387 let @xmath388 be the vertex which minimizes @xmath389 .
note that we may have @xmath390 .
there are at most @xmath391 @xmath13-neighbors of @xmath42 on the path @xmath392 from @xmath42 to @xmath393 . then note that there are at most @xmath394 @xmath13-neighbors of @xmath42 on @xmath395 that are within @xmath304 of @xmath42 .
so the maximum number of @xmath396 such that @xmath397 is bounded by @xmath398 . define an ordering on @xmath142 by saying that @xmath403 if @xmath404 was discovered before @xmath405 during @xmath331 , or if @xmath406 . if @xmath407 , and @xmath408 for all @xmath409 , write @xmath410 .
similarly define @xmath411 and @xmath412 .
let each @xmath413 choose @xmath1 neighbors in the color of epoch @xmath309 .
we say that @xmath42 is _ good _ if it chooses @xmath414 and @xmath415 such that @xmath416 where @xmath417 .
for each @xmath387 define @xmath418 , @xmath419 and @xmath420 .
since @xmath421 we have @xmath422 since the @xmath423 vertices of @xmath424 are on a path , we must have @xmath425 , otherwise @xmath42 has @xmath426 neighbors at distance at least @xmath427 , contradicting @xmath428 .
this implies @xmath429 .
similarly , @xmath430 . fix a vertex @xmath387 and define @xmath431 and @xmath432 , @xmath433 as follows .
@xmath434 is the set of the first @xmath435 vertices of @xmath436 discovered during @xmath331 .
@xmath437 is the set of the last @xmath435 vertices of @xmath436 discovered before any vertex of @xmath323 .
lastly , @xmath438 consists of the @xmath435 last vertices discovered in @xmath439 . since @xmath440 and @xmath441 , the sets @xmath442 exist and are disjoint .
since @xmath385 , the probability that @xmath42 chooses @xmath443 and @xmath444 is at least @xmath445 .
if this happens , we have @xmath446 in other words , @xmath413 is good with probability at least @xmath445 .
since @xmath400 , w.h.p .
there exist two good vertices @xmath447 .
since @xmath448 , the shortest path from @xmath449 to @xmath42 does not contain @xmath275 , and the shortest path from @xmath450 to @xmath275 does not contain @xmath42 .
also , by choice of @xmath451 we have @xmath452 .
suppose @xmath275 and @xmath42 pick @xmath453 and @xmath454 , and w.l.o.g .
suppose @xmath455 .
the cycle @xmath456 has length @xmath457 99 t. bohman and a.m. frieze , hamilton cycles in 3-out _ random structures and algorithms _ 35 ( 2010 ) 393 - 417 .
p. erds and a. rnyi , on the evolution of random graphs , _ publ .
_ 5 ( 1960 ) 17 - 61 .
a.m. frieze , n. goyal , l. rademacher and s. vempala , expanders via random spanning trees .
a.m. frieze , maximum matchings in a class of random graphs , _ journal of combinatorial theory b _ 40 ( 1986 ) 196 - 212 .
t. fenner and a.m. frieze , on the connectivity of random m - orientable graphs and digraphs , _ combinatorica _ 2 ( 1982 ) 347 - 359 . a. frieze and m. krivelevich , on the non - planarity of a random subgraph .
combinatorics , probability and computing _ 22 ( 2013 ) 722 - 732 .
m. krivelevich , c. lee and b. sudakov , long paths and cycles in random subgraphs of graphs with large minimum degree .
m. krivelevich , c. lee and b. sudakov , robust hamiltonicity of dirac graphs , _ transactions of the american mathematical society _ 366 ( 2014 ) 3095 - 3130 .
m. krivelevich and w. samotij , optimal packings of hamilton cycles in sparse random graphs , _ siam journal on discrete mathematics _ 26 ( 2012 ) 964 - 982 .
o. riordan , long cycles in random subgraphs of graphs with large minimum degree .
e. shamir and e. upfal , one - factor in random graphs based on vertex choice , _ discrete mathematics _ 41 ( 1982 ) 281 - 286 . d.w .
walkup , matchings in random regular bipartite graphs , _ discrete mathematics _ 31 ( 1980 ) 59 - 64 . | we consider random subgraphs of a fixed graph @xmath0 with large minimum degree .
we fix a positive integer @xmath1 and let @xmath2 be the random subgraph where each @xmath3 independently chooses @xmath1 random neighbors , making @xmath4 edges in all . when the minimum degree @xmath5 then @xmath2 is @xmath1-connected w.h.p . for @xmath6 ; hamiltonian for @xmath1 sufficiently large .
when @xmath7 , then @xmath2 has a cycle of length @xmath8 for @xmath9 . by w.h.p
we mean that the probability of non - occurrence can be bounded by a function @xmath10 ( or @xmath11 ) where @xmath12 . |
astronomy at long infrared wavelengths ( 201000 ) is a relatively young field due to the need for space missions to avoid the absorption of the atmosphere in this waveband .
this waveband , however , is vital for astronomical studies as this is where dust in the interstellar medium ( ism ) radiates .
this is important for studies of galaxy evolution as star formation regions tend to be dusty , and therefore the use of uv and optical measurements to trace the star formation rate can lead to it being underestimated ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
calibrating the relationship between infrared emission and star formation rate has been difficult due to uncertainties from the contribution of the general stellar population to heating the dust , the fact that not all optical / uv emission is absorbed , and uncertainties in the properties of the dust .
the dust emission is affected by the composition of the dust , and the proportion of non - equilibrium to equilibrium heating .
studies with previous space missions _ iras _ , _ iso _ , _ spitzer _ and _ akari _ tried to address these questions ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
however , as they were limited to wavelengths less than 160 , they were not sensitive to the cold dust ( @xmath415k ) and missed up to 50% of the dust mass in galaxies .
the continuum emission from the dust has been + + @xmath5__h _ _ esa e p i nasa .
+ proposed as a potential method of measuring the total mass of the ism @xcite ; traditionally the amount of gas has been measured using hi and co surveys , but due to sensitivity and resolution issues this method is limited to low redshift and small numbers of galaxies .
@xcite found that for early - type galaxies ( e / s0 ) the ism was detected for 50% of objects through its dust emission but only 22% through its co emission .
in addition , the conversion of the co tracer to molecular gas mass is highly uncertain and is believed to vary with metallicity ( see * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
this is a topical area as recent studies using the planck all - sky survey and _ iras _
maps have made an estimate of the amount of ` dark gas ' @xcite in the milky way .
the ` dark gas ' is thought to be molecular gas which is traced by dust , but not detected with the standard co method .
previous works have also suggested the presence of ` dark gas ' by using @xmath6-ray emission from cosmic - ray interactions with clouds of gas @xcite and by the kinematics of recycled dwarf galaxies @xcite .
_ herschel _ is one of the european space agency s flagship observatories and observes in the far - infrared ( fir ) in the range 55671 ( see * ? ? ?
due to the large space mirror and cryogenic instruments , it has a high sensitivity and unprecedented angular resolution at these wavelengths .
_ herschel _ has the ability to target both large numbers of galaxies and map large areas of sky .
it has two photometric instruments : pacs @xcite which can observe in 3 broad bands around 70 , 100 and 160 ( 70 and 100 can not be used simultaneously ) and spire @xcite which observes simultaneously in filter bands centered at 250 , 350 and 500 .
spire provides flux measurements on the longer wavelength side of the fir dust peak ( @xmath0160 ) , allowing us to obtain a full census of dust in nearby galaxies ( e.g. , * ? ? ?
andromeda ( m31 ) and the milky way ( mw ) are the only two large spirals in the local group .
studies of andromeda therefore provide the best comparison to observations of our own galaxy with the advantage that we get a ` global ' picture , whereas investigations of the milky way are limited by our location within the galaxy .
the total size of m31 and the scalelength of its disk are both approximately twice those of the mw ( see * ? ? ?
* and references therein )
. however , the star - formation rate of the mw is @xmath036@xmath7 @xcite compared to only @xmath00.31.0 in m31 ( * ? ? ?
* ; * ? ? ?
* g. p. ford et al . , in preparation ) , despite similar amounts of gas present @xcite .
for this reason , m31 is often labelled as ` quiescent ' . the dust emission from m31
is dominated by a dusty star - forming ring at a radius of 10kpc , and was first observed in the infrared by _ iras _ @xcite .
many projects to map the ism in m31 , have been undertaken using observations in the mid - infrared ( mir ) with _ spitzer _
@xcite , in the fir with _ spitzer _
@xcite , in the hi atomic line @xcite and in the co(@xmath8=1 - 0 ) line @xcite .
studies of the gas kinematics and dust emission show m31 has a complicated structure ( e.g. , * ? ? ?
* ; * ? ? ?
* ) ; @xcite attribute many of the features observed to density waves from a possible head - on collision with m32 . @xcite investigated the relation between gas , dust and star formation using fir _ spitzer _ data .
@xcite used _
spitzer _ and gas observations to investigate the conversion between co(@xmath8=1 - 0 ) line flux and molecular hydrogen column - density in a sample of nearby galaxies including m31 . in their analysis
they found m31 has the lowest dust temperatures in their sample and therefore would benefit most by including _
herschel _ data .
the _ herschel _ exploitation of local galaxy andromeda ( helga ) is a survey covering a @xmath9 area centered on m31 .
we use pacs - spire parallel mode , observing at 100 , 160 , 250 , 350 and 500 simultaneously .
further details of the observations can be found in section [ sec : firobs ] and in @xcite .
helga observations have been used in other works to investigate structures in dust and hi at large radii @xcite , the relationship between gas and star formation ( g. p. ford et al .
, in preparation ) , and the structure of m31 and the cloud - mass function ( j. kirk et al .
, in preparation ) .
in this paper we use the _ herschel _ data combined with the wealth of ancillary data to investigate the distribution , emission properties and the processes heating the dust in m31 on spatial scales of @xmath0140kpc .
there have still been relatively few attempts to map the dust within a galaxy ; recent studies with _ herschel _ include @xcite , @xcite , @xcite , @xcite , and @xcite , but they are often limited to small numbers of independent pixels .
the advantage of m31 over these other studies is the close proximity which allows us to investigate the dust at higher spatial scale and with many more independent pixels .
we also apply the _ planck _
method for finding ` dark gas ' @xcite to m31 and use this method to determine a value of the x - factor the relationship between the molecular hydrogen column density and the observed co tracer . in section [ sec : data ] we present the data used for this analysis and section [ sec : firsed ] describes our method for fitting the spectral energy distribution of dust .
section [ sec : res ] presents our maps of the dust properties , including the distribution of dust surface density , temperature and spectral index .
this section also describes a comparison of the distributions of gas and dust and of the search for an excess emission at long wavelengths . in section [ sec : disc ]
we discuss the dust properties including the composition of the dust and the processes heating the dust . in this section
we also search for ` dark gas ' and determine the value of the x - factor .
the conclusions are presented in section [ sec : conc ] .
_ herschel _ observations of m31 were taken using both pacs and spire in parallel mode covering an area of @xmath10 centered on m31 . to observe an area this large ,
the observations were split into two halves with a cross - scan on each half , which produced data at 100 , 160 , 250 , 350 and 500 simultaneously ( observation ids : 1342211294 , 1342211309 , 1342211319 , 1342213207 ) .
full details of the observing strategy and data reduction can be found in @xcite .
the pacs data reduction was performed in two stages .
the initial processing up to level-1 ( i.e. , to the level where the pointed photometer timelines have been calibrated ) is performed in hipe v8.0 @xcite using the standard pipeline . to remove low - frequency noise ( or drifts ) in the arrays , residual glitches and create the final maps we use scanamorphos ( v15 , * ? ? ?
the final maps have a pixel scale of 2 and 3 and a spatial resolution of 12.5 and 13.3 full - width half maximum ( fwhm ) at 100 and 160 respectively .
the spire data were processed up to level-1 with a custom pipeline script adapted from the official pipeline .
we apply the latest flux correction factors from the spire instrument control centre to update the maps to the latest calibration product @xcite . for the baseline subtraction
we use a custom method called brigade ( m. w. l. smith et al .
, in preparation ) which uses an alternative technique for correcting temperature drifts .
the final maps were created using the naive mapper with pixel sizes of 6 , 8and 12 with spatial resolution of 18.2 , 24.5 , 36.0 fwhm for the 250 , 350 and 500 maps , respectively .
all _ herschel _ images are shown in figure [ fig : her - maps ] .
in addition to the _ herschel _ data , we make use of the 70 _ spitzer _ mips @xcite map published in @xcite . this observation covers a region of m31 approximately @xmath11 in size and has an exposure time of @xmath12 s pixel@xmath13 .
the data was processed using the mips data analysis tool , version 2.90 @xcite and full details of the reduction can be found in @xcite . to investigate the atomic hydrogen in andromeda we use the hi moment - zero map presented in @xcite .
the observations were made with the westerbork synthesis radio telescope ( wrst ) covering a region of @xmath14 with a resolution of 18@xmath15 15 . in this work we present our results using a map which has not been corrected for opacity effects since this correction is uncertain .
for the results which make use of the hi map we also test how our results are affected by using a hi map corrected for self - opacity using the prescription outlined in @xcite .
the uncorrected hi map has a sensitivity of @xmath16 .
for the molecular hydrogen gas content we use co(@xmath8=1 - 0 ) observations presented in @xcite made with the iram 30 m telescope .
this covers an area of @xmath17 with a sensitivity of @xmath00.35kkms@xmath13 .
all maps other than the _ herschel _ images used for this analysis are shown in figure [ fig : oth - maps ] .
to investigate how the dust properties vary across m31 we undertake a pixel - by - pixel dust analysis , using the _ spitzer _ 70 , _ herschel _ pacs and spire images .
we first convolve the data to the same resolution as the 500map ( our largest fwhm ) by using a kernel which matches the point - spread function ( psf ) in a particular band to the 500 band .
this procedure is described in detail in @xcite .
the images are then re - binned into 36 pixels , chosen to be about the same size as the 500 beam so that each pixel is approximately independent from each other . for each band
we subtract a background value for the whole map , estimated from regions around the galaxy .
the uncertainties on the flux in each pixel are found by measuring the standard deviation of the pixels in these background regions and adding this in quadrature with the calibration uncertainty .
the flux errors in the majority of pixels are dominated by the calibration uncertainty which we take as 7% for _ spitzer _ 70 @xcite , 10% for pacs
( see * ? ? ? * ) and 7% for spire ( see section [ sect : sedfit ] for more details ) . for each pixel we fit the far - infrared
sub - millimetre sed with a one - temperature modified black - body model described by : @xmath18 where @xmath19 is the dust mass with dust temperature @xmath20 , @xmath21 is the planck function and @xmath22 is the distance to the galaxy .
@xmath23 is the dust absorption coefficient , described by a power law with dust emissivity index @xmath1 such that @xmath24 .
we assume a typical value for the ism of @xmath25 @xcite .
while the absolute value of @xmath23 is uncertain , its value will not affect trends with other parameters as it is a fixed scaling constant .
the distance to andromeda was taken in this work to be @xmath26 @xcite .
we initially used a fixed value of @xmath1 across the whole galaxy , but we found that with a fixed value it was impossible to adequately fit the seds ; we therefore let @xmath1 be a free parameter . to ensure the simplex fitting routine
did not get stuck in a local minimum , we ran the sed - fitter in two ways : first with all three parameters ( m@xmath27 , t@xmath27 , @xmath1 ) free to vary ; second by fixing the value of @xmath1 while allowing m@xmath27 and t@xmath27 to vary and repeating the process for all values of @xmath1 in the range 0.20 to 5.90 in 0.01 intervals , selecting the result with the lowest @xmath28 .
both methods gave consistent results but we created the final maps of the dust properties by choosing the result with the lowest @xmath28 for each pixel .
the spire calibration has an overall systematic uncertainty of 5% due to the uncertainty in the prime calibrator neptune , and a statistical uncertainty of 2% determined from instrumental reproducibility ; the spire observer s manual recommends linearly adding these to give an overall uncertainty of 7% . to implement this in practice in our fitting algorithm
we increased the uncorrelated uncertainty to give the same overall uncertainty when the errors are added in quadrature .
this is implemented in the sed fitting by using the full covariance matrix in the @xmath28 calculation .
we apply our own color correction to the _ herschel _ maps by removing the _ herschel _ pipeline ` k4 ' parameter and then convolving the sed with the filter transmission in the fitting process ( for spire the filters appropriate for extended sources are used ) .
this method takes full account of all the wavelength - dependant effects associated with pacs and spire . in previous work @xcite we found that there is a significant contribution from a warmer component of dust at wavelengths @xmath470and so the _ spitzer _ flux at 70 is treated as an upper limit in the fitting ( i.e , if the flux value is higher than the model it does not contribute to @xmath28 ) .
the omission of the warm component in the fitting process has a negligible effect on the derived dust mass as the cold component dominates the total dust mass .
@xcite suggest that a warmer component could influence the dust emission up to wavelengths of 160 ; to test this we repeated the sed - fitting by treating the _ spitzer _ and pacs fluxes as upper limits and found it made a negligible difference to our results . to estimate the uncertainties on the results of our fits we use a monte - carlo technique .
for each pixel we create a set of 1000 artificial seds , created by taking the original flux densities and adding a random value selected from a normal distribution with a mean of zero and a standard deviation equal to the uncertainty in the measured flux ( the correlations between the calibration uncertainties for spire are also included ) .
we estimate the error in the derived parameters for each pixel from the 1000 fits .
we find that for each pixel there is an uncertainty of 20% in our estimate of the surface density of the dust , of @xmath291.4k in our estimate of the dust temperature , and @xmath290.31 in our estimate of @xmath1 .
values from pixels fitted with the modified - blackbody model .
the red line represents the 10% significance value for one degree of freedom.,scaledwidth=49.0% ] in producing the final dust mass , temperature and @xmath1 images in this work , we only used pixels where the fluxes in all six bands ( five _ herschel _ & mips 70 ) have a signal - to - noise greater than 5@xmath30 .
while this potentially causes us to miss the very coldest dust due to weak emission in the shortest wavelength bands , we choose it as a conservative approach to ensure we have accurate estimates of temperature . in practice
it is the 100 map , which has the lowest sensitivity , which limits the number of pixels in our selection . despite this very conservative selection
, there are still @xmath31 pixels in our resultant maps . to see if our model is a statistically reasonable fit to the data , we created a histogram of the @xmath28 values for all pixels ( figure [ fig : chisq ] ) . as the 70 flux
is used as an upper limit ( see section [ sect : sedfit ] ) and is usually higher than the best - fit model it does not usually contribute to @xmath28 and therefore we only have 1 degree of freedom ( @xmath32 ) .
the 10% significance level for @xmath28 occurs at 2.71 which is shown by the red line in figure [ fig : chisq ] .
we find 9.8% of our fits have @xmath28 above this level showing our model is an adequate representation of the data .
we have also checked to see if our radial gradients in temperature , @xmath1 , and dust surface density are affected by lowering the criteria to include fluxes greater than 3@xmath30 and find no significant changes .
recent results from the key insight on nearby galaxies : a far - infrared survey with _
herschel _
( kingfish ) presented in @xcite have suggested that the one - temperature modified black - body model underestimates the dust mass compared to the @xcite prescription .
they attribute this difference to the contribution of warm dust emitting at shorter wavelengths . for our analysis
this does not appear to be the case .
first when we set the pacs fluxes ( @xmath4160 ) as upper limits there is no significant change in our results .
second , if multiple temperature components were present this would bias our @xmath1 values to lower values ( since each temperature component peaks at a different wavelength , * ? ? ?
* ) , but we mostly find higher @xmath1 values than expected ( see section [ sec : res ] ) .
third the @xmath28 analysis suggests the model is consistent with the data .
in addition , @xcite , using a similar analysis , have found in m51 that the dust mass distribution is similar when using the @xcite or the single modified blackbody prescription .
we applied the same fitting technique as outlined in section [ sect : sedfit ] to the global flux densities , measured in an elliptical aperture with a semi - major axis of 2.0@xmath33 and semi - minor axis of 0.73@xmath33 .
this produces a total dust mass of @xmath34 with a dust temperature of @xmath35k and @xmath36 .
the total dust mass from summing all the pixels in our pixel - by - pixel analysis gives a value @xmath37 , a factor of @xmath02.5 lower .
this is expected as the fraction of 500 flux in the pixels which satisfy our signal to noise criterion is approximately half the global value . combining our global dust mass with gas measurements ,
we find a global gas - to - dust mass ratio of 72 .
the global temperature is consistent with these for other spiral galaxies obtained using similar methods ( e.g. , * ? ? ?
the pixel - by - pixel analysis shows a large range of temperatures and @xmath1 values as discussed in detail in section [ sec : res ] .
examples of fits for individual pixels are shown in figure [ fig : exampleseds ] . to help understand the significance of our results and any potential biases or degeneracies in the parameters , we ran a monte - carlo simulation . assuming the dust emits as a single - component modified blackbody , we generated synthetic flux values for a range of temperatures and @xmath1 values , with the same wavelengths as our real data .
noise was then added to the simulated fluxes with a value equal to the errors in the real fluxes ( excluding the calibration errors ) for 2000 repetitions per t , @xmath1 combination .
the calibration error was not included as it would systematically shift the fluxes for all pixels .
we chose an input mass surface density of 0.5@xmath38 as this roughly corresponds to the values found in the 10kpc ring .
the quantity that is most important for our work is the dust mass , which from equation [ equ : sed ] is just a multiplicative term .
figure [ fig : sim - mass ] shows the mean and median mass returned for the 2000 repetitions as the input temperature is varied between 12 and 30k for a @xmath1 of 2 ; the error bars show the error on the mean . between 15 and 30k ,
the mean and median dust mass returned matches within the errors the input dust surface density of 0.50 @xmath39 . at dust temperatures of 15k and below
, there are large errors on the returned mass , which is due to the fluxes at the pacs wavelengths not reaching the required sensitivity to be included in the fit . for the actual data , we only included pixels
in which there is at least a 5@xmath30 detection in all bands , which will avoid the wildly incorrect estimates of the dust mass seen in the simulation . to fully estimate the dust mass and temperature of very cold dust ( t @xmath40 15k ) , we need flux measurements at @xmath41500 .
nevertheless , the lack of an excess at 500 ( section [ sec : excess ] ) is circumstantial evidence that andromeda does not contain very cold dust .
we investigated if the results in figure [ fig : sim - mass ] were different for a different choice of @xmath1 but found no systematic differences . by plotting the difference in the resultant temperature in each pixel compared to the input model temperature
, we find that between 15k and 25k , the temperature uncertainty is @xmath00.6k . above 25k
the uncertainty increases , although in the simulation we did not include a 70 point which would likely provide a constraint to our fits if the dust temperature was @xmath4125k .
the simulation suggests that for input @xmath1 between 1.5 and 2.4 , the uncertainty in the returned value of @xmath1 for each pixel is @xmath00.1 .
as expected these uncertainties are lower than returned by the monte - carlo technique in section [ sect : sedfit ] as we have not included a calibration uncertainty . with @xmath42 modified - blackbody .
the mean of the returned masses is shown by the blue points and the median is shown by the black points .
the input mass surface density of 0.50 @xmath38 is shown by the red dashed line . to avoid overlapping data points the blue
have been shifted by -0.1k and the black by + 0.1k.,scaledwidth=49.0% ] a degeneracy is known to exist between temperature and @xmath1 ; figure [ fig : sim - delt - delt ] shows that if there is an error in one parameter this is anti - correlated with the error in the other parameter .
as the distribution is centered on the origin , there is no systematic offset in the returned value of t or @xmath1 . as there is no systematic offset
, the error in the mean values over many pixels will therefore be much smaller than the error for a single pixel .
this simulation is based on the dust emission arising from a one - component modified blackbody . fitting a one - component modified blackbody to pixels for which there are multiple temperature components would produce a bias towards smaller values of @xmath1 @xcite , although we would hope to detect this by finding high @xmath28 values . in section [
sec : beta - t ] we show that different regions of m31 have different @xmath43 relations and discuss why this is unlikely to be due to multiple temperature components . to summarise , while there is a @xmath1t degeneracy from the fitting algorithm this does not create any systematic offsets in the value returned . if the dust temperature falls below 15k we are unable to constrain the sed due to the lack of data points beyond 500 . and temperature .
the uncertainties are taken from the simulated data with @xmath44k and @xmath45 . a clear anti - correlation is observed with the distribution centered on the correct values . to fully populate this graph we increased the simulation to 20,000 repetitions for this @xmath1 , t combination.,scaledwidth=49.0% ]
by fitting seds to every pixel , we have created maps of dust surface density , temperature , @xmath1 and gas - to - dust ratio which are shown in figure [ fig : sed - maps ] ( for details on how the gas surface density is calculated see section [ sec : radgd ] ) .
the dust surface density distribution , un - surprisingly , is more similar to the maps of gas and star - formation rate than to the 3.6 image ( see figure [ fig : her - maps ] & [ fig : oth - maps ] ) , which traces the stellar mass distribution .
the @xmath1 and temperature maps show significant radial variations .
figure [ fig : sed - rad ] shows how the dust column density , temperature and @xmath1 vary with radius ( the physical radius is calculated assuming an inclination of 77@xmath46 and pa of 38@xmath46 , @xcite ) . in the star - forming ring at 10kpc , @xmath1 has an average value of @xmath01.8 but increases with decreasing radius reaching a maximum of @xmath02.5 at a radius of @xmath03kpc .
this is higher than found in global studies of galaxies @xcite .
however similarly high values have been reported recently in @xcite and @xcite for dust within galaxies .
the value for the ring agrees well with early results from _
@xcite for dust in the galactic disk and the solar neighbourhood .
the 10kpc ring has an average dust surface density of @xmath00.6@xmath47 with a dust temperature of 18k . towards the very center of the galaxy the dust column density decreases to @xmath00.04@xmath47 , @xmath1 values fall to @xmath01.9 and
the dust temperature increases to @xmath030k . )
, the dashed green line represents the transition radius ( 3.1kpc ) found when fitting the @xmath1 results.,scaledwidth=49.0% ] .
the solid red line represents the best - fit exponential profile to the gas - to - dust ratio.,scaledwidth=49.0% ] in figure [ fig : sed - rad ] we see a clear break in the radial variation for both temperature and @xmath1 at a radius @xmath03kpc .
we fit a model with two straight lines and a transition radius ( using a simplex routine ) to the @xmath1 results .
the same method is used with the temperature values but we set the transition radius to the value obtained from the fit to the @xmath1 values , which occurs at 3.1kpc ( shown by the dashed green line in figure [ fig : sed - rad ] or black ellipse in figure [ fig : sed - maps ] ) . at radii smaller than the transition radius ,
the temperature decreases with radius from @xmath030k in the center to @xmath017k , with an associated increase in @xmath1 from @xmath01.8 to @xmath02.5 . at radii larger than the transition radius the dust temperature slowly increases with radius while @xmath1 decreases with radius to @xmath01.7 at 13kpc .
the best - fit relationships between @xmath20 , @xmath1 and @xmath48 are shown in figure [ fig : sed - rad ] and listed below : @xmath49 @xmath50 we discuss the cause of the temperature and @xmath1 variations in section [ sec : disc ] . for sections [ sec : heating ] & [ sec : beta - t ] we consider the inner ( @[email protected]@xmath52 ) and outer regions separately . in figure
[ fig : gas - rad ] the radial variations of the atomic gas , the molecular gas and the gas - to - dust ratio are shown .
we assume an x - factor of @xmath53^{-1}$ ] @xcite to convert a co line flux to a h@xmath54 column density , although this value is notoriously uncertain and has been found to vary with metallicity @xcite . for our analysis of m31 , the choice of x - factor makes very little difference as the total gas is dominated by the atomic phase . out of the 3974 pixels plotted in the gas - to - dust figure ,
only 101 have a molecular fraction of @xmath4150% and globally the molecular gas only constitutes @xmath07% of the atmoic gas @xcite . to estimate the total gas surface density we include the contribution of the atomic or molecular gas in each pixel if the value is greater than 3@xmath30 in their respective maps
. only 86 of our 5@xmath30 dust pixels are not covered by the co(@xmath8=1 - 0 ) observations ; these pixels are in the outskirts of the galaxy and we assume the contribution of molecular gas is negligible . we find a tight relation between the gas - to - dust ratio and radius ( spearman rank coefficient of 0.80 ) as shown in figure [ fig : gas - rad ] , which is described by an exponential profile , shown by the red line . the gas - to - dust ratio increases exponentially from low values of @xmath020 in the center of the galaxy to @xmath0110 at 15kpc typical of the milky way in the local environment ( see * ? ? ?
* and references therein ) .
note that the values are not corrected for helium in the ism .
metallicity gradients have been estimated from oxygen line ratios [ oiii]/h@xmath1 , [ oii]/[oiii ] and r@xmath55 by @xcite .
they infer a radially decreasing metallicity gradient of @xmath56 from the r@xmath55 parameter .
@xcite calculate oxygen abundance gradients based on a set of 11 hii regions from @xcite and find gradients in the range of -0.027 to -0.013@xmath57 depending on the calibration used .
if a constant fraction of the metals in the ism is incorporated into dust grains @xcite , one would expect the gas - to - dust gradient to be @xmath58 metallicity gradient .
we find a gas - to - dust gradient of @xmath59 consistent within the uncertainties to the gradient measured by @xcite ( if the hi map corrected for self - opacity is used the gradient slightly increases to @xmath60 ) .
this supports the claim that gas can be well traced by dust at a constant metallicity . to see
if the uncertainty in the choice of x - factor could affect this result , we carried out the same procedure but limited the analysis to pixels where the molecular hydrogen contribution is less than 10% of the total gas mass .
this produced only a small change in the gas - to - dust gradient to @xmath61 which is still consistent with the metallicity gradient .
searches for a long - wavelength sub - mm excess ( i.e. @xmath41500 ) has been made in both the milky way @xcite and for nearby galaxies .
a sub - mm excess is important as it could suggest very cold dust is present ( @xmath4015k ) which would dominate the dust mass in galaxies .
other possible explanations of an excess include variations in the dust emissivity index with wavelength @xcite , different dust populations or contamination from a synchrotron radio source .
sub - mm excesses have been reported from observations of low - metallicity dwarf galaxies ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and spiral galaxies ( see * ? ? ?
* ; * ? ? ?
most detections have been made by combining fir data with ground - based data at 850 or 1 mm data rather than from data only at @xmath62500 .
we searched for a sub - millimetre excess in m31 by comparing the 500 flux to our best - fit models .
a 500 excess is defined to be any observed 500 flux that is greater than the expected flux at 500 from our modified blackbody fit to the data .
figure [ fig : excess ] is a histogram of the ratio of the excess at 500 and the noise on the 500 map .
the distribution of the excess is consistent with a gaussian function with a mean of 0.54@xmath30 and standard deviation of 0.31@xmath30 .
the fact the distribution of the histogram is a gaussian suggests that what we are seeing is random noise with a fixed offset ( not centered on zero ) .
two non - astronomical sources could explain a fixed offset , either an incorrect background subtraction or a calibration error .
the background correction applied is quite small ( 0.2@xmath30 ) , and thus an error in this is unlikely to be the entire explanation of the offset .
the distribution of the excess is consistent with the 2% random spire calibration uncertainty , so both factors together could explain the small offset . if the excess is generated from dust with a 10k temperature and @xmath1 of 2 , the dust mass is @xmath010@xmath63@xmath64 in our @xmath415@xmath30 pixels , which corresponds to only 13% of the mass of the warmer dust .
if we used a model containing dust at more then one temperature we would not get a useful constraint on the colder dust component without additional data at longer wavelengths
. in particular observations at @xmath0850 ( e.g. with scuba2 ) would be useful to determine if any cold dust present .
recent studies by @xcite and @xcite have used _ herschel _ colors to confirm previous works ( e.g. , * ? ? ? * ; * ? ? ?
* ) that emission from dust in nearby galaxies at wavelengths longer than 160 is mostly from dust heated by the general stellar population .
these authors conclude that at wavelengths shorter than 160 , the emission tends to be dominated from warmer dust heated by newly formed stars .
@xcite in a study of m31 using _
spitzer_mir / fir , uv and optical data , also concluded that the dust is mostly heated by an old stellar population ( a few gyr old ) .
to investigate the relation between our derived dust properties and the sfr and the general stellar population , we have used the 3.6 map presented in @xcite to trace the general stellar population ( rather than the luminous newly formed stars that dominate the uv ) and a map of star - formation rate ( sfr ) .
the sfr map is created from far - uv and 24 images which have been corrected for the contribution of the general stellar population to the emission at these wavelengths ( for details see g. p. ford et al . in preparation ) .
these maps were all convolved to the same resolution and binned to the same pixel size as all the other maps in our analysis .
we have plotted the fluxes from these maps against the results of our sed - fits ( see figure [ fig : heating ] ) . in section [ sec : spatial - dist ] we showed there is a clear break in the dust properties at a radius of 3.1kpc .
therefore in figure [ fig : heating ] the pixels at a radius less than 3.1kpc are shown in blue and those at a radius above 3.1kpc in green .
for each graph ( and both sets of pixels ) the spearman rank coefficient is computed and an estimate made of its significance ( see table[table : spearman ] ) .
with such a large number of pixels , all but one of our correlations are formally significant . in the discussion below
we have concentrated on the strongest correlations as measured by the spearman correlation coefficient .
there are strong correlations in figure [ fig : heating ] between @xmath1 and 3.6 emission .
we believe these are most likely to be caused by radial variations in @xmath1 seen in figure [ fig : sed - rad ] and the decrease in 3.6 brightness with radius .
we discuss the possible cause of the radial variation in @xmath1 in section [ sec : beta - t ] .
we find a strong correlation between dust surface - density and the sfr in the outer regions , but not with the surface density of total stars traced by the 3.6 emission .
this correlation is expected as stars are formed in clouds of gas and dust . in the inner region
there is an anti - correlation between the dust surface density and the 3.6 flux , which seem most likely to be explained by both quantities varying with radius : the 3.6 emission from the bulge increasing towards the center and the dust surface density decreasing towards the center .
ccccc & & inner & -0.73 & @xmath65 + & & outer & -0.16 & @xmath66 + & & inner & 0.31 & @xmath67 + & & outer & 0.74 & 0.00 + ( rl)1 - 5 & & inner & 0.90 & @xmath68 + & & outer & -0.09 & @xmath69 + & & inner & 0.04 & @xmath70 + & & outer & 0.14 & @xmath71 + ( rl)1 - 5 & & inner & -0.52 & @xmath72 + & & outer & 0.56 & @xmath73 + & & inner & 0.16 & @xmath74 + & & outer & -0.24 & @xmath75 [ table : spearman ] the strongest correlation is seen between the dust temperature and the 3.6 flux in the inner region .
figure [ fig : ftcorr ] shows a log - log graph of the two quantities and a linear fit to the points .
the gradient represents the power _
n _ where @xmath76 where @xmath77 .
while the correlation suggests the dust in the bulge is heated by the general stellar population , for a modified blackbody with @xmath42 we would expect a gradient of 6 .
the difference is probably explained by the simplicity of our assumptions : that there is only a single stellar population in the bulge and that the bulge has a constant depth in the line of sight .
if these assumptions are incorrect , the 3.6 surface brightness of m31 will only be an imperfect tracer of the intensity of the interstellar radiation field ( isrf ) .
looking at the temperature beyond 3.1kpc , we find a weak , but still highly significant , correlation with sfr , suggesting that the isrf has a significant contribution from star - forming regions .
as most of the star formation in m31 occurs in the 10kpc ring this is to be expected .
this correlation can be seen in figure [ fig : sfrcont ] where most but not all of the temperature peaks in the 10kpc ring appear aligned with the peaks in the sfr map . in the same region there is a slight anti - correlation of temperature with the 3.6 flux which could be explained by the radial decrease in the 3.6 flux whilst the dust temperature increases slightly with radius .
the fact that dust temperature increases slightly with radius , while the number - density of stars , traced by the 3.6 , is falling with radius suggests that outside the bulge the dust is mainly heated by young stars . nevertheless , the lack of a strong correlation between dust temperature and either sfr or 3.6 flux suggest the optical / uv light absorbed by a dust grain is from photons from a large range of distances ( e.g. , photons from the bulge heating dust in the disk ) .
@xcite have studied fir colour ratios in m81 , m83 and ngc4203 .
they find the 250/350 color ratio has the strongest positive correlation with 1.6 emission .
an increase in the 250/350 ratio would indicate either an increase in dust temperature or a decrease of @xmath1 .
@xcite conclude that the most likely explaination is the temperature effect , with the dust being heated by the general stellar population traced by the 1.6 emission .
we find a similar correlation to @xcite , only we trace the stellar radiation field with the 3.6 band .
however , our sed - fitting results suggest that this is caused by a combination of changes in temperature and @xmath1 .
note that since @xcite only use color ratios they are unable to discriminate between changes of temperature and @xmath1 . at radii greater than 3.1kpc
our results suggest the variation in the 250/350 is mainly caused by a change in @xmath1 .
@xcite found that the 70/160 color ratio has the greatest correlation with star - formation rate , which is evidence that there is dust at more than one temperature contributing to the 70500 emission .
one possible explanation of our failure to find a correlation between the 3.6 emission and dust temperature outside 3.1kpc , instead of the negative correlation between 3.6 and @xmath1 we find , might be if the _ herschel _ emission at short wavelengths contains a contribution from a warmer dust component .
this would mean our fits of a one - component modified blackbody would produce misleading results .
however , as stated in section [ sect : sedfit ] when we attempted the same sed fitting process but using all flux densities @xmath4160 as upper limits , we see little difference in our results . the dust emissivity index ( @xmath1 ) is related to the physical properties of the dust grains , including the grain composition , grain size , the nature of the absorption process and the equilibrium temperature of the dust . we would also expect to see a change in @xmath1 due to environment from the processing of the grains via grain growth ( e.g. coagulation , mantle accretion ) or destruction through surface sputtering by ions / atoms or shattering by shocks . in m31
, we detect an apparent inverse correlation between @xmath78 and @xmath1 for the inner and outer regions of m31 , as shown in figure [ fig : b - tmodels ] .
we find the form of the relation is different for the two regions .
such an inverse relationship has been observed in the milky way with previous fir - submm experiments and surveys including archeops @xcite , which showed @xmath1 ranging from 4 to 1 with the dust temperature varying between 7 and 27k , and pronaos @xcite , which shows a variation of @xmath1 from 2.40.8 for dust temperatures between 11 and 80k .
@xcite used ir - mm data of galactic high latitude clouds and found a similar trend , and more recently @xcite with _ herschel _ found a similar inverse relationship with @xmath79 for @xmath80 for galactic longitude @xmath81 ( at longitude @xmath82 @xmath83 for 1823k )
. recently @xcite used _
herschel_-atlas observations to investigate @xmath1 variations in low - density , high - latitude galactic cirrus , measuring values of @xmath1 ranging from 4.51.0 for @xmath84
. these @xmath85 relationships could be indicative of a problem with the temperature@xmath1 degeneracy arising from the sed fitting , the presence of dust with a range of temperatures along the line of sight @xcite , or real variations of the properties of the dust grains . on the assumption that the inverse correlations between @xmath1 and @xmath78 in figure [ fig : b - tmodels ] are not simply caused by the two variables being separately correlated with radius
, we looked for other possible causes of the relationships : 1 .
the fitting can lead to a spurious inverse correlation between @xmath1 and @xmath78 ( * ? ? ?
* section [ sec : simulation ] ) .
the most striking feature in figure [ fig : b - tmodels ] is the clear separation in points between the inner 3.1kpc and the outer regions . to test whether these different distributions might be produced by the fitting artifact
, we used the monte - carlo simulations from section [ sec : simulation ] to simulate the effect of fitting a modified blackbody for various combinations of @xmath1 and @xmath86 .
the grey lines in figure [ fig : b - tsim ] show the best - fit relationships for different input @xmath78 and @xmath1 combinations and clearly show that the two different relationships in the two regions can not be obtained from a single population of dust grains .
the green and blue data points represent the range of output @xmath78 and @xmath1 for an input modified blackbody with @xmath87k ( green ) and @xmath88k ( blue ) , with @xmath89 . a comparison of figures [ fig : b - tmodels ] and [ fig : b - tsim ] shows that in both regions of m31 there is a larger range of temperature and dust emissivity for the real data than that found in the monte - carlo simulation , indicating there are genuine variations of @xmath78 and @xmath1 in both regions .
moreover , the fitting artifact can not explain the observed relationships of @xmath1 and temperature with radius ( figure [ fig : sed - rad ] ) .
2 . artificial inverse @xmath90 relationships can also be produced if a one - component modified blackbody model is used to fit dust which contains a range of dust temperatures @xcite . since we are averaging through the disk of a galaxy along the line of sight ( los ) , it is obviously possible that the dust contains a range of dust temperatures . while we can not fully address this issue , our @xmath1 values are higher than expected , which is the opposite of what happens from a los averaging of temperatures .
we also find no statistical evidence from our fits that there is more than one component of dust .
also @xcite and @xcite show inverse @xmath85 relationships still exist in places where it is unlikely there is dust at more than one temperature .
3 . variation of @xmath1 with wavelength has been reported by some authors both from theoretical models and laboratory experiments and from observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , with a transition around 500 . in section [ sec : excess ]
we show that there is no evidence for excess 500 emission , suggesting this is not an explanation of our results . to describe the @xmath85 relationships we use an empirical model of the form @xmath91 , commonly used in the literature ( e.g. , * ?
* ; * ? ? ?
* ; * ? ? ?
* ) to fit the @xmath85 anti - correlation . the observed anti - correlation between t and @xmath1 may arise due to a change in the physical properties of grains including the grain optical constants changing with temperature for amorphous grains or changes in the dust emissivity with wavelength ( see * ? ? ? * and references therein ) .
other possibilities include grain growth or quantum mechanical effects ( though these latter grain properties only arise at lower dust temperatures than observed in m31 ) .
the best - fit relationship which describes @xmath85 for r@xmath92 and for @xmath93 are ( shown in figure [ fig : b - tmodels ] ) : @xmath94 where the steeper @xmath95 relationship at r@xmath96 agrees well with the relationship found in the plane of the milky way at longitudes of @xmath81 @xcite and in low - density , high - latitude cirrus @xcite .
there is some evidence that the @xmath95 relationship in m31 is slightly steeper , so that for the same temperature compared to the galactic plane , m31 has a lower @xmath1 ( but this is only at @xmath97 level ) . what could be a physical ( or chemical ) explanation of the different @xmath98 relationships in the two regions
? typical values of @xmath1 are in the range 1.52.0 for interstellar dust grains , and have been found in global extra - galactic studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and _ average _ global values measured in the milky way ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
low values of @xmath1 for large grains would typically represent freshly - formed dust grains in circumstellar disks or stellar winds .
alternatively , @xmath99 has been observed in regions where small grains dominate @xcite .
high values of @xmath1 ( @xmath100 ) might occur due to grain coagulation or to the growth of icy mantles on the surface of the grains in denser regions @xcite .
studies have also suggested high values of @xmath1 are associated with very cold dust ( @xmath101 e.g. , * ? ? ?
* ) possibly caused by a change in the absorption properties due to quantum effects at low temperatures , increasing self - absorption in amorphous grains via tunnelling @xcite .
the highest value of @xmath1 is seen at the 3.1kpc boundary between the two regions .
this can not be caused by changes in the quantum mechanical absorption , since this is only thought to be important for cold dust at temperatures @xmath102 .
the high @xmath1 values could be due to efficient grain coagulation or mantle growth in dense molecular clouds , although this too seems unlikely as little co(@xmath103 ) is observed in this region . while there is no obvious explanation for the high @xmath1 values at this radius , there are many indications that this 3kpc `` boundary '' is an interesting regime , we discuss this further in section [ sec : transition ] . in the inner region
, we suggest that the decrease in @xmath1 with corresponding increase in @xmath86 might be caused by the increased intensity of the isrf . towards the center of m31 ,
we would expect increased sputtering or sublimation of mantles from the increased isrf , shown by the increased temperature of the dust and the increased x - ray emission observed in the center @xcite .
the lack of gas in the central regions ( figure [ fig : gas - rad ] ) also suggests that dust is less likely to be shielded and thus more efficiently sputtered and leading to smaller grain sizes .
as we mentioned above , a problem with this analysis is that it is difficult to determine which are the causal relationships .
for example , we have argued that the radial variation in @xmath86 is due to the radial variation in the isrf . the radial variation in @xmath1 might then be due to a physical relationship between @xmath104 and @xmath1 or it might be the case that there is no causal relationship between these parameters but the radial variation in @xmath1 is caused by a different effect .
for example , an interaction between m32 and m31 might have caused a wave of star formation which has moved out through the galaxy , which might have led ( by a number of processes ) to the radial variations in @xmath1 .
therefore , we can rule out some hypotheses but we can not conclusively determine which is the true explanation using this data set . interpreting
the transition in dust properties seen at 3.1kpc ( figure [ fig : sed - rad ] ) is difficult .
one possible clue comes from previous gas kinematics studies .
@xcite found that the hi rotation curve inside a 4kpc radius is warped with respect to the rest of disk .
@xcite suggest the inner hi data is consistent with a bar extended to 3.2kpc , while a newer analysis by @xcite explains the hi distribution as the result of a triaxial rotating bulge .
@xcite , using _ spitzer _
irac observations , identified a new inner dust ring with dimensions of 1.5@xmath151kpc . by using the stellar and gas distributions and from the presence of the 10kpc ring , they conclude that an almost head - on collision has occurred between m31 and m32 around 210 million years ago .
this collision could explain the perturbation of the gas observed in the central 4kpc .
these other observations all show that the inner 3kpc of m31 is an intriguing region , although it it not clear what are the causes of the difference in the dust properties . the perturbation of the gas may have lead to the processing of dust grains , or potentially material from m32 could have been deposited after the interaction .
the total dust mass for the pixels in our selection within 3.1kpc is @xmath105 which is a plausible amount to be deposited as recently dust masses of @xmath0@xmath106 have been reported in virgo dwarf ellipticals @xcite .
another possibility is the dust properties are affected by the differences between conditions in the bulge and disk .
@xcite decomposed the luminosity profile of _ spitzer _ irac data into a bulge , disk and halo . from their figure 16 we can see that our transition radius of @xmath03kpc is approximately where the bulge emission begins to become a significant fraction of the optical disk emission .
whether the transition in the dust parameters is due to the changing contribution to the isrf from the general stellar population and star formation or if there is another influence in the bulge is unknown .
the detection of ` dark gas ' in the milky way was a surprising early result from _ planck _
@xcite , obtained by combining _ iras _ 100 data and the six - band _ planck _ data from 3503 mm .
the _ planck _
team compared the dust optical - depth with the total column density of hydrogen ( @xmath107 ) , where the optical depth at each wavelength is given by @xmath108 where @xmath109 is the flux density in that band and @xmath110 is the blackbody function .
they assumed that at low @xmath107 the atomic hydrogen dominates over the molecular component while at high column density the molecular hydrogen dominates the emission .
for these two regimes they found a constant gas - to - dust ratio , but at intermediate column densities they found an excess of dust compared to the gas .
this excess is attributed to gas traced by dust but not by the usual hi and co lines , and is found to be the equivalent 28% of the atomic gas or 118% of the molecular gas .
the excess dust emission was typically found around molecular clouds , suggesting that the most likely cause is the presence of molecular gas not traced by the co line .
we attempted the same analysis as the @xcite for m31 using our sed fitting results from section [ sec : res ] . instead of using equation [ equ : planck ]
, we compare the column density of gas estimated from the hi and co with the column density of dust ( for convenience we call this @xmath111 ) .
we use this parameter as it is calculated with data from all wavelengths , whereas if we used equation [ equ : planck ] , small errors in temperature would cause large uncertainties in @xmath112 for wavelengths close to the peak of emission .
the _ planck _ team found no radial variation in the gas - to - dust ratio in the milky way ( see figure 10 * ? ? ? * ) .
in andromeda we show that the gas - to - dust ratio does vary radially ( section [ sec : radgd ] and figure [ fig : gas - rad ] ) , as expected from the metallicity gradient . to determine if there is an excess at intermediate column density in dust compared to that expected from the gas we have to correct for the radial change in gas - to - dust ratio . to remove this dependence ,
we adjust dust column density by using the exponential fit ( shown by the red line in figure [ fig : gas - rad ] ) so the gas - to - dust ratio at all radii has the same value as the center of m31 . to avoid
biasing this correction by assuming a value for the x - factor which is highly uncertain , we estimate this relationship from pixels where the atomic hydrogen column density is @xmath113% of @xmath107 ( 1569 pixels out of the 3600 ) . in figure [
fig : optdep - gas ] we show that the relationship between corrected dust column density ( @xmath111 ) and gas column density is well represented by assuming the two quantities are directly proportional with no excess in dust column - density that could be attributed to ` dark gas ' .
although only a small proportion of the gas is molecular , we still need to use a value for the x - factor .
we can estimate this quantity from the data itself by finding the values of the x - factor and the constant of proportionality between the gas column density and the corrected @xmath111 that gives the minimum @xmath114 value ( the fitted line is shown in figure [ fig : optdep - gas ] . )
we find a best value for the x - factor of @xmath115^{-1}$ ] ( or expressed as @xmath116^{-1}$ ] ) , where the random error is estimated using a monte - carlo technique ( similar to one used in section [ sect : sedfit ] ) . for the dust column densities in each pixel we use the uncertainties provided by the sed fitter
as explained in section [ sect : sedfit ] , which is on average @xmath022% .
@xcite quote a calibration error of 15% for the co observations ( which directly results in at least a 15% uncertainty in the x - factor ) which we combine with the noise in each pixel of our processed moment - zero co map . for the hi observations we use an uncertainty map provided by robert braun ( private communication ) which has an average uncertainty of 12% on the raw 10 map .
we also include a 5% systematic uncertainty ( e.g. , calibration uncertainties ) .
if the opacity corrected hi map is used our best fit x - factor is @xmath117^{-1}$ ] .
@xcite find values of the x - factor between @xmath118 to @xmath119^{-1}$ ] when analysing the southern , northern and inner regions for m31 with _ spitzer
our average value of the x - factor falls within their range of values .
a full analysis of the spatial variations of the x - factor with our data will be undertaken in a future paper .
there are two main problems with this method which could lead us to miss ` dark gas ' .
first , unlike the milky way we have to average through the whole disk of m31 and , second , m31 has a significantly lower molecular gas fraction .
the latter prevents us from fitting the model to pixels with just very high and low values of @xmath107 as the pixels with highest molecular gas fraction are not clustered to high @xmath120 values ( this is illustrated in figure [ fig : codist ] ) .
this suggests andromeda may not be the best galaxy for this analysis as the molecular contribution to the overall column density is quite low .
however , we can also try an alternative method of looking for ` dark gas ' because we have one important advantage over the _ planck _ team : we can see m31 from the outside .
we can therefore make a map of the ratio of radially corrected dust column density ( @xmath111 ) to gas column density to look for regions of enhancements in this ratio ( figure [ fig : tauimage ] ) .
the image clearly shows spatial variations which could either suggest regions of ` dark gas ' or local variations in the metallicity or emissivity of dust . to distinguish between these scenarios an independent measurement of the ` dark gas '
is required . on the outskirts of molecular clouds
, co could be photodissociated and the carbon gas would therefore reside in c or c@xmath121 @xcite .
planned observations of the c[ii ] 158 line in andromeda with _ herschel _ could then be a potential test for investigating whether ` dark gas ' exists in m31 .
in this paper we present the results of an analysis of dust and gas in andromeda using new _
herschel _
observations from the helga survey .
we have @xmath04000 independent pixels with observations in the range of 70500 .
we find the following results : 1 .
we find that a variable dust emissivity index , @xmath1 , is required to adequately fit all the pixels in andromeda .
when a variable @xmath1 is used , the modified blackbody model with a single temperature is found to be a statistically reasonable fit to the data in the range 100500 .
there is no significant evidence of an excess of dust emission at 500 above our model .
2 . there are two distinct regions with different dust properties , with a transition at @xmath122kpc . in the center of andromeda , the temperature peaks with a value of @xmath030k and a @xmath1 of @xmath01.9 .
the temperature then declines radially to a value of @xmath017k at 3.1kpc with a corresponding increase in @xmath1 to @xmath02.5 . at radii larger than 3.1kpc @xmath1 declines but only with a small associated increase in temperature .
the drop in @xmath1 towards the center of the galaxy may be caused by increased sputtering or sublimation of mantles from an increased isrf .
the origin of the high @xmath1 values at 3.1kpc from the center is less clear but may be indicative of either grain coagulation or an increase in the growth of icy mantles .
the dust surface density for our pixels in which flux is detected at @xmath123 in all herschel bands range is between @xmath00.12.0@xmath38 .
we find the gas - to - dust ratio increases exponentially with radius .
the gradient matches that predicted from the metallicity gradient assuming a constant fraction of metals are included into dust grains .
the dust surface density is correlated with the star formation rate rather than with the stellar surface density .
5 . in the inner 3.1kpc
the dust temperature is correlated with the 3.6 flux .
this suggests the heating of the dust in the bulge is dominated by the general stellar population . beyond 3.1kpc
there is a weak correlation between dust temperature and the star formation rate .
we find no evidence for ` dark gas ' , using a similar technique as the _ planck _ team .
however , we find this technique may not be as effective for m31 due to poor angular resolution and line of sight effects .
we have used an alternative technique by constructing a gas - to - dust map after correcting for the radial gradient .
we do find regions with enhancements ( i.e. , higher values of @xmath124 ) , which may show places where ` dark gas ' exists , or may be due to local variations in the gas - to - dust ratio .
a detection of a potential component of co - free molecular gas will be possible with future observations to measure the c[ii ] line planned with _
herschel_. 7 . by minimising the scatter between our
radially corrected dust column - density and the column - density of gas inferred from the hi and co line we find a value for the x - factor of @xmath115^{-1}$ ] ( or expressed as @xmath116^{-1}$ ] ) .
our results of andromeda represent the largest resolved analysis of dust and gas in a single galaxy with _
the results of this analysis on m31 is strikingly different from those obtained by the planck team in the milky way , since we find no clear evidence for ` dark gas ' , a radial gradient in the gas - to - dust ratio and evidence for radial variation in the dust emissivity index ( @xmath1 ) . in future work
it will be important to understand these differences between the two big spirals in the local group .
we thank everyone involved with the _ herschel _ observatory .
pacs has been developed by a consortium of institutes led by mpe ( germany ) and including uvie ( austria ) ; ku leuven , csl , imec ( belgium ) ; cea , lam ( france ) ; mpia ( germany ) ; inaf- ifsi / oaa / oap / oat , lens , sissa ( italy ) ; iac ( spain ) .
this development has been supported by the funding agencies bmvit ( austria ) , esa - prodex ( belgium ) , cea / cnes ( france ) , dlr ( germany ) , asi / inaf ( italy ) , and cicyt / mcyt ( spain ) .
cvn , pr , dh , gr , yn and ke acknowledge support from the belgian federal science policy office via the prodex programme of esa .
spire has been developed by a consortium of institutes led by cardiff univ .
( uk ) and including : univ .
lethbridge ( canada ) ; naoc ( china ) ; cea , lam ( france ) ; ifsi , univ .
padua ( italy ) ; iac ( spain ) ; stockholm observatory ( sweden ) ; imperial college london , ral , ucl - mssl , ukatc , univ .
sussex ( uk ) ; and caltech , jpl , nhsc , univ .
colorado ( usa ) .
this development has been supported by national funding agencies : csa ( canada ) ; naoc ( china ) ; cea , cnes , cnrs ( france ) ; asi ( italy ) ; mcinn ( spain ) ; snsb ( sweden ) ; stfc , uksa ( uk ) ; and nasa ( usa ) . | we present an analysis of the dust and gas in andromeda , using _ herschel
_ images sampling the entire far - infrared peak .
we fit a modified - blackbody model to @xmath04000 quasi - independent pixels with spatial resolution of @xmath0140pc and find that a variable dust - emissivity index ( @xmath1 ) is required to fit the data .
we find no significant long - wavelength excess above this model suggesting there is no cold dust component .
we show that the gas - to - dust ratio varies radially , increasing from @xmath020 in the center to @xmath070 in the star - forming ring at 10kpc , consistent with the metallicity gradient . in the 10kpc ring
the average @xmath1 is @xmath01.9 , in good agreement with values determined for the milky way ( mw ) .
however , in contrast to the mw , we find significant radial variations in @xmath1 , which increases from 1.9 at 10kpc to @xmath02.5 at a radius of 3.1kpc and then decreases to 1.7 in the center .
the dust temperature is fairly constant in the 10kpc ring ( ranging from 1720k ) , but increases strongly in the bulge to @xmath030k .
within 3.1kpc we find the dust temperature is highly correlated with the 3.6 flux , suggesting the general stellar population in the bulge is the dominant source of dust heating there . at larger radii , there is a weak correlation between the star formation rate and dust temperature .
we find no evidence for ` dark gas ' in m31 in contrast to recent results for the mw .
finally , we obtained an estimate of the co x - factor by minimising the dispersion in the gas - to - dust ratio , obtaining a value of @xmath2@xmath3^{-1}$ ] . |
according to the recent announcements by atlas and cms experiments , a new particle ( @xmath1 ) could exist at around 750 gev by the observation of the diphoton invariant mass spectrum from the run - ii data in 13 tev @xcite .
this can typically be interpreted as the following 13 tev data in terms of the production cross section of @xmath1 and its branching ratio of two photons , @xmath2 also the atlas experiment group @xcite reported @xmath3 gev that is the best fit value of the total decay width of @xmath1 while the cms data indicate a smaller decay width @xcite .
it suggests that @xmath1 could be a scalar ( or pseudoscalar ) and further new fields with nonzero electrical charges are in favor of being introduced because @xmath1 must interact with such charged fields strongly . in order to give reasonable explanations or interpretations ,
a vast of paper along this line of issue has been recently arisen in ref .
@xcite .
radiative seesaw model could provide one of the elegant scenarios to include such a new field @xmath1 , since exotic fields with nonzero electric charges ( bosons or fermions ) are naturally introduced in order not only to explain the diphoton excess but also to explain the tiny neutrino masses such as zee - babu model @xcite .
moreover there often exist dark matter ( dm ) candidates , which also play a role in generating tiny neutrino masses . in our paper , we propose a new type of radiative neutrino model with global hidden @xmath0 symmetry , in which neutrino masses are induced at the four loop level , three singly charged bosons that dominantly play a role in explaining the diphoton excess , and there exist dm candidates with fermion or boson .
also we explain the discrepancy of the muon anomalous magnetic moment to the standard model ( sm ) sizably , by using the exotic charged fermions .
notice here that all these new fields play a role in generating the four loop neutrino masses .
diphoton excess is explained by introducing another isospin doublet boson with the same quantum numbers , i.e. two higgs doublet model ( 2hdm ) .
if we choose the cp - odd neutral boson as @xmath1 , it can not couple to the charged bosons but to the charged fermions . in this case ,
sizable excess can not be obtained since there do not exist enough contribution from charged fermions . on the other hand ,
if we choose the cp - even boson as @xmath1 , it can couple to both the charged bosons and charged fermions .
we find that charged fermion contribution is small in our setup and charged scalar bosons play a role to enhance diphoton branching ratio .
indeed we have to take care of the sm model results such as branching ratio , since the cp - even scalar must influence to these observables . here
we identify @xmath1 as a heavier cp - even neutral boson , which couples to the new charged bosons already introduced in our model . in sec .
ii , we show our model , including neutrino sector , muon anomalous magnetic moment , and the dm sector ( fermion and boson case ) . in sec .
iii , we discuss the diphoton excess . in sec .
iv , we have numerical analyses .
we conclude and discuss in sec .
.contents of fermion and scalar fields and their charge assignments under @xmath4 .
we introduce @xmath5 sets of charged scalar fields , i.e. @xmath6 . [ cols="^,^,^,^,^,^,^,^,^,^,^,^ " , ] in this section , we explain our model with a hidden @xmath0 symmetry .
the particle contents and their charges are shown in tab .
[ tab:1 ] .
we add vector - like exotic su(2 ) doublet charged fermion @xmath7 with @xmath8 hypercharge and singlet @xmath9 with @xmath10 hypercharge , majorana fermions @xmath11 , @xmath5 sets of three singly charged scalars @xmath12 , and @xmath13 ( @xmath14 ) with different quantum numbers , two neutral scalars @xmath15 and @xmath16 , and one additional higgs doublet to the sm .
we assume that only the two higgs doublets @xmath17 and @xmath16 have vacuum expectation values ( vevs ) , which are respectively symbolized by @xmath18 and @xmath19 .
the quantum number @xmath20 under the hidden @xmath0 symmetry is arbitrary , but its assignment for each field is unique to realize our four loop neutrino model . the @xmath21 symmetry assures the stability of the dm candidates , therefore the lightest majorana fermion @xmath22 for fermion field and/or the lightest isospin singlet of @xmath23 for boson field . here
we identify the first generation of @xmath11 or @xmath24 as a dark matter candidate respectively .
we also introduce an additional softly - broken @xmath25 symmetry where second higgs doublet @xmath26 and right - handed up - type quarks are assigned to parity odd under this symmetry .
thus the yukawa coupling for two higgs doublets with sm fermions is that of type - ii thdm .
the relevant lagrangian and higgs potential under these symmetries are given by @xmath27 \nonumber \\ & + \lambda_{\phi_1 k_{1}}^{ab } ( \phi_1^\dagger \phi_1)(k_{1}^{a+ } k_{1}^{b- } ) + \lambda_{\phi_2 k_{1}}^{ab } ( \phi_2^\dagger \phi_2)(k_{1}^{a+ } k_{1}^{b- } ) \nonumber \\ & + \lambda_{\phi_1 k_{2}}^{ab } ( \phi_1^\dagger \phi_1)(k_{2}^{a+ } k_{2}^{b- } ) + \lambda_{\phi_2 k_{2}}^{ab } ( \phi_2^\dagger \phi_2)(k_{2}^{a+ } k_{2}^{b- } ) \nonumber \\ & + \lambda_{\phi_1 s^+}^{ab } ( \phi_1^\dagger \phi_1)(s^{a+ } s^{b- } ) + \lambda_{\phi_2 s^+}^{ab } ( \phi_2^\dagger \phi_2)(s^{a+ } s^{b- } ) \nonumber \\ & + \lambda_{\phi_1 s } ( \phi_1^\dagger \phi_1)|s|^2 + \lambda_{\phi_2 s } ( \phi_2^\dagger \phi_2)|s|^2 + \lambda_{\phi_1 \varphi } ( \phi_1^\dagger \phi_1)|\varphi|^2 + \lambda_{\phi_2 \varphi } ( \phi_2^\dagger \phi_2)|\varphi|^2 + \cdots \label{eq : potential } , \end{aligned}\ ] ] where the flavor indices are abbreviate for brevity , and we omitted some quartic terms containing only @xmath28 which are irrelevant in our analysis .
after the global @xmath0 spontaneous breaking by @xmath29 , we obtain the majorana masses @xmath30 .
the first term of @xmath31 generates the sm charged - lepton masses @xmath32 after the spontaneous breaking of electroweak symmetry by @xmath33 .
we work on the basis where all the coefficients are real and positive for simplicity . in our analysis
, we assume @xmath34 is negligibly small so that mixing between @xmath16 and neutral components of the doublets are ignored .
then vevs and masses of higgs doublets are obtained same as type - ii 2hdm .
the isospin doublet scalar fields can be parameterized as @xmath35^t$ ] where @xmath36 gev is vev of the higgs doublet , and one component of @xmath37 and @xmath38 are respectively absorbed by the longitudinal component of @xmath39 and @xmath40 boson .
the isospin singlet scalar field can be parameterized by @xmath41 where mixing between @xmath42 and @xmath43 is negligible in our assumption .
then we focus on the cp - even higgs where mass eigenstates are @xmath44 where @xmath45 and @xmath46 denote sm higgs and heavier cp - even higgs respectively . in our analysis of @xmath46 production via gluon fusion
, we focus on the yukawa interactions of @xmath46 and top quark @xmath47 where @xmath48 as usual .
the @xmath49 and @xmath50 couplings are respectively proportional to @xmath51 and @xmath52 in the 2hdm @xcite . in this paper , we assume alignment limit @xcite , @xmath53 , to suppress @xmath54 decay channel . _
exotic charged fermion mass matrix _ : the singly exotic charged fermion mass matrix is given by @xmath55 where we define @xmath56^t$ ] , @xmath57 and assume @xmath58 .
the mass eigenstates @xmath59 and @xmath60 are defined by the bi - unitary transformation : @xmath61 where @xmath62 and @xmath63 .
the mass eigenvalues and the mixing angles @xmath64 are respectively given by @xmath65 notice here that the mass of the doubly charged fermion @xmath66 is given by @xmath67 . _ neutrino mass matrix _ : the leading contribution to the active neutrino masses @xmath68 is given at four - loop level as shown in figure [ fig : neut1 ] , and its formula is given as follows : @xmath69,\\ d&=\frac{a x_4(x_4+x_3 - 1)}{(x_3 - 1)^2-x_3}-b , \label{mnu1}\end{aligned}\]]where each of @xmath70 and @xmath71 is the mass of @xmath72 and @xmath24 , and satisfies @xmath73 , and we assume the coupling constants are same for different charged scalar sets so that @xmath74 is multiplied for simplicity . here
we define @xmath75 $ ] , @xmath68 should be @xmath76 from the neutrino oscillation data @xcite .
_ muon anomalous magnetic moment _ : the muon anomalous magnetic moment ( muon @xmath77 ) has been measured at brookhaven national laboratory that suggests there is a discrepancy between the experimental data and the prediction in the sm .
the difference @xmath78 is respectively calculated in ref .
@xcite and ref .
@xcite as @xmath79 the above results given in eq .
( [ dev12 ] ) correspond to @xmath80 and @xmath81 deviations , respectively .
our formula of muon @xmath77 is given by @xmath82 \nonumber \\
& \qquad \qquad \qquad \qquad + \frac{|y_s|_{22}}{2 } \sum_{\alpha,\beta = r , i } \left[c^2_{\theta_e } f(s_{\alpha},e_1^{- } ) + s^2_{\theta_e } f(s_{\alpha},e_2^{-})\right ] \biggr],\label{eq : g-2 } \\ & f(x , y)\approx \frac { 2 m_x^6 + 3 m_x^4 m_y^2 - 6 m_x^2 m_y^4 + m_y^6 + 6 m_x^4 m_y^2 \ln\left[\frac{m_y^2}{m_x^2}\right ] } { 12(m_x^2- m_y^2)^4 } \label{damu},\end{aligned}\ ] ] where we assume @xmath83 is same value for different charged scalar set .
it is worth mentioning the lepton flavor violating ( lfv ) processes that are always arisen in generating the muon anomalous magnetic moment . in our case ,
lfvs are generated from the terms proportional to @xmath84 and @xmath85 at the one - loop level , and these couplings or masses related to exotic fermions or bosons are constrained .
the stringent bound is given by the @xmath86 process with penguin diagram @xcite .
however once we can take @xmath84 and @xmath85 to be diagonal , such lfvs can simply be evaded .
even it is the case , the neutrino mixings are expected to be induced via the coupling of @xmath87 .
hence we retain the consistency of the lfv constraints without conflict of the neutrino oscillation data and the muon anomalous magnetic moment , applying this assumption to the numerical analysis . _
dark matter _ : + _ case 1 .
fermion dm _ + first of all , assuming the lightest majorana particle of @xmath11 as our dm candidate , which is denoted by @xmath88
, we find the dominant mode to explain the observed relic density @xmath89 @xcite .
the non - relativistic cross section comes from @xmath90 with @xmath91-channel , and its formula is given by @xmath92},\end{aligned}\ ] ] where @xmath93 is the decay rate of @xmath46 and its concrete formulae are found in ref .
notice here we neglect that the mixing between @xmath45 and @xmath42 so that @xmath88 does not interact with quark sector
. therefore , the spin independent scattering cross section vanishes at the tree level .
the measured relic density is obtained at around the pole of @xmath94 . in this case , @xmath91 should be directly integrated out from @xmath95 to infinity .
here we have followed the formula of refs .
@xcite and @xcite to get the relic density in our numerical analysis .
+ next , we consider the bosonic dm , assuming @xmath24 as the dm candidate .
we find three modes to explain the observed relic density @xmath89 @xcite .
our dominant non - relativistic cross sections come from @xmath96 with @xmath97-channels , @xmath98 with contact interaction , and @xmath99 with contact interaction and @xmath97-channels .
their formulae are given by @xmath100 v^4_{\rm rel},\\ \sigma v_{\rm rel}(2x\to 2h ) & \approx \frac{|\lambda_{h s}|^2}{64\pi m^2_x}\sqrt{1-\frac{m^2_h}{m_x^2 } } , \\
\sigma v_{\rm rel}(2x\to 2 g ) & \approx \frac{n^4 \mu'^2}{8 m_x^2 } \left [ \left(\frac2{v ' } - \frac{2\mu'}{m_r^2+m_x^2}\right)^2 + \frac{4 m_r^2 m_x^2 \mu ' ( m_r^2+m_x^2 -v ' \mu')}{(m_r^2 + m_x^2)^4v'}v_{\rm rel}^2 \right],\end{aligned}\ ] ] where @xmath101 is the combination of quartic coupling of @xmath102 .
the spin independent scattering cross section @xmath103 is also given by @xmath104,\end{aligned}\ ] ] where @xmath105 gev is the neutron mass , @xmath106 , @xmath107 is determined by the lattice simulation , and @xmath108 gev is the sm - like higgs .
the latest bound on the spin - independent scattering process was reported by the lux experiment as an upper limit on the spin - independent ( elastic ) dm - nucleon cross section , which is approximately @xmath109 @xmath110 ( when @xmath111 gev ) with the 90 % confidence level @xcite . in our numerical analysis
below , we set the allowed region for all the mass range of dm to be @xmath112
in this section , we discuss the production of heavier cp - even higgs @xmath46 and its decays at the lhc 13 tev . in our analysis we adopt the alignment limit @xmath113 to suppress @xmath54 partial decay width and to make lighter cp - even higgs sm - like .
the production of @xmath46 is given by gluon fusion process via top yukawa coupling .
the relevant effective interaction is given by @xcite @xmath114 in the alignment limit , where @xmath115 - i \pi ] ^2 $ ] with @xmath116 .
then the production cross section for @xmath117 gev is @xmath118 pb at @xmath119 tev @xcite .
the partial decay width for @xmath120 is obtained as @xmath121 the total decay width of @xmath46 is dominantly given by @xmath122 channel where @xmath123 @xcite .
we thus note that the width of @xmath46 tends to large for small @xmath124 and @xmath125 gev is given by @xmath126 .
the @xmath127 decay channel is induced through top quark loop and charged scalar loops . here
we note that contribution from charged higgs boson from doublets is small @xcite .
the contribution from exotic fermion is also small since interaction @xmath128 provide cancellation between @xmath129 and @xmath130 contributions .
thus we focus on the contribution from the loop diagram which contains singlet charged scalars .
the charged scalar loops are induced by the interactions @xmath131 which are obtained from eq .
( [ eq : potential ] ) . since @xmath132 branching ratio
is consistent with sm prediction , we require @xmath133 to suppress extra charged scalar contributions .
taking into account the alignment limit , we obtain the relevant interactions of @xmath46 and charged scalars such that @xmath134 where only diagonal terms contribute to @xmath127 process .
the partial decay width is then given by @xmath135 where @xmath136 ^ 2]$ ] and @xmath137 .
for simplicity , we apply same value for all @xmath138 in calculating the branching ratio .
the masses of the charged scalars are chosen as @xmath139 gev(@xmath140 ) in order to enhance the value of @xmath141 . in fig .
[ fig : diphoton ] , we show the parameter region in @xmath142 plane which provides products of @xmath46 production cross section and branching ratio for diphoton channel as @xmath143 fb .
thus we find that several sets of charged scalar bosons are required to obtain desired value of @xmath144 and @xmath145 is preferred .
( in unit of fb ) in @xmath142 plane . , title="fig:",width=302 ] ( in unit of fb ) in @xmath142 plane
. , title="fig:",width=302 ]
the fermion dm case in the upper figures has @xmath146 random sampling points , and the boson dm case in the lower figures has a million random sampling points where left(right ) side figures correspond to @xmath147 . , title="fig:",width=302 ] ) .
the fermion dm case in the upper figures has @xmath146 random sampling points , and the boson dm case in the lower figures has a million random sampling points where left(right ) side figures correspond to @xmath147 . , title="fig:",width=302 ] ) .
the fermion dm case in the upper figures has @xmath146 random sampling points , and the boson dm case in the lower figures has a million random sampling points where left(right ) side figures correspond to @xmath147 . , title="fig:",width=302 ] ) .
the fermion dm case in the upper figures has @xmath146 random sampling points , and the boson dm case in the lower figures has a million random sampling points where left(right ) side figures correspond to @xmath147 .
, title="fig:",width=302 ] in this section , we perform numerical analysis and show our model can explain neutrino mass , muon @xmath77 , relic density of dm and the diphoton excess simultaneously .
firstly , we set masses of all charged singlet scalar as @xmath148 gev , which is required to explain the diphoton excess .
now we randomly select values of the fifteen parameters within the corresponding ranges @xmath149\ \text{tev } , \quad \mu=\mu ' = v ' \in [ 0.5\ , , 1]\ \text{tev } , \quad m_\phi \in [ 2 m_x -0.1\text{gev } , 2 m_x + 0.1\text{gev}],{\nonumber}\\ & m_{e^{1/2}}=m_l = m_n = m_{r}=m_{i } \in [ 0.5\,\text { } , 1\,\text{tev } ] , { \nonumber}\\ & f = y_s \in [ 1,\sqrt{4\pi } ] , \quad s_{\theta_e } = g \in [ 0,1 ] , \quad \lambda_{h s } \in [ 0 , 0.01 ] , \label{range_scanning}\end{aligned}\ ] ] to reproduce the neutrino mass scale @xmath150 , the anomalous magnetic moment @xmath151 in eq .
( [ dev12 ] ) , the measured relic density @xmath152 in eq .
( [ eq : relicexp ] ) , and @xmath153 , where we universalily apply the parameter ranges to the fermion dm and boson dm case , and fix to be @xmath154 , @xmath155 , and @xmath156 for simplicity . rather small set of quartic coupling @xmath157 is to satisfy the constraint of the direct detection result .
then we show an example of each result for the fermion dm case and the boson dm case , as can be seen in fig .
[ fig : scan - fb ] , in which our solutions are represented as relic density and the muon anomalous magnetic moment within our range in eq.([range_scanning ] ) . here
the fermion case in the upper figures has @xmath146 random sampling points , and the boson case in the lower figure has a million random sampling points .
the left(right ) side figures correspond to @xmath147 where lower @xmath5 provides more solutions .
the fermion dm case has more solutions than the boson dm case in spite that the boson case are more free parameters , however the boson case tends to obtain larger value of the muon anomalous magnetic moment than the fermion case .
since the number of solutions determine the relic density , we could have more solutions for the boson case if some parameters related to the relic density @xmath158 and @xmath159 are fine - tuned .
we have proposed a new type of radiative neutrino model with global hidden @xmath0 symmetry , in which neutrino masses are induced at the four loop level , the discrepancy of the muon anomalous magnetic moment to the standard model ( sm ) is sizably obtained by using the exotic charged fermions , and both the fermion dm and boson dm candidate can satisfy the observed relic density without conflict of the direct detection searches .
diphoton excess has been investigated by adopting two higgs doublets where the yukawa coupling with sm fermions is same as type - ii 2hdm . here
we identify @xmath1 as a heavier cp - even neutral boson , which couples to the new charged bosons .
we find that several sets of charged scalar bosons are required to obtain desired value of @xmath144 and @xmath145 is preferred .
finally we have done the numerical analysis satisfying all these physical values or constraints , and shown allowed solutions in terms of the relic density and the muon anomalous magnetic moment as can be seen in fig .
[ fig : scan - fb ] .
h.o . thanks to prof .
shinya kanemura , prof .
seong chan park , dr .
kenji nishiwaki , dr .
yuta oriaksa , dr .
ryoutaro watanabe , and dr .
kei yagyu for fruitful discussions . h. o. is sincerely grateful for all the kias members , korean cordial persons , foods , culture , weather , and all the other things . the atlas collaboration , atlas - conf-2015 - 081 .
cms collaboration [ cms collaboration ] , collisions at 13tev , cms - pas - exo-15 - 004 .
k. harigaya and y. nomura , arxiv:1512.04850 [ hep - ph ] .
y. mambrini , g. arcadi and a. djouadi , arxiv:1512.04913 [ hep - ph ] .
m. backovic , a. mariotti and d. redigolo , arxiv:1512.04917 [ hep - ph ] .
a. angelescu , a. djouadi and g. moreau , arxiv:1512.04921 [ hep - ph ] .
y. nakai , r. sato and k. tobioka , arxiv:1512.04924 [ hep - ph ] .
s. knapen , t. melia , m. papucci and k. zurek , arxiv:1512.04928 [ hep - ph ] .
d. buttazzo , a. greljo and d. marzocca , arxiv:1512.04929 [ hep - ph ] .
a. pilaftsis , arxiv:1512.04931 [ hep - ph ] .
r. franceschini _ et al .
_ , arxiv:1512.04933 [ hep - ph ] . s. di chiara , l. marzola and m. raidal , arxiv:1512.04939 [ hep - ph ] .
t. higaki , k. s. jeong , n. kitajima and f. takahashi , arxiv:1512.05295 [ hep - ph ] .
s. d. mcdermott , p. meade and h. ramani , arxiv:1512.05326 [ hep - ph ] .
j. ellis , s. a. r. ellis , j. quevillon , v. sanz and t. you , arxiv:1512.05327 [ hep - ph ] . m. low , a. tesi and l. t. wang , arxiv:1512.05328 [ hep - ph ] .
b. bellazzini , r. franceschini , f. sala and j. serra , arxiv:1512.05330 [ hep - ph ] .
r. s. gupta , s. jager , y. kats , g. perez and e. stamou , arxiv:1512.05332 [ hep - ph ] . c. petersson and r. torre , arxiv:1512.05333 [ hep - ph ] .
e. molinaro , f. sannino and n. vignaroli , arxiv:1512.05334 [ hep - ph ] .
b. dutta , y. gao , t. ghosh , i. gogoladze and t. li , arxiv:1512.05439 [ hep - ph ] . q. h. cao , y. liu , k. p. xie , b. yan and d. m. zhang , arxiv:1512.05542 [ hep - ph ] .
s. matsuzaki and k. yamawaki , arxiv:1512.05564 [ hep - ph ] .
a. kobakhidze , f. wang , l. wu , j. m. yang and m. zhang , arxiv:1512.05585 [ hep - ph ] .
r. martinez , f. ochoa and c. f. sierra , arxiv:1512.05617 [ hep - ph ] .
p. cox , a. d. medina , t. s. ray and a. spray , arxiv:1512.05618 [ hep - ph ] .
d. becirevic , e. bertuzzo , o. sumensari and r. z. funchal , arxiv:1512.05623 [ hep - ph ] . j. m. no , v. sanz and j. setford , arxiv:1512.05700 [ hep - ph ] .
s. v. demidov and d. s. gorbunov , arxiv:1512.05723 [ hep - ph ] .
w. chao , r. huo and j. h. yu , arxiv:1512.05738 [ hep - ph ] .
s. fichet , g. von gersdorff and c. royon , arxiv:1512.05751 [ hep - ph ] .
d. curtin and c. b. verhaaren , arxiv:1512.05753 [ hep - ph ] .
l. bian , n. chen , d. liu and j. shu , arxiv:1512.05759 [ hep - ph ] .
j. chakrabortty , a. choudhury , p. ghosh , s. mondal and t. srivastava , arxiv:1512.05767 [ hep - ph ] .
a. ahmed , b. m. dillon , b. grzadkowski , j. f. gunion and y. jiang , arxiv:1512.05771 [ hep - ph ] .
agrawal , j. fan , b. heidenreich , m. reece and m. strassler , arxiv:1512.05775 [ hep - ph ] . c. csaki , j. hubisz and j. terning , arxiv:1512.05776 [ hep - ph ] .
a. falkowski , o. slone and t. volansky , arxiv:1512.05777 [ hep - ph ] .
d. aloni , k. blum , a. dery , a. efrati and y. nir , arxiv:1512.05778 [ hep - ph ] .
y. bai , j. berger and r. lu , arxiv:1512.05779 [ hep - ph ] .
e. gabrielli , k. kannike , b. mele , m. raidal , c. spethmann and h. veermae , arxiv:1512.05961 [ hep - ph ] .
r. benbrik , c. h. chen and t. nomura , arxiv:1512.06028 [ hep - ph ] .
j. s. kim , j. reuter , k. rolbiecki and r. r. de austri , arxiv:1512.06083 [ hep - ph ] .
a. alves , a. g. dias and k. sinha , arxiv:1512.06091 [ hep - ph ] .
e. megias , o. pujolas and m. quiros , arxiv:1512.06106 [ hep - ph ] . l. m. carpenter , r. colburn and j. goodman , arxiv:1512.06107 [ hep - ph ] .
j. bernon and c. smith , arxiv:1512.06113 [ hep - ph ] .
w. chao , arxiv:1512.06297 [ hep - ph ] .
m. t. arun and p. saha , arxiv:1512.06335 [ hep - ph ] . c. han , h. m. lee , m. park and v. sanz , arxiv:1512.06376 [ hep - ph ] .
s. chang , arxiv:1512.06426 [ hep - ph ] .
i. chakraborty and a. kundu , arxiv:1512.06508 [ hep - ph ] .
r. ding , l. huang , t. li and b. zhu , arxiv:1512.06560 [ hep - ph ] .
h. han , s. wang and s. zheng , arxiv:1512.06562 [ hep - ph ] .
x. f. han and l. wang , arxiv:1512.06587 [ hep - ph ] .
m. x. luo , k. wang , t. xu , l. zhang and g. zhu , arxiv:1512.06670 [ hep - ph ] . j. chang , k. cheung and c. t. lu , arxiv:1512.06671 [ hep - ph ] .
d. bardhan , d. bhatia , a. chakraborty , u. maitra , s. raychaudhuri and t. samui , arxiv:1512.06674 [ hep - ph ] .
t. f. feng , x. q. li , h. b. zhang and s. m. zhao , arxiv:1512.06696 [ hep - ph ] .
o. antipin , m. mojaza and f. sannino , arxiv:1512.06708 [ hep - ph ] . f. wang , l. wu , j. m. yang and m. zhang , arxiv:1512.06715 [ hep - ph ] . j. cao , c. han , l. shang , w. su , j. m. yang and y. zhang , arxiv:1512.06728 [ hep - ph ] .
f. p. huang , c. s. li , z. l. liu and y. wang , arxiv:1512.06732 [ hep - ph ] .
w. liao and h. q. zheng , arxiv:1512.06741 [ hep - ph ] .
j. j. heckman , arxiv:1512.06773 [ hep - ph ] .
m. dhuria and g. goswami , arxiv:1512.06782 [ hep - ph ] .
x. j. bi , q. f. xiang , p. f. yin and z. h. yu , arxiv:1512.06787 [ hep - ph ] .
j. s. kim , k. rolbiecki and r. r. de austri , arxiv:1512.06797 [ hep - ph ] .
l. berthier , j. m. cline , w. shepherd and m. trott , arxiv:1512.06799 [ hep - ph ] .
w. s. cho , d. kim , k. kong , s. h. lim , k. t. matchev , j. c. park and m. park , arxiv:1512.06824 [ hep - ph ]
. j. m. cline and z. liu , arxiv:1512.06827 [ hep - ph ] .
m. bauer and m. neubert , arxiv:1512.06828 [ hep - ph ] .
m. chala , m. duerr , f. kahlhoefer and k. schmidt - hoberg , arxiv:1512.06833 [ hep - ph ] .
d. barducci , a. goudelis , s. kulkarni and d. sengupta , arxiv:1512.06842 [ hep - ph ] . s. m. boucenna , s. morisi and a. vicente , arxiv:1512.06878 [ hep - ph ] . c. w. murphy , arxiv:1512.06976 [ hep - ph ] .
a. e. c. hernandez and i. nisandzic , arxiv:1512.07165 [ hep - ph ] .
u. k. dey , s. mohanty and g. tomar , arxiv:1512.07212 [ hep - ph ] .
g. m. pelaggi , a. strumia and e. vigiani , arxiv:1512.07225 [ hep - ph ] .
j. de blas , j. santiago and r. vega - morales , arxiv:1512.07229 [ hep - ph ] .
a. belyaev , g. cacciapaglia , h. cai , t. flacke , a. parolini and h. serodio , arxiv:1512.07242 [ hep - ph ] .
p. s. b. dev and d. teresi , arxiv:1512.07243 [ hep - ph ] .
w. c. huang , y. l. s. tsai and t. c. yuan , arxiv:1512.07268 [ hep - ph ] .
s. moretti and k. yagyu , arxiv:1512.07462 [ hep - ph ] .
k. m. patel and p. sharma , arxiv:1512.07468 [ hep - ph ] .
m. badziak , arxiv:1512.07497 [ hep - ph ] .
s. chakraborty , a. chakraborty and s. raychaudhuri , arxiv:1512.07527 [ hep - ph ] .
q. h. cao , s. l. chen and p. h. gu , arxiv:1512.07541 [ hep - ph ] .
w. altmannshofer , j. galloway , s. gori , a. l. kagan , a. martin and j. zupan , arxiv:1512.07616 [ hep - ph ] .
m. cvetic , j. halverson and p. langacker , arxiv:1512.07622 [ hep - ph ] .
j. gu and z. liu , arxiv:1512.07624 [ hep - ph ] .
b. c. allanach , p. s. b. dev , s. a. renner and k. sakurai , arxiv:1512.07645 [ hep - ph ] . h. davoudiasl and c. zhang , arxiv:1512.07672 [ hep - ph ]
. n. craig , p. draper , c. kilic and s. thomas , arxiv:1512.07733 [ hep - ph ] .
k. das and s. k. rai , arxiv:1512.07789 [ hep - ph ] .
k. cheung , p. ko , j. s. lee , j. park and p. y. tseng , arxiv:1512.07853 [ hep - ph ] .
j. liu , x. p. wang and w. xue , arxiv:1512.07885 [ hep - ph ] .
j. zhang and s. zhou , arxiv:1512.07889 [ hep - ph ] .
j. a. casas , j. r. espinosa and j. m. moreno , arxiv:1512.07895 [ hep - ph ] .
l. j. hall , k. harigaya and y. nomura , arxiv:1512.07904 [ hep - ph ] .
h. han , s. wang and s. zheng , arxiv:1512.07992 [ hep - ph ] .
j. c. park and s. c. park , arxiv:1512.08117 [ hep - ph ] .
a. salvio and a. mazumdar , arxiv:1512.08184 [ hep - ph ] .
d. chway , r. dermisek , t. h. jung and h. d. kim , arxiv:1512.08221 [ hep - ph ] .
g. li , y. n. mao , y. l. tang , c. zhang , y. zhou and s. h. zhu , arxiv:1512.08255 [ hep - ph ]
. m. son and a. urbano , arxiv:1512.08307 [ hep - ph ] .
y. l. tang and s. h. zhu , arxiv:1512.08323 [ hep - ph ] . h. an , c. cheung and y. zhang , arxiv:1512.08378 [ hep - ph ]
. j. cao , f. wang and y. zhang , arxiv:1512.08392 [ hep - ph ] .
f. wang , w. wang , l. wu , j. m. yang and m. zhang , arxiv:1512.08434 [ hep - ph ] . c. cai , z. h. yu and h. h. zhang , arxiv:1512.08440 [ hep - ph ] .
q. h. cao , y. liu , k. p. xie , b. yan and d. m. zhang , arxiv:1512.08441 [ hep - ph ] .
j. e. kim , arxiv:1512.08467 [ hep - ph ] .
j. gao , h. zhang and h. x. zhu , arxiv:1512.08478 [ hep - ph ]
. w. chao , arxiv:1512.08484 [ hep - ph ] .
x. j. bi _ et al .
_ , arxiv:1512.08497 [ hep - ph ] .
f. goertz , j. f. kamenik , a. katz and m. nardecchia , arxiv:1512.08500 [ hep - ph ] . l. a. anchordoqui , i. antoniadis , h. goldberg , x. huang , d. lust and t. r. taylor , arxiv:1512.08502 [ hep - ph ] .
s. b. dev , r. n. mohapatra and y. zhang , arxiv:1512.08507 [ hep - ph ] .
n. bizot , s. davidson , m. frigerio and j .- l .
kneur , arxiv:1512.08508 [ hep - ph ] .
l. e. ibanez and v. martin - lozano , arxiv:1512.08777 [ hep - ph ] . c. w. chiang , m. ibe and t. t. yanagida , arxiv:1512.08895 [ hep - ph ] .
s. k. kang and j. song , arxiv:1512.08963 [ hep - ph ] .
y. hamada , t. noumi , s. sun and g. shiu , arxiv:1512.08984 [ hep - ph ] .
x. j. huang , w. h. zhang and y. f. zhou , arxiv:1512.08992 [ hep - ph ] .
y. jiang , y. y. li and t. liu , arxiv:1512.09127 [ hep - ph ] .
k. kaneta , s. kang and h. s. lee , arxiv:1512.09129 [ hep - ph ] .
l. marzola , a. racioppi , m. raidal , f. r. urban and h. veermae , arxiv:1512.09136 [ hep - ph ] .
e. ma , arxiv:1512.09159 [ hep - ph ] .
a. dasgupta , m. mitra and d. borah , arxiv:1512.09202 [ hep - ph ] .
s. jung , j. song and y. w. yoon , arxiv:1601.00006 [ hep - ph ] .
c. t. potter , arxiv:1601.00240 [ hep - ph ] .
e. palti , arxiv:1601.00285 [ hep - ph ] . t. nomura and h. okada , arxiv:1601.00386 [ hep - ph ] .
x. f. han , l. wang , l. wu , j. m. yang and m. zhang , arxiv:1601.00534 [ hep - ph ] .
p. ko , y. omura and c. yu , arxiv:1601.00586 [ hep - ph ] .
k. ghorbani and h. ghorbani , arxiv:1601.00602 [ hep - ph ] .
d. palle , arxiv:1601.00618 [ physics.gen-ph ] .
u. danielsson , r. enberg , g. ingelman and t. mandal , arxiv:1601.00624 [ hep - ph ] .
w. chao , arxiv:1601.00633 [ hep - ph ] . c. csaki , j. hubisz , s. lombardo and j. terning , arxiv:1601.00638 [ hep - ph ] .
a. karozas , s. f. king , g. k. leontaris and a. k. meadowcroft , arxiv:1601.00640 [ hep - ph ] .
a. e. c. hernandez , i. d. m. varzielas and e. schumacher , arxiv:1601.00661 [ hep - ph ] .
t. modak , s. sadhukhan and r. srivastava , arxiv:1601.00836 [ hep - ph ] .
b. dutta , y. gao , t. ghosh , i. gogoladze , t. li , q. shafi and j. w. walker , arxiv:1601.00866 [ hep - ph ] . f. f. deppisch , c. hati , s. patra , p. pritimita and u. sarkar , arxiv:1601.00952 [ hep - ph ] . h. ito , t. moroi and y. takaesu , arxiv:1601.01144 [ hep - ph ] . h. zhang , arxiv:1601.01355 [ hep - ph ] .
a. berlin , arxiv:1601.01381 [ hep - ph ] .
s. bhattacharya , s. patra , n. sahoo and n. sahu , arxiv:1601.01569 [ hep - ph ] .
f. deramo , j. de vries and p. panci , arxiv:1601.01571 [ hep - ph ] .
i. sahin , arxiv:1601.01676 [ hep - ph ] .
s. fichet , g. von gersdorff and c. royon , arxiv:1601.01712 [ hep - ph ] .
d. borah , s. patra and s. sahoo , arxiv:1601.01828 [ hep - ph ] .
d. stolarski and r. vega - morales , arxiv:1601.02004 [ hep - ph ] .
m. fabbrichesi and a. urbano , arxiv:1601.02447 [ hep - ph ] . c. hati , arxiv:1601.02457 [ hep - ph ] .
ko and t. nomura , arxiv:1601.02490 [ hep - ph ] .
j. cao , l. shang , w. su , y. zhang and j. zhu , arxiv:1601.02570 [ hep - ph ] . j. h. yu , arxiv:1601.02609 [ hep - ph ] .
r. ding , z. l. han , y. liao and x. d. ma , arxiv:1601.02714 [ hep - ph ] .
s. alexander and l. smolin , arxiv:1601.03091 [ hep - ph ] .
j. h. davis , m. fairbairn , j. heal and p. tunney , arxiv:1601.03153 [ hep - ph ] .
i. dorsner , s. fajfer and n. kosnik , arxiv:1601.03267 [ hep - ph ] .
a. e. faraggi and j. rizos , arxiv:1601.03604 [ hep - ph ] .
a. djouadi , j. ellis , r. godbole and j. quevillon , arxiv:1601.03696 [ hep - ph ] .
f. jegerlehner and a. nyffeler , phys .
* 477 * , 1 ( 2009 ) [ arxiv:0902.3360 [ hep - ph ] ] .
m. benayoun , p. david , l. delbuono and f. jegerlehner , eur .
j. c * 72 * , 1848 ( 2012 ) [ arxiv:1106.1315 [ hep - ph ] ] .
j. adam _ et al . _
[ meg collaboration ] , phys .
lett . * 110 * , 201801 ( 2013 ) [ arxiv:1303.0754 [ hep - ex ] ] .
k. nishiwaki , h. okada and y. orikasa , phys .
d * 92 * , no .
9 , 093013 ( 2015 ) [ arxiv:1507.02412 [ hep - ph ] ] .
p. gondolo and g. gelmini , nucl .
b * 360 * , 145 ( 1991 ) .
d. s. akerib _ et al . _
[ lux collaboration ] , phys .
lett . * 112 * , 091303 ( 2014 ) [ arxiv:1310.8214 [ astro-ph.co ] ] .
a. djouadi , l. maiani , g. moreau , a. polosa , j. quevillon and v. riquer , eur .
j. c * 73 * , 2650 ( 2013 ) [ arxiv:1307.5205 [ hep - ph ] ] .
v. khachatryan _ et al . _
[ cms collaboration ] , jhep * 1410 * , 160 ( 2014 ) [ arxiv:1408.3316 [ hep - ex ] ] .
the numerical analyses on the higgs decays are performed using the program hdecay : a. djouadi , j. kalinowski and m. spira , comput .
commun . 108
( 1998 ) 56 ; a. djouadi , m. muhlleitner and m. spira , acta .
phys . polon .
b38 ( 2007 ) 635 . | we propose a new type of radiative neutrino model with global hidden @xmath0 symmetry , in which neutrino masses are induced at the four loop level , and we discuss the discrepancy of the muon anomalous magnetic moment to the standard model , and the fermion or boson dm candidate .
we also discuss diphoton excess within our model and show how to explain the sizable excess . |
the study of the edit distance in graphs initially appeared in a paper by axenovich , kzdy and the first author @xcite and , independently , by alon and stav @xcite .
it has several potential applications , such as to biological consensus trees @xcite and property testing problems in theoretical computer science @xcite .
more recently , interest has been shown in determining the value of the _ edit distance function _ , introduced in @xcite by balogh and the first author .
strategies for determining this function appear in @xcite , by marchant and thomason , as well as in @xcite . given a hereditary property ( that is , a set of graphs closed under vertex deletion and isomorphism ) , what is the least number of edge additions or deletions sufficient to make any graph on @xmath6 vertices a member of the property ? what is the behavior of this value as @xmath7 ? in @xcite , the binary chromatic number of a graph @xmath8 is used as a means of bounding the maximum number of edge additions and deletions ( edits ) sufficient to ensure that every @xmath6-vertex graph has no induced copy of a single graph @xmath8 .
a hereditary property which consists of the graphs with no induced copy of @xmath8 is denoted @xmath9 and is called a .
the methods in @xcite give an asymptotically exact result in some cases , most notably when @xmath8 is self - complementary .
let @xmath10 denote the random graph on @xmath6 vertices with edge probability @xmath1 . in @xcite , a version of szemerdi s regularity lemma
is applied to show that , as @xmath7 , the number of edits necessary to make @xmath10 a member of a given hereditary property approaches , with high probability , the maximum possible number over all @xmath6-vertex graphs within @xmath11 , so long as @xmath1 is chosen correctly with respect to the given hereditary property , @xmath0 .
in fact , the maximum number of edits sufficient to change a density-@xmath1 , @xmath6-vertex graph into a member of @xmath0 is asymptotically the same as the expected number of edits required to put @xmath10 into @xmath0 .
the edit distance function , @xmath12 , from @xcite , computes the limit of the maximum normalized edit distance of a density-@xmath1 , @xmath6-vertex graph from @xmath13 as @xmath7 for all probabilities @xmath1 .
see definition [ def : distfunc ] below .
not surprisingly , the maximum value of this function occurs at the same @xmath1 value described in @xcite .
marchant and thomason also explore the value of @xmath14 for various hereditary properties in @xcite , developing some insightful results for determining the value of the function in general .
one discovery from @xcite of particular interest is a relationship between the problem of determining the edit distance function for @xmath15 and constructions by brown in @xcite for @xmath16-free graphs , related to the zarankiewicz problem .
more generally , edit distance has been discussed as a potential metric for graph limits , and also as a parameter for property testing techniques to which graph limits have been applied ( see , for example , borgs , et al .
@xcite ) . in this paper , we explore what can be said about the edit distance function for the hereditary property @xmath2 , the set of all graphs that do not contain a complete bipartite graph with cocliques of @xmath17 and @xmath3 vertices as an induced subgraph .
in particular , we * compute the entire edit distance functions for the properties @xmath18 and @xmath19 . *
explore constructions that arise from strongly regular graphs and provide good upper bounds for @xmath20 .
one in particular , derived from the @xmath21-vertex generalized quadrangle @xmath22 , defines @xmath23 for @xmath24 . *
compute the edit distance function for the property @xmath2 for @xmath25 $ ] . *
show that , for odd @xmath3 , @xmath26 for @xmath27 $ ] .
these are the only known principal hereditary properties for which the maximum of the edit distance function is achieved on a nontrivial interval . *
examine the relationship between constructions by fredi @xcite related to the zarankiewicz problem and the trivial bound @xmath28 , which achieves the value of the function for small values of @xmath1 when @xmath5 .
when @xmath29 the constructions by fredi @xcite improve upon this bound for small values of @xmath1 .
* look at constructions derived from powers of cycles that give a general upper bound for some values of @xmath1 and @xmath3 .
* derive a lower bound for @xmath20 that is nontrivial and is achieved for some value of @xmath1 if a specified strongly regular graph exists . *
summarize the known bounds for @xmath20 for @xmath30 .
prior results and notation come primarily from @xcite , as well as previous work : @xcite , @xcite , @xcite , @xcite ; however , there are a number of other excellent resources on related topics . for a more extensive review of this literature
, the reader may wish to consult thomason @xcite .
we now introduce some important definitions and state our results more rigorously .
+ + we begin by recalling the definitions of graph edit distance and the edit distance function .
let @xmath31 and @xmath8 be simple graphs on the same labeled vertex set , and let @xmath13 be a hereditary property , then 1 .
@xmath32 is the edit distance from @xmath31 to @xmath8 , 2 .
@xmath33 is the edit distance from @xmath31 to @xmath13 and 3 .
@xmath34 is the maximum edit distance from the set of all @xmath6-vertex graphs to the hereditary property @xmath13 . since @xmath13 is by definition closed under isomorphism , vertex labels
may be ignored when considering @xmath35 .
in fact , @xmath35 could be defined equivalently as the minimum number of edge changes necessary to make @xmath31 a member of @xmath13 . the limit of the maximum edit distance from an @xmath6-vertex graph to a hereditary property @xmath13 normalized by the total number of potential edges in an @xmath6-vertex graph , @xmath36 is demonstrated in @xcite to exist and to be realized asymptotically with high probability by the random graph @xmath37 , where @xmath38 $ ] is a probability that depends on @xmath13 and is not necessarily unique .
[ def : distfunc ] the of a hereditary property @xmath13 is defined as follows : @xmath39 this function has also been denoted as @xmath40 in , for example , @xcite .
it was also proven in @xcite that , if @xmath41 denotes the expectation , then @xmath42/{\textstyle \binom{n}{2 } } .\ ] ] the limits above were proven to exist in @xcite , and furthermore , @xmath43 is both continuous and concave down . as a result , the edit distance function attains a maximum value that is equal to @xmath44 .
the point , or interval , at which @xmath45 is attained is denoted @xmath46 , and when it is evident from context , the subscript @xmath13 may be omitted from both . *
colored regularity graphs ( crgs ) * are the building blocks for determining the edit distance function for specific hereditary properties .
we leave the formal definition of crgs as well as basic facts about them to section [ sec : crgs ] for the reader who is unfamiliar with these objects .
a few of the new results given in section [ sec : results ] are structural in nature and so an understanding of crgs is useful in order to put the results in full context .
+ + [ sec : results ] in this paper , we prove the following results for the hereditary properties @xmath18 and @xmath19 .
the case of @xmath47 is mentioned in section 5.3 of @xcite .
[ mt1 ] let @xmath48 .
then @xmath49 with @xmath50 and @xmath51 .
[ mt2 ] let @xmath52 .
then @xmath53 with @xmath54 and @xmath55 . .
the point @xmath56 is indicated.,width=240 ] .
the point @xmath56 is indicated.,width=240 ] it should be noted that @xmath57 and @xmath58 from theorem [ mt1 ] could be found using alternative methods from previous literature as well .
in fact , they are a direct result of lemma 5.14 in @xcite , as is the value of @xmath43 for @xmath59 in both theorems .
the @xmath57 and @xmath58 values in theorem [ mt2 ] , however , are not so easily found .
the techniques used to prove both theorems also have the potential to yield some results for @xmath60 when @xmath61 and @xmath62 , as discussed in sections [ 6 ] and [ 7 ] . in @xcite
, it is established that for @xmath59 and @xmath63 , the edit distance function @xmath64 .
we extend this result to hold true for @xmath65 .
[ thm : bdext ] let @xmath66 , @xmath65 and @xmath63 , then @xmath64 .
this extension along with a new crg construction results in the determination of @xmath67 , the maximum value of the edit distance function , for all odd @xmath3 . using the general lower bound in theorem [ thm : genlb ] below ,
we also demonstrate , via theorem [ thm : oddt ] , that this maximum value occurs on a nondegenerate _
interval _ of values for @xmath1 .
that is , @xmath68 is not a single value for all odd @xmath62 . [
thm : genlb ] let @xmath69 and @xmath70 .
if @xmath71 is a black - vertex , @xmath1-core crg with white and gray edges such that the gray edges have neither a @xmath72 nor a @xmath73 ( as defined in lemma [ lem : forbsub ] ) , then @xmath74 .\ ] ] [ thm : oddt ] for odd @xmath62 and @xmath63 , @xmath75.\ ] ] for small @xmath1 and @xmath3 large enough , we demonstrate a similar result for zarankiewicz type constructions by fredi @xcite to that discovered in @xcite for @xmath15 and rejected for @xmath76 and @xmath77 . [ thm : tg9 ] for @xmath63 , the edit distance function @xmath78 for any prime power @xmath79 such that @xmath80 divides @xmath81 .
[ cor : tg9 ] for @xmath82 , there exists a value @xmath83 , so that if @xmath84 , then @xmath85 for some values of @xmath1 , which approach @xmath86 as @xmath79 increases .
that is , arbitrarily close to @xmath87 , there is some value for @xmath1 such that @xmath88 .
a strongly regular graph construction provides the upper bound @xmath89 for @xmath23 .
such constructions continue to be relevant for larger @xmath3 values .
[ thm : srgform ] for any @xmath90-strongly regular graph , there exists a corresponding crg , @xmath71 , such that @xmath91 if @xmath92 and @xmath93 , then @xmath71 forbids @xmath72 embedding , and when equality holds for both @xmath94 and @xmath95 , @xmath96 there is a very close connection between the result in theorem [ thm : genlb ] and strongly regular graphs .
if we take the expression in ( [ eq : srgfbound ] ) and minimize it with respect to @xmath97 ( see expression ( [ eqn : theor ] ) ) , then we obtain the expression on the right - hand side of ( [ eq : genlb ] ) . in particular
, we show that if a @xmath98-strongly regular graph exists , then the corresponding crg , @xmath71 , has @xmath99 tangent to the curve @xmath100 $ ] for some value of @xmath1 .
thus , we obtain the exact value of @xmath101 for that value of @xmath1 .
the following general upper bound arises from a crg construction involving the second power of cycles .
it is superseded by the strongly regular graph constructions for small values of @xmath3 , but not necessarily for large @xmath3 .
[ thm : cycc ] for @xmath63 , @xmath102 observe that this bound is better than that of the trivial @xmath103 for @xmath104 when @xmath105 .
below are known upper bounds for @xmath30 .
it should be noted that as our knowledge of existing strongly regular graphs increases , new upper bounds are also likely to be discovered .
[ thm : sum ] let @xmath63 . * if @xmath106 , then @xmath107 * if @xmath108 , then @xmath109 * if @xmath110 , then @xmath111 * if @xmath112 , then @xmath113 we compare these upper bounds to the lower bound in theorem [ thm : genlb ] via the figures in appendix [ sec : figs ] .
+ + [ sec : crgs ] to help describe how @xmath60 may be calculated , some definitions from @xcite are required .
a , @xmath71 , is a complete graph with vertices colored black or white , and with edges colored black , white or gray . at times
, it may be convenient to refer to the graph induced by edges of a particular color in a crg , @xmath71 .
we shall refer to these graphs as the black , white and gray subgraphs of @xmath71 .
the investigation of edit distance in @xcite and in @xcite uses a different paradigm with an analogous structure called a _
type_. essentially , our black , white and gray are their red , blue and green , respectively .
a from a ( simple ) graph @xmath8 to a colored regularity graph @xmath71 is a mapping @xmath114 , which satisfies the following : 1 .
if @xmath115 , then either @xmath116 and @xmath117 is colored black , or @xmath118 and the edge @xmath119 is colored black or gray .
2 . if @xmath120 , then either @xmath116 and @xmath117 is colored white , or @xmath118 and the edge @xmath119 is colored white or gray . basically , a colored homomorphism is a map from a simple graph to a crg so that black is only associated with adjacency , white is only associated with nonadjacency and gray is associated with adjacency , nonadjacency or both
. we will refer to a colored homomorphism from a simple graph @xmath8 to a crg @xmath71 as _ an embedding of @xmath8 in @xmath71 _ , and we denote the set of all crgs that only allow the embedding of simple graphs in a hereditary property @xmath13 as @xmath121 or merely @xmath122 when @xmath13 is clear from the context .
since any hereditary property may be described by a set of forbidden induced subgraphs , an equivalent description of @xmath121 is the set of all crgs that do not permit the embedding of any of the forbidden induced subgraphs associated with @xmath13 .
for instance , @xmath123 is the set of all crgs that do not admit @xmath124 embedding . in order to calculate @xmath60 , colored regularity graphs
are used in @xcite in order to define the following functions : @xmath125 \label{eqn : f } \\
g_k(p ) & = & \min\{\mathbf{u}^t m_k(p ) \mathbf{u}:\mathbf{u}^t\mathbf{1}=1\text { and } \mathbf{u}\geq0\}\label{eqn : g } .\end{aligned}\ ] ] here @xmath71 is a crg with @xmath126 vertices . @xmath127 and @xmath128 represent the sets of white vertices , black vertices , white edges and black edges in @xmath71 respectively .
@xmath129 is essentially a weighted adjacency matrix for @xmath71 with black vertices and edges receiving weight @xmath130 , white vertices and edges receiving weight @xmath1 and gray edges receiving weight @xmath86 . from @xcite
, it is known that @xmath131 .
moreover , alon and stav @xcite show that if @xmath132 is the binary chromatic number of @xmath13 , then @xmath133 .
marchant and thomason , demonstrate in @xcite that @xmath134 .
that is , given @xmath1 there exists at least one crg , @xmath135 , such that @xmath136 .
if we say a crg , @xmath71 , is a _ sub - crg _ of another crg , @xmath137 , when @xmath138 , @xmath139 , @xmath140 and @xmath141 , then it may be observed that @xmath142 . furthermore , as was noted in @xcite , if @xmath143 , then there is no need to consider both @xmath71 and @xmath137 when attempting to determine @xmath144 .
thus , a special subset of crgs is defined as follows .
a is a crg @xmath137 such that for no nontrivial sub - crg @xmath71 of @xmath137 is it the case that @xmath145 . in other words ,
if @xmath137 is a @xmath1-core crg , and @xmath71 is a nontrivial sub - crg of @xmath137 , then @xmath146 .
it can be shown ( see @xcite ) that a crg , @xmath71 , is @xmath1-core if and only if @xmath147 for a unique vector @xmath148 with positive entries summing to @xmath149 .
any crg , @xmath71 , that is not @xmath1-core contains at least one @xmath1-core sub - crg @xmath137 so that @xmath150 .
thus we could also say that @xmath151 that is , when looking for crgs to determine @xmath60 , the search may be limited to the important subset of crgs , @xmath1-cores .
this observation is especially helpful for determining lower bounds for the edit distance function . to prove the upper bounds for the edit distance function in this paper , we show that for each @xmath1 there exists a crg , @xmath135 , so that @xmath99 gives the bound for the value of @xmath60 . for the lower bounds , we employ so - called symmetrization , due to sidorenko @xcite , previously used for the computation of edit distance functions in @xcite and elsewhere such as marchant and thomason @xcite .
we also exploit features of the graphs @xmath72 and the concavity of the edit distance function to demonstrate that for no @xmath1-core crg , @xmath135 , can @xmath152 be less than the value in the theorem . by the continuity of the edit distance function ,
if we know the value of the function on an open interval , then we also know the value on its closure .
hence , for convenience , most of our proofs will only address the value of the function on the interior of a given interval .
we also note that , in @xcite , whenever the edit distance function of a hereditary property is computed , there is an attempt to determine all of the @xmath1-core crgs that achieve the value of the edit distance function . in this paper
we only concern ourselves with the value of the edit distance function itself and do not address the issue of multiple defining constructions .
+ + one reason for our interest in the edit distance function for forb@xmath153 is its relation to the zarankiewicz problem .
this problem addresses the question of how many edges a graph can have before it must contain a @xmath154 subgraph for fixed @xmath155 and @xmath3 . in an intriguing result from @xcite , a construction from brown @xcite for @xmath16-free graphs
is applied to construct an infinite set of new crgs that improve upon the previously known bounds for @xmath156 on certain intervals for arbitrarily small @xmath1 .
marchant and thomason @xcite establish that it is sufficient to consider only @xmath1-core crgs for which the gray subgraph has neither a @xmath157 nor a @xmath16 .
brown s constructions are not @xmath157-free , but a bipartite graph can be created from the construction that has no copy of @xmath16 .
similarly , for @xmath72 , it is sufficient to have no gray @xmath72 or book @xmath73 to forbid @xmath72 embedding , where the graph @xmath73 is a `` book '' as defined in @xcite and is defined precisely in lemma [ lem : forbsub ] .
although the brown constructions show that the edit distance function for @xmath15 is strictly less than @xmath158 for sufficiently small @xmath1 , known constructions for dense @xmath72-free graphs do not play a role in the computation of the edit distance function for @xmath18 or @xmath19 in the same way .
however , the edit distance function for @xmath19 is achieved over the interval @xmath159 $ ] by a construction formed from a strongly regular graph , namely a generalized quadrangle , often denoted @xmath160 , and similar results to those from the brown constructions do reemerge when @xmath29 and fredi s constructions from @xcite are considered .
+ + in section [ 2 ] , we discuss some results from @xcite and @xcite for the edit distance function and how they may be applied to the problem of determining the function for forb@xmath153 .
we then proceed to some general results and observations in section [ 3 ] that will be useful throughout the paper . sections [ 4 ] and [ 5 ] contain the proofs of our results for forb@xmath161 and forb@xmath162 , respectively .
section [ 6 ] addresses the proofs of theorems [ thm : bdext ] , [ thm : genlb ] and [ thm : oddt ] . in section
[ 7 ] , we present several new crg constructions that yield upper bounds for @xmath20 in general .
the remaining sections are reserved for conclusions and acknowledgements .
if a crg is @xmath1-core , one can say some interesting things about its overall structure . from @xcite , for instance , we have the following useful result .
[ edges ] if @xmath71 is a @xmath1-core crg , then all edges of @xmath71 are gray except * if @xmath70 , some edges joining two black vertices might be white or * if @xmath163 , some edges joining two white vertices might be black .
the crgs with all edges gray are useful in bounding the edit distance function , as we see in @xcite .
let @xmath164 denote the crg with @xmath165 white vertices , @xmath166 black vertices and only gray edges .
in particular : 1 .
let @xmath167 be the crg consisting of a white and black vertex joined by a gray edge .
2 . let @xmath168 be the crg consisting of @xmath80 black vertices all joined by gray edges .
theorem [ thm : k22 ] settled the case of @xmath47 , and theorem [ thm : trivbds ] permits us to focus on black - vertex crgs .
[ thm : k22 ] let @xmath169forb@xmath170 .
then @xmath171 with @xmath50 and @xmath51 .
[ thm : trivbds ] let @xmath63 , @xmath172
. then 1 . for @xmath173 , @xmath174 and 2 . for @xmath175 , either * @xmath176 or * @xmath177 where @xmath71 is a @xmath1-core crg with only black vertices and , consequently , no black edges .
the following lemma is about the structure of the @xmath1-core crgs described in the second part of theorem [ thm : trivbds ] .
it was originally observed in example 5.16 of @xcite .
the proof is a straightforward case analysis .
[ lem : forbsub ] a crg , @xmath71 , on all black vertices with only white and gray edges forbids @xmath72 embedding if and only if its gray subgraph contains no @xmath72 or @xmath178 as a subgraph , where @xmath178 is a book as described in @xcite .
that is , the graph @xmath178 is defined to be the graph consisting of @xmath179 triangles that all share a single common edge .
as demonstrated in @xcite , for a @xmath1-core crg , @xmath71 , there is a _ unique _ vector @xmath180 so that @xmath181 . for a @xmath1-core crg @xmath71 with optimal weight vector @xmath180 , the entry of @xmath180 corresponding to a vertex , @xmath182 ,
is denoted by @xmath183 .
this is the , and the function @xmath183 is the . with this in mind
, we have two propositions from @xcite , which follow easily from @xcite .
[ prop : dgformula ] let @xmath71 be a p - core crg with all vertices black .
then for any @xmath184 and optimal weighting @xmath180 , @xmath185 , where @xmath186 is the sum of the weights of the vertices adjacent to @xmath187 via a gray edge .
[ prop : xbound ] let @xmath71 be a p - core crg with all vertices black , then for @xmath188 $ ] and optimal weighting @xmath180 , @xmath189 because of theorem [ thm : trivbds ] , we may restrict our attention to those crgs , @xmath71 , for which @xmath190 . as a result , proposition [ prop : dgformula ] gives the lower bound @xmath191
. meanwhile , proposition [ prop : xbound ] restricts the optimal weights of all vertices in @xmath71 to be no more than @xmath1 .
these two restrictions are useful when attempting to prove lower bounds for @xmath192 .
we begin with some notation used throughout the paper .
let @xmath71 be a black - vertex , @xmath1-core crg with @xmath193 and optimal weight function @xmath180 : * @xmath194 , * @xmath195 is a fixed vertex in @xmath71 such that @xmath196 , for all @xmath184 , and @xmath197 is its weight , * @xmath198 and @xmath199 , * @xmath200 is a fixed vertex with maximum weight in @xmath201 , and @xmath202 , * @xmath203 is the set of all vertices in @xmath71 that are neither @xmath195 , nor contained in @xmath201 ; or equivalently , @xmath203 is the set of all vertices in the white neighborhood of @xmath195 , and * @xmath204 for some set @xmath205 . partitioning the vertices in a black - vertex , @xmath1-core crg that forbids a @xmath72 embedding into the three sets @xmath206 and @xmath203 as seen in figure [ partition ]
, illustrates some interesting features of its optimal weight function when the gray neighborhoods of these vertices are examined .
one such feature is the upper bounds in proposition [ starting prop ] for @xmath207 .
-core crg , @xmath71 .
dashed lines and gray background represent gray edges .
white edges are omitted , as are edges within subsets.,width=226 ] [ starting prop ] let @xmath208 $ ] be a black - vertex , @xmath1-core crg .
if either @xmath209 or both @xmath70 and the gray sub - crg of @xmath71 is triangle - free , then @xmath210 where @xmath197 is the maximum weight of a vertex in @xmath71 , and @xmath202 is the maximum weight of a vertex in that vertex s gray neighborhood .
the inequality @xmath211 follows directly from definitions of @xmath207 and @xmath212 , since @xmath212 is the greatest weight in @xmath71 . to justify the inequality @xmath213 , we break the problem into two cases : * case 1 : * @xmath195 and @xmath200 have no common gray neighbor . + recall that @xmath200 is a vertex with maximum weight in the gray neighborhood of @xmath195 , a vertex with maximum weight in all of @xmath71 , and assume that @xmath214 .
then applying proposition [ prop : dgformula ] and theorem [ thm : trivbds ] , @xmath215+\left[p+\frac{1 - 2p}{p}x_1\right ] = 2p+\left(\frac{1 - 2p}{p}\right)(x+x_1 ) > 2p+(1 - 2p ) .\ ] ] this is a contradiction because in case 1 , @xmath216 .
thus , @xmath217 , since the sum of the weights of the vertices in @xmath71 must be @xmath149 .
this completes the proof of proposition [ starting prop ] for @xmath218 , since , in this case , no @xmath135 contains a gray triangle .
so we may assume that @xmath219 .
* case 2 : * @xmath195 and @xmath200 have a common gray neighbor and @xmath209 . + in this case , @xmath200 has a single neighbor @xmath220 in @xmath201 because any more such neighbors would result in a gray book @xmath221 . furthermore , we note that in order to avoid a gray book @xmath221 , the common neighborhood of @xmath200 and @xmath220 in @xmath203 must be empty .
consequently , @xmath222 .
applying similar reasoning to that in case 1 , @xmath223+\left[p+\frac{1 - 2p}{p}x_1\right]+ \left[p+\frac{1 - 2p}{p}{{\mathbf{x}}}(u_2)\right ] .\ ] ] so , @xmath224+\left[p+\frac{1 - 2p}{p}x_1\right]+ \left[p+\frac{1 - 2p}{p}{{\mathbf{x}}}(u_2)\right ] \\ { { \mathbf{x}}}(u)+\left({{\mathbf{x}}}(w)+2x+x_1+{{\mathbf{x}}}(u_2)\right ) & \geq & 3p+\frac{1 - 2p}{p}\left(x+x_1+{{\mathbf{x}}}(u_2)\right ) \\ { { \mathbf{x}}}(u)+{{\mathbf{x}}}(w)+x & \geq & 3p+\frac{1 - 3p}{p}\left(x+x_1+{{\mathbf{x}}}(u_2)\right ) \\ 1 & \geq & 3p+\frac{1 - 3p}{p}\left(x+x_1+{{\mathbf{x}}}(u_2)\right ) .
\\\end{aligned}\ ] ] with @xmath209 and @xmath225 , we have a contradiction . applying the pigeon - hole principle
, we also have the following lower bound for @xmath226 : [ ph ] in a crg , if @xmath195 is a vertex with maximum weight , @xmath197 , the maximum weight in the gray neighborhood of @xmath195 is @xmath207 , and the order of the gray neighborhood of @xmath195 is @xmath226 , then @xmath227 .
while simple , when combined with propositions [ prop : dgformula ] and [ starting prop ] along with the observation that @xmath228 , this fact forces a delicate balance between the weights of the vertex @xmath195 , the vertices in @xmath201 , and the vertices in @xmath203 .
in this section , we establish the value of @xmath229 for @xmath230 , determining the entire function via continuity and theorem [ thm : trivbds ] , which gives that @xmath231 for @xmath232 $ ] . for the following discussion , we assume that @xmath71 is a @xmath1-core crg on all black vertices into which @xmath124 may not be embedded and that @xmath193 . the following lemma yields a useful restriction of the order of @xmath201 .
[ upperbounds for u ] let @xmath71 be a black - vertex , @xmath1-core crg with @xmath230 , no gray triangles , no gray @xmath124 and @xmath193 .
if @xmath195 is a vertex of maximum weight , @xmath212 , in @xmath71 , and @xmath233 , then @xmath234 let @xmath235 be an enumeration of the vertices in @xmath201 , the gray neighborhood of @xmath195 .
observe that @xmath71 can not contain a @xmath157 with all gray edges , and so @xmath201 contains no gray edges .
therefore , with the exception of @xmath195 , the entire gray neighborhood of each @xmath236 is contained in @xmath203 .
furthermore , if any three vertices in @xmath201 had a common gray neighbor in @xmath203 , then @xmath71 would contain a gray @xmath124 .
that is , each vertex in @xmath203 is adjacent to at most @xmath17 vertices in @xmath201 via a gray edge .
applying these observations , @xmath237 using proposition [ prop : dgformula ] and the assumption that @xmath238 , @xmath239 the fact that @xmath240 , gives @xmath241 the following technical lemma is an important tool in the proof of the theorem . [ k23 thrm ] let @xmath71 be a black - vertex , @xmath1-core crg for @xmath230 with no gray triangles , no gray @xmath124 and @xmath193 . if @xmath212 and @xmath207 are defined as in proposition [ starting prop ] , then @xmath242\left[\frac{1}{x_1}+\frac{1}{p(p - x)}\right ] \leq \frac{2(1-x)}{p - x } .\ ] ] by fact [ ph ] , @xmath243 , and by lemma [ upperbounds for u ] , @xmath244 .
therefore , @xmath245 after combining the @xmath246 terms we get , @xmath247&\leq&\frac{2(1-x)}{p - x},\end{aligned}\ ] ] and then applying proposition [ prop : dgformula ] , @xmath248\left[\frac{1}{x_1}+\frac{1}{p(p - x)}\right]&\leq&\frac{2(1-x)}{p - x}.\end{aligned}\ ] ] we now turn to the proof of the main theorem for this section .
let @xmath249 , and @xmath71 be a black - vertex , @xmath1-core crg with @xmath250 and no gray triangle ( i.e. , the book @xmath251 ) or gray @xmath124 . with the above assumptions
, we will show that there is no possible value for @xmath212 , the value of the largest vertex - weight .
to do so , we break the problem into 2 cases : @xmath252 and @xmath253 .
* case 1 : * @xmath254 .
+ we start with the inequality from lemma [ k23 thrm ] , @xmath248\left[\frac{1}{x_1}+\frac{1}{p(p - x)}\right]&\leq&\frac{2(1-x)}{p - x},\end{aligned}\ ] ] and apply the bound @xmath213 from proposition [ starting prop ] to get @xmath248\left[\frac{1}{p - x}+\frac{1}{p(p - x)}\right]&\leq&\frac{2(1-x)}{p - x}.\\\end{aligned}\ ] ] from proposition [ prop : xbound ] , @xmath255 , and so @xmath248\left[1+\frac{1}{p}\right ] & \leq & 2(1-x)\\ x\left(\frac{1-p}{p^2}\right ) & \leq & 1-p\\ x & \leq & p^2,\\\end{aligned}\ ] ] a contradiction , since @xmath256 for @xmath230 . * case 2 : * @xmath257 .
+ we again apply lemma [ k23 thrm ] , only now we employ the trivial bound @xmath211 from proposition [ starting prop ] : @xmath248\left[\frac{1}{x}+\frac{1}{p(p - x)}\right]&\leq&\frac{2(1-x)}{p - x}\\ \left[p+\frac{1 - 2p}{p}x\right]\left[p(p - x)+x\right]&\leq&2px(1-x)\\ ( 4p^2 - 3p+1)x^2-(3p^3)x+p^4&\leq & 0.\\\end{aligned}\ ] ] observe that @xmath258 is always positive , and therefore the parabola @xmath259 , in the variable @xmath212 , is concave up , so the range of @xmath212 values for which this inequality is satisfied is @xmath260 $ ] where @xmath261 if @xmath262 , then neither @xmath263 nor @xmath264 is real , and so the inequality is never satisfied . for @xmath265 , routine calculations show that @xmath266 , a contradiction to the assumption that @xmath253 . hence , there is no possible value for @xmath212 if @xmath267 , so the proof is complete .
this section addresses the case of @xmath23 .
+ + recall that from theorem [ thm : trivbds ] we already know that @xmath268 . for the remaining upper bound
, we turn to strongly regular graphs : a is a graph on @xmath126 vertices such that each vertex has degree @xmath97 , each pair of adjacent vertices has exactly @xmath94 common neighbors , and each pair of nonadjacent vertices has exactly @xmath95 common neighbors .
[ lem : srg ub ] let @xmath63 .
if there exists a @xmath90-strongly regular graph with @xmath92 and @xmath93 , then @xmath269 let @xmath31 be the aforementioned strongly regular graph .
we construct a crg , @xmath71 , on @xmath126 black vertices with gray edges in @xmath71 corresponding to adjacent vertices in @xmath31 and white edges in @xmath71 corresponding to nonadjacent vertices in @xmath31 .
no pair of adjacent vertices has @xmath270 common neighbors , so there is no book @xmath73 in the gray subgraph , and no pair of vertices has @xmath271 common neighbors , so there is no @xmath72 in the gray subgraph .
thus , by lemma [ lem : forbsub ] , @xmath272 . furthermore , @xmath273 = \frac{1}{k}+\frac{k - d-2}{k}p .\ ] ] in fact , there is a @xmath274-strongly regular graph @xcite .
it is a so - called `` generalized quadrangle , '' @xmath275 . as a result
, @xmath276 a list of known strongly regular graphs and their parameters has been compiled by andries brouwer @xcite .
their implications for @xmath192 in general are explored further in section [ 7.1 ] . + + because the edit distance function is both continuous and concave down , it is sufficient to verify that @xmath277 for @xmath278 and that @xmath279 for @xmath280 .
this is because the line determined by the bound @xmath281 passes through the points @xmath282 and @xmath283 .
furthermore , by theorem [ thm : trivbds ] , we need only consider crgs that have black vertices and white and gray edges .
lemmas [ p large ] and [ p small ] settle the cases where @xmath284 and where @xmath278 , respectively .
[ p large ] let @xmath284 . if @xmath71 is a black - vertex , @xmath1-core crg that does not contain a gray book @xmath221 or a gray @xmath285 , then @xmath286 , with equality occurring only if @xmath71 is a gray triangle ( i.e. , @xmath287 ) .
see appendix [ sec : lemproofs ] . before we prove lemma [ p small ] , there are two propositions that are necessary and used in several cases .
[ p small 1 ] let @xmath249 , and let @xmath71 be a black - vertex , @xmath1-core crg with no gray book @xmath221 and no gray @xmath285
. if @xmath288 , @xmath198 , @xmath289 and @xmath290 is the set of vertices in @xmath201 that are incident to a gray edge in @xmath201 , then @xmath291 see appendix [ sec : lemproofs ] .
+ [ p small 2 ] let @xmath249 , and let @xmath71 be a black - vertex , @xmath1-core crg with no gray book @xmath221 and no gray @xmath285 . if @xmath193 , then both @xmath292\geq\frac{1}{25 } .\ ] ] we begin with proposition [ p small 1 ] and then use @xmath293 from fact [ ph ] : @xmath294 \left[\frac{p - g}{p}+\frac{1-p}{p}x\right ] & \leq & 3x-3x^2 .
\label{eq : remark } \end{aligned}\ ] ] recall that @xmath295 because @xmath296 , so @xmath297 \left[p+\frac{1-p}{p}x\right ] & \leq & 3x-3x^2 \nonumber
\\ p^2-(1 + 3p)x+\frac{1 - 3p+5p^2}{p^2}x^2 & \leq & 0 . \nonumber\end{aligned}\ ] ] the quadratic formula gives that not only must the discriminant be nonnegative ( requiring @xmath298 ) , but also @xmath299 .\ ] ] some routine but tedious calculations demonstrate that , for @xmath300 , this expression is at least @xmath301 , achieving that value uniquely at @xmath302 .
[ p small ] let @xmath278 . if @xmath71 is a black - vertex , @xmath1-core crg that does not contain a gray book @xmath221 or a gray @xmath285 , then @xmath303 .
we assume that @xmath193 .
+ * case 1 : * @xmath304 . + according to proposition [ p small 1 ] , @xmath305 and since @xmath306 and @xmath307 , @xmath308 a contradiction .
so , @xmath309 .
+ * case 2 : * @xmath310 and @xmath311 .
+ using fact [ ph ] , and then proposition [ prop : dgformula ] @xmath312 a contradiction . + * case 3 : * @xmath310 and @xmath313 .
+ first we bound @xmath226 : @xmath314 so , @xmath315 . since @xmath226 is odd , @xmath316 . by proposition [ p small 1 ] , @xmath317 and applying proposition [ prop : dgformula ] , @xmath318 + 6x \\
\frac{1-p-12p^2}{p^2}x & \leq & 3-\frac{1 + 8p}{p}\cdot\frac{p - g}{p } \\ \frac{1-p-12p^2}{p^2}\left[\frac{p^2}{9p-1}\right ] & \leq & 3-\frac{1 + 8p}{p}\cdot p \\ \frac{(1 - 4p)(1 + 3p)}{9p-1 } & \leq & 2(1 - 4p )
\\ \frac{1 + 3p}{9p-1 } & \leq & 2 , \end{aligned}\ ] ] which implies @xmath319 , a contradiction .
+ * case 4 : * @xmath310 and @xmath320 .
+ now we compute a stronger bound on @xmath321 .
let @xmath200 and @xmath220 be vertices in @xmath201 that are adjacent via a gray edge , and let their weights be @xmath207 and @xmath322 , respectively .
then @xmath323 because @xmath200 and @xmath220 have no common gray neighbor other than @xmath195 and because they can have no additional gray neighbor in @xmath201 .
applying proposition [ prop : dgformula ] , @xmath324 and since @xmath325 , @xmath326 we can bound the number of vertices in @xmath327 by using the fact that @xmath328 .
returning to proposition [ p small 1 ] , @xmath329\left(\frac{p - g}{p}-x\right ) & \leq & 3 - 3x-\frac{1+p}{p}{{\mathbf{x}}}(u)+{{\mathbf{x}}}(u_1 ) \\ { { \mathbf{x}}}(u)\left(\frac{p - g}{px}-1+\frac{1+p}{p}\right)-3 + 3x & \leq & { { \mathbf{x}}}(u_1)\left(1+\frac{p - g}{px}-1\right)-\ell_1\left(\frac{p - g}{p}-x\right).\end{aligned}\ ] ] if @xmath330 , then @xmath331 .
of course , @xmath332 is lower - bounded by proposition [ prop : dgformula ] .
@xmath333\left(\frac{p - g}{px}+\frac{1}{p}\right)-3 + 3x & \leq & \frac{\ell_1}{2}(p - x)\left(\frac{p - g}{px}\right)-\ell_1\left(\frac{p - g}{p}-x\right ) \\ \left[\frac{p - g}{p}+\frac{1 - 2p}{p}x\right ] \left(\frac{p - g}{px}+\frac{1}{p}\right)-3 + 3x & \leq & \ell_1\left[x-\frac{p - g}{p}\cdot\frac{3x - p}{2x}\right ]
\\ \left[p+\frac{1 - 2p}{p}x\right]\left(\frac{p}{x}+\frac{1}{p}\right)-3 + 3x & \leq & \ell_1\left[x-\frac{p(3x - p)}{2x}\right ] \\ p^2-(1 + 2p)x+\frac{1 - 2p+3p^2}{p^2}x^2 & \leq & \ell_1\frac{(p - x)(p-2x)}{2 } .
\\\end{aligned}\ ] ] now , we bound @xmath334 , depending on the sign of @xmath335 , requiring two more cases .
+ * case 4a : * @xmath310 and @xmath320 and @xmath336 .
+ here we use the bound @xmath337 : @xmath338 by proposition [ p small 2 ] , we may restrict our attention to @xmath339 and so we may substitute the smallest possible value for @xmath212 , which still maintains the inequality . @xmath340 a contradiction . + * case 4b : * @xmath310 and @xmath320 and @xmath341 .
+ here we use the bound @xmath342 and then replace @xmath212 with @xmath343 , the value that minimizes the left - hand side : @xmath344 this , too , is a contradiction for @xmath278 , completing the proof of lemma [ p small ] .
in this section we extend the generally known interval for @xmath20 from @xmath232 $ ] to @xmath345 $ ] . with a new crg construction , this extension is sufficient to determine @xmath67 and a subset of @xmath68 for odd @xmath3 .
subsection [ 6.1 ] contains the proof of theorem [ thm : bdext ] , while the remaining subsections address theorems [ thm : genlb ] and [ thm : oddt ] .
let @xmath71 be a @xmath1-core crg for @xmath25 $ ] that does not permit @xmath72 embedding for @xmath105 . if we assume that @xmath346 , then by theorem [ thm : trivbds ] , @xmath71 has only black vertices and no black edges .
again , we partition the vertices of @xmath71 into three sets @xmath347 , @xmath348 and @xmath203 , where @xmath195 is a fixed vertex with maximum weight @xmath212 ; @xmath201 is the set of all vertices in the gray neighborhood of @xmath195 with @xmath200 a vertex of maximum weight @xmath207 in @xmath201 ; and @xmath203 is the set of all remaining vertices , or those vertices adjacent to @xmath195 via white edges . finally , let @xmath349 signify the sum of the weights of all vertices in the gray neighborhood of @xmath236 . then by lemma
[ lem : forbsub ] , the total weight of the vertices in @xmath203 is at least @xmath350 since no vertex in @xmath201 can be adjacent to more than @xmath351 other vertices in @xmath201 without forming a book @xmath73 gray subgraph with @xmath195 .
thus , @xmath352\leq1.\ ] ] applying proposition [ prop : xbound ] and letting @xmath353 , @xmath354x_1 & \leq & 1 \\ 2(p - g)-p+(1 - 2p)x & \leq & \left[p(t-1)-1\right]x_1 \\ 2(p - g)-p+(1 - 2p)x & \leq & \left[p(t-1)-1\right]x \\
p-2 g & \leq & \left[p(t+1)-2\right]x.\end{aligned}\ ] ] since @xmath355 and @xmath356 by proposition [ prop : xbound ] , @xmath357\frac{g}{1-p } \\
\frac{1-p}{t-1 } & \leq & g.\end{aligned}\ ] ] by theorem [ thm : trivbds ] , @xmath358 , for @xmath345 $ ] , so @xmath359 .
we will now show that this result is enough to determine the maximum value of @xmath20 for odd @xmath3 .
[ prop : t+1 ] let @xmath63 for odd @xmath3
. then @xmath360 .
let @xmath71 be the crg consisting of @xmath361 black vertices with white subgraph forming a perfect matching and all other edges gray .
the crg , @xmath71 , does not contain a gray @xmath72 or book @xmath73 , and so by lemma [ lem : forbsub ] , @xmath71 forbids a @xmath72 embedding .
the crg , @xmath71 , contains exactly @xmath362 white edges , so by equation ( [ eqn : f ] ) , @xmath363=\frac{1}{t+1}.\ ] ] therefore , @xmath364 .
since by theorem [ thm : bdext ] , @xmath365 , and by proposition [ prop : t+1 ] , @xmath366 , we have that @xmath367 for odd @xmath62 .
+ + we conclude this section by determining a general lower bound for the edit distance function of @xmath2 .
it is the lower bound from theorem [ eq : genlb ] , and it allows us to make the claim in theorem [ thm : oddt ] that , in the case of odd @xmath3 , there is a nondegenerate interval @xmath68 that achieves the maximum value of the function .
here we use the standard bounds from propositions [ prop : dgformula ] and [ prop : xbound ] .
let @xmath288 , where @xmath71 is a black - vertex , @xmath1-core crg , and let @xmath368 denote the gray neighborhood of a given vertex @xmath187 in @xmath71 .
then if @xmath235 are the vertices in the gray neighborhood , @xmath201 , of a fixed vertex of maximum weight , @xmath195 , @xmath369 \leq(t-1)(1-x - d_g(u_0)).\ ] ] the left - hand side of this inequality calculates the weight of the total gray neighborhood of each vertex in @xmath201 that must be contained in @xmath203 , the set of all vertices not in @xmath201 or @xmath195 . on the right - hand side
we make use of the facts that @xmath240 and that no vertex in @xmath203 may be adjacent to more than @xmath370 vertices in @xmath201 without violating lemma [ lem : forbsub ] by forming a gray @xmath72 with @xmath195 .
thus , applying proposition [ prop : dgformula ] , @xmath371 -\sum_{i=1}^{\ell}{{\bf x}}\left(n_g(u_i)\cap n_g(u_0)\right)\leq(t-1)(1-x - d_g(u_0)).\ ] ] again considering lemma [ lem : forbsub ] reveals that no vertex @xmath372 can have more than @xmath351 gray neighbors in @xmath201 without inducing a gray book @xmath73 with @xmath195 .
therefore , @xmath373+\frac{1 - 2p}{p}d_g(u_0)-(t-3)d_g(u_0 ) & \leq & ( t-1)(1-x - d_g(u_0))\\ \ell\left[\frac{p - g}{p}-x\right ] & \leq & ( t-1)(1-x)-\frac{1}{p}d_g(u_0 ) .\end{aligned}\ ] ] recalling that by proposition [ prop : xbound ] , @xmath374 , we use the pigeon - hole bound from fact [ ph ] @xmath375 to get @xmath376 & \leq & ( t-1)(1-x)-\frac{1}{p}d_g(u_0 ) \\
d_g(u_0)\left[\frac{p - g}{p}-x\right ] & \leq & ( t-1)x(1-x)-\frac{x}{p}d_g(u_0 ) \\
d_g(u_0)\left[\frac{p - g}{p}+\frac{1-p}{p}x\right ] & \leq & ( t-1)x(1-x ) .\end{aligned}\ ] ] by proposition [ prop : dgformula ] , @xmath377\left[\frac{p - g}{p}+\frac{1-p}{p}x\right ] \leq ( t-1)x(1-x ) .\ ] ] collecting terms yields , @xmath378x + \left[\left(\frac{1 - 2p}{p}\right)\left(\frac{1-p}{p}\right)+(t-1)\right]x^2\leq 0 , \ ] ] and so minimizing the left - hand side of the inequality with respect to @xmath212 , we have @xmath379 ^ 2}{4\left[\left(\frac{1 - 2p}{p}\right)\left(\frac{1-p}{p}\right)+(t-1)\right ] } & \leq & 0 \\ \left(\frac{p - g}{p}\right)^2\left(4t-5\right ) + 2\left(\frac{p - g}{p}\right)(t-1)\left(\frac{2 - 3p}{p}\right)-(t-1)^2 & \leq & 0 .\end{aligned}\ ] ] using the quadratic formula , @xmath380 \\
g & \geq & p-\frac{t-1}{4t-5}\left[3p-2 + 2\sqrt{1 - 3p+(t+1)p^2}\right ] .\end{aligned}\ ] ] the function in ( [ eq : genlb ] ) achieves its maximum at @xmath381 , and that maximum is , in fact , @xmath382 .
hence @xmath20 is at least @xmath382 at @xmath383 and is at least @xmath382 at @xmath384 . as a result of concavity , @xmath385 .\ ] ] equality holds whenever @xmath3 is odd because , in that case , proposition [ prop :
t+1 ] gives that @xmath386 , so @xmath68 must be an interval .
this concludes the proof of theorem [ thm : oddt ] .
if @xmath105 , then we can analyze the first and second derivatives , with respect to @xmath1 , of @xmath387\right ) .\ ] ] the maximum difference between @xmath388 and the lower bound in theorem [ thm : genlb ] on the interval @xmath389 $ ] is @xmath382 and occurs when @xmath383 .
we can also see that ( [ eq : psmalllb ] ) is bounded below by @xmath390 .
that we have been able to determine the entire edit distance function for @xmath18 and @xmath19 raises the question of whether it might be possible to do something similar for @xmath2 when @xmath62 .
that is , can we always find a few simple upper bounds that determine the entire edit distance function ? in this section we show that when @xmath62 , the number and types of known upper bounds for the function increases significantly , though this does not necessarily preclude the possibility that a few , yet to be discovered , crg constructions could determine the entire function . in section [ 7.1 ]
, we revisit the work in section [ 5.1 ] on strongly regular graphs . in section
[ 7.2 ] , we give some constructions inspired by the analysis of triangle - free graphs in @xcite . + + recall that a _ strongly regular graph _ with parameters @xmath90 is a @xmath97-regular graph on @xmath126 vertices such that each pair of adjacent vertices has @xmath94 common neighbors , and each pair of nonadjacent vertices has @xmath95 common neighbors . here
we develop a function based on the existence of a strongly regular graph .
suppose that @xmath71 is a crg with all vertices black and all edges white or gray that is derived from a @xmath90-strongly regular graph so that the edges of the strongly regular graph correspond to gray edges of @xmath71 .
in such a case we recall from section [ 5.1 ] that @xmath391 as is commonly known ( see @xcite , for instance ) , if a strongly regular graph with parameters @xmath90 exists then it is necessary , though not sufficient , for @xmath392 if we substitute @xmath393 and @xmath394 in this equation and then solve for @xmath126 , we find that @xmath395 and substituting these values into @xmath396 yields @xmath397 fixing @xmath1 and minimizing @xmath396 with respect to @xmath97 gives the following expression : @xmath398 which is equal to the lower bound from ( [ eq : genlb ] ) in theorem [ thm : genlb ] .
of course , in order to even have a chance of actually attaining ( [ eqn : theor ] ) with a strongly regular graph construction , both @xmath97 and @xmath399 must be integers .
this equation , however , provides something of a best case scenario for strongly regular graphs , and if there is a crg , @xmath71 , derived from a @xmath98-strongly regular graph that realizes equation ( [ eqn : theor ] ) , then @xmath99 is tangent to the lower bound in ( [ eq : genlb ] ) at @xmath400 determining the value of @xmath20 exactly .
the remaining upper bounds in theorem [ thm : sum ] are the result of checking constructions from the known strongly regular graphs listed at @xcite .
figure [ fig : chart ] ( see appendix a ) is a chart of the relevant parameters and @xmath99 functions for @xmath30 . to our knowledge , it is not known whether , for fixed @xmath3 , there are a finite or infinite number of @xmath98-strongly regular graphs .
see elzinga @xcite for values of @xmath94 and @xmath95 for which the number of strongly regular graphs with parameters @xmath90 is known to be finite or infinite .
there is an additional construction defining the upper bound for @xmath112 in theorem [ thm : cycc ] , described in the following section . for two vertices @xmath401 , where @xmath31 is a simple connected graph , let @xmath402 denote the length of the minimum path from @xmath212 to @xmath403 .
the @xmath404 power of @xmath31 , @xmath405 , is the graph with vertex set @xmath406 , and edge set @xmath407 .
let @xmath408 be the cycle on @xmath126 vertices raised to the @xmath409th power .
define @xmath410 to be the crg on @xmath126 black vertices with white edges corresponding to those in @xmath408 and gray edges corresponding to those in the complement of @xmath408 . recall that @xmath411 denotes the set of white edges for a given crg . then @xmath412 \\ & = & \frac{1}{k^2}[(1-p)k+2p(rk ) ] \\ & = & \left(\frac{2r-1}{k}\right)p+\frac{1}{k}.\end{aligned}\ ] ] @xmath413 forbids a @xmath72 embedding , and therefore @xmath414 .
first , we check that @xmath413 does not contain a gray @xmath415 . if @xmath200 and @xmath220 are any two vertices in @xmath416 , then @xmath417 .
this inequality is justified by observing that two vertices @xmath200 and @xmath220 that are neighbors in @xmath418 have the smallest possible number of total neighbors in @xmath416 , and this common neighborhood has order @xmath77 .
it then follows that @xmath419 in the complement of @xmath420 .
thus , @xmath413 does not contain a gray @xmath415 .
second , we check that @xmath413 does not contain a gray @xmath178 . if @xmath200 and @xmath220 are any two nonadjacent vertices in @xmath416 , then @xmath421 . therefore , by reasoning similar to above , @xmath422 in the complement of @xmath416 , implying @xmath413 does not contain a gray @xmath178 .
thus , by lemma [ lem : forbsub ] , @xmath413 forbids a @xmath72 embedding , and therefore @xmath423 .
while there are several other orders and powers of cycles that would also lead to a construction forbidding @xmath72 embedding , none of them have a corresponding @xmath99 value that beats the upper bound @xmath424 , so we restrict our interest to this one . for @xmath105 , @xmath425 is always an improvement on the bound @xmath426 from theorem [ thm : trivbds ] , though it is improved upon or made irrelevant by bounds from strongly regular graphs , for @xmath427 .
when @xmath428 , the function @xmath429 is tangent to the edit distance function at @xmath430 , where the edit distance function achieves its maximum value .
+ + as is observed in lemma [ lem : forbsub ] and used in the exploration of the past two constructions , graphs that forbid @xmath72 and @xmath73 as subgraphs are of interest when looking for crgs that forbid @xmath72 embedding .
the following results come from examining the bipartite versions of @xmath72-free graph constructions described by fredi @xcite .
this strategy mimics the one used in @xcite with brown s @xmath16-free construction .
we take the construction described in @xcite for a @xmath72-free graph @xmath31 on @xmath431 vertices , each with degree @xmath79 , where @xmath79 is a prime power so that @xmath80 divides @xmath81 .
we should note here that in the original construction from @xcite , loops were omitted , reducing the degree of some vertices to @xmath81 .
it is to our advantage , however , to leave the loops in so that the final construction will be @xmath79-regular . by the same proof as in @xcite , the graph with loops still retains the property that no two vertices have a common neighborhood greater than @xmath80 even when a looped vertex is considered to be in its own neighborhood .
next , we create a crg , @xmath71 , by taking two copies of the vertex set @xmath432 from the @xmath72-free graph with loops described above : @xmath433 , @xmath434 . color
all of these @xmath435 vertices black , and let @xmath436 with all edges not in @xmath437 white .
the gray subgraph of @xmath71 is bipartite , so it can not contain a @xmath73 , and since no two vertices @xmath438 and @xmath439 from the original construction have more than @xmath80 common neighbors , the common neighborhood of two vertices in the gray subgraph of @xmath71 is also at most @xmath80 .
thus by lemma [ lem : forbsub ] , @xmath71 forbids a @xmath72 embedding .
the crg , @xmath71 , has @xmath440 vertices and @xmath441 gray edges , so by equation ( [ eqn : f ] ) , @xmath99 is as described in the statement of theorem [ thm : tg9 ] .
though the property of being bipartite is sufficient to exclude a @xmath73 subgraph , using a bipartite @xmath72-free construction may not be the optimal choice .
a more efficient crg may be constructed from another graph that has a gray subgraph that is both @xmath72- and @xmath73-free , but , for instance , still contains triangles .
nevertheless , we can discover more about the potential for these constructions to improve upon the bounds for @xmath20 by fixing @xmath1 and considering the general formula in theorem [ thm : tg9 ] as a continuous function with respect to @xmath79 .
[ lem : q_0 ] let @xmath69 , and let @xmath442 be prime powers such that @xmath80 divides both @xmath443 and @xmath81 . if the crg , @xmath444 , is constructed according to the proof of theorem [ thm : tg9 ] with parameter @xmath83 and if the crg , @xmath71 , is constructed according to the proof of theorem [ thm : tg9 ] with parameter @xmath79 , then @xmath445 for @xmath446 .
we begin the proof by fixing @xmath1 and @xmath3 and analyzing @xmath447 .
note that @xmath448 , and @xmath449 .
consider when the derivative @xmath450 is positive and , therefore , @xmath451 is increasing .
since the greater value of @xmath79 that makes @xmath452 ( note that the leading term is nonnegative ) occurs at @xmath453 , it follows that @xmath454 when @xmath455 .
if @xmath209 and @xmath456 , then @xmath457 thus , @xmath454 for @xmath458 .
therefore @xmath459 for @xmath1 in this interval .
additionally , we can make some statements about when we can expect constructions that originate from the @xmath72-free graphs described by fredi @xcite to improve upon the bound @xmath388 for any @xmath79 .
[ qvalues9 ] fix @xmath82 , and let @xmath79 be a prime power such that @xmath80 divides @xmath81 .
let @xmath71 be the crg with parameter @xmath79 described in the proof of theorem [ thm : tg9 ] , hence @xmath460 .
then for any sufficiently large prime power @xmath79 and corresponding @xmath71 , there is an interval of values of @xmath1 on which @xmath461 .
moreover as @xmath462 the left - hand endpoints of these open intervals approach @xmath86 . that is , we can find an infinite sequence of crg constructions that improve upon the known bounds for @xmath2 when @xmath29 , and the intervals on which these improvements occur get arbitrarily close to @xmath86 .
we begin by observing that @xmath463 .
thus if @xmath461 , @xmath464 the minimum value of @xmath465 occurs when @xmath466 .
therefore , the inequality above is satisfied for some @xmath79 and @xmath1 values if and only if @xmath467
^ 2(q^2 - 1)-\left[\frac{(t-1)(q+2)}{4(q^2 - 1)}\right](t-1)(q+2)+(t-1 ) & < & 0 \\ ( t-1)\left(1-\frac{(t-1)(q+2)^2}{8(q^2 - 1)}\right ) & < & 0 .\end{aligned}\ ] ] that is , @xmath99 from the constructions in @xcite is less than @xmath388 for some value of @xmath1 if and only if @xmath468 . for positive @xmath79 ,
it is always the case that @xmath469 , and so any @xmath79 satisfying the constraints of the original construction will improve upon the upper bound established by @xmath388 for some @xmath1 when @xmath29 .
furthermore , for a fixed prime power @xmath79 for which @xmath80 divides @xmath81 , it is a definite improvement for some open neighborhood around @xmath466 .
this value approaches @xmath86 as @xmath462 , and there are an infinite number of prime powers @xmath79 such that @xmath80 divides @xmath81 ( see @xcite ) .
thus , it is the case that for arbitrarily small @xmath1 , we can find some @xmath79 such that @xmath461 .
lemma [ qvalues8 ] is an analysis of the fredi constructions when @xmath470 .
we then show that our bounds from these constructions do not have an effect on the value of @xmath20 for @xmath470 .
[ qvalues8 ] fix @xmath471 , and let @xmath79 be a prime power such that @xmath80 divides @xmath81 .
let @xmath71 be the crg with parameter @xmath79 described in the proof of theorem [ thm : tg9 ] , hence @xmath460 .
then @xmath472 returning to inequality ( [ eqn : fur ] ) and performing a similar analysis to that in the proof of lemma [ qvalues9 ] , we see that if @xmath473 , then @xmath474 for some value of @xmath1 if and only if @xmath475 the lower bound for @xmath79 described above is immaterial since for @xmath473 it is always negative .
the upper bound completes the proof of lemma [ qvalues8 ] . using lemma [ qvalues8 ] , we generated the following table of possible @xmath79 values that obey the inequality in ( [ eq : qbound ] ) .
since we have already determined the entire edit distance function for @xmath18 and @xmath19 , only @xmath476 needed to be considered : [ cols="^,^",options="header " , ] + @xmath477 .,width=240 ] .,width=235 ] .,width=235 ] .,width=240 ] .,width=235 ] .,width=240 ] .,width=235 ] .,width=235 ]
_ * proof of lemma [ p large ] : * _ we break this into two cases : when @xmath71 does and does not have a gray triangle . + * case 1 : * @xmath71 has a gray triangle . +
let the gray subgraph of @xmath71 contain a triangle whose vertices are @xmath478 , @xmath479 and @xmath480 with optimal weights @xmath481 and @xmath482 , respectively . because @xmath71 has no gray @xmath221 , we know that no pair of the vertices @xmath483 have a common gray neighbor other than the remaining vertex in the triangle .
letting @xmath288 , we have the following because the sum of the optimal weights on all vertices in @xmath71 is @xmath149 : consequently , @xmath487 with equality if and only if @xmath488 ; i.e. , @xmath71 itself is a gray triangle . +
* case 2 : * @xmath71 has no gray triangle .
+ let @xmath195 be a vertex of largest weight , @xmath197 , and let @xmath198 .
the absence of a gray triangle means that there are no gray edges between pairs of vertices in @xmath201 .
furthermore , no vertex in @xmath203 can be adjacent to more than three vertices in @xmath201 via a gray edge , since by lemma [ lem : forbsub ] , the gray subgraph of @xmath71 does not contain a @xmath285 .
let @xmath235 be an enumeration of the vertices in @xmath201 with weights @xmath489 , respectively , and @xmath288 .
then @xmath490 and applying proposition [ prop : dgformula ] to compute @xmath349 , @xmath491 if @xmath495 , then we may use the fact that @xmath496 , @xmath497 if @xmath498 , then we use proposition [ prop : xbound ] and substitute @xmath499 , @xmath500 regardless of the value of @xmath284 , if @xmath492 , then @xmath501 .
therefore , we may assume that @xmath502 .
@xmath504 and so bounding @xmath332 using proposition [ prop : dgformula ] , @xmath505 using proposition [ prop : xbound ] , @xmath506 yields @xmath507 and so if @xmath503 , then @xmath487 , with equality if and only if @xmath499 , and consequently , @xmath71 is a gray triangle .
so , we may further assume that @xmath508 .
third , suppose @xmath509 .
then @xmath510 returning to inequality ( [ eqn:3 ] ) , we have @xmath511x \\ p(1+p)-g(1 + 4p ) & \leq & -(1+p)(1 - 2p)\left(\frac{p - g}{5p-1}\right ) \\
p(1+p)(5p-1)+p(1+p)(1 - 2p ) & \leq & g\left[(1+p)(1 - 2p)+(1 + 4p)(5p-1)\right ] \\
\frac{1+p}{6 } & \leq & g \\ \frac{1-p}{3}+\frac{3p-1}{6 } & \leq & g .\end{aligned}\ ] ] if @xmath509 , then @xmath501 . if @xmath515 , then we use the fact that @xmath516 : @xmath517\left[\frac{p - g}{6p-1}\right ] \\ p(1 + 2p)-\frac{p(3p^2+p-1)}{6p-1 } & \leq & g\left[1 + 5p-\frac{3p^2+p-1}{6p-1}\right ] \\ \frac{1 + 3p}{9 } & \leq & g \\ \frac{1-p}{3}+\frac{2(3p-1)}{9 } & \leq & g .\end{aligned}\ ] ] if @xmath518 , then we use fact [ ph ] to bound @xmath356 , @xmath517\left[\frac{g}{1-p}\right ] \\ p(1 + 2p ) & \leq & g\left[1 + 5p+\frac{3p^2+p-1}{1-p}\right ] \\ \frac{(1-p)(1 + 2p)}{5 - 2p } & \leq & g \\ \frac{1-p}{3}+\frac{2(4p-1)(1-p)}{3(5 - 2p ) } & \leq & g .\end{aligned}\ ] ] regardless of the value of @xmath284 , if @xmath512 , then @xmath501 .
let @xmath235 be an enumeration of the vertices of @xmath201
. then @xmath519 and applying proposition [ prop : dgformula ] , @xmath520 simplification yields the first inequality . the second inequality results from observing that @xmath521 . | the edit distance between two graphs on the same vertex set is defined to be the size of the symmetric difference of their edge sets .
the edit distance function of a hereditary property , @xmath0 , is a function of @xmath1 , and measures , asymptotically , the furthest graph of edge density @xmath1 from @xmath0 under this metric . in this paper
, we address the hereditary property @xmath2 , the property of having no induced copy of the complete bipartite graph with 2 vertices in one class and @xmath3 in the other . employing an assortment of techniques and colored regularity graph constructions , we are able to determine the edit distance function over the entire domain @xmath4 $ ] when @xmath5 and extend the interval over which the edit distance function for @xmath2 is known for all values of @xmath3 , determining its maximum value for all odd @xmath3 .
we also prove that the function for odd @xmath3 has a nontrivial interval on which it achieves its maximum .
these are the only known principal hereditary properties for which this occurs . in the process of studying this class of functions ,
we encounter some surprising connections to extremal graph theory problems , such as strongly regular graphs and the problem of zarankiewicz . |
the internet of things ( iot ) and 5 g architectures require connecting heterogeneous devices including machine - to - machine ( m2 m ) and wireless sensor networking ( wsn ) units with potentially lower data rate but low latency , reliable , energy efficient and secure mechanisms @xcite .
these applications require short packets and simple modulation / demodulation mechanisms for the widespread utilization of iot .
the massive number of devices require not only low hardware complexity schemes but also methods of energy harvesting such as simultaneous wireless information and power transfer ( swipt ) transmitting data and energy by using radio frequency ( rf ) or magneto - inductive ( mi ) methods @xcite .
however , the existing literature generally concentrates on the energy versus capacity trade - off including power allocation schemes @xcite .
the recent studies analyze rf based swipt modulation methods including spatial domain ( sd ) and intensity based energy pattern changes @xcite .
rf solutions are not mature today to achieve swipt and require specialized circuits for rf to direct current ( dc ) conversion . in this article , novel and practical network topology modulation and demodulation architectures
are presented for mi communications ( mic ) networks by changing the spatial pattern of power transmitting active coils . the proposed mic based iot architecture ( mi - iot ) provides reliable , simple and secure mechanisms with low cost and low latency performances for connecting everyday objects with direct power transmission capability .
mic is an alternative method with the advantage of uniformity for varying environments without medium specific attenuation , multi - path and high propagation delay in challenging environments including underwater , underground and nanoscale medium with in - body and on - chip applications @xcite . in @xcite and @xcite , a trade - off analysis is presented for the problem of information and power transfer on a coupled - inductor circuit with power allocation policies . in @xcite , a nanoscale communication architecture with graphene coils is presented satisfying both power and data transmissions for in - body and on - chip applications . on the other hand , existing studies on mic networks
treat other coils as sources of interference including multiple - input multiple - output ( mimo ) and diversity architectures @xcite .
however , swipt architectures , mac schemes utilizing the same time - frequency resources , and modulation methods other than classical signal waveform approaches are not discussed . in this article , the information is embedded to coil positions by _ modulating the frequency selective mi channel _ or _ network topology _ instead of classical signal waveform modulation .
the proposed scheme fulfills the idea of fully coupled information and power transfer for widespread utilization of mi - iot applications @xcite .
it eliminates signal modulation and does not waste any power for data transmission .
furthermore , it does not require transmitter pre - coding , channel state information and intensity pattern modulation as in rf counterpart utilizing generalised pre - coding aided spatial modulation ( gpsm ) scheme @xcite .
moreover , mac sharing problem is intrinsically solved due to including whole network topology as data symbol .
the solution requires no receiver feedback and provides mac sharing without sacrificing resources for transmitter synchronization or interference prevention .
the contributions achieved , for the first time , are summarized as follows : * network topology modulation mechanism directly modulating the frequency selective mi channel . * practical topology demodulation combined with swipt . * topology modulating macs fully utilizing time - frequency bands without synchronization or contention . * reliable , capable of energy harvesting , low cost , low hardware complexity and low latency iot networking .
the proposed method supports the widespread adoption of mic networks for iot satisfying the following requirements : 1 . _
low latency _ : continuous energy and data transmission in a mac topology without the overhead of resource sharing , synchronization or receiver feedback .
high reliability _ : the robustness of the mic channel to fading and low probability of symbol error .
low hardware complexity _ : simple and low - cost swipt transceiver with already available mi circuits , and without separate structures for data and power transmission .
energy harvesting capability _ : intrinsically by mi coils without requiring separate rf to dc conversion circuits .
_ security _ : immunity to rf fields and radiative effects ; potential detection of intruder coils as the changes in network topology symbol and power transmission levels
. 6 . _ energy efficiency _ : no extra energy for signal waveform based data transmission ; the crowded set of mi coils forming a waveguide to enhance the communication range without consuming any active transmission power .
the remainder of the paper is organized as follows . in section [ sysmodel ] , mi network system model for mac topologies
is presented . in sections
[ stm ] and [ stdm ] , topology modulation and demodulation mechanisms are introduced , respectively .
then , in section [ numersim ] , the proposed methods are simulated for two - user mac network .
finally , in sections [ openis ] and [ conclusion ] , open issues and conclusions are discussed , respectively .
coils at each transceiver with varying topologies consisting of @xmath0 different spatial symbols.,title="fig:",width=158 ] coils at each transceiver with varying topologies consisting of @xmath0 different spatial symbols.,title="fig:",width=158 ] + ( a ) ( b ) + in this paper , there are @xmath1 transmitters denoted by @xmath2 for @xmath3 $ ] to transmit data to a single receiver , i.e. , indexed as the first user and denoted by @xmath4 , in a mac topology with random positions and orientations as shown in fig . [ fig12](a ) .
each user has a grid of closely spaced @xmath5 coils where each coil is either in open circuit condition without any effect to the network or actively resonating as shown in fig .
[ fig12](b ) .
the number of concurrently active coils in each transceiver with the ability to change its spatial coil distribution is denoted by @xmath6 .
the grid can be realized on a flexible paper suitable to attach to daily objects and to be recycled considering the billions of devices .
circuit theoretical equivalents of transceivers are presented in detail in @xcite .
each coil has series rlc type circuit with resistance @xmath7 , capacitance @xmath8 and inductance @xmath9 for @xmath10 $ ] where @xmath11 with resonance frequency @xmath12 .
furthermore , each coil has the same properties to simplify the analysis , i.e. , @xmath13 , @xmath14 and @xmath15 for @xmath10 $ ] .
the receiver loads denoted by @xmath16 for the coils of @xmath4 are set to @xmath17 without requiring any calibration with a trade - off in the received power performance .
transmitter and receiver diversity are provided by @xmath18 loaded coils in @xmath4 ( the remaining ones as passive coils ) and all transmitter coils with the same resonance voltage .
voltage sources and coil currents in the network denoted by the column vectors @xmath19 and @xmath20 , respectively , satisfy @xmath21 where @xmath22 is mutual inductance matrix , @xmath23 $ ] , @xmath24 is the operating frequency , @xmath25 denotes the voltage level in @xmath26th coil of @xmath27th user , @xmath28 $ ] , @xmath29 $ ] and @xmath30 denotes transpose .
@xmath31 has the elements @xmath32 for @xmath33 , @xmath34 for @xmath35 $ ] and @xmath36 for @xmath37 $ ] where
@xmath38 , @xmath39 is the complex unity and @xmath40 is the mutual inductance between @xmath27th and @xmath26th coil .
next , the network topology modulation is introduced .
the spatial grid topologies of each transmitter are changed to form varying mutual couplings .
a signal modulating voltage waveform is not required but only varying spatial patterns of actively resonating coils .
therefore , the network is continuously transmitting power to the receiver by also embedding information into the spatial structure that the energy is transmitted , however , without any signal modulation complexity or energy reservation for data .
the network topology modulation is mainly realized by changing @xmath41 denoted by @xmath42 which is easily calculated by using matrix identities as follows : @xmath43 where the elements @xmath44 of @xmath45 are equal to @xmath46 for @xmath33 and @xmath47 otherwise , diagonal @xmath48 includes eigenvalues , @xmath49 , @xmath50 , @xmath51 is the identity matrix , @xmath52 is the @xmath27th row of @xmath53 , @xmath54 for @xmath55 $ ] and zero otherwise , @xmath56 , @xmath57 and @xmath58 .
the proof is given in appendix [ proof1 ] .
it is observed in simulation studies that different network topologies provide different @xmath48 allowing to realize _
eigenvalue modulation_. eigenvalue modulation symbols @xmath59 are either pre - determined for each @xmath60 $ ] by using pilot training phase or they are estimated based on the preliminary knowledge of spatial modulation patterns and physical size of lan topology .
it is observed that @xmath48 is robust to angular disorientations while the detailed analysis is left as a future work . in the following discussions ,
four different modulation methods are proposed in terms of oscillation frequencies of each coil and time sharing .
[ 1]>m#1 .network topology modulation types and properties [ cols="^,^,^,^,^,^,^,^",options="header " , ] [ tab1 ] assume that each coil with the index @xmath61 in a grid transmits power at a different frequency @xmath62 for @xmath63 $ ] . furthermore , voltage level @xmath64 is set to unity with equal phase for all coils for simplicity .
then , the current at @xmath62 in @xmath27th coil of the receiver , i.e. , @xmath65 , is given as follows : @xmath66 where @xmath67 is assumed to be independent and identically distributed complex valued additive white gaussian noise ( awgn ) with variance of @xmath68 and @xmath69 for @xmath70 or @xmath71 .
moreover , if time - division multiple access ( tdma ) channel sharing is utilized , then the induced current at a time slot is due to some specific @xmath72 for @xmath3 $ ] and @xmath65 becomes equal to the following : @xmath73 where @xmath74 is obtained by deleting the rows and columns of @xmath75 with the indices not in @xmath76 $ ] .
if all the coils utilize the same frequency set in parallel , i.e. , @xmath77 \rbrace $ ] , where @xmath78 is the number of concurrent frequencies then , the transmitter diversity is realized and the received current is given as follows : @xmath79 if tdma is utilized then , @xmath80 due to @xmath72 is as follows : @xmath81 assume that a set of symbols consisting of @xmath0 different topologies for each transceiver is utilized .
the total number of different symbol combinations excluding @xmath4 is denoted by @xmath82 and @xmath27th network symbol is denoted by @xmath83 .
the possible set of modulation types indexed with @xmath84 for @xmath85 $ ] defined in ( [ eq5a ] ) , ( [ eq5 ] ) , ( [ eq6 ] ) and ( [ eq7 ] ) , respectively , and their properties are shown in table [ tab1 ] .
@xmath86 and @xmath87 have the constellation size of @xmath88 while it becomes @xmath89 for @xmath90 and @xmath91 .
the best time - frequency diversity is achieved with @xmath91 given in ( [ eq6 ] ) which is discussed in the following sections .
the computational complexity of the proposed modulation scheme includes the calculation of total transmit power at the sources and normalizing the power resulting in @xmath92 multiplications and @xmath93 additions where @xmath94 denotes the number of random frequency sets ( @xmath95 ) to include more diverse frequency effects .
next , demodulation mechanisms are discussed .
the demodulation is achieved by finding the spatial modulation matrix @xmath96 for different symbols and at different frequencies , and then comparing with the received current at varying frequencies .
furthermore , the symbols are defined for the whole network by solving mac sharing problem without requiring synchronization or time - frequency resource sharing .
if we assume that @xmath91 is used , the network topology symbol has the index @xmath97 for @xmath60 $ ] and @xmath24 is the power transmission frequency then , the measured noisy current denoted by @xmath98 at the @xmath27th coil receiver is found by combining ( [ eq3 ] ) and ( [ eq6 ] ) as follows : @xmath99 where @xmath100 is the @xmath27th row of @xmath53 in ( [ eq3 ] ) , @xmath101 , @xmath102 , @xmath103 and @xmath104 . the proof is given in appendix [ proof2 ] .
it is further simplified with @xmath105 , @xmath106 and by using perturbation equality @xmath107 where @xmath108 and the proof is given in appendix [ proof3 ] .
then , @xmath109 becomes as follows : @xmath110 then , inserting into ( [ demod1 ] ) and dropping @xmath111 , it becomes @xmath112 where @xmath113 , @xmath114 is the kronecker product and @xmath115 for @xmath116 $ ] is given by the following : @xmath117 where @xmath118 , @xmath119 , @xmath120 and @xmath121 . if it is assumed that eigenvalues are close to each other then , @xmath115 is approximated by dropping the @xmath122 terms due to the unitary magnitude property of @xmath123
then , the received current vector @xmath124^t$ ] is approximated by @xmath125 where @xmath126 $ ] , @xmath127^t$ ] is the complex receiver noise vector having @xmath128 spectral density for real and imaginary parts at each frequency , @xmath129 satisfying @xmath130 and @xmath131 equalizes the received or transmitted power for the symbol .
+ ( a ) + + ( b ) there are two different demodulation methods with respect to the amount of knowledge about @xmath132 for each symbol @xmath133 such that it is either perfectly known by pilot aided transmission or @xmath134 is estimated based on pre - calculation for the communication ranges without any information about @xmath135 .
it is observed in section [ numersim ] that the eigenvalues in @xmath59 have low variances for varying positions and orientations , and they are specific for each symbol @xmath133 .
demodulation schemes are realized by utilizing this uniqueness as shown in the following .
the receiver is assumed to have the full knowledge of @xmath131 , @xmath134 and @xmath136 for any @xmath97 with the aid of pilot symbols .
the modulation schemes are shown in fig . [ fig5](a ) where @xmath137 in pilot training case .
the demodulation mechanism is shown in fig . [
fig5](b ) . each symbol
@xmath133 is the vector @xmath138 with period @xmath139 , transmission delay @xmath140 and @xmath141 is the awgn component with independent and identically distributed elements having variance @xmath142 for @xmath143 $ ] .
the optimum estimation is achieved with maximum likelihood ( ml ) decision using the distance metric , i.e. , @xmath144 , and choosing the minimum based on decision regions @xmath145 for each @xmath133 @xcite .
the computational complexity includes finding the maximum of the correlations and the term @xmath146 for @xmath91 due to @xmath147 multiplications and additions in @xmath148 with @xmath149 .
transmit powers are adapted with @xmath131 to equalize total received power for each @xmath97 for simplicity .
then , with equally probable symbols , the probability of error per transfer is given by @xmath150 where @xmath151 is approximated by @xmath152 and @xmath153 is the @xmath26th element of @xmath154 .
the upper bound is given by @xmath155 where @xmath156 is the distance between @xmath83 and @xmath133 , and @xmath157 is the complementary error function .
+ it is assumed that there is no knowledge about the symbol set or @xmath135 .
an average @xmath158 is pre - calculated based on the observation of the small variance of @xmath59 and by considering the approximation in ( [ demoduoneuser3 ] ) without @xmath122 terms for the estimated range and orientations of the transmitters .
the proposed demodulation scheme is shown in fig .
[ fig6 ] . the transmit power for @xmath154
is equalized with @xmath159 instead of the receiver power since there is no feedback .
then , the estimation of @xmath154 is achieved by finding the index @xmath97 minimizing @xmath160 where @xmath161 , @xmath162 denotes parameter estimation regarding symbol @xmath133 when symbol @xmath154 is transmitted , @xmath163 and @xmath164 denotes moore - penrose pseudoinverse of @xmath158 .
then , @xmath165 equals to @xmath166 where @xmath167 , @xmath168 , @xmath169 and @xmath170 is the row vector of all ones .
@xmath171 is a generalized non - central chi - square distribution with the number of degrees of freedom of one where @xmath172 is complex gaussian random variable with the expectation @xmath173 and variance @xmath174 , and @xmath175 denotes euclidean norm . if the probability @xmath176 is denoted by @xmath177 then , the probability of error @xmath178 is bounded by @xmath179 .
the probability distribution of the difference of two correlated non - central chi - square random variables is complex to formulate explicitly .
in addition , random sets of energy transmission frequencies are utilized to improve detection performance by averaging the true decisions given for each set @xmath180 \rbrace$ ] for @xmath181 $ ] as shown in fig .
furthermore , the noise is smaller in each detection procedure where @xmath78 different frequencies are utilized instead of @xmath182 .
the computational complexity includes finding the minimums of @xmath94 different vectors of length @xmath82 , finding the maximum of a vector of size @xmath82 and the term @xmath183 for @xmath91 due to @xmath184 multiplications and @xmath185 additions .
+ ( a ) ( b ) the proposed system is simulated for two - user mac scheme as a proof of concept .
it is quite easy to extend to crowded iot networks .
the receiver is chosen to have a single coil to simplify the analysis .
the orientations of the transmitter coils are chosen in parallel with @xmath4 with equal distances to simplify the analysis as shown in fig . [
fig34](a ) .
the positions of the coils are represented in cartesian coordinate system as @xmath186^t$ ] and @xmath187^t$ ] .
their equal normal vector is @xmath188 $ ] .
@xmath189 and @xmath6 are set to three and five , respectively .
modulation topology for a single coil is assumed to be one of two schemes , i.e. , either the coils with the indices @xmath190 or @xmath191 are active at any time .
therefore , there is a total of three different network topology symbols as shown in fig . [ fig34](b ) .
single channel use transmits @xmath192 bits of total mac data .
@xmath82 can be easily increased , however , three symbols better clarify the proposed system .
[ 1]>m#1 @xmath193 & the coil radius and inter - coil distance in the grids & @xmath194 cm , @xmath194 mm + @xmath195 & the resistance , inductance and capacitance of a single coil unit & @xmath196 nh , @xmath197 - @xmath198 nf + @xmath199 & operating resonance frequency & @xmath194 mhz , @xmath200 mhz + @xmath201 & transmission frequency interval & @xmath202 - @xmath203 , @xmath204 - @xmath205 + @xmath94 , @xmath78 , @xmath206 & the number of random freq .
sets , freqs . in a set , and
the symbols & @xmath207 , @xmath208 , @xmath197 + @xmath209 & noise power spectral density & @xmath210 w / hz + @xmath211 & average trans .
power per symbol & @xmath212 mw + [ tab3 ] the simulation parameters are shown in table [ tab3 ] . the coil radius and the boundary distance between the coils
are set to @xmath213 cm and @xmath214 mm , respectively , so that a transceiver with a planar grid of coils has the area of @xmath215 @xmath216 to be easily attached to daily objects as shown in fig .
[ fig12](a ) .
the number of turns is one for simplicity .
@xmath17 is calculated by using @xmath217 for square cross - section copper wire of width and height @xmath218 mm compatible with @xcite and @xcite where the resistivity is @xmath219 , skin depth is @xmath220 and @xmath221 @xmath222 .
@xmath47 is equal to @xmath223 nh where @xmath224 @xcite .
the capacitance is @xmath225 nf or @xmath197 nf for @xmath226 mhz or @xmath200 mhz , respectively , with @xmath227 mhz to reduce parasitic effects @xcite .
noise spectral density is simulated for @xmath228 \times \mbox{n}_{th } $ ] to analyze worse cases than the thermal noise where @xmath229 w / hz , @xmath230 j / k and @xmath231 k @xcite .
total transmit power for @xmath82 symbols is set to @xmath194 mw in low power regime with an average @xmath232 ( mw ) for massive utilization in everyday objects .
mutual inductance calculation is based on the general model in @xcite and @xcite including position , orientation and size dependency .
the demodulation mechanisms with and without any training are simulated next . the case with pilot training
is simulated for @xmath233 mhz , @xmath234 , @xmath235 and @xmath137 due to perfectly known symbol family allowing to decrease @xmath94 .
the received power for each symbol is equalized to simplify the system and the analysis where unequal transmit powers are used for each symbol with the average @xmath236 mw and distributed equally among @xmath78 frequencies . for different symbols and ( b ) the upper bound on the symbol error probability for varying @xmath237 and noise levels.,title="fig:",width=268 ] + ( a ) + for different symbols and ( b ) the upper bound on the symbol error probability for varying @xmath237 and noise levels.,title="fig:",width=268 ] + ( b ) eigenvalues are plotted for varying distances reaching to @xmath238 50 meters as shown in fig . [ matlabfig12](a ) .
they are robust with respect to the distance which is utilized in the next section to demodulate by directly using the eigenvalues .
the upper bound on the probability of symbol error is shown in fig .
[ matlabfig12](b ) where it drops to @xmath239 at several meter distances including high - noise regimes allowing for reliable iot communications .
it is possible to realize a lan having the range of tens of meters with low cost , planar and small coils with the potential to tune system performance by changing system parameters . the small variations of eigenvalues shown in fig .
[ matlabfig12](a ) for varying distances are utilized to practically demodulate without any training or synchronization among the transmitters .
the case is simulated for @xmath240 mhz , @xmath241 and @xmath242 .
@xmath211 is distributed among @xmath243 frequencies equally while equalizing only the transmit power for each symbol without any receiver feedback . and noise levels.,width=307 ] the calculation of error probability
is realized through monte carlo simulations with @xmath244 trials due to complexity of obtaining explicit expressions .
therefore , the only errors higher than @xmath245 are visible in the simulation results .
it is observed in fig .
[ matlabfig3 ] that the proposed system achieves reliable communication in several meters without any pilot training or receiver feedback . in comparison with far field rf based swipt systems
, far field directive power beaming achieves several meters of ranges for indoor and outdoor applications achieving to transmit several mws of power @xcite .
however , high efficiency rectifiers require input powers between @xmath246 - @xmath247 mw with much higher transmission powers although the radiative rf range is not limited . on the other hand
, it is possible to achieve mi swipt with several mws transmission power with the possibility to tune and further improve the range performance by optimizing coil dimensions , the number of coils , constellations and operating frequency , and by utilizing passive waveguides or relays , e.g. , enabling ranges reaching kilometers with larger radius coils @xcite .
the problems of synchronization , channel access and resource sharing are solved by utilizing time - frequency resources together by all the transmitters and by only changing the network topology .
the following topics promise to further improve the system : * the computational complexity exponentially dependent on @xmath248 due to multi - user ml detection needs to be reduced with modern methods , e.g. , sphere decoding , by analyzing the structures of @xmath131 , @xmath134 , @xmath136 and @xmath164 . *
analysis and experimentation of performance for varying the following : types and numbers of constellations ; numbers , positions and orientations of coil grids ; coil and grid dimensions ; frequency sets , voltage sources and synchronization ; passive relaying and waveguides . *
the effects of intruder coils due to change in the network topology need to be analyzed in more detail and a protocol needs to be developed to detect the changes in the received data and the consumed powers in the coils .
in this paper , novel modulation and demodulation mechanisms are proposed with energy efficient , low cost and low hardware complexity mechanisms for swipt in a wireless mi mac network .
the network topology of power transmitting coils is changed to modulate information symbols and practical demodulation schemes are presented . the designed scheme is numerically simulated for two - user mac topology with small size coils suitable to attach to daily objects .
the spatial modulation method provides reliable lan performances with tens of meters of ranges and low power operation of mw level transmit powers .
swipt obtained with mic networks is future promising for iot applications and macs with energy efficient , simple and low cost solutions .
[ [ proof1 ] ] @xmath249 is equal to @xmath250 .
it is converted to @xmath251 where @xmath252 .
then , @xmath253 is given by @xmath254 .
furthermore , @xmath255 is simplified as @xmath256 .
the inverse @xmath257 is found by inserting the simplification of @xmath255 as follows : @xmath258 where @xmath259 and @xmath52 is the @xmath27th row of @xmath53 .
the simplification while passing from @xmath260 to @xmath261 is realized by using the equality @xmath262 which can be obtained by using simple matrix calculations .
[ [ proof2 ] ] the expression @xmath263 to obtain the received current is simplified by using ( [ app_eq1 ] ) as follows : @xmath264 where it is assumed that @xmath265 symbol is transmitted , @xmath266 is the oscillating voltage for the @xmath267 coil at @xmath24 for @xmath268 $ ] and zero otherwise , the equality @xmath260 is obtained by using @xmath269 , and @xmath270 is utilized in the equality @xmath261 .
then , the current at @xmath271 receiver coil is found by replacing @xmath53 with @xmath272 .
[ [ proof3 ] ] @xmath273 is calculated by using ( [ app_eq2 ] ) as follows : @xmath274 where @xmath275 . if @xmath276 , then the perturbation equality , i.e. , @xmath107 with @xmath108 , is utilized where @xmath277 , @xmath278 and @xmath279 .
then , ( [ demoduoneuser2 ] ) is easily obtained .
this work is supported by vestel electronics inc . ,
manisa , 45030 turkey .
b. gulbahar and o. b. akan , `` a communication theoretical modeling and analysis of underwater magneto - inductive wireless channels , '' _ ieee trans . on wireless communications _ ,
11 , no . 9 , pp . 33263334 , 2012 .
s. li , y. sun and w. shi , `` capacity of magnetic - induction mimo communication for wireless underground sensor networks , '' _ international journal of distributed sensor networks _ , article i d 426324 , 2015
. b. gulbahar , `` theoretical analysis of magneto - inductive thz wireless communications and power transfer with multi - layer graphene nano - coils , '' to appear in _ ieee transactions on molecular , biological , and multi - scale communications _ , pp . 11 , 2017 .
h. nguyen , j. i. agbinya and j. devlin , `` fpga - based implementation of multiple modes in near field inductive communication using frequency splitting and mimo configuration , '' _ ieee transactions on circuits and systems i _ , vol .
62 , no . 1 ,
pp . 302310 , 2015 .
s. babic et al . , `` mutual inductance calculation between circular filaments arbitrarily positioned in space : alternative to grover s formula , '' _ ieee transactions on magnetics _ , vol .
46 , no . 9 , pp . 3591 - 3600 , 2010 | internet - of - things ( iot ) architectures connecting a massive number of heterogeneous devices need energy efficient , low hardware complexity , low cost , simple and secure mechanisms to realize communication among devices .
one of the emerging schemes is to realize simultaneous wireless information and power transfer ( swipt ) in an energy harvesting network .
radio frequency ( rf ) solutions require special hardware and modulation methods for rf to direct current ( dc ) conversion and optimized operation to achieve swipt which are currently in an immature phase . on the other hand , magneto - inductive ( mi )
communication transceivers are intrinsically energy harvesting with potential for swipt in an efficient manner . in this article , novel
modulation and demodulation mechanisms are presented in a combined framework with multiple - access channel ( mac ) communication and wireless power transmission .
the network topology of power transmitting active coils in a transceiver composed of a grid of coils is changed as a novel method to transmit information .
practical demodulation schemes are formulated and numerically simulated for two - user mac topology of small size coils .
the transceivers are suitable to attach to everyday objects to realize reliable local area network ( lan ) communication performances with tens of meters communication ranges .
the designed scheme is promising for future iot applications requiring swipt with energy efficient , low cost , low power and low hardware complexity solutions .
ps .
@headingsoddheadevenheadoddfootevenfoot simultaneous wireless information and power transfer , magneto - inductive communication , network topology modulation , internet - of - things |
while radial velocity surveys and other techniques have yielded over 200 extrasolar planets , there are to date no convincing images of an extrasolar planet around a star .
a few candidate `` planetary mass objects '' have been identified co - moving with pre - main sequence stars and young brown dwarfs ( e.g. gq lup , oph 162225 - 240515 , etc . ; e.g. * ? ? ?
* ; * ? ? ?
however further observations have shown these objects to be higher mass ( e.g. * ? ? ?
a possible exception is the companion to 2m1207a ( * ? ? ?
* j1207334 - 393254 = twa 26 ) , for which mass estimates have ranged between @xmath03 - 8 @xmath9 . although its mass inferred from evolutionary models is well below the deuterium
burning limit , it is unlikely that the object could have formed through classical core accretion in a circumstellar disk . given its estimated mass ratio ( @xmath0 0.20.3 ) , and separation ( @xmath0 50 au ) it is thought to have formed via gravitational fragmentation , similar to a binary star system @xcite .
the properties of the 2m1207 system are discussed in detail in @xcite .
the most striking result from that study is that while both the colors and near - ir spectrum of 2m1207b are consistent with models of an unreddened @xmath01600k object , the inferred luminosity for the object is @xmath02.5 mag ( factor of @xmath010 ) below that expected for a @xmath05 - 10 myr object at all wavelengths . even more remarkable , the object is slightly fainter than the observed sequence of k and l absolute magnitudes for older field objects with @xmath10 @xmath11 1600k @xcite .
@xcite rule out a handful of simple resolutions to explain the apparent underluminosity of 2m1207b , and ultimately settle for an unlikely but testable hypothesis : that the object is obscured by an edge - on disk of large circumstellar dust grains producing 2.5@xmath6 of gray extinction . in this contribution
, we propose an alternative explanation for the low luminosity of 2m1207b : namely that it has a small radius . in this scenario 2m1207b
is the hot , long - lived afterglow of a recent collision between two protoplanets ( cf .
the observational consequences of protoplanet collisions have been discussed elsewhere @xcite . here
we review the problem of 2m1207b , introduce a new hypothesis , discuss its merits and deficiencies , and offer an observational test that can rule it out .
we first review some observational properties of the 2m1207 system .
2m1207a is classified as m8 with emission
line activity characteristic of t tauri stars @xcite .
the system appears to harbor a circumstellar accretion disk as evidenced by broad h@xmath12 emission @xcite , mid - ir excess @xcite , and outflow activity @xcite .
the motion of 2m1207a is consistent with membership in the tw hya association ( twa ) , a loose group of @xmath020 stars with mean age @xmath08-myr - old situated at a mean distance of @xmath050pc @xcite .
the optical / near - ir colors of 2m1207a are consistent with no reddening given its spectral type .
comparison of its position in the h
r diagram with theoretical models suggests a mass of @xmath025@xmath9 and age consistent with its membership in the twa @xcite .
2m1207b shares common proper motion and parallax with 2m1207a within the astrometric uncertainties ( milliarcsecond - level measurements over a few years ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the idea that b could be a foreground or background field l dwarf has been ruled out . at a distance
compared to ( * ? ? ?
* pc ) due to the effects of a improved proper motion for 2m1207 @xcite , and a revised estimate of the twa group velocity ( 22.4@xmath21 kms@xmath13 ; which follows mamajek 2005 , but omits the deviant parallax for twa 9 ) .
the updated velocity increases the distances to the other twa members in @xcite by 7% . ] of 66@xmath25 pc , the companion is at a projected separation of 51@xmath24 au from the primary .
@xcite concentrate their analysis of the temperature for 2m1207b on comparison of data to the lyon group model atmospheres . from the available h and k band spectra
, they determine the best fit model atmosphere has @xmath10 = 1600 @xmath2 100 k from the dusty grid of @xcite .
a range of surface gravity from log(@xmath14 ) = [ 3.54.5 ] was explored , but the value is poorly constrained .
they also compare the available photometry of 2m1207b from 0.94.0 @xmath4 m with predictions from the dusty models , and conclude that they are consistent with this temperature estimate . while @xcite demonstrate that dusty models are a somewhat poor fit for old late l - type field dwarfs , @xcite have shown that they produce adequate spectral fits for the young objects 2m1207b and ab pic b. 2m1207b is significantly redder and dustier than typical l dwarfs , as predicted for low surface gravity objects .
we explore a complementary approach , comparing the spectra published in @xcite as well as the available photometry with template objects drawn from wide field surveys ( predominantly older objects ) .
the low resolution h and k band spectrum available at signal to noise ratio of 310 is morphologically similar to other l dwarfs suspected of having low gravity : 2mass j01415823 - 4633574 ( 2 m j0141 ; * ? ? ?
* ) , 2mass j22244381 - 0158521 @xcite , and sdss j22443167 + 2043433 ( sdss j2244 ; * ? ? ?
all of these objects exhibit weak metal resonance lines ( e.g. * ? ? ?
* ; * ? ? ?
* ) , stronger than expected co for their spectral type @xcite , and unusual pseudo continua in the h band spectra attributed to collision induced molecular hydrogen absorption ( e.g. * ? ? ?
* ) , all indications of low surface gravity . given the morphological correspondence between the spectrophotometry of 2m1207b and these other low gravity l dwarfs , it is reasonable to assume that 2m1207b has a similar nature .
further , the spectrum of 2m1207b shows no signs of ch@xmath15 absorption which would indicate @xmath10 below 1400 k. the colors of 2m1207b are very red compared to observed sequences of field l dwarfs ( e.g. * ? ? ?
* ; * ? ? ?
the anomalous l dwarfs listed above also exhibit this behavior which can be attributed to low gravity ( e.g. * ? ? ?
* ; * ? ? ?
taking the observed ( @xmath16-@xmath17 ) colors of 2m1207b and several plausible intrinsic colors matches , we searched for reddening solutions that would fit the sed of 2m1207b . adopting the colors of 2 m j0141 @xcite as a low gravity early l template , we derived a@xmath18 @xmath11 9 mag , which matches the @xmath19 photometry well .
however , 2 m j0141 has @xmath10 = 2000k @xcite , and a@xmath18 @xmath11 9 mag implies a@xmath20 @xmath11 1 mag , which would be insufficient to move 2m1207b above the old dwarf sequence for that @xmath10 .
alternatively , adopting the colors of sdss j2244 @xcite as a late - l low - gravity template , we arrive at a@xmath18 @xmath11 3.7 , again reproducing the colors of 2m1207b within the ( rather large ) errors from @xmath01 - 4@xmath4 m . however , the required reddening ( a@xmath20 @xmath11 0.4 mag ) is insufficient to solve the underluminosity of 2m1207b .
if we assume that 2m1207b has the ( @xmath16-@xmath17 ) colors of a late type m dwarf , we derive a@xmath18 @xmath11 11 mag , but can not reconcile the observed colors without invoking excess emission in the k and l bands . in the limit of zero extinction , one can find models of extremely low gravity that fit the sed ( e.g. * ? ? ? * ) , so we take that as the simplest assumption consistent with the observed properties of known l dwarfs and informed by model atmospheres . in order to estimate the bolometric luminosity of 2m1207b
, we must also estimate an appropriate bolometric correction to apply to the observed absolute magnitude . given the distance to the source , the lack of evidence for interstellar reddening , the available photometry , and the temperature estimate discussed above
, we can apply a bolometric correction to any flux estimate from 0.94.0 @xmath4 m .
@xcite demonstrates that bc@xmath20 varies little as a function of spectral type for l dwarfs , so we apply the @xmath21-band bc to 2m1207b to minimize the uncertainties in the estimate of m@xmath22 ( bc@xmath20 = [email protected] mag ; same as in * ? ? ?
this results in a luminosity estimate of log(l / l@xmath23 ) = [email protected] dex . for @xmath10
= 1600@xmath2100 k , the dusty models predict bc@xmath20 = 3.56 @xmath2 0.07 mag , and a luminosity of log(l / l@xmath23 ) = 4.66 @xmath2 0.08 dex .
the difference between adopting the empirical or theoretical bc values is within the errors , so we conservatively adopt the empirically derived log(l
/ l@xmath23 ) as an upper limit . given its luminosity , 2m1207b lies 47 times below where expected for an age range of 510 myr given its inferred temperature range of 15001700 k. @xcite derive an effective temperature for the young ( @xmath00.3 gyr ) l / t transition object hd 203030b that is considerably lower ( by @xmath0230 k ) than predicted for its spectral type ( l7.5 ) .
if 2m1207b were in fact a late - l spectral type with a temperature of 1100 k , we could reconcile its position in the h - r diagram given its age .
however , this is _
@xmath0500 k _ cooler than the @xmath10 derived from spectral synthesis models according to @xcite .
@xcite suggest that the temperature of l - t transition objects should only weakly depend on temperature . exploring this solution to the 2m1207b problem
requires higher snr spectra and further study .
can the theoretical evolutionary tracks be wrong in luminosity by such a large factor ?
as pointed out by @xcite , another well - studied young , low - mass binary ( ab pic ) exhibits hrd positions consistent with the age of the group ( tuc - hor ; @xmath030 myr ) . @xcite
have argued that the luminosities of young planets formed through core
accretion are likely to be over - predicted in models that initially start objects in high entropy states . however , this argument does not apply if the object formed through gravitational fragmentation , as has been suggested @xcite .
the @xcite models predict that for ages @xmath8myr after accretion has ceased , all planets that formed through core - accretion with masses of @xmath710@xmath9 will be colder than @xmath10 @xmath11 800k and with log(l / l@xmath23 ) @xmath7 5.2 dex . hence , while 2m1207b is vastly underluminous for its @xmath10 compared to the `` hot - start '' evolutionary models of @xcite , it is overluminous and too hot for the fiducial `` cold - start '' core accretion models for @xmath710@xmath9 from @xcite
. observationally , there seems to be no trend that would suggest an error in the evolutionary tracks that could account for the hrd position of 2m1207b .
as pointed out in @xcite one can not reconcile the observed spectrum and sed of 2m1207b with its apparent low luminosity given the available models .
here we take a different approach , adopting the derived temperature of 1600 k , and postulate that the observed source flux is a reflection of its actual luminosity .
we hypothesize that 2m1207b is the result of a recent collision of two protoplanets ( cf . * ? ? ?
adopting the temperature and luminosity in 2 , one derives a radius of 48,700 @xmath2 8,800 km (= [email protected] @xmath24 , [email protected] @xmath25 ) . for a range of densities ,
the inferred mass and gravity are given in table [ tab : pred ] .
we hypothesize that the object is a hot protoplanet of density @xmath01 g@xmath26 . with this radius and density0.7 - 1.4 g@xmath26 , and the known transiting hot jupiters have densities of @xmath00.3 - 1.3 g@xmath26 @xcite .
the post - accretion evolutionary track of a 1 @xmath9 object by @xcite has density @xmath00.5 - 0.6 g@xmath26 in its first 10 myr . ]
, the fiducial planet would have a mass of @xmath27 = 81 m@xmath28 ( 0.85 m@xmath29 = 4.7 m@xmath30 ) . throughout this discussion
we will refer to hypothetical protoplanets @xmath31 and @xmath32 which merged to produce body @xmath33 ( 2m1207b ) . in our solar system ,
the largest planetesimals to impact the planets during the late stages of accretion appear to have had mass ratios of order @xmath34 = @xmath35/@xmath36 @xmath0 0.1 , including the bodies responsible for producing the obliquities of saturn , uranus , and neptune @xcite .
thus we assume that our @xmath081 @xmath37 object was the product of a collision between protoplanets with masses @xmath36 = @xmath38 and @xmath35 = @xmath39 , or @xmath36
= 74@xmath37 and @xmath35 = 7 @xmath37 . following @xcite ,
the _ minimum _ impact velocity for two planets will be their mutual escape velocity , defined as : @xmath40 in the simplified case of identical densities @xmath41 for bodies @xmath31 , @xmath32 , and @xmath33 , one can simplify this equation in terms of mass and radius of the final planet @xmath33 : @xmath42 where the last factor is within @xmath715% of unity for all @xmath34 @xmath43 10@xmath13 . in our fiducial model ,
the radii of the fiducial impactors are 3.0 and 6.4 @xmath25 , respectively .
this leads to a fiducial impact velocity @xmath8 30.6 kms@xmath13 .
following @xcite , one can calculate the radiative timescale of a long - lived afterglow from the collision of bodies @xmath31 and @xmath32 : @xmath44 where @xmath45 is the impact velocity of the impactor in units of 10 kms@xmath13 .
@xmath46 is the radius of the planet after collision , and @xmath47 is the radius of earth .
@xmath48 is the temperature of the emitting photosphere in units of 1000k .
we can rewrite the radiative timescale in terms of the impactor mass ratio @xmath34 , the properties of the final body @xmath33 , and assuming impact velocity equals @xmath49 : @xmath50 where @xmath51 = @xmath52 . for @xmath34 @xmath53 1 , @xmath51 @xmath0 @xmath34 ( to within @xmath740% accuracy for @xmath34 @xmath7 10@xmath13 ) . for our fiducial model , @xmath54 @xmath8 59 kyr or @xmath01% the age of the tw hya association .
these radiative timescales are not negligible , and suggest that a hot afterglow could be visible for an appreciable fraction of the system lifetime . in modeling the collision of a jupiter and an earth - like protoplanet , @xcite estimate that less than 1% of the impact energy is radiated away in the initial prompt flash .
they further argue that most of the collision - deposited energy is locked up deep in the post - collision planet , and radiated over a long timescale as an afterglow that peaks in the ir .
we propose that 2m1207b is such a long - lived afterglow .
could a plausible circum(sub)stellar disk form a planetary system with the required properties ? the formation and collision of two such large planetesimals at radii @xmath8 10 au in a protoplanetary disk surrounding a brown dwarf is very unlikely given the mass surface density and orbital timescales expected @xcite .
perhaps 2m1207b formed at smaller radii as the ice line in the disk of 2m1207a swept through a large range of inner radii from 10 to 0.1 au ( cf . * ? ? ?
* ) as the young brown dwarf evolved ( see * ? ? ?
* for an alternate scenario ) .
if we consider a primordial disk surrounding 2m1207a that is marginally gravitationally stable ( @xmath55 @xmath0 0.1 ) it would have a total gas+dust mass of 23 @xmath9 . adopting the protoplanetary core mass scenario of ( * ? ?
* see also lodato et al .
2005 ) , the time evolution of the mass of a planet accreting 10@xmath56 g planetesimals is : @xmath57 where @xmath58 is time , @xmath59 is disk surface density of solids , @xmath60 is the orbital distance , and @xmath61 is the mass of 2m1207a . assuming a disk with the above mass , mass surface density profile @xmath62 @xmath63 @xmath64 , and an outer radius of 15 au , the total disk mass surface density at 3 au is 250 g@xmath65 . if we consider a gas to dust+plus ice ratio of 25 ( 100/4 ) , we arrive at a mass surface density in solids of 10 g@xmath65 at 3 au . using the above equation for a brown dwarf of mass 0.025 m@xmath66
, we estimate that a core of 5 - 10 @xmath37 can form within @xmath03 myr at this radius . in 10 myr ,
a similar mass core could form at a distance of 5 au from the brown dwarf . assuming the above disk model parameters , roughly half of the total disk mass ( solid and gas ) resides inside of 7.5 au and half outside .
it is at least plausible that 24 cores of 510 @xmath37 could form between 110 au within 310 myr in this system utilizing the bulk of available solids in the system .
if two of those cores accreted enough gas to form neptune - to - saturn mass protoplanets , we can envision a scenario where : a ) failed cores of 510 @xmath37 could collide with a successfully formed gas / ice giant protoplanet , creating the observed hot collisional afterglow ; and b ) another gas / ice giant , along with the presence of the remnant primordial disk , could eject 2m1207b to its observed orbital radius of 50 au .
motivated by evidence for gas giants at large separations having created observed structure in debris disks , @xcite have investigated gas giant migration / ejection scenarios in disks . @xcite
have also proposed that neptune and uranus formed closer to the sun ( between jupiter and saturn ) and were ejected to larger orbital radii through dynamical processes .
we note that the remnant disk surrounding 2m1207a has a mass comparable to that we propose for the ejected 2m1207b @xcite though its outer radius is unconstrained from current observations .
this is , of course , a highly improbable series of events .
the hypotheses of whether or not 2m1207b is a hot protoplanet collision afterglow or is obscured by a dense disk of large dust grains can be tested . in the scenario proposed here
, 2m1207b is actually a @xmath080@xmath37 object with radius @xmath049,000 km .
the surface gravity of such an object in cgs units would be log(@xmath14 ) @xmath0 3 ( table 1 ) .
this is significantly lower than than for 5 - 10 myr - old 38 @xmath9 objects which have log(@xmath14 ) @xmath67 4 @xcite .
if 2m1207b possesses an edge - on disk exhibiting grey extinction , then spatially resolved ground - based observations should reveal : a ) an infrared excess at @xmath68 @xmath8 4
@xmath4 m from the disk ; b ) a 10 @xmath4 m silicate absorption feature consistent with the disk being edge - on ; c ) polarized emission from scattered light at shorter wavelengths ; and/or d ) resolved scattered light emission consistent with an edge - on dust disk system ( cf .
if 2m1207b is actually a physically smaller ( and therefore lower mass ) companion , it should exhibit near - infrared spectra : a ) consistent with the 1600 k temperature advocated by @xcite ; and b ) low surface gravity ( log(g ) @xmath0 3 ) in high s / n spectra . as mentioned above
, surface gravity affects the spectra of very cool objects in ways that can be observed through analysis of atomic and molecular features .
* see also kirkpatrick et al .
2006 ) suggest that log(@xmath14 ) can be estimated to within 0.3 - 0.5 dex from high s / n near - infrared spectra of m and l dwarfs . @xcite
specifically investigate the gravity dependence of the feature at 1.14 while gorlova et al . provide a preliminary calibration of the surface gravity effects of at 1.25
( see also * ? ? ?
. these effects should be clear in modest s / n spectra ( 20 - 30 ) easily distinguishing between the log(@xmath14 ) @xmath0 4 model of @xcite and the log(@xmath14 ) @xmath0 3 model proposed here .
further , we anticipate that our protoplanetary collision remnant would be metal - rich compared to the primary .
models from @xcite as well as @xcite demonstrate the significant differences in brown dwarf and gas giant planet atmospheric models by varying the metallicity .
such effects would be easily observable in s / n @xmath0 2030 spectra , obtainable in 1 - 2 nights of observing time on a 6 - 10 meter telescope equipped with adaptive optics .
perhaps future surveys will uncover additional hot protoplanet collision afterglow candidates with even smaller inferred masses .
eem is supported through a clay postdoctoral fellowship from the smithsonian astrophysical observatory .
mrm is supported by laplace .
we thank phil hinz , scott kenyon , matt kenworthy , stan metchev , subu mohanty , and steve strom for discussions .
we thank adam burrows for bringing the @xcite paper to our attention , kevin zahnle for an inspiring colloquium regarding the earth - moon impact , and the anonymous referee for a thoughtful review .
allard , f. , et al .
2001 , , 556 , 357 allers , k. n. , et al .
2007 , , 657 , 511 anic , a. , alibert , y. , & benz , w. 2007 , , 466 , 717 bakos , g. .
, et al . 2007 , , 656 , 552 baraffe , i. , et al . 2003 , , 402 , 701 borysow , a. , jorgensen , u. g. , & zheng , c. 1997 , , 324 , 185 boss , a. p. 2006
, , 637 , l137 burrows , a. , sudarsky , d. , & hubeny , i. 2006 , , 640 , 1063 chabrier , g. , et al.2000 , , 542 , 464 chauvin , g. , et al .
2004 , , 425 , l29 chauvin , g. , et al .
2005 , , 438 , l25 cushing , m. c. , rayner , j. t. , & vacca , w. d. 2005 , , 623 , 1115 fortney , j. j. , et al.2006 , , 642 , 495 gizis , j. e. 2002 , , 575 , 484 goldreich , p. , lithwick , y. , & sari , r. 2004 , , 614 , 497 golimowski , d. a. , et al .
2004 , , 127 , 3516 gorlova , n. i. , et al.2003 , , 593 , 1074 ida , s. , & lin , d. n. , 2004 , , 604 , 388 jayawardhana , r. & ivanov , v. d. , 2006 , science , 313 , 1279 kennedy , g. m. , kenyon , s. j. , & bromley , b. c. , 2006 , , 650 , l139 kirkpatrick , j. d. , et al .
2006 , , 639 , 1120 knapp , g. r. , et al .
2004 , , 127 , 3553 leggett , s. k. , et al.2001 , , 548 , 908 lissauer , j. j. , & safronov , v. s. 1991 , icarus , 93 , 288 lodato , g. , delgado - donate , e. , & clarke , c. j. 2005 , , 364 , l91 luhman , k. l. , et al .
2007 , , 659 , 1629 luhman , k. l. , et al .
2007 , , in press ( arxiv:0706.0279v1 ) mamajek , e. e. 2005 , , 634 , 1385 marley , m. s. , et al .
2007 , , 655 , 541 mcgovern , m. r. , et al.2004 , , 600 , 1020 mclean , i. s. , et al .
2003 , , 596 , 561 metchev , s. a. , & hillenbrand , l. a. 2006 , , 651 , 1166 mohanty , s. , jayawardhana , r. , & barrado y navascus , d. 2003 , , 593 , l109 mohanty , s. , et al .
2007 , , 657 , 1064 neuhuser , r. , et al .
2005 , , 435 , l13 riaz , b. , & gizis , j. e. 2007 , , 661 , 354 song , i. , et al .
, 2006 , , 652 , 724 stevenson , d. j. 1987 , ann .
earth and planet .
, 15 , 271 stern , s. a. 1994 , , 108 , 2312 sterzik , m. f. , et al.2004 , , 427 , 245 thommes , e. w. , duncan , m. j. , & levison , h. f. 2003 , icarus , 161 , 431 veras , d. , & armitage , p. j. 2004 , icarus , 172 , 349 vrba , f. j. , et al .
2004 , , 127 , 2948 webb , r. a. , et al . 1999 , , 512 , l63 wetherill , g. w. 1980 , , 18 , 77 whelan , e. t. , et al . 2007 , , 659 , l45 zhang , b. , & sigurdsson , s. 2003 , , 596 , l95 | we introduce an alternative hypothesis to explain the very low luminosity of the cool ( l - type ) companion to the @xmath025 m@xmath1 @xmath08 myr - old brown dwarf 2m1207a .
recently , mohanty et al . ( 2007 ) found that effective temperature estimates for 2m1207b ( 1600@xmath2100 k ) are grossly inconsistent with its lying on the same isochrone as the primary , being a factor of @xmath010 underluminous at all bands between @xmath3 ( 0.8@xmath4 m ) and @xmath5 ( 3.6@xmath4 m ) .
mohanty et al .
explain this discrepency by suggesting that 2m1207b is an 8 m@xmath1 object surrounded by an edge - on disk comprised of large dust grains producing 2.5@xmath6 of achromatic extinction .
we offer an alternative explanation : the apparent flux reflects the actual source luminosity .
given the temperature , we infer a small radius ( @xmath049,000 km ) , and for a range of plausible densities , we estimate a mass @xmath7 m@xmath1 .
we suggest that 2m1207b is a hot protoplanet collision afterglow and show that the radiative timescale for such an object is @xmath8 @xmath01% the age of the system . if our hypothesis is correct , the surface gravity of 2m1207b should be an order of magnitude lower than predicted by mohanty et al .
( 2007 ) . |
the _ solar dynamics observatory ( sdo ) _
@xcite is a solar satellite launched by nasa on february 11 2010 .
the scientific goal of this mission is a better understanding of how the solar magnetic field is generated and structured and how solar magnetic energy is stored and released into the helio- and geo - sphere , thus influencing space weather .
_ sdo _ contains a suite of three instruments : * the _ helioseismic and magnetic imager ( sdo / hmi ) _
@xcite has been designed to study oscillations and the magnetic field at the solar photosphere . *
the _ atmospheric imaging assembly ( sdo / aia ) _
@xcite is made of four telescopes , providing ten full - sun images every twelve seconds , twenty four hours a day , seven days a week .
* the _ extreme ultraviolet variability experiment ( sdo / eve ) _
@xcite measures the solar extreme ultraviolet ( euv ) irradiance with unprecedented spectral resolution , temporal cadence , accuracy , and precision . the present paper deals with an important aspect of the image reconstruction problem for _ sdo / aia _ @xcite .
the four telescopes of such instrument capture images of the sun s atmosphere in ten separate wave bands , seven of which centered at euv wavelengths .
each image is a @xmath2 square array with pixel width in the range @xmath3 arcsec and is acquired according to a standard ccd - based imaging technique .
in fact , each _ aia _ telescope utilizes a @xmath4-megapixel ccd divided into four @xmath5 quadrants .
as typically happens in this kind of imaging , _ aia _ ccds are affected by primary saturation and blooming , which degrade both quantitatively and qualitatively the _ aia _ imaging properties . _
primary saturation _
@xcite refers to the condition where a set of pixel cells reaches the full well capacity , i.e. these pixels store the maximum number possible of photon - induced electrons . at saturation ,
pixels lose their ability to accommodate additional charge , which therefore spreads into neighboring pixels , causing either erroneous measurements or second - order saturation . such spread of charge
is named _ blooming _
@xcite and typically shows up as a bright artifact along a privileged axis in the image .
figure [ fig : saturation - blooming ] shows a notable example of combined saturation and blooming effects in an _ sdo / aia _ image captured during the september 6 , 2011 event .
the recovery of information in the primary saturation region by means of an inverse diffraction procedure is the main goal of the present paper .
further , we also introduce here an interpolation approach that allows a robust estimate of the background in the diffraction region as well as reasonable estimate of the flux in the central blooming region . [ cols="^,^ " , ]
_ sdo / aia _ images are strongly affected by both primary saturation and blooming , that may occur at all different wavelengths and acquisition times , even in the case of flaring events characterized by a moderate peak flux .
this paper describes the first mathematical description of a robust method for the de - saturation of such images at both a primary and secondary ( blooming ) level .
the method relies on the description of de - saturation in terms of inverse diffraction and utilizes correlation and expectation - maximization for the recovery of information in the primarily saturated region .
this approach requires to compute a reliable estimate of the image background which , for this paper , has been obtained by means of interpolation in the fourier space .
the knowledge of the background permits to recover information in the blooming region in a very natural way .
the availability of an automatic procedure for image de - saturation in the _ sdo / aia _ framework may potentially change the extent with which euv information from the sun can be exploited .
in fact , armed with our computational approach , many novel problems can be addressed in _
sdo / aia _ imaging .
for example , one can study the impact of the choice of the model for the diffraction psf on the quality of the de - saturation . in this paper
we used a synthetic estimate of the diffraction psf provided by solar software ( ssw ) but other empirical or semi - empirical forms can be adopted .
furthermore , this technique can be extended to account for the dependance of the psf from the passband wavelengths .
finally , the routine implementing this approach is fully automated and this allows the systematic analysis of many events recorded by _ aia _ and their integration with data provided by other missions such as _ rhessi _ @xcite or , in the near future _ stix _ @xcite .
this work was supported by a grant of the italian indam - gncs and by the nasa grant nnx14ag06 g .
the _ sdo / aia _ hardware is equipped with a feedback system that reacts to saturation in correspondence of intense emissions by reducing the exposure time . as a result , for a typical _ aia _ acquisition along a time range of some minutes , during which saturation occurs , the telescope always provides some unsaturated frames that can be utilized to estimate the background . a possible way to realize such an estimate
is based on the following scheme .
let us denote with @xmath6 and @xmath7 two unsaturated images acquired at times @xmath8 and @xmath9 , respectively and with @xmath10 a saturated image acquired at @xmath11 with @xmath12 ( note that @xmath6 , @xmath10 , and @xmath7 are normalized at the same exposure time ) .
the algorithm for the estimate of the background is : 1 .
@xmath6 and @xmath7 are deconvolved with em ( using the global psf ) to obtain the reconstructions @xmath13 and @xmath14 ( the kl - kkt rule can be used to stop the iterations ) .
both @xmath13 and @xmath14 are fourier transformed by means of a standard fft - based procedure to obtain @xmath15 and @xmath16 .
a low - pass filter is applied to both @xmath15 and @xmath16 to obtain @xmath17 and @xmath18 , respectively .
4 . for each corresponding pair of pixels in @xmath17 and @xmath18 that are not negligible , an interpolation routine
is applied , both for the real and imaginary part .
this provides @xmath19 in correspondence of @xmath20 .
the resulting vector @xmath21 is fourier inverted to obtain the interpolated reconstruction @xmath22 .
the core psf @xmath23 is finally applied to @xmath22 to obtain @xmath24 in the image domain .
@xmath24 is a reliable estimate , for time @xmath11 , of the background @xmath25 introduced in section 2 . on the other hand ,
a reliable estimate of the image in the bloomed region is provided by the restriction @xmath26 of @xmath24 onto @xmath27 determined as in section 2 .
we finally observe that , in this algorithm , the interpolation step is applied in the fourier domain because , after filtering , a lot of pixels are negligible and therefore the computational burden of the procedure is notably decreased .
9999 pesnell w d , thompson b j and chamberlin p c 2012 the solar dynamics observatory ( sdo ) _ sol .
phys . _ * 275 * 3 scherrer p h et al 2012 the helioseismic and magnetic imager ( hmi ) investigation for the solar dynamics observatory ( sdo ) _ solar phys . _ * 275 * 207 lemen j r et al 2012 the _ atmospheric imaging assembly ( aia ) _ on the _ solar dynamics observatory ( sdo ) _ _ solar phys .
_ * 275 * 17 woods t n et al 2012 extreme ultraviolet variability experiment ( eve ) on the solar dynamics observatory ( sdo ) : overview of science objectives , instrument design , data products , and model developments _ solar phys .
_ * 275 * 115 grigis p , su y and weber m 2012 _ aia psf characterization and image deconvolution _ , sdo documentation ( http://www.lmsal.com/sdodocs ) boerner p et al 2012 initial calibration of the atmospheric imaging assembly ( aia ) on the solar dynamics observatory ( sdo ) _ solar phys . _
* 275 * 41 poduval b , deforest c e , schmelz j t and pathak s 2013 point - spread functions for the extreme - ultraviolet channels of sdo / aia telescopes _ astrophys
. j. _ * 765 * 144 martinez p and klotz a 1997 _ a practical guide to ccd astronomy _
( cambridge : cambridge university press ) bekefi g and barrett a h 1978 _ electromagnetic vibrations , waves , and radiation _ ( boston : mit press ) raftery c l , krucker s and lin r p 2011 imaging spectroscopy using _ aia _ diffraction patterns in conjunction with _ rhessi _ and _ eve _ observations _ astrophys . j. lett . _ * 743 * l27 schwartz r a , torre g and piana m 2014 systematic de - saturation of images from the _ atmospheric imaging assembly _ in the _ solar dynamics observatory _ _ astrophys . j. lett .
_ * 793 * l23 shepp l a and vardi y 1982 maximum likelihood reconstruction for emission tomography ieee _ trans . med . imaging _
* 1 * 113 benvenuto f and piana m 2014 regularization of multiplicative iterative algorithms with non - negative constraint _ inverse problems _ * 30 * 035012 benvenuto f , schwartz r , piana m and massone a m 2013 expectation maximization for hard x - ray count modulation profiles _ astron . astrophys . _
* 555 * a61 de pierro a r 1987 on the convergence of the iterative image space reconstruction algorithm for volume ect ieee _ trans . med
. imaging _ * 6 * 2 lin r p et al 2002 the reuven ramaty high energy solar spectroscopic imager ( rhessi ) _ solar phys . _
* 210 * 3 benz a et al 2012 the spectrometer telescope for imaging x - rays on board the solar orbiter mission _ proc . _
spie conference on space telescopes and instrumentation 2012 - ultraviolet to gamma ray * 8443 * 84433l | the _ atmospheric imaging assembly _ in the _ solar dynamics observatory _ provides full sun images every @xmath0 seconds in each of @xmath1 extreme ultraviolet passbands .
however , for a significant amount of these images , saturation affects their most intense core , preventing scientists from a full exploitation of their physical meaning . in this paper
we describe a mathematical and automatic procedure for the recovery of information in the primary saturation region based on a correlation / inversion analysis of the diffraction pattern associated to the telescope observations .
further , we suggest an interpolation - based method for determining the image background that allows the recovery of information also in the region of secondary saturation ( blooming ) . |
in response to discrepancies that have persistently been reported between predicted and measured neutrino fluxes , some authors have suggested a violation of lorentz invariance , either via different couplings of neutrinos to gravity [ 1 ] , or by different indices of refraction of the vacuum to different types of neutrinos[2 ] . in this letter
we consider another such hypothesis , in which lorentz invariance is violated by a foam - like vacuum that is granulated near the planck scale .
we consider that neutrinos are mixed over a scale of @xmath7 , where @xmath8 is the neutrino energy in a cosmic rest frame and is conserved by the mixing process , @xmath9 @xmath10 is the planck length ( energy ) and @xmath11 is a dimensionless number of order of 10@xmath12 .
we note that the hypothesis mitigates the discrepancy between ssm predictions and the gallex / sage , homestake , and superkamiokande ( sk ) data for solar neutrinos .
it also mitigates the reported anomaly in atmospheric neutrinos as well .
it is also marginally consistent with limits set by post - beam dump oscillation experiments , which reveal no significant oscillation at neutrino energies of 30 gev over distances of order of 1/2 km .
physical justification for the hypothesis is not quite known to us at present , and the results are presented here in the spirit of suggestive numerical coincidence : that the energy scale in the expression for @xmath13 happens to be near the planck scale . a brief discussion of possible physics that could provide some theoretical justification for this hypothesis is given in section iv .
we will , for convenience , assume that the mixing redistributes neutrinos uniformly and irreversibly among three flavors .
reversible mixing , that is , oscillations , could also be considered at a phenomenological level with an oscillation length that scales as @xmath14 , and would give similar results ( except for monochromatic neutrinos ) , if the coupling among all three flavors were the same .
the major feature of the present hypothesis ( although we shall refer to it for convenience as an irreversible mixing [ i m ] hypothesis ) is the energy dependence of the mixing length @xmath15 , and not so much the irreversibility . consider a case of @xmath16 neutrino flavors , all interacting with the foam with the same mean free - path length @xmath13 . in this case
the evolution of each flavor number @xmath17 , ( @xmath18 ) , as a function of distance @xmath19 from the source is governed by the expression : @xmath20 note that , if only a single flavor is emitted ( as in the solar neutrino case ) , then the spectrum of the emitted flavor first softens with time ( that is , with distance from source ) , then returns to its original form at a lower amplitude .
the solar neutrino problem may be described as a deficit of the observed neutrino fluxes compared to predicted ones : + a ) the superkamiokanda experiment[3 ] presently sets the upper limit to the boron neutrino flux at @xmath21 , more than two times less than the ssm bp98[4 ] `` theoretical '' value @xmath22 ; + b ) gallex and sage experiments[5 ] measure the reaction rate in @xmath23 detector as @xmath24 snu ( figure reported at neutrino98 ) which is @xmath25 times less than ssm bp98[4 ] value 129 snu , the gallium rate being caused mainly by low - energy @xmath6 neutrinos ; + c ) the homestake experiment[6 ] with @xmath26 reports a reaction rate as @xmath27 snu which is three times less than ssm bp98 value 7.56 snu , here the detected @xmath28 s being mostly energetic boron neutrinos .
+ we also note that the @xmath6 neutrino energy spectrum and total flux are claimed[4 ] to be known much better than those of boron neutrinos .
table i presents the results of applying the i m hypothesis to the solar chlorine case .
the number of neutrino flavors was taken be equal to 3 with no sterile flavor .
individual yields in the total rate were calculated using files from web site [ 7 ] .
spectral distortion effects are small in the chlorine case as @xmath6 neutrinos give no yield while boron neutrinos ( due to their large energies and strong dependence of i m effect on the neutrino energy ) give almost exactly the limiting value @xmath29 .
that is , all @xmath30 electronic neutrinos have been mixed into three active flavors equally .
table ii presents the results of applying the i m hypothesis to the solar gallium case .
the number of active mixing flavors of neutrinos was taken to be equal to 3 with no sterile flavor .
spectral distortion effects are included in the calculation of rates with the use of files in [ 7 ] . again , as in the chlorine case , spectral distortion effects are small as @xmath6 neutrinos have low energy ( and hence large free - path length ) , while boron neutrinos ( due to their large energies and strong dependence of i m effect on the neutrino energy ) give almost exactly the limiting value @xmath31 .
i m predicts variation in the solar neutrino fluxes caused by the variation of the earth - sun distance .
seasonal variation of the earth - sun distance , @xmath32 , ( @xmath33 is the mean eccentricity of earth orbit ) , leads to relative variation of the total @xmath34-flux , @xmath35 without i m ; the i m leads to the following relative variation of the initial flavor @xmath28 s flux : @xmath36.\ ] ] numerically , taking @xmath37 and @xmath38 for @xmath39 , we get @xmath40 . this value may be approximately applied to @xmath41 and @xmath42 solar neutrinos with mean energies of about @xmath43 and @xmath44 , respectively . for @xmath5 neutrinos ,
the mean energy is @xmath45 , and , as free - path @xmath13 is proportional to @xmath46 , the variation of the flux is reduced to @xmath47 .
in this field there are two observed `` deficits '' : + a ) the flux of upward - going neutrinos ( mainly muonic ) is about .6 of flux of downward - going neutrinos .
+ b ) the muon - to - electron neutrino relation is observed to be about .6 of the predicted value which is slightly larger than 2 . + we discuss both cases .
up - down asymmetry in muonic and electronic events was observed in a number of experiments and was suggested in [ 8 ] as a simple test for discriminating among neutrino oscillation / mixing mechanisms and their modifications . at @xmath48 . ] in fig .
1 we present the @xmath46 mixing - induced up - down asymmetry parameter as a function of @xmath49 for both electronic and muonic neutrinos , with @xmath50 at @xmath51 . we note that the energy dependence of the muon asymmetry and electron asymmetry differs significantly from the cases considered in [ 8 ] .
we use atmospheric neutrino fluxes from [ 9 ] and calculate only fluxes at the sk site but not reaction rates for the sk detector as our aim is to show only main characteristics of up - down asymmetry in the i m model .
note that atmospheric neutrino fluxes are calculated with rather large uncertainties that depend on the solar activity cycle phase , geomagnetic effects etc .
these uncertainties may be especialy important at low neutrino energies .
inclusion of these and other factors would not drastically change the characteristics of muonic and electronic up - down asymmetries .
sk experiment[10 ] shows a deficit of muonic events compared to predicted values which was claimed as evidence of neutrino oscillations .
i m with values of free - path length about @xmath50 or less at @xmath53 , as it is shown in fig .
2 , may reduce initial muonic - to - electronic - neutrino ratio down to observed values . at @xmath48 , .1 ( + ) and .02 ( * ) .
ordinates are ratio - of - ratios , that is , @xmath52 ratio with account of i m divided by @xmath52 ratio without i m . also shown are sk[10 ] data on muonic - to - electronic events ratio . ]
as lorentz invariance seems to be a symmetry in our low energy world , it is reasonable to suppose that if it is broken , such breaking takes place at high energy , e.g. by a foam - like condensate of quantum black holes , or a spaghetti - like condensate of flux tubes at the gut scale .
low energy neutrino scattering off small scale inhomogeneities so formed would most likely have an scattering cross section @xmath54 e@xmath55 , as for any low energy neutrino scattering[11 ] .
also , it could be argued that because lorentz invariant equations work as well as they do in describing the observed world , with the modest price of containing infinities , then any violation of lorentz invariance probably enters as an anomaly via the choice of regularization procedure .
the difficulty is that a foam - like condensate that breaks lorentz invariance could also break translational invariance , and this would imply a scattering that is apparently ruled out by observations of the neutrinos from supernova 1987a .
also , at least for charged particles , it is ruled out by the survival times of particle beams in storage rings .
we have no solution to this difficulty at present ; any such solution , it appears to us , would probably require a geometric interpretation of the different neutrino flavors as well as a model for why momentum and angular momentum are exactly conserved quantities whereas neutrino flavor is not . alternatively , the foam may be entirely virtual , that is , unable to increase the entropy of real particles , in which case the mixing would be reversible .
but then it is harder to justify the @xmath46 dependence of mixing length .
neutrino mixing considered in this letter has only one free parameter ( mean free - path length @xmath13 ) given the number of flavors . in spite of this
, @xmath56 mixing leads to a marginally acceptable account of various deficits observed for the solar and atmospheric neutrinos . taking @xmath57 and @xmath58 at energy @xmath59 characteristic of the solar neutrino problem , then at energy @xmath60 , characteristic of the atmospheric neutrino problem , we have @xmath61 . at @xmath62 ,
characteristic of the post - beam neutrino experiments , we have @xmath63 , so the fermilab mixing experiments[12 ] probably imply @xmath64 larger about 2 . the calculations presented in this letter show that for @xmath64 about 2 , the predicted count rate for the homestake detector is reduced from 7.5 to 3.5 snu , the predicted rate for gallex is reduced from 129 to 113 snu , and the atmospheric @xmath52 ratio is reduced by a factor of .6 at energies above .6 gev .
although the reductions of the predicted count rate for the radiochemical detectors , homestake , gallex and sage are not quite enough to completely reconcile the reported observations with the standard solar model , the overall trend in these reductions for both atmospheric and solar neutrino experiments is perhaps suggestive enough to make the hypothesis worthy of consideration at present .
a clean prediction , however , is that longer terrestrial lengths for post beam dump neutrino experiments must reveal mixing .
a null result over the cern - macro distance or the fermilab - wisconsin distance would unequivocally rule out the hypothesis .
de acknowldeges the ibs for partial support and zs is grateful to bgu and particularly to de for the generous hospitality .
j. n. bahcall , astro - ph/9808162 ; j.n.bahcall , s.basu and m.h.pinsonneat , phys .
b 433 ( 1998 ) 1 . +
we refer to the ssm bp98 as a representative of a number of more or less `` concordant '' ssm s ( see references in bp98 ) .
however we should mention that there is also the `` discordant '' ssm ds98 : a. dar , g. shaviv , astro - ph/9808098 ; we do not attempt any judgment of their relative merits .
.irreversible mixing , with ssm bp98[4 ] predictions for chlorine reaction rates , at n=3 flavors for some values of the free - path length @xmath66 at @xmath67 .
case @xmath68 corresponds to ssm bp98 predictions . [
cols="^,^,^,^,^,^ " , ] | it is hypothesized that neutrinos are massless and mixed among 3 ( @xmath0 , @xmath1 and @xmath2 ) flavors with a mixing length of order of @xmath3 , where @xmath4 is a universal coupling constant .
it is suggested that this hypothesis mostly reconciles standard solar models with the observations of solar @xmath5 neutrinos , solar @xmath6 neutrinos , and the ratio of atmospheric muon neutrinos to electron neutrinos .
solar neutrinos ; atmospheric neutrinos ; planck scale mixing + |
low temperature phase transitions in condensed matter systems are usually driven by the cooperative actions of the system s electronic degrees of freedom . in many cases the exchange interactions between the spins of electrons on neighbouring ions
causes them to order below a characteristic transition temperature .
however , in lanthanide compounds , the localised @xmath1-electrons on each ionic site may have non - spherical charge or current distributions , which may be described by a multipole expansion of their electric or magnetic fields @xcite .
these multipoles may interact , and in certain cases these interactions may be stronger than the spin exchange interactions , driving a phase transition to a multipolar ordered phase . there have been many examples of electric quadurpolar order observed , such as in ceb@xmath2 @xcite , prpb@xmath0 @xcite , and tmte @xcite , whilst higher order electric multipoles were observed in dyb@xmath3c@xmath3 @xcite .
in contrast to the lanthanides , there have not been many cases of multipolar order observed in actinide compounds .
this is due partly to the larger spatial extent of the @xmath4 wavefunctions , which causes them to become delocalised , and hence invalidates any description of their electric or magnetic fields about a particular point .
the insulating actinide oxides uo@xmath3 and npo@xmath3 , however , have been studied extensively , and exhibit electric quadrupolar and high order magnetic multipolar order respectively @xcite .
in contrast upd@xmath0 is one of the few metallic actinide compounds which has well localised @xmath4 electrons , and it was one of the first compounds to be shown to exhibit quadrupolar ordering .
anomalies were observed in the heat capacity @xcite , thermal expansion @xcite , magnetic susceptibility @xcite and electrical resistivity @xcite of upd@xmath0 at low temperatures which are indicative of phase transitions .
polarised neutron diffraction measurements revealed superlattice peaks at @xmath5 which appear only in the non - spin - flip channel , and thus can be attributed to a structural transition @xcite , from hexagonal to orthorhombic symmetry .
x - ray resonant scattering ( xrs ) measurements @xcite showed that this transition arises from the ordering of the electric quadrupole moment of the 5@xmath6 electrons .
in addition to this phase transition at @xmath7 k , there are three further transitions to different antiferroquadrupolar ( afq ) ordered states at @xmath8 k , @xmath9 k and @xmath10 k. below @xmath11 a superlattice peak at @xmath12 where @xmath13 is odd is observed , whilst below @xmath14 there are additional peaks at @xmath12 where @xmath13 is even .
the @xmath13 odd peaks show that there is antiferroquadrupolar ordering along the @xmath15-direction , also denoted as an anti - phase stacking of quadrupoles .
the @xmath13 even peaks show an additional ordering of quadrupole moments in - phase along @xmath15 .
measurements of the order parameter using x - ray resonant scattering show that the @xmath13 odd order is associated mainly with @xmath16 quadrupoles @xcite , whilst the @xmath13 even order is associated with @xmath17 quadrupoles @xcite .
the @xmath16 ordering is accompanied by a component from the @xmath18 quadrupoles , whilst the @xmath17 is accompanied by a @xmath19 component .
the directions @xmath20 and @xmath21 are equivalent to the @xmath22 and @xmath15 crystallographic directions and @xmath23 is perpendicular to both . in addition , there is also an additional ordering of the @xmath19 quadrupoles in anti - phase below @xmath24 .
this sequence of phase transitions is summarised in figure [ fg : ordertemps ] .
upd@xmath0 adopts the double hexagonal close packed tini@xmath0 crystal structure ( space group @xmath25 , @xmath26 , no .
194 ) @xcite with lattice parameters @xmath27 and @xmath28 .
this means that the nearest neighbour u - u distance , 4.1 , is larger than the hill limit ( @xmath29 ) @xcite and thus the 5@xmath6 electrons are well localised .
the u@xmath30 ions are in the @xmath31 and @xmath32 sites , which have respectively @xmath33 ( @xmath34 ) and @xmath35 ( @xmath36 ) point symmetry which we shall refer to as _ quasi - cubic _ and _ hexagonal_. the xrs data shows that the ordering involves primarily the ions on the quasi - cubic sites @xcite .
knowledge of the crystal field ( cf ) interactions is essential in determining the more complex two - ion interactions . in particular
it is crucial to know the cf ground state .
moreover , the excited states and the matrix elements of the angular momentum operators , @xmath37 , between these and the ground state determine the intensities of excitations observed by inelastic neutron scattering and also to some extent the magnitude of the magnetisation in the ordered phase , as explained by mcewen et al .
thus we shall first consider in section [ sec - upd3-cf ] the cf level scheme deduced from inelastic neutron scattering measurements in the paramagnetic phase .
quasi - cubic site crystal field level scheme .
_ the crystal field energy levels and corresponding wavefunctions in the paramagnetic phase expressed in the @xmath38 basis .
arrows denote transitions from the ground state with non - negligible dipole matrix elements whose squared values are shown as numbers near the arrow . ]
we then present high field magnetisation and magnetostriction measurements in section [ sec - highfield ] , from which the critical fields and magnetic phase diagrams were determined experimentally . finally , a mean field model of the quadrupolar ordering is constructed in section [ sec - upd3-mfcub ] .
the order parameters of each of the quadrupolar phases ( determined by resonant x - ray scattering ) and their transition temperatures , were used to constrain the two - ion quadrupolar interaction parameters for the model , whilst the dipolar exchange parameters were determined from the measured critical fields .
the results of the model are then compared to the measured high field data . whilst the model which we shall present
is empirically based on the measured physical properties of upd@xmath0 rather than directly on measured exchange constants , we note that it is the first attempt to explain comprehensively this fascinating compound with its many competing ordered phases from a microscopic point of view , and hope to stimulate further _ ab initio _ studies of the exchange interactions involved .
the quadrupolar interactions which drive the many phase transitions in upd@xmath0 are difficult to measure directly , because neutrons couple only to the magnetic dipoles in the system , and the energies are far too low for inelastic x - ray scattering to resolve .
however , the wealth of physical property measurements available on upd@xmath0 has encouraged us to try to synthesise this into a mean - field model which explains , to a large extent , these varied measurements .
we hope that this may encourage the construction of models to explain the properties of similar quadrupolar ( or higher multipolar ) ordered compounds thus deepening the understanding of what lies behind these phenomena .
the single - ion properties of uranium are generally found to be close to the @xmath39-coupling limit @xcite .
so , in order to simplify the analysis , we shall ignore any mixing with higher order multiplets in determining the crystal field ( cf ) parameters . as mentioned in section [ sec - upd3-intro ]
, there are two inequivalent sites for the u@xmath30 ions in the crystal structure of upd@xmath0 .
the different point symmetry of these sites gives rise to different crystal fields , but it happens that both split the 9-fold degenerate @xmath40 ground multiplet into three singlets and three doublets .
the energies and wavefunctions of these levels , however , are different for the two sites . the nature of the ground state may be deduced from single crystal susceptibility measurements and estimates of the magnetic entropy determined from heat capacity measurements .
these results indicate a singlet ground state on the hexagonal sites and a doublet on the quasi - cubic sites @xcite . the cf split energy levels were determined from previously reported inelastic neutron scattering measurements made on the time - of - flight spectrometer het at the isis facility , uk @xcite , and on the triple - axis - spectrometer in8 at ill , grenoble @xcite .
we identified magnetic excitations at 4.1 , 9.7 , 12.3 , 16.8 , 20.4 and 30 mev .
the 16 mev peak exhibits considerable dispersion and is assigned to the transition between the @xmath41 singlet ground state and the @xmath42 doublet excited state of the hexagonal site ions .
its dispersion was used to determine the exchange interactions between hexagonal sites @xcite .
the remaining peaks are assigned to the quasi - cubic site ions .
henceforth we shall be concerned mainly with the quasi - cubic sites , as the quadrupolar ordering primarily involves the uranium ions on these sites .
these sites have trigonal , @xmath34 ( @xmath33 ) , point symmetry , so that the crystal field hamiltonian is @xmath43 where @xmath44 are crystal field parameters and @xmath45 are stevens operators .
the quantisation ( @xmath21 ) axis is taken to be the trigonal axis , which in this case is parallel to @xmath15 . from the measured transition energies and with the restriction of a doublet ground state , a crystal field fitting program @xcite was used to obtain initial estimates of the cf parameters for the quasi - cubic sites .
this program relies on the orthogonality of the spherical harmonic functions from which the cf operators are constructed .
it allows one to find a set of cf parameters , @xmath44 , given the energy levels and wavefunctions produced by the crystal field .
the fitting algorithm may thus vary either the wavefunctions to fit a particular set of energy levels , or vice versa . in this case , however , we also face constraints on the wavefunctions .
._crystal field parameters in stevens normalisation in mev_. the parameters were deduced from fitting to inelastic neutron spectra and the constraints on the wavefunctions of the lowest three energy levels of the quasi - cubic sites as described in the text . [ cols= " > , < , > , < " , ] the calculated magnetostriction fits the data well in the high field regions , but does not reproduce the low field behaviour , particular for the case where the field is applied along @xmath20 .
it may be that at low fields , some of the measured magnetostriction is due to domain rotation , which is not considered in the mean field model .
we have deduced a mean field model including up to four nearest neighbour dipolar exchange and quadrupolar interactions between the 5@xmath6 electrons of upd@xmath0 which is in generally good qualitative agreement with a broad range of experimental results .
the interactions between electrons on the quasi - cubic sites were deduced from resonant x - ray scattering measurements of the quadrupolar order parameters of each of the four low temperature ordered phases , and from the measured transition temperatures and critical fields .
the model was then used to calculate the high field magnetisation and magnetostriction , and the magnetic phase diagram up to 30 t. with an applied magnetic field in the basal plane , the calculations for the @xmath20 and @xmath23 directions agree with the data .
the calculated magnetisation with the field parallel to @xmath21 is a factor of 3 too large , however , and the calculated critical field in this direction is much smaller than that measured . despite this , the @xmath21-direction calculated magnetic phase diagrams qualitatively reproduces that measured . in conclusion ,
we have developed a nearest neighbour mean - field model which reproduces many of the main features observed in upd@xmath0 .
the authors thank jens jensen , gillian gehring , amanda gipson and helen walker for fruitful discussions .
m.d.l . thanks the uk engineering and physical sciences research council for a research studentship , and ucl graduate school for a project grant .
j.g.p . acknowledges assistance by g. chouteau and the support provided by the grenoble high magnetic field laboratory .
m.d.l . and m.r .
acknowledge the support of the european science foundation s cooperation in science and technology ( cost ) program , contract number cost - stsm - p16 - 02275 .
experiments performed at the nhmfl were supported by nsf cooperative agreement dmr-0084173 , by the state of florida , and the u.s .
department of energy .
support by the deustche forschungsgemeinschaft ( sfb463 ) , and the austrian fwf ( p17226 ) is gratefully acknowledged .
35ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/revmodphys.81.807 [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9601(01)00338 - 3 [ * * , ( ) ] link:\doibase 10.1143/jpsj.70.1857 [ * * , ( ) ] link:\doibase 10.1016/j.physb.2005.01.268 [ * * , ( ) ] link:\doibase 10.1143/jpsj.79.034703 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.024417 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.137203 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/20/39/395221 [ * * , ( ) ] link:\doibase 10.1016/0921 - 4526(94)00499-l [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/15/28/304 [ * * , ( ) ] link:\doibase 10.1143/jpsj.70.1731 [ * * , ( ) ] link:\doibase 10.1016/0304 - 8853(92)91395-a [ * * , ( ) ] link:\doibase 10.1103/physrevlett.87.057201 [ * * , ( ) ] link:\doibase 10.1107/s0365110x55001527 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/revmodphys.81.235 [ * * , ( ) ] link:\doibase 10.1016/j.jmmm.2006.10.520 [ * * , ( ) ] _ _ , @noop ph.d
. thesis , ( ) link:\doibase 10.1088/0953 - 8984/24/3/036002 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1016/s0304 - 8853(97)00993 - 1 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1063/1.1149009 [ * * , ( ) ] link:\doibase 10.1016/0921 - 4526(94)91855 - 4 [ * * , ( ) ] link:\doibase 10.1016/j.jmmm.2003.12.1394 [ * * , ( ) ] @noop link:\doibase 10.1088/0953 - 8984/4/48/021 [ * * , ( ) ] @noop _ _ ( , ) @noop in link:\doibase 10.1109/icnn.1995.488968 [ _ _ ] , vol .
link:\doibase 10.1016/0921 - 4526(93)90669-w [ * * , ( ) ] link:\doibase 10.1103/physrevb.60.r8430 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/22/2/022202 [ * * , ( ) ] link:\doibase 10.1080/00018730500037264 [ * * , ( ) ] in @noop _ _ , ( , , ) pp . | upd@xmath0 is known to exhibit four antiferroquadrupolar ordered phases at low temperatures .
we report measurements of the magnetisation and magnetostriction of single crystal upd@xmath0 , along the principal symmetry directions , in fields up to 33 t. these results have been combined with recent inelastic neutron and x - ray resonant scattering measurements to construct a mean field model of upd@xmath0 including up to fourth nearest neighbour interactions . in particular we find that anisotropic quadrupolar interactions must be included in order to explain the low temperature structures derived from the scattering data . |
constructing martingales with given marginal distributions has been an active area of research over the last decade ( e.g. @xcite ) .
( here and in the entire paper , marginal distributions ( also marginals ) refer to the 1-dimensional distributions . ) a condition for the existence of such martingales is given by kellerer @xcite ( see hirsch and roynette @xcite for a new and improved proof ) .
three constructions of markov martingales with pre - specified marginal distributions were given by madan and yor @xcite , namely the skorokhod embedding method , the time - changed brownian motion and the continuous martingale approach pioneered by dupire @xcite .
recently , hirsch _
et al_. @xcite gave six different methods for constructing martingales whose marginal distributions match those of a given family of probability measures .
they also tackle the tedious task of finding sufficient conditions to ensure that the chosen family is indeed increasing in the convex order , or as they coined it , a peacock . in this paper , we deal with a different , albeit related , scenario
. we do not start with a family of probability distributions , rather we start with a given martingale ( the existence of which is assumed ) and produce a large family ( as opposed to just a handful ) of new martingales having the same marginal distributions as the original process .
we say that these martingales `` mimic '' the original process .
this same task was undertaken in @xcite for the brownian motion .
it gave rise to the papers @xcite and @xcite ( who coined the term faked brownian motion ) .
albin @xcite and oleszkiewicz @xcite answered the question of the existence of a continuous martingale with brownian marginals .
however , their constructions yield non - markov processes .
et al_. @xcite then generalised albin s construction and produced a sequence of ( non - markov ) martingales with brownian marginals . in this paper , we extend the construction of @xcite to a much larger class of processes , namely self - similar markov martingales . before formulating a solution to this problem
we give a brief account on the origin and relevance of the mimicking question to finance , and more specifically to option pricing ; that is the pricing of a contract that gives the holder the right to buy ( or sell ) the instrument ( a stock ) at a future time @xmath2 for a specified price @xmath3 .
the theoretical valuation of an option is performed in such a way as not to allow arbitrage opportunities
arbitrage occurs when riskless trading results in profit .
the first fundamental theorem ( e.g. , @xcite , page 231 ) states that the absence of arbitrage in a market with stock price @xmath4 , @xmath5 , is essentially equivalent to the existence of an equivalent probability measure under which the stock price is a martingale .
( here without loss of generality , we let the riskless interest rate be zero . )
the second fundamental theorem ( e.g. , @xcite , page 232 ) implies that the arbitrage - free price of an option is given by @xmath6 $ ] , where the expectation is taken under the equivalent martingale measure . in the classical model of black and
scholes the stock price @xmath4 is given , under the martingale measure , by the exponential brownian motion @xmath7 .
the parameter @xmath8 is known as the volatility , and is assumed to be a constant .
the resulting expectation produces the well - known black scholes formula for option prices .
however , empirical evidence shows that in order to match the black
scholes formula to market prices of options one needs to vary @xmath8 . as a consequence it is natural to ask whether there exist alternative models of stock prices that result in a prescribed option pricing formula , such as black scholes .
finally , it is easy to see that the collective knowledge of @xmath9,k\geq0\}$ ] determines the distribution of @xmath10 or the marginal distribution , for example @xcite . therefore , if one wants to keep the option prices given by the original formula but without the limitations of the original process ( such as constant volatility ) one has to look for martingales ( to have the model arbitrage free ) with given marginals ( to keep the same option prices ) .
this question received much attention in the last ten years , see the pioneering work of madan and yor @xcite . throughout this paper , we assume that all processes are cdlg and progressively measurable
. we will use the notation @xmath11 to mean equal in distribution for random variables or equal in finite - dimensional distributions for processes , and this will be clear in the context . for a given random measure @xmath12 , the measure @xmath13 for @xmath14
is defined by @xmath15 .
we will also write @xmath16 $ ] to mean the measure defined by @xmath17 = \mathbb{e}[\int g(x ) m(\mathrm{d}x)]$ ] for any positive function @xmath18 .
we start with a martingale @xmath19 which is also a markov process , that is , for any bounded measurable function @xmath18 , @xmath20 = \mathbb{e}[g(z_t)|z_s]$ ] for @xmath21 , where @xmath22 denotes the natural filtration of @xmath23 .
we aim to construct new processes that will have the same marginal distributions as @xmath19 while retaining the martingale and markov properties .
we assume further that @xmath19 is self - similar , that is , there exists a ( strictly ) positive function @xmath24 such that @xmath25 for all @xmath14 . for @xmath19 to be non - trivial and stochastically continuous at @xmath26 , we must have ( see , e.g. , @xcite ) that @xmath27 for some @xmath28 , i.e. , @xmath29 ( which implies that @xmath30 ) .
we then say that @xmath19 is self - similar with exponent @xmath31 , or simply , @xmath31-self - similar .
denote the transition function of @xmath19 by @xmath32 for @xmath21 .
then the self - similarity of @xmath19 translates to the following scaling property on @xmath33 : @xmath34 if @xmath35 is the infinitesimal generator associated with @xmath33 , this scaling property is equivalent to @xmath36 where @xmath37 is the operator defined by @xmath38 . from this
, we see that @xmath39 in the following , we present a mimicking scheme to the process @xmath19 by randomising the transition functions .
we will see that this is equivalent to time - changing the process together with an appropriate scaling .
we then obtain some properties of the resulting processes and give some examples .
let @xmath19 be a markov process with transition function @xmath33 , which is self - similar with exponent . note that if @xmath19 is a martingale , then it has a cdlg version ;
if @xmath19 is not a martingale , there are conditions for a markov process to be cdlg . in this section ,
we construct new markov processes that possess the same marginal distributions as @xmath19 and show that the resulting processes are martingales under certain conditions .
[ lemmap ] let @xmath40 . for any @xmath41
$ ] and a measurable set @xmath42 , @xmath43 and @xmath44 here we prove only the second equality , the first one is proved similarly .
suppose that @xmath45 $ ] . by the scaling property
, we have @xmath46 notice that when @xmath47 or @xmath48 is @xmath49 , the two equalities in the lemma trivially hold true .
let @xmath50 be a family of probability distributions on the set @xmath51 $ ] , where @xmath52 , the dirac measure at @xmath53 .
suppose that for any bounded measurable function @xmath54 and @xmath55 , we have @xmath56 then @xmath57 defined as follows is a transition function , @xmath58 clearly , for each @xmath59 , @xmath60 is a probability measure and for each @xmath61 , @xmath62 is measurable in @xmath63 .
note also that @xmath64 .
next , using lemma [ lemmap ] , we obtain , for @xmath40 , @xmath65 and @xmath66 in other words , @xmath57 satisfies the chapman kolmogorov equations .
if @xmath67 depends on @xmath68 and @xmath69 only through @xmath70 ( i.e. @xmath71 ) , then the scaling property of @xmath33 carries over to @xmath57 : @xmath72 this follows immediately from the definition of @xmath57 and the scaling of @xmath33 .
let , for @xmath73 , @xmath74 be a random variable having distribution @xmath67 .
property ( [ e : propg ] ) is equivalent to the property that if , for @xmath55 , @xmath74 and @xmath75 are independent random variables , then @xmath76 .
further , if we let @xmath77 and write @xmath78 for the distributions of @xmath79 ( with @xmath80 ) , then property ( [ e : propg ] ) translates to the convolution identity @xmath81 as we seek to retain the scaling property of the original process , we assume that @xmath82 and immediately reduce property ( [ e : propg ] ) to @xmath83 the family @xmath84 defines a subordinator ( process with positive , independent and stationary increments ) and by lvy
khintchine it has laplace transforms of the form @xmath85 the function @xmath86 , known as the laplace exponent , takes the form @xmath87 with drift @xmath88 and lvy measure @xmath89 satisfying @xmath90 and @xmath91 .
conversely , to each @xmath86 ( i.e. , to each pair @xmath92 as above ) corresponds a convolution semigroup @xmath84 and in turn a family @xmath93 which satisfies property ( [ e : propg ] ) .
for details see , for example , @xcite , section 1.2 .
this ensures the existence of @xmath93 and a process with transition function @xmath57 .
[ t : transition ] let @xmath19 be a @xmath31-self - similar markov process . to each @xmath86 ,
laplace exponent of a subordinator , corresponds a @xmath31-self - similar markov process @xmath94 , starting from 0 and having the marginals of @xmath19 .
furthermore , if @xmath19 is a martingale and @xmath95 , then @xmath94 is also a martingale .
writing in terms of @xmath96 , @xmath21 , the new transition function @xmath97,\quad\quad s\le t\ ] ] can be seen as a randomisation of @xmath98 .
furthermore , the condition on @xmath99 for @xmath94 to be a martingale can be written as @xmath100 = ( s / t)^{\kappa}$ ] . by the kolmogorov extension theorem , there exists a markov process @xmath94 with transition function @xmath60 . as for the martingale property
, we first observe that @xmath101 so that @xmath95 translates to @xmath102 for @xmath103 .
then we have @xmath104 using the definition of @xmath57 and the martingale property of @xmath19 .
the process @xmath94 can also be obtained using subordination ( with a suitable scaling in the state space and an appropriate time - change ) .
the idea of subordination was suggested by bertoin in the context of brownian marginals .
however , subordination alone ( albeit with a suitable scaling in the state space ) is not sufficient . a further logarithmic change of time is needed .
this naturally creates an issue at 0 . to deal with this
, we follow oleszkiewicz @xcite .
[ t : timechange ] let @xmath105 be a subordinator independent of @xmath19 .
let , for @xmath106 , @xmath107 then the process @xmath108 has the same marginal distributions as @xmath109 and there exists a process @xmath110 such that for any @xmath111 , @xmath112 .
the process @xmath110 is a @xmath31-self - similar markov process with transition function @xmath113,\quad\quad s\le t.\end{aligned}\ ] ] moreover , @xmath94 is a martingale provided that @xmath19 is a martingale and @xmath114 = ( s / t)^{\kappa}$ ] .
since @xmath115 is measurable in @xmath116 and right - continuous in @xmath69 , it is measurable as a function of @xmath117 .
hence , for each @xmath118 , @xmath119 is a random variable . for @xmath14 ,
let @xmath120 and @xmath121 , so that @xmath122 and @xmath123
. then we have , for @xmath124 and with @xmath125 , @xmath126 for @xmath127 , let @xmath128 be the law of @xmath129 .
then the measures ( @xmath130 are consistent and by the kolmogorov extension theorem , there exists a process @xmath131 with finite - dimensional distributions ( @xmath132 .
a similar argument shows that @xmath133 , from which we deduce that @xmath131 is @xmath31-self - similar . as such
, @xmath94 extends by continuity to @xmath134 by letting @xmath135 .
the equality of marginal distributions of @xmath136 and @xmath19 follows from the scaling property of @xmath19 as @xmath137 for any fixed @xmath118 . using successively lemmas [ l : lamperti ] , [ l : genbochner ] , [ l : genproduct ] and [ l : gentimechange ] ( see the ) , we see that @xmath136 is markovian . by the scaling property of @xmath33 and the stationarity of subordinator @xmath138
, we obtain , for @xmath139 , the transition function of @xmath136 as @xmath140.\ ] ] notice that the transition function @xmath141 does not depend on @xmath47 and it is the same as @xmath57 defined earlier with @xmath142 being the distribution of @xmath143 .
the rest of the assertions of the proposition then follows immediately from theorem [ t : transition ] .
( alternatively , we can carry out the proof independently , without referring to @xmath57 , following oleszkiewicz @xcite . )
the process constructed in proposition [ t : timechange ] is identical ( in law ) to the process obtained in theorem [ t : transition ] with @xmath142 being the distribution of @xmath143 , or @xmath144 . for @xmath94 to be a martingale , @xmath145 in ( [ e : laplaceexp ] ) must be at most 1 , and is 1 if and only if @xmath146 ( @xmath147 and @xmath148 for any @xmath149 ) and @xmath150 .
in proposition [ t : timechange ] , we can not replace @xmath151 with a two - sided subordinator @xmath152 for @xmath153 , where @xmath154 and @xmath155 are independent subordinators .
this is because by doing that , we will not have independent increments . in particular , since @xmath156 , then for @xmath157 , the increment @xmath158 , where @xmath159 denotes the filtration generated by @xmath138 .
in this section , we obtain the infinitesimal generators of the process @xmath94 and display some of their path properties .
we will work within the martingale framework , that is , unless otherwise stated , we will assume that our initial process @xmath19 is a martingale and we will use a subordinator @xmath138 with drift @xmath145 , lvy measure @xmath89 and lapace exponent satisfying @xmath95 . [
t : generator ] suppose that @xmath19 has infinitesimal generator @xmath160 .
then the infinitesimal generator of the process @xmath94 is given by @xmath161 for @xmath162 differentiable and in the domain of @xmath35 .
first , from lemma [ l : lamperti ] , @xmath163 is time - homogeneous with generator @xmath164 and transition semigroup @xmath165 .
next , applying lemma [ l : genbochner ] , the generator of the process @xmath166 is @xmath167 then , let @xmath168 and using lemma [ l : genproduct ] , the generator of @xmath169 is @xmath170 since @xmath171 for @xmath172 .
finally , we time - change the process @xmath173 with @xmath174 to get @xmath175 .
thus , by lemma [ l : gentimechange ] , the generator of @xmath175 is @xmath176 due to a change of variable , the scaling property of @xmath33 and identity ( [ e : ssgen1 ] ) .
the generator of @xmath177 is established by noting that @xmath178 does not depend on @xmath47 . since @xmath19 is self - similar
, @xmath179 is in the domain of @xmath35 for all @xmath14 whenever @xmath162 is in the domain of @xmath35 .
therefore , @xmath162 is in the domain of @xmath180 , if @xmath162 is also differentiable .
note that when @xmath181 and @xmath182 , we recover the process @xmath19 and @xmath183 . for a measurable function @xmath162 ,
if there exists a measurable function @xmath18 such that for each @xmath69 , @xmath184 almost surely and @xmath185 is a martingale , then @xmath162 is said to belong to the domain of the extended infinitesimal generator of @xmath94 and the extended infinitesimal generator @xmath186 . if @xmath187 belongs to the domain of the extended infinitesimal generator of @xmath94 , then @xmath94 has predictable quadratic variation @xmath188 see examples in section [ s : examples ] for the computation @xmath189 in some specific cases .
suppose that @xmath19 is continuous in probability .
then the process @xmath94 is also continuous in probability , that is , for every @xmath69 , @xmath190 we have , for @xmath191 , @xmath192\\[-8pt ] & & { } + \mathbb{p } \biggl ( \bigl| s^{\kappa } \mathrm{e}^{-{\kappa}\zeta_{a+\ln s } } ( z_{\mathrm{e}^{\zeta_{a+\ln t } } } - z_{\mathrm{e}^{\zeta_{a+\ln s } } } ) \bigr| > \frac{c}{2 } \biggr).\nonumber\end{aligned}\ ] ] however , the first term @xmath193 which converges to 0 as @xmath194 , since @xmath138 is continuous in probability as a subordinator . to deal with the last term in ( [ e : contprob1 ] )
we first observe that a process that is continuous in probability does not jump at fixed points so that @xmath195 .
further , since @xmath19 is also continuous in probability , @xmath196 as soon as @xmath197 .
therefore , for @xmath198 , @xmath199 = 0,\end{aligned}\ ] ] where @xmath200 . if @xmath19 is continuous in probability with finite second moments and @xmath138 has no drift @xmath201 , then @xmath94 is a purely discontinuous martingale .
let @xmath202 so that @xmath203 for @xmath118 .
first , we observe that with probability one , @xmath19 does not jump at @xmath204 if @xmath205 jumps at @xmath69 . indeed ,
if @xmath206 is a countable set of points in @xmath207 , then , as @xmath19 is continuous in probability , @xmath208 .
let @xmath209 and @xmath210 where @xmath211 , then @xmath212 = 0,\ ] ] where @xmath213 denotes the filtration of @xmath205 .
taking @xmath47 to infinity , we obtain the desired result . to show that @xmath94 is purely discontinuous , that is , @xmath214 , we compute the sum of the square of jumps of @xmath94 . in general
, we have @xmath215 & = & \mathbb { e } \biggl [ \sum _ { \mathrm{e}^{-a}<s\le t } ( \delta x_s)^2 \mathbf{1}_{\delta u_s > 0 , \delta z_{(u_{s-})}\neq0 } \biggr ] \\ & & { } + \mathbb{e } \biggl [ \sum_{\mathrm{e}^{-a}<s\le t } ( \delta x_s)^2 \mathbf { 1}_{\delta u_s > 0 , \delta z_{(u_{s-})}=0 } \biggr ] \\ & & { } + \mathbb{e } \biggl [ \sum_{\mathrm{e}^{-a}<s\le t } ( \delta x_s)^2 \mathbf{1}_{\delta u_s = 0 } \biggr].\end{aligned}\ ] ] as @xmath19 is continuous in probability , the first term is zero due to the observation at the start of the proof .
we write @xmath216 $ ] for the third term . for the second term
, we have , on the set @xmath217 , @xmath218 let @xmath219 $ ] . since @xmath19 is @xmath31-self - similar , @xmath220 . as @xmath19 is a martingale , @xmath221=0 $ ] and @xmath222=\mathbb { e}[z_t^2]-\mathbb{e}[z_s^2]$ ]
thus , we obtain @xmath223 \\ & & \quad= s^{2\kappa } \bigl ( u_s^{-2\kappa } ( \theta_{u_s}-\theta_{(u_{s- } ) } ) + \theta_{(u_{s- } ) } \bigl(u_s^{-\kappa}-u_{s-}^{-\kappa } \bigr)^2 \bigr ) \\ & & \quad = 2 s^{2\kappa } \theta_1 \bigl ( 1 - u_s^{-\kappa}u_{s-}^{\kappa } \bigr).\end{aligned}\ ] ] since @xmath224 has probability one on the set @xmath225 , it follows that @xmath226 & = & \mathbb{e } \biggl [ \sum_{\mathrm{e}^{-a}<s\le t } 2 s^{2\kappa } \theta_1 \bigl ( 1 - u_s^{-\kappa}u_{s-}^{\kappa } \bigr ) \mathbf{1}_{\delta u_s>0 } \biggr ] \\ & = & \mathbb{e } \biggl [ \sum_{0<r\le a+\ln t } 2 \mathrm{e}^{2\kappa(r - a ) } \theta_1 \bigl ( 1 - \mathrm{e}^{-{\kappa}\delta\zeta_r } \bigr ) \mathbf{1}_{\delta\zeta _ r>0 } \biggr].\end{aligned}\ ] ] writing in terms of @xmath89 and @xmath86 , and using ( [ e : laplaceexp ] ) with @xmath227 , @xmath228 & = & \theta_1 \int_0^{a+\ln t } 2 \mathrm{e}^{2\kappa(r - a ) } \,\mathrm{d}r \int_{(0,\infty ) } \bigl(1-\mathrm{e}^{-z\kappa}\bigr ) \nu(\mathrm{d}z ) \\ & = & \theta_1 \bigl(t^{2\kappa}-\mathrm{e}^{-2a\kappa}\bigr ) ( 1-\beta).\end{aligned}\ ] ] adding those three terms and taking limit as @xmath229 , we have @xmath230 = \theta_1 t^{2\kappa } ( 1-\beta ) + \lim_{a\to\infty } l(a , t).\ ] ] since @xmath94 is square integrable on any finite interval if @xmath19 has finite second moments , it has quadratic variation with expectation @xmath231_t ] = \mathbb{e } [ x_t^2 ] = \mathbb { e } [ z_t^2 ] = t^{2\kappa } \theta_1 $ ] .
therefore , @xmath232 & = & \mathbb{e } \bigl [ [ x , x]_t \bigr ] - \mathbb{e } \biggl [ \sum_{0<s\le t , \delta x_s\neq0 } ( \delta x_s)^2 \biggr ] \\ & = & \theta_1 t^{2\kappa } \beta- \lim_{a\to\infty } l(a , t).\end{aligned}\ ] ] note that both @xmath233 $ ] and @xmath234 are non - negative .
thus , when @xmath235 , we must have @xmath236 , which gives @xmath233 = 0 $ ] and hence @xmath214 .
given any self - similar markov martingale @xmath19 with transition function @xmath33 and infinitesimal generator @xmath35 , we can mimic @xmath19 as per section [ s : scheme ] .
we construct a new markov martingale @xmath94 that has the same marginal distributions as @xmath19 and possesses the same self - similarity @xmath19 enjoys from each @xmath138 .
we assume that the subordinator @xmath138 has drift @xmath145 and lvy measure @xmath89 with laplace exponent @xmath86
satisfying @xmath237 , or @xmath238 = ( s / t)^\kappa$ ] . examples of subordinators include poisson process , compound poisson process with positive jumps , gamma process and stable subordinators .
for example , we can take @xmath138 as a poisson process with rate @xmath239 to satisfy @xmath227 . in the following ,
we provide some examples of mimicking with the infinitesimal generators and the predictable quadratic variations computed explicitly to have a better understanding of the processes .
we finish this section with a discussion on modifying our construction to mimic some brownian related martingales and its limitation . for any @xmath240 , define the process @xmath23 by @xmath241 , where @xmath242 is a brownian motion .
note that @xmath243 this is a gaussian process with zero mean and covariance function @xmath244 .
it is a markov process and also a martingale .
moreover , it is @xmath245-self - similar ( @xmath246 ) since for all @xmath14 , @xmath247 a key aspect of the construction in @xcite is the following representation of the mimic @xmath177 when @xmath248 : @xmath249 where @xmath74 has distribution @xmath142 , @xmath250 is standard normal and , @xmath74 , @xmath250 and @xmath251 are independent .
this representation extends to the case of other gaussian continuous martingales .
in fact , it also extends to the case of stable processes see proposition [ p : stablerep ] .
[ p : gauscontmartrep ] with @xmath252 , the mimic @xmath110 has the representation @xmath253 where @xmath254 , @xmath255 and , @xmath74 , @xmath250 and @xmath251 are independent . since @xmath19 and @xmath138 have independent and stationary increments , for @xmath256 , @xmath257 where @xmath258 is an independent copy of @xmath138 , and @xmath250 is a random variable distributed as @xmath259 .
note that this representation holds also for @xmath260 .
knowing that @xmath19 has generator @xmath261 for @xmath262 , we can compute the generator of @xmath94 following theorem [ t : generator ] and obtain , for @xmath263 , @xmath264 taking @xmath187 , then equation ( [ e : quadvar ] ) , along with routine calculations and equation ( [ e : laplaceexp ] ) , gives the following result .
the predictable quadratic variation of @xmath94 is @xmath265 since @xmath266 , we can also write @xmath267 when @xmath248 , @xmath19 is a brownian motion and our results agree with @xcite .
a process @xmath268 is a squared bessel process of dimension @xmath269 , for some @xmath270 , if it satisfies @xmath271 , where @xmath42 denotes a brownian motion .
the squared bessel process @xmath272 started at 0 is a continuous markov process satisfying the self - similarity with @xmath273 .
let @xmath274 .
then @xmath23 is a 1-self - similar markov process and satisfies the sde @xmath275 note that @xmath272 is stochastically dominated by the square of the norm of an @xmath276-dimensional brownian motion , where @xmath277 , thus @xmath278\leq nt$ ] and @xmath279\leq2nt^2 $ ] .
it follows that @xmath19 is a true martingale ( see , e.g. , @xcite , theorem 7.35 ) .
the infinitesimal generator of @xmath19 is @xmath280 , @xmath262 , thus , that of @xmath94 is @xmath281 the predictable quadratic variation of @xmath94 is @xmath282 where @xmath86 is the laplace exponent of @xmath138 . from @xmath283
, we have for @xmath284 , @xmath285 and taking conditional expectation ( @xmath286 is a true martingale see above ) , @xmath287 = z_{t\mathrm{e}^{-u}}^2 + 4t\bigl(1-\mathrm{e}^{-u}\bigr)z_{t\mathrm{e}^{-u } } + 2\delta t^2\bigl(1-\mathrm{e}^{-2u}\bigr).\ ] ] thus , using equation ( [ e : laplaceexp ] ) we obtain , with @xmath187 , @xmath288 however @xmath289 .
the result then follows from equation ( [ e : quadvar ] ) .
note that if @xmath181 , @xmath148 for any @xmath149 and we recover @xmath290 .
suppose @xmath23 is an @xmath291-stable process with @xmath292 .
then @xmath19 is a markov process and @xmath293 that is , @xmath19 is @xmath31-self - similar with @xmath294 .
it is a lvy process with lvy triplet @xmath295 , where @xmath296 for some positive constants @xmath180 and @xmath42 .
assume that @xmath19 is a martingale , in which case the lvy triplet must satisfy @xmath297 [ p : stablerep ] the mimic @xmath110 has the representation @xmath298 where @xmath254 , @xmath255 and , @xmath74 , @xmath250 and @xmath251 are independent .
see proposition [ p : gauscontmartrep ] .
the stable process @xmath19 has infinitesimal generator @xmath299 where @xmath300 . to distinguish the characteristics of @xmath138 from that of @xmath19
, we add the subscript @xmath138 to the drift and lvy measure of the subordinator @xmath138
. then theorem [ t : generator ] gives the infinitesimal generator of @xmath94 for @xmath263 as @xmath301 let @xmath1 be an integrable , symmetric random variable ( i.e. @xmath302 ) and @xmath242 be a brownian motion independent of @xmath1 . following and extending @xcite , page 283 , for any @xmath31 let @xmath303 and @xmath304 . then @xmath23 is a markov martingale such that for each @xmath262 , @xmath305 .
moreover , @xmath19 is @xmath31-self - similar . indeed , using the brownian motion @xmath306 , we have
@xmath307 it follows that @xmath308 and @xmath309 .
hence , @xmath310 since @xmath19 is a markov process with transition semigroup @xmath311 where @xmath312 is the cumulative distribution function of @xmath1 , it has infinitesimal generator @xmath313 using theorem [ t : generator ] , we obtain the infinitesimal generator of the mimic @xmath94 , for @xmath263 , @xmath314 the predictable quadratic variation of @xmath94 is @xmath315 let @xmath187 . since @xmath316 and @xmath317 , the result follows immediately from equation ( [ e : quadvar ] ) .
note that @xmath318 .
the predictable quadratic variations of @xmath94 and @xmath19 are given by the same functional of the process .
now we discuss how we can ( and can not ) alter our martingale condition to mimic some brownian related processes , including the martingales associated with the hermite polynomials and the exponential martingale of brownian motion .
consider the hermite polynomials @xmath319 which are defined by @xmath320 equivalently , @xmath321 .
let @xmath322 and @xmath323 .
then @xmath324 , where @xmath42 denotes a brownian motion , is a local martingale for every @xmath276 since @xmath325 is space time harmonic , that is , @xmath326 .
take @xmath327 , the process @xmath328 is markovian and 1-self - similar , thus can be mimicked using our mimicking scheme with any subordinator that satisfies @xmath289 . for @xmath329 ,
@xmath324 is @xmath330-self - similar , but it is not markovian ( see @xcite ) .
so we are not able to mimic this process by a direct application of the method described above
. however , a slight modification of our construction proves sufficient to achieve our aim .
let @xmath177 be a mimic of the brownian motion as in section [ s : gauscontmart ] with @xmath248 , but without the requirement that @xmath177 be a martingale .
then we have the following result .
[ p : hermite ] for each @xmath276 , the process @xmath331 has the same marginal distributions as @xmath324 and is a martingale if and only if @xmath332 , or @xmath333 = ( s / t)^{n/2}$ ] .
it is obvious that @xmath331 and @xmath324 have the same marginals .
here we prove the martingale condition .
the transition function of @xmath94 is @xmath334 where @xmath33 is the transition function of @xmath42 and @xmath67 is the distribution of @xmath335 . writing @xmath336 as the natural filtrations of @xmath331 and
since @xmath324 is a martingale , we have @xmath337 & = & \mathbb{e } \biggl[\int h_n(y , t)\widetilde{p}(s , t , x_s,\mathrm{d}y ) \big|\mathcal { f}_s \biggr ] \\ & = & \mathbb{e } \biggl[\int h_n \bigl((t / s)^{1/2}r^{1/2}x_s , rt \bigr ) g_{s , t}(\mathrm{d}r ) \big|\mathcal{f}_s \biggr].\end{aligned}\ ] ] using that @xmath338 for any @xmath339 , we then obtain @xmath340 = ( t / s)^{n/2 } h_n(x_s , s ) \mathbb { e } \bigl[\mathrm{e}^{-(n/2)\zeta_{\ln(t / s ) } } \bigr].\ ] ] hence , @xmath331 is a martingale if and only if @xmath341 = ( s / t)^{n/2}$ ] .
therefore , in order to mimic the process @xmath324 , we can mimic @xmath342 , with the martingale requirement @xmath343 changed to @xmath344 , and then apply the function @xmath325 to the resulting process .
it is of interest to ask whether the above trick extends to other space time harmonic functions .
in particular , could this enable us to mimic the geometric brownian motion @xmath345 .
unfortunately , this is not the case .
in fact , @xmath325 , @xmath346 , are the only analytic functions for which this trick works .
suppose that @xmath347 and there exists @xmath348 such that @xmath349 , in other words , @xmath350 is analytic on the set @xmath351 .
suppose further that @xmath352 is space time harmonic , so that @xmath353 is a martingale .
suppose that @xmath177 mimics @xmath342 with the martingale requirement replaced with @xmath354 for a positive integer @xmath355 .
then @xmath356 has the same marginals as @xmath353 and is a martingale if and only if @xmath357 where @xmath358 is the @xmath355th hermite polynomial . for @xmath356 to be a martingale , we must have for any @xmath63 and @xmath21 , @xmath359 that is , @xmath360 = h(x , s)$ ] for any @xmath63 , @xmath68 and @xmath361 .
letting @xmath362 and @xmath363 , this is equivalent to @xmath364 = h(x , s)$ ] for all @xmath63 and @xmath68 , or @xmath365
= \sum _ { m , n=0}^\infty a_{m , n } x^ms^n.\ ] ] therefore , for all @xmath366 , @xmath276 and @xmath367 , we must have @xmath368 = a_{m , n}$ ] .
thus , either @xmath369 or @xmath370=1 $ ] . recall that @xmath371 = \mathbb{e}[\exp(\lambda t-\lambda \zeta_t ) ] = \exp(-t(\psi(\lambda)-\lambda))$ ] and there is at most one @xmath149 satisfying @xmath148 .
now , choose @xmath372 such that @xmath373=1 $ ] for a @xmath374 . then , for all @xmath375 such that @xmath376 , @xmath369 .
therefore , @xmath377 furthermore , @xmath378 by pluciska @xcite and fitzsimmons @xcite , @xmath379 is the @xmath355th hermite polynomial .
let @xmath177 be any mimic of @xmath342 in the sense of section [ s : scheme ] but without the martingale requirement .
although the process @xmath380 has the same marginal distributions as @xmath345 , it is not a martingale unless @xmath381 , in which case @xmath382 .
[ l : gentimechange ] let @xmath383 be a markov process with infinitesimal generator @xmath384 and @xmath385 be a deterministic , differentiable , increasing function in @xmath69 with derivative @xmath386 .
then the time - changed process @xmath387 is also a markov process with infinitesimal generator @xmath388 .
furthermore , if @xmath162 is in the domain of @xmath180 , then @xmath162 is in the domain of @xmath389 .
let @xmath390 be the filtration of @xmath391 and @xmath392 be the filtration of @xmath393 .
for any bounded measurable function @xmath18 , we have , for @xmath21 , @xmath394 = \mathbb { e}\bigl[g(y_{c_t})|\widetilde{\mathcal{f}}_s \bigr ] = \mathbb { e}\bigl[g(y_{c_t})|\mathcal{f}_{c_s}\bigr ] = \mathbb{e}\bigl[g(y_{c_t})|y_{c_s}\bigr ] = \mathbb{e}\bigl[g ( \widetilde{y}_t)|\widetilde{y}_s\bigr].\ ] ] for @xmath69 where the function @xmath395 is strictly increasing , the infinitesimal generator of @xmath396 is @xmath397 - f(x)}{c_u - c_t } \frac{c_u - c_t}{u - t } = a_{c_t}f(x)c'_t.\ ] ] if @xmath398 in a small neighbourhood of @xmath69 , then @xmath399 and @xmath400 let @xmath383 be a markov process with infinitesimal generator @xmath384 and @xmath385 be a deterministic , differentiable function in @xmath69 with derivative @xmath386 and @xmath401 for any @xmath69 . then the process @xmath402 is also a markov process and has generator @xmath403 where @xmath37 is an operator defined by @xmath404 .
furthermore , if @xmath405 is in the domain of @xmath180 for any @xmath395 and @xmath162 is differentiable , then @xmath162 is in the domain of @xmath389 .
let @xmath390 be the filtration of @xmath391 and @xmath392 be the filtration of @xmath393 .
let @xmath406 be a function such that @xmath407 .
since @xmath406 is one - to - one , @xmath408 .
therefore , for any bounded measurable function @xmath18 , we have @xmath394 = \mathbb { e}\bigl[g\circ h_t ( y_t)|\widetilde { \mathcal{f}}_s\bigr ] = \mathbb{e}\bigl[g \circ h_t ( y_t)|\mathcal{f}_s\bigr ] = \mathbb{e}\bigl[g\circ h_t(y_t)|y_s\bigr ] = \mathbb { e}\bigl[g ( \widetilde{y}_t)|\widetilde{y}_s\bigr].\ ] ] the infinitesimal generator of @xmath393 is @xmath409 - \pi_{c_u}f\biggl(\frac{1}{c_t}x\biggr ) + f\biggl(\frac{c_u}{c_t}x\biggr ) - f(x)\biggr)\big/(u - t ) \\ & = & a_t\pi_{c_t}f \biggl(\frac{1}{c_t}x \biggr ) + \frac{c'_t}{c_t}xf'(x).\end{aligned}\ ] ] [ l : lamperti ] suppose @xmath23 is a @xmath31-self - similar markov process .
suppose @xmath98 and @xmath160 are , respectively , the transition function and infinitesimal generator of @xmath19 .
let @xmath410 .
then @xmath411 is a time - homogeneous markov process with transition semigroup @xmath412 and infinitesimal generator @xmath413 furthermore , if @xmath162 is in the domain of @xmath35 and differentiable , then it is in the domain of @xmath414 . by the scaling property of @xmath33
, we have @xmath415 it follows that @xmath416 is time - homogeneous and @xmath417 .
the generator of @xmath416 can be obtained by applying lemma [ l : gentimechange ] and lemma [ l : genproduct ] , and seeing that @xmath418 from equation ( [ e : ssgen ] ) with @xmath419 and @xmath420 .
note that for all @xmath14 , @xmath421 is in the domain of @xmath35 by the scaling property of @xmath19 .
thus , writing @xmath422 as the generator of @xmath423 , @xmath421 is in the domain of @xmath422 by lemma [ l : gentimechange ] .
[ l : genbochner ] suppose @xmath424 is a time - homogeneous markov process with semigroup @xmath425 and generator @xmath35 , and @xmath426 is a subordinator independent of @xmath427 with drift @xmath145 and lvy measure @xmath89 .
set @xmath428 .
then the process @xmath429 is a time - homogeneous markov process with generator @xmath180 where @xmath430 furthermore , if @xmath162 is in the domain of @xmath35 , then it is in the domain of @xmath180 .
if @xmath431 has zero drift , then @xmath391 is a pure jump process .
see sato @xcite , theorem 32.1 .
the authors are grateful to two referees and an associate editor for their careful reading of an earlier version of the paper and a number of suggestions and improvements . | we construct a family of self - similar markov martingales with given marginal distributions
. this construction uses the self - similarity and markov property of a reference process to produce a family of markov processes that possess the same marginal distributions as the original process .
the resulting processes are also self - similar with the same exponent as the original process .
they can be chosen to be martingales under certain conditions . in this paper , we present two approaches to this construction , the transition - randomising approach and the time - change approach .
we then compute the infinitesimal generators and obtain some path properties of the resulting processes .
we also give some examples , including continuous gaussian martingales as a generalization of brownian motion , martingales of the squared bessel process , stable lvy processes as well as an example of an artificial process having the marginals of @xmath0 for some symmetric random variable @xmath1 . at the end , we see how we can mimic certain brownian martingales which are non - markovian . ./style / arxiv - general.cfg , |
theoretical and experimental investigations in low - dimensional systems lead to the opportunity of visualizing various novel quantum mechanical effects @xcite in a tunable way .
persistent current being one such exotic quantum mechanical phenomenon observed in normal metal mesoscopic rings and nanotubes pierced by aharonov - bohm ( ab ) flux @xmath3 .
prior to its experimental evidence , the possibility of a non - decaying current in normal metal rings was first predicted by bttiker , imry and landauer @xcite in a pioneering work , and , in the sub - sequent years theoretical attempts were made @xcite to understand the actual mechanism behind it .
the experimental realization of this phenomenon of non - decaying current in metallic rings / cylinders has been established quite in late .
it has been first examined by levy _
et al . _
@xcite and later many other experiments @xcite have confirmed the existence of non - dissipative currents in such quantum systems . although the studies involving simple mesoscopic rings have already generated a wealth of literature there is still need to look deeper into the problem to address several important issues those have not yet been explored , as for example the understanding of the behavior of persistent current in multiply connected quantum network , specially in presence of disorder .
it is well known that in presence of random site - diagonal disorder in an one - dimensional ( @xmath4d ) mesoscopic ring all the energy levels are localized @xcite , and accordingly , the persistent current gets reduced enormously in presence of disorder compared to that of an ordered ring .
but there are some @xmath4d disordered systems which support extended eigenstates along with the localized energy levels , and these materials may provide several interesting issues , mainly to provide a localization to delocalization transition and vice versa .
for example , in a pioneering work dunlap _
et al . _
@xcite have shown that even in 1d disordered systems extended eigenstates are possible for certain kind of topological correlations among the atoms .
they have proposed that any physical system which can be described by the random dimer model should exhibit the transmission resonances and a huge enhancement in the transmission takes place when the fermi level coincides with the unscattered states . in a consecutive
et al . _
@xcite have argued that the random dimer model can also be used to explain the insulator to metal transition in polyaniline as a result of the movement of the fermi level to extended region .
later , several other works @xcite have also been carried out in such type of materials to exhibit many important physical results .
the existence of localized energy eigenstates together with the extended states in a simple ring geometry has been explored in some recent works by jiang _
et al . _
they have analyzed the nature of these states by evaluating persistent current and the wave amplitudes at different sites of the ring . in these systems
localized states appear by virtue of disorder .
but , in our present work we make an attempt to establish localized eigenstates , even in the absence of disorder , along with extended states simply by considering the effect of topology of the system . to the best of our knowledge
, this behavior has not been addressed earlier in the literature .
here we consider a three - arm mesoscopic ring in which two outer arms are subjected to binary alloy lattices and the middle one contains identical lattice sites , and , we show that due to the presence of the middle arm quasi - localized energy eigenstates are observed within the band of extend regions .
it leads to the possibility of getting switching action from a high conducting state to a low one and vice versa as a result of the movement of the fermi level .
we illustrate this behavior by studying persistent current in the quantum network and explore the nature of energy eigenstates in terms of the probability amplitude in different lattice sites of the geometry .
our present analysis can be utilized to study magnetic response in any complicated quantum network and we believe that this work offers an excellent opportunity to study the simultaneous effects of topology and the magnetic fields threaded by two sub - rings in our three - arm ring system .
with an introduction in section i we organize the paper as follows . in section ii , first we present the model , then describe the theoretical formalism which include the hamiltonian and the formulation of persistent currents in individual branches of the network . in section iii we analyze the results and finally in section
iv we draw our conclusions .
let us refer to fig .
[ ring ] . a three - arm mesoscopic ring where the upper and lower sub - rings are threaded by ab fluxes @xmath1 and @xmath2 , respectively .
the outer arms are subjected to binary alloy lattices ( consisting of a and b types of atoms ) and the middle arm contains identical lattice sites ( atomic sites labeled by c ) except those on the boundaries .
the filled colored circles correspond to the positions of the atomic sites . within a tight - binding framework the hamiltonian for such a network reads as , @xmath5 where , @xmath6 represents the site energy for the outer arms , while for the middle arm it is assigned by @xmath7 . in the outer ring @xmath8 or @xmath9 alternately so that it forms a binary alloy . on the other hand , @xmath10 for the atomic sites
those are referred by c atoms . @xmath11 and @xmath12 are the nearest - neighbor hopping integrals in the outer and middle arms , respectively . due to
the presence of magnetic fluxes @xmath1 and @xmath2 in two sub - rings , phase factors @xmath13 and @xmath14 appears into the hamiltonian .
they are expressed as follows : @xmath15 and @xmath16 . here
the fluxes are measured in units of the elementary flux - quantum @xmath17 ( @xmath18 ) , and , @xmath19 , @xmath20 and @xmath21 represent the total number of single bonds ( each single bond is formed by connecting two neighboring lattice sites ) in the upper , middle and lower arms , respectively .
it reveals that @xmath22 number of atomic sites in the quantum network .
@xmath23 ( @xmath24 ) corresponds to the creation ( annihilation ) operator of an electron at the @xmath25 site , and , a similar definition also goes for the atomic sites @xmath26 . in the second quantized notation
the general expression of charge current operator is in the form @xcite , @xmath27 here , @xmath28 is the length of the arm in which we are interested to calculate the current and @xmath29 represents the nearest - neighbor hopping strength .
the nearest - neighbor hopping strength ( @xmath29 ) is equal to @xmath11 for the outer arms , while for the middle arm it becomes identical to @xmath12 .
therefore , for a particular eigenstate @xmath30 the persistent current becomes , @xmath31 , where @xmath32 . here
s are the wannier states and @xmath34 s are the corresponding coefficients
. following the above relations , now we can write down the expressions for persistent currents in the individual branches for a given eigenstate @xmath30 .
they are as follows .
+ 0.1 cm for the upper arm : @xmath35 where , summation over @xmath36 spans from @xmath4 to @xmath19 .
0.1 cm in the case of middle - arm : @xmath37 here , the net contribution comes from @xmath20 bonds .
0.1 cm finally , for the case of lower arm : @xmath38 in this case the net contribution comes from the lower bonds .
the lattice constant @xmath39 is set equal to @xmath4 . at absolute zero temperature ( @xmath40k ) , the net persistent current in any branch of the quantum network for a particular electron filling can be obtained by taking sum of the individual contributions from the energy levels with energies less than or equal to fermi energy @xmath41 .
therefore , for @xmath42 electron system total persistent in any branch becomes @xmath43 , where @xmath44 , @xmath45 and @xmath28 , for the upper , middle and lower arms , respectively .
once @xmath46 , @xmath47 and @xmath48 are known , the net persistent current for the full network can be easily obtained simply adding the contributions of the individual arms , and hence the total current is given by @xmath49 .
the net persistent current ( @xmath50 ) can also be determined in some other ways as available in the literature .
most probably the easiest way of calculating persistent current is to take first order derivative of ground state energy with respect to magnetic flux @xcite .
however in this method it is not possible to find the distribution of persistent current in individual arms of the network with a high degree of accuracy . on the other hand in our present scheme , the so - called second - quantized approach ,
there are certain advantages compared to other available procedures .
firstly , we can easily calculate persistent current in any branch of a network .
secondly , the determination of individual responses in separate arms provides much deeper insight to the actual mechanism of electron transport in a transparent way . in the present work
we investigate all the essential feature of magneto - transport at absolute zero temperature and choose the units where @xmath51 . throughout the numerical work we set @xmath52 and measure the energy scale in unit of @xmath11 .
we first start with a perfect quantum system where @xmath53 and @xmath9 and @xmath54 are all identical to each other and we set @xmath55 . to have a clear idea about the magnetic response of the model quantum system , first we illustrate the behavior of energy spectra as a function of flux @xmath1 for different values of flux @xmath2 threaded by the lower sub - ring .
the results are presented in fig .
[ energy1 ] , where ( a ) and ( b ) correspond to @xmath56 and @xmath57 , respectively . in the absence of flux @xmath2 ,
energy levels near the edges of the spectrum become more dispersive than those lying in the central region ( see fig .
[ energy1](a ) ) and near the center of the spectrum the energy levels are almost non - dispersive with respect to flux @xmath1 .
this feature implies that the persistent current amplitude becomes highly sensitive to the electron feeling i.e. , the fermi energy @xmath41 of the system , since the current is directly proportional to the slope of the energy levels @xcite .
the situation becomes much more interesting when we add a magnetic flux in the lower sub - ring . here
, the energy levels near the central region of the spectrum becomes more dispersive in nature than the energy levels near the edges ( fig .
[ energy1](b ) ) , and it increases gradually with flux @xmath2 , which gives a possibility of getting higher current amplitude with increasing the total number of electrons @xmath42 in the system . a similar kind of energy spectrum is also observed if we plot the energy levels as a function of flux @xmath2 instead of @xmath1 , keeping @xmath1 as a constant .
all these energy levels exhibit @xmath17 ( @xmath58 , in our choice of units @xmath51 ) flux - quantum periodicity .
thus , for such a simple quantum network persistent current amplitude might be regulated for a particular filling simply by tuning the magnetic flux threaded by anyone of two such sub - rings , and , its detailed descriptions are available in the sub - sequent parts . in fig .
[ current1 ] we present the variation of persistent current in individual arms of the three - arm quantum network as a function of flux @xmath1 for some fixed values of @xmath2 .
the panels from the top correspond to the results for the upper , middle and lower arms , respectively , and in all these cases the current is determined for the half - filled band case i.e. , @xmath59 . the left column represents the current for @xmath56 , and the right column gives the current when @xmath2 is fixed at @xmath57 . from the spectra we notice that in some cases current shows continuous like behavior while in some other cases it exhibits saw - tooth like nature as a function of flux @xmath1 threaded by the upper sub - ring .
this saw - tooth or continuous like feature solely depends on the behavior of the ground state energy for a particular filling ( @xmath42 ) .
it is to be noted that in a conventional ordered ab ring we always get saw - tooth like behavior of persistent current irrespective of the filling of the electrons @xcite . in the saw - tooth variation
a sudden change in direction of persistent current takes place across a particular value of magnetic flux which corresponds to a phase reversal from the diamagnetic nature to the paramagnetic one or vice versa . in our three - arm geometry
we also observe that though the current in the upper arm or in the lower arm is not so sensitive to the flux @xmath2 , but the current amplitude in the middle arm changes remarkably , even an order of magnitude , in presence of flux @xmath2 , which leads to a net larger current since the total current is obtained by adding the contributions from the individual arms . to explore the filling dependent behavior of persistent current , in fig .
[ currentne ] we display persistent currents for three different arms as a function of flux @xmath1 for a typical value of @xmath2 .
here , @xmath2 is set at @xmath57 .
the red , green and blue lines represent the currents for @xmath60 , @xmath61 and @xmath62 , respectively .
the current in different arms shows quite a complex structure which strongly depends on the electron filling as well as magnetic flux @xmath2 . in all these cases
persistent current provides @xmath17 flux - quantum periodicity , like a traditional single - channel mesoscopic ring or a multi - channel cylinder .
now we focus our attention to the geometry where site energies in the outer arms are no longer identical to each other i.e. , @xmath63 . in this case energy spectrum
gets modified significantly compared to the previous one where site energies are uniform ( @xmath64 ) . to illustrate it in fig .
[ energy2 ] we plot the energy - flux characteristics for a three - arm quantum network considering @xmath65 , where ( a ) and ( b ) correspond to the identical meaning as given in fig .
[ energy1 ] . since the upper and lower arms of the network are subjected to the binary alloy lattices we get two sets of discrete energy levels spaced by a finite gap around @xmath66 ( see fig .
[ energy2 ] ) .
quite interestingly we see that the energy levels near the two extreme edges of the spectrum are more dispersive in nature than those situated along the inner region . with increasing the difference in site energies ( @xmath67 )
, we get more less dispersive energy levels in the inner region and for large enough value of @xmath67 these levels become almost non - dispersive and they practically contribute nothing to the current . thus , for such a system a mixture of quasi - extended and quasi - localized energy levels are found out and it can provide a very large or almost zero current depending on the electron filling . for a very large system size , the energy separation between two successive levels in each set of discrete energy levels gets reduced and we get two quasi - band of energies separated by a finite gap , where the gap is controlled by the parameter values .
it is important to note that , unlike the previous one ( fig .
[ energy1 ] ) , for this case the energy spectrum is not so sensitive to flux @xmath2 ( fig .
[ energy2 ] ) .
the presence of c - type of atoms in the middle arm which divides the binary alloy ring into two sub - rings is responsible for the existence of quasi - localized energy levels near the inside edges of two quasi - band of energies .
thus we get more non - dispersive energy levels with increasing the length of the middle arm .
the existence of nearly extended and localized states becomes much more clearly visible from our current - flux spectra . as illustrative example , in fig .
[ current2 ] we display the variation of persistent current in individual arms including the total current of a three - arm ring considering @xmath65 for the quarter - filled ( @xmath68 ) band case . the flux @xmath2 is set equal to @xmath57 . from the spectra it is clearly observed that the current in each arm provides a non - zero value ( figs .
[ current2](a)-(c ) ) , and accordingly , the system supports a finite current as shown in fig .
[ current2](d ) .
the situation becomes completely opposite when the filling factor is changed .
quite remarkably we notice that persistent current almost vanishes in three separate branches which provides almost vanishing net current in the half - filled band case .
the results are illustrated in fig .
[ current3 ] , where ( a)-(d ) correspond to the identical meaning as in fig .
[ current2 ] .
the vanishing nature at half - filling and the non - vanishing behavior of current when the system is quarterly filled can be easily understood from the following argument .
the total current in any branch or in the complete system mainly depends on the contributions coming from the higher occupied energy levels , while the contributions from the other occupied energy levels cancel with each other .
therefore , for the quarter - filled band case , the net contribution comes from the energy levels which are quasi - extended in nature and a non - zero current appears . on the other hand , for the half - filled band case , the net contribution arises from the levels those are almost localized , and hence , nearly vanishing current is obtained .
thus , we can emphasize that the three - arm ring leads to a possibility of getting high - amplitude to low - amplitude ( almost zero ) persistent current simply by tuning the filling factor @xmath42 i.e. , the fermi energy @xmath41 , and , hence the network can be used as a mesoscopic switch .
the high - conducting to low - conducting switching action with the change of electron filling @xmath42 in our topology can also be very well explained from the spectrum given in fig .
[ drude ] , where we measure the conducting nature by calculating charge stiffness constant , the so - called drude weight ( @xmath69 ) , in accordance with the idea originally put forward by kohn @xcite . the drude weight for the system
can be easily determined by taking the second order derivative of the ground state energy for a particular filling with respect to flux @xmath1 ( @xmath70 ) threaded by the ring @xcite .
kohn has shown that @xmath69 decays exponentially to zero for an insulating system , while it becomes finite for a conducting system . a nice feature of the result shown in fig .
[ drude ] is that , the charge stiffness constant almost drops to zero around the half - filled region which reveals the insulating phase , while away from this region it ( @xmath69 ) has a finite non - zero value that indicates a conducting nature .
this feature corroborates the findings presented in figs .
[ current2 ] and [ current3 ] . to ensure the extended or localized nature of energy eigenstates ,
finally we demonstrate the variation of probability amplitude ( p.a . ) of the eigenstates as a function site index @xmath71 .
the probability amplitude of getting an electron at any site @xmath71 for a particular eigenstate @xmath30 is obtained from the factor @xmath72 .
here we analyze the localization behavior for two different energy eigenstates , viz , @xmath73-st and @xmath74-nd states .
for the first one the energy is located well inside a quasi - band , while for the other the energy is placed at the edge of this band .
the results are given in fig .
[ probability ] for a three - arm ring with @xmath75 atomic sites .
the red dashed lines are used to separate the three distinct regions of the network . in fig .
[ probability](a ) we present the probability amplitudes of the 21st eigenstates and see that the probability amplitude becomes finite for any site @xmath71 which indicates that the energy eigenstate is quasi - extended . while , for the other state ( @xmath74-nd ) the probability amplitude almost vanishes at every site of the upper and lower arms of the network . only at the atomic sites of the middle arm we have finite probability amplitudes .
this state does not contribute anything to the current and we can refer the state as a localized one .
to summarize , in the present work we have explored the magneto - transport properties of a @xmath0 shaped three - arm quantum ring in the non - interacting electron framework .
the upper and lower sub - rings of the network are threaded by magnetic fluxes @xmath1 and @xmath2 , respectively .
we have used a single - band tight - binding hamiltonian to illustrate the model quantum system , where the outer arms are subjected to the binary alloy lattices and the middle arm has identical lattice sites . in the absence of the middle arm ,
all the energy eigenstates are extended , but the inclusion of the middle arm produces some quasi localized states within the band of extended states and provides a possibility of getting a high conducting state to the low conducting one upon the movement of the fermi energy .
thus , the system can be used a mesoscopic switch .
we have verified the switching action from high- to low - conducting state and vice versa by investigating the persistent current and charge stiffness constant in the network for different band fillings .
we have numerically computed both the total current and the individual currents in separate branches by using second - quantized approach .
we hope our present analysis may be helpful for studying magneto - transport properties in any complicated quantum network .
finally , we have also examined the nature of the energy eigenstates in terms of the probability amplitude in different sites of the geometry . | we investigate magneto - transport properties of a @xmath0 shaped three - arm mesoscopic ring where the upper and lower sub - rings are threaded by aharonov - bohm fluxes @xmath1 and @xmath2 , respectively , within a non - interacting electron picture .
a discrete lattice model is used to describe the quantum network in which two outer arms are subjected to binary alloy lattices while the middle arm contains identical atomic sites .
it is observed that the presence of the middle arm provides localized states within the band of extended regions and lead to the possibility of switching action from a high conducting state to a low conducting one and vice versa .
this behavior is justified by studying persistent current in the network .
both the total current and individual currents in three separate branches are computed by using second - quantized formalism and our idea can be utilized to study magnetic response in any complicated quantum network .
the nature of localized eigenstates are also investigated from probability amplitudes at different sites of the quantum device . |
idealizations are often employed , when modeling real magnetization dynamics . far from equilibrium , magnetic fluctuations , modeled by white noise
are examples of such an idealization and the magnetization processes predicted in this way have been observed in ultrafast reversal magnetization experiments . however real fluctuations always have finite auto - correlation time and the corresponding spin system dynamics is , rather , described by non - markovian stochastic processes @xcite . despite the practical relevance of non - markovian processes , however , these have received much less attention than markovian processes , chiefly , because of the simplicity of the theory and the familiarity of the corresponding mathematical tools @xcite .
the study of the magnetization dynamics of super - paramagnetic systems , however , is a suitable framework and provides motivation to describe magnetic fluctuations , induced by contact to a heat bath@xcite , that are sensitive to non markovian effects .
the effective magnetization dynamics of a spin system and a noise is described by a stochastic differential equation of langevin type , also known as the stochastic landau - lifshitz - gilbert ( sllg ) equation : @xmath0 the einstein summation convention is adopted , with latin indices standing for vector components , @xmath1 is the anti - symmetric levi - civita pseudo - tensor and @xmath2 is a magnetization 3-vector , whose norm is fixed and can be set to unity . here
@xmath3 is the precession frequency which derives from a spin hamiltonian and is such connected to the spin vector .
@xmath4 is the damping coefficient .
the stochastic properties of @xmath5 describe the fluctuations of external and/or internal degrees of freedom , random perturbations and noise .
its correlation functions are assumed to satisfy the relations @xmath6 with @xmath7 and @xmath8 is the time - ordered statistical average product of stochastic variables @xcite and @xmath9 is a single characteristic constant time . with proper conditions gaussian process ,
poisson process and two - state ( dichotomy ) markov process enjoy this `` time derivative property '' @xcite .
in general a detailed description of the stochastic processes in the extended phase space @xmath10 is not easy to establish @xcite .
stochastic processes are completely described by an infinite set of multivariate probability distributions . in practice ,
not all multivariate distributions are not desirable but only a few , mainly the one - variable distribution @xmath11 and the two - variable distribution , @xmath12 .
the former is used to determine the average @xmath13 of any functional @xmath14 of the process @xmath15 , e.g. its statistical moments .
the latter is used to determine two - time statistics , e.g. , the correlation function @xmath16 . in the white noise limit ,
a fokker - planck equation on @xmath17 has been established only for a thermally interacting spin system subject to a constant external field @xcite . beyond the mean - field approximation , the explicit spin dependance of the effective field , including a magnetic anisotropy , on the probability distribution @xmath18 , may be addressed .
in contact with a thermostat , described by the stochastic vector , @xmath19 , the spin system precesses about the axis of magnetic effective field , defined by both a uniform external magnetic field with @xmath20 and an anisotropic single - axis .
this can be observed in super - paramagnetic nano - magnets which behave as effective , two - state , anisotropic spin systems @xcite .
the deterministic frequency vector is given by @xmath21 , where @xmath22 is the anisotropy single - axis and @xmath23 the anisotropy frequency .
the noise vector , @xmath19 , is assumed to be described by a gaussian stochastic process , that is completely defined by its 1-point function @xmath24 and 2point function @xmath25 .
therefore the only non trivial information is contained in the these , that we assume to define an ornstein - uhlenbeck process @xcite , which means that @xmath26 and @xmath27 it can be shown that this process does converge to the centered white noise process , when @xmath28 .
moreover , it is assumed that the thermostat can be described by a fluctuation - dissipation theorem that relates the strength of the noise , @xmath29 , with the temperature @xmath30 by the relationship @xmath31 , where @xmath4 is the parameter that appeared in eq.([llg0 ] ) .
these microscopic degrees of freedom give rise to average quantities , that probe collective properties .
such quantities can be operationally defined by taking statistical averages over the noise of eq.([llg0 ] ) at equal times . in the limit of low damping ( i.e. @xmath32 ) ,
the equation for the magnetization , @xmath33 , is : @xmath34 and implicates higher order correlation functions of noise and spin that have to be determined in turn . by applying the shapiro - loginov formulae of differentiation @xcite to the components of the @xmath35 at equal times
, we obtain the equation @xmath36 introducing the right - hand side of eq.([llg0 ] ) into eq.([shapiro ] ) , evolution equations for the mixed moments @xmath37 are obtained .
due to the non - linearity of the sllg equation , an infinite hierarchy arises @xcite : the equations for the second - order moments depend on third - order moments , and so on . in order to solve the system ,
closure relationships need to be found .
so the equations for the moments @xmath38 are now required and are provided by @xmath39 . assuming that this identity holds , the time derivative of @xmath40 is replaced by the right - hand side of eq.([llg0 ] ) inside the statistical average .
this leads to an additional set of equations for the second - order moments of the microscopic spin degrees of freedom , that are closely related to those written by garanin _
et al _ ( see eq.(6)@xcite ) and garcia - palacios _ et al .
_ ( see eq.(2.10 ) @xcite ) . as long as the stochastic variables follow individual gaussian processes , their connected moments ( labelled as @xmath41 ) , beyond the second moment are zero @xcite .
the case of the mixed connected moments of spin and noise is more subtle .
if a linear coupling between the spin and the noise is assumed , the mixed cumulants , above the second will , also , vanish .
if the gaussian hypothesis is dropped , the general relations , at equal times , between the desired moments are : @xmath42 these considerations lead to the following system of equations : @xmath43 these expressions are closed by considering first the independence of gaussian processes , the form of the effective frequency vector and the expression of the autocorrelation function of the noise .
one has explicitly : @xmath44 combined with eqs.([sys1],[sys2],[sys3 ] ) , the derived system of equations is called the dynamical landau - lifshitz - bloch model ( dllb ) . in the limit of @xmath28 , eq.([sys2 ] ) gives a direct expression for the mixed moment of the noise and the magnetization as @xmath45 , which makes the @xmath46 matrix antisymmetric .
once introduced into eq.([sys1 ] ) , a longitudinal contribution for the spin dynamics appears in which the amplitude of the noise is proportional to the square root of the temperature . for finite values of @xmath9 , eq.([sys1 ] ) exhibits a `` memory effect '' which shows that the value of @xmath47 for the time @xmath48 depends on all the previous values of @xmath49 before the time @xmath48 .
numerical experiments were performed to test the consistency of the closure hypothesis . for `` low '' values of @xmath9 ,
the dynamics is close to markovian and eqs.([sys1],[sys2],[sys3 ] ) are expected to match the results obtained from an averaged markovian stochastic dynamics . in order to realize these comparisons , a reference model is constructed from eq.([llg0 ] ) with a white noise random vector @xmath50 .
multiple spin trajectories with the same initial conditions are generated and followed in time according to a previously detailed symplectic integration scheme @xcite .
a statistical average is taken over all the realizations .
eqs.([sys1],[sys2],[sys3 ] ) are integrated with an eight - order runge - kutta integration scheme , with adaptive steps , constructed from the jacobian of the whole system .
the figure ( [ fig1]-left ) ( resp . [ fig1]-right ) exhibits the average magnetization ( resp .
spin - spin autocorrelation function ) from the stochastic llg equation and our dllb model for the same value of the amplitude @xmath29 in the white noise limit . for low value of @xmath29 ( i.e. in the normal phase ) ,
the average stochastic dynamics follows our dllb model .
a similar situation has been already investigated for an external constant field only @xcite .
when @xmath9 increases , the noise is no longer white and does not follow a markov s process .
a complete stochastic system of equations for both the spin and the noise is built by supplementing eq.([llg0 ] ) with @xmath51 where @xmath52 is a gaussian centered stochastic process @xcite , but with @xmath53 and @xmath54 .
formally , this system is integrated as @xmath55 with @xmath56 is the liouville operator corresponding to eq.([stochnoise ] ) . because the liouville operators of the spin and the noise commute , an integration of the colored system is performed by the combination of the two operators as @xmath57 .
this is performed for larger values of @xmath9 for which the figure ( [ fig2 ] ) exhibits @xmath47 and the diagonal part of @xmath58 .
it is observed that , even for `` large '' values of @xmath9 , our dllb model continues to track the dynamics of the colored stochastic llg system .
the influence of the higher than second - order cumulants on the dynamics of the average magnetization and spin autocorrelation function is mitigated by a large damping factor @xmath4 which prevents detailed descriptions of the magnetization dynamics .
moreover , because of the link between the amplitude @xmath29 and the damping factor @xmath4 , through a to - be - derived fluctuation - dissipation theorem in the special case of an explicit dependance of the spin in the effective field , the domain of validity of our dllb model in temperature could be significantly controlled .
when the intensity of the noise increases , the spin dynamics is no longer gaussian .
it is observed numerically , in the stochastic dynamics , by calculating the cumulant function @xmath59 .
the ten independent spin cumulants are depicted , for several values of the amplitude @xmath29 in the figure ( [ fig3 ] ) .
as shown , for this particular situation , some terms go to zero in time , whereas some remain finite , depending on the underlying symmetries . at the expense of higher order expressions , the eqs .
( [ sys1],[sys2],[sys3 ] ) can be supplemented by an expansion of @xmath60 and integrated together .
this represents an option where the deviation off the gaussian character of the spin distribution is explicitly treated at the third - order .
the third - order correlation functions for the dynamical variables @xmath61 satisfying a stochastic langevin - like equation are generally determined by first solving this equation in the presence of the noise and then averaging the product of these variables over the ensemble of the random noise .
a path integral formalism can be established for deducing identities between the correlation functions .
this strategy has been tested in various approximation schemes , including the quenched approximation ( i.e. @xmath62 ) .
a shortcut might be to look for an alternate description of an effective system of equations in a closed time path formalism of schwinger allowing to describe a general non - equilibrium field theory at finite temperature @xcite .
this is why a closure approximation was proposed by martin , siggia and rose @xcite which represents the classical limit of the closed time path formalism and will be the subject of further investigation . | a stochastic approach for the description of the time evolution of the magnetization of nanomagnets is proposed , that interpolates between the landau - lifshitz - gilbert and the landau - lifshitz - bloch approximations , by varying the strength of the noise .
its finite autocorrelation time , i.e. when it may be described as colored , rather than white , is , also , taken into account and the consequences , on the scale of the response of the magnetization are investigated .
it is shown that the hierarchy for the moments of the magnetization can be closed , by introducing a suitable truncation scheme , whose validity is tested by direct numerical solution of the moment equations and compared to the averages obtained from a numerical solution of the corresponding colored stochastic langevin equation .
this comparison is performed on magnetic systems subject to both an external uniform magnetic field and an internal one - site uniaxial anisotropy . |
dropout is a technique to regularize artificial neural networks it prevents overfitting @xcite . a fully connected network with two hidden layers of 80 units each can learn to classify the mnist training set perfectly in about 20 training epochs unfortunately the test error is quite high , about 2% . increasing the number of hidden units by a factor of 10 and using dropout results in a lower test error , about 1.1% .
the dropout network takes longer to train in two senses : each training epoch takes several times longer , and the number of training epochs needed increases too .
we consider a technique for speeding up training with dropout it can substantially reduce the time needed per epoch . consider a very simple @xmath0-layer fully connected neural network with dropout . to train it with a minibatch of @xmath1 samples ,
the forward pass is described by the equations : @xmath2\times w_{k}\qquad k=0,\dots,\ell-1.\ ] ] here @xmath3 is a @xmath4 matrix of input / hidden / output units , @xmath5 is a @xmath4 dropout - mask matrix of independent bernoulli(@xmath6 ) random variables , @xmath7 denotes the probability of dropping out units in level @xmath8 , and @xmath9 is an @xmath10 matrix of weights connecting level @xmath8 with level @xmath11 .
we are using @xmath12 for ( hadamard ) element - wise multiplication and @xmath13 for matrix multiplication .
we have forgotten to include non - linear functions ( e.g. the rectifier function for the hidden units , and softmax for the output units ) but for the introduction we will keep the network as simple as possible . the network can be trained using the backpropagation algorithm to calculate the gradients of a cost function ( e.g. negative log - likelihood ) with respect to the @xmath9 : @xmath14^{\mathsf{t}}\times\frac{\partial\mathrm{cost}}{\partial x_{k+1}}\\ \frac{\partial\mathrm{cost}}{\partial x_{k } } & = \left(\frac{\partial\mathrm{cost}}{\partial x_{k+1}}\times w_{k}^{\mathrm{\mathsf{t}}}\right)\cdot d_{k}.\end{aligned}\ ] ] with dropout training , we are trying to minimize the cost function averaged over an ensemble of closely related networks .
however , networks typically contain thousands of hidden units , so the size of the ensemble is _ much _ larger than the number of training samples that can possibly be ` seen ' during training .
this suggests that the independence of the rows of the dropout mask matrices @xmath5 might not be terribly important ; the success of dropout simply can not depend on exploring a large fraction of the available dropout masks .
some machine learning libraries such as pylearn2 allow dropout to be applied batchwise instead of independently ] .
this is done by replacing @xmath5 with a @xmath15 row matrix of independent bernoulli@xmath16 random variables , and then copying it vertically @xmath1 times to get the right shape .
to be practical , it is important that each training minibatch can be processed quickly . a crude way of estimating the processing time is to count the number of floating point multiplication operations needed ( naively ) to evaluate the @xmath13 matrix multiplications specified above :
@xmath17 however , when we take into account the effect of the dropout mask , we see that many of these multiplications are unnecessary .
the @xmath18-th element of the @xmath9 weight matrix effectively ` drops - out ' of the calculations if unit @xmath19 is dropped in level @xmath8 , or if unit @xmath20 is dropped in level @xmath11 . applying 50% dropout in levels @xmath8 and @xmath11 renders 75% of the multiplications unnecessary .
if we apply dropout independently , then the parts of @xmath9 that disappear are different for each sample .
this makes it effectively impossible to take advantage of the redundancy
it is slower to check if a multiplication is necessary than to just do the multiplication .
however , if we apply dropout batchwise , then it becomes easy to take advantage of the redundancy .
we can literally drop - out redundant parts of the calculations .
the binary @xmath15 batchwise dropout matrices @xmath5 naturally define submatrices of the weight and hidden - unit matrices .
let @xmath21 $ ] denote the submatrix of @xmath3 consisting of the level-@xmath8 hidden units that survive dropout .
let @xmath22 $ ] denote the submatrix of @xmath9 consisting of weights that connect active units in level @xmath8 to active units in level @xmath11 .
the network can then be trained using the equations : @xmath23 the redundant multiplications have been eliminated .
there is an additional benefit in terms of memory needed to store the hidden units : @xmath24 needs less space than @xmath3 . in section [ sec : implementation ]
we look at the performance improvement that can be achieved using cuda / cublas code running on a gpu .
roughly speaking , processing a minibatch with 50% batchwise dropout takes as long as training a 50% smaller network on the same data .
this explains the nearly overlapping pairs of lines in figure [ fig : training - time ] .
we should emphasize that batchwise dropout only improves performance during training ; during testing the full @xmath9 matrix is used as normal , scaled by a factor of @xmath25 .
however , machine learning research is often constrained by long training times and high costs of equipment . in section [ sec
: results - for - fully - connected ] we show that all other things being equal , batchwise dropout is similar to independent dropout , but faster .
moreover , with the increase in speed , all other things do not have to be equal . with the same resources ,
batchwise dropout can be used to * increase the number of training epochs , * increase the number of hidden units , * increase the number of validation runs used to optimize `` hyper - parameters '' , or * to train a number of independent copies of the network to form a committee .
these possibilities will often be useful as ways of improving generalization / reducing test error . in section [ sec : convolutional - networks ]
we look at batchwise dropout for convolutional networks .
dropout for convolutional networks is more complicated as weights are shared across spatial locations .
a minibatch passing up through a convolutional network might be represented at an intermediate hidden layer by an array of size @xmath26 : 100 samples , the output of 32 convolutional filters , at each of @xmath27 spatial locations .
it is conventional to use a dropout mask with shape @xmath26 ; we will call this independent dropout .
in contrast , if we want to apply batchwise dropout efficiently by adapting the submatrix trick , then we will effectively be using a dropout mask with shape @xmath28 .
this looks like a significant change : we are modifying the ensemble over which the average cost is optimized . during training , the error rates are higher .
however , testing the networks gives very similar error rates .
we might have called batchwise dropout _ fast dropout _ but that name is already taken _ _ @xcite .
fast dropout is very different approach to solving the problem of training large neural network quickly without overfitting .
we discuss some of the differences of the two techniques in the appendix .
in theory , for @xmath29 matrices , addition is an @xmath30 operation , and multiplication is @xmath31 by the coppersmithwinograd algorithm .
this suggests that the bulk of our processing time should be spent doing matrix multiplication , and that a performance improvement of about 60% should be possible compared to networks using independent dropout , or no dropout at all .
in practice , sgemm functions use strassen s algorithm or naive matrix multiplication , so performance improvement of up to 75% should be possible .
we implemented batchwise dropout for fully - connected and convolutional neural networks using cuda / cublas ] .
we found that using the highly optimized function to do the bulk of the work , with cuda kernels used to form the submatrices @xmath32 and to update the @xmath9 using @xmath33 , worked well .
better performance may well be obtained by writing a sgemm - like matrix multiplication function that understands submatrices . for large networks and minibatches , we found that batchwise dropout was substantially faster , see figure [ fig : training - time ] .
the approximate overlap of some of the lines on the left indicates that 50% batchwise dropout reduces the training time in a similar manner to halving the number of hidden units .
the graph on the right show the time saving obtained by using submatrices to implement dropout .
note that for consistency with the left hand side , the graph compares batchwise dropout with dropout - free networks , _ not _ with networks using independent dropout .
the need to implement dropout masks for independent dropout means that figure 1 slightly undersells the performance benefits of batchwise dropout as an alternative to independent dropout . for smaller networks ,
the performance improvement is lower bandwidth issues result in the gpu being under utilized .
if you were implementing batchwise dropout for cpus , you would expect to see greater performance gains for smaller networks as cpus have a lower processing - power to bandwidth ratio .
if you have @xmath34 hidden units and you drop out @xmath35% of them , then the number of dropped units is approximately @xmath36 , but with some small variation as you are really dealing with a binomial@xmath37 random variable
its standard deviation is @xmath38 .
the sizes of the submatrices @xmath32 and @xmath24 are therefore slightly random . in the interests of efficiency and simplicity ,
it is convenient to remove this randomness .
an alternative to dropping each unit independently with probability @xmath39 is to drop a subset of exactly @xmath40 of the hidden units , uniformly at random from the set of all @xmath41 such subsets .
it is still the case that each unit is dropped out with probability @xmath39 .
however , within a hidden layer we no longer have strict independence regarding which units are dropped out .
the probability of dropping out the first two hidden units changes very slightly , from @xmath42 also , we used a modified form of nag - momentum minibatch gradient descent @xcite .
after each minibatch , we only updated the elements of @xmath32 , not all the element of @xmath9 . with @xmath43 and @xmath44 denoting the momentum matrix / submatrix corresponding to @xmath9 and @xmath32 , our update was @xmath45 the momentum still functions as an autoregressive process , smoothing out the gradients , we are just reducing the rate of decay @xmath46 by a factor of @xmath47 .
dropout networks trained using a restricted the number of dropout patterns ( each @xmath13 is from an independent experiment ) .
the blue line marks the test error for a network with half as many hidden units trained without dropout .
, scaledwidth=45.0% ]
the fact that batchwise dropout takes less time per training epoch would count for nothing if a much larger number of epochs was needed to train the network , or if a large number of validation runs were needed to optimize the training process .
we have carried out a number of simple experiment to compare independent and batchwise dropout . in many cases
we could have produced better results by increasing the training time , annealing the learning rate , using validation to adjust the learning process , etc .
we choose not to do this as the primary motivation for batchwise dropout is efficiency , and excessive use of fine - tuning is not efficient .
for datasets , we used : * the mnist set of @xmath48 pixel handwritten digits . * the cifar-10 dataset of 32x32 pixel color pictures ( @xcite ) . * an artificial dataset designed to be easy to overfit .
following @xcite , for mnist and cifar-10 we trained networks with 20% dropout in the input layer , and 50% dropout in the hidden layers .
for the artificial dataset we increased the input - layer dropout to 50% as this reduced the test error . in some cases ,
we have used relatively small networks so that we would have time to train a number of independent copies of the networks .
this was useful in order to see if the apparent differences between batchwise and independent dropout are significant or just noise .
our first experiment explores the effect of dramatically restricting the number of dropout patterns seen during training .
consider a network with three hidden layers of size 1000 , trained for 1000 epochs using minibatches of size 100 .
the number of distinct dropout patterns , @xmath49 , is so large that we can assume that we will never generate the same dropout mask twice . during independent dropout training
we will see 60 million different dropout patterns , during batchwise dropout training we will see 100 times fewer dropout patterns . for both types of dropout , we trained 12 independent networks for 1000 epochs , with batches of size 100 . for batchwise dropout we got a mean test error of 1.04% [ range ( 0.92%,1.1% ) , s.d .
0.057% ] and for independent dropout we got a mean test errors of 1.03% [ range ( 0.98%,1.08% ) , s.d . 0.033% ] .
the difference in the mean test errors is not statistically significant . to explore further
the reduction in the number of dropout patterns seen , we changed our code for ( pseudo)randomly generating batchwise dropout patterns to restrict the number of distinct dropout patterns used
. we modified it to have period @xmath50 minibatches , with @xmath51 ; see figure [ fig : limited - dropout - patterns ] . for @xmath52
this corresponds to only ever using one dropout mask , so that 50% of the network s 3000 hidden weights are never actually trained ( and 20% of the 784 input features are ignored ) . during training
this corresponds to training a dropout - free network with half as many hidden units
the test error for such a network is marked by a blue line in figure [ fig : limited - dropout - patterns ] .
the error during testing is higher than the blue line because the untrained weights add noise to the network .
if @xmath50 is less than thirteen , is it likely that some of the networks 3000 hidden units are dropped out every time and so receive no training . if @xmath50 is in the range thirteen to fifty , then it is likely that every hidden unit receives some training , but some pairs of hidden units in adjacent layers will not get the chance to interact during training , so the corresponding connection weight is untrained . as the number of dropout masks increases into the hundreds , we see that it is quickly a case of diminishing returns .
to test the effect of changing network size , we created an artificial dataset .
it has 100 classes , each containing 1000 training samples and 100 test samples .
each class is defined using an independent random walk of length 1000 in the discrete cube @xmath53 . for each class we generated the random walk , and then used it to produce the training and test samples by randomly picking points along the length of walk ( giving binary sequences of length 1000 ) and then randomly flipping 40% of the bits .
we trained three layer networks with @xmath54 hidden units per layer with minibatches of size 100 .
see figure [ fig : artificial - dataset.-100 ] .
looking at the training error against training epochs , independent dropout seems to learn slightly faster .
however , looking at the test errors over time , there does not seem to be much difference between the two forms of dropout . note that the @xmath55-axis is the number of training epochs , not the training time .
the batchwise dropout networks are learning much faster in terms of real time .
learning cifar-10 using a fully connected network is rather difficult .
we trained three layer networks with @xmath56 hidden units per layer with minibatches of size 1000 .
we augmented the training data with horizontal flips .
see figure [ fig : cifar - fc ] .
dropout for convolutional networks is more complicated as weights are shared across spatial locations .
suppose layer @xmath8 has spatial size @xmath57 with @xmath58 features per spatial location , and if the @xmath8-th operation is a convolution with @xmath59 filters .
for a minibatch of size @xmath1 , the convolution involves arrays with sizes : @xmath60 dropout is normally applied using dropout masks with the same size as the layers .
we will call this independent dropout independent decisions are mode at every spatial location .
in contrast , we define batchwise dropout to mean using a dropout mask with shape @xmath61 .
each minibatch , each convolutional filter is either on or off across all spatial locations .
these two forms of regularization seem to be doing quite different things .
consider a filter that detects the color red , and a picture with a red truck in it .
if dropout is applied independently , then by the law of averages the message `` red '' will be transmitted with very high probability , but with some loss of spatial information .
in contrast , with batchwise dropout there is a 50% chance we delete the entire filter output .
experimentally , the only substantial difference we could detect was that batchwise dropout resulted in larger errors during training . to implement batchwise dropout efficiently , notice that the @xmath61 dropout masks corresponds to forming subarrays @xmath32 of the weight arrays @xmath9 with size @xmath62 the forward - pass is then simply a regular convolutional operation using @xmath32
; that makes it possible , for example , to take advantage of the highly optimized @xmath63 function from the https://developer.nvidia.com/cudnn[nvidia cudnn ] package . for mnist , we trained a lenet-5 type cnn with two layers of @xmath64 filters , two layers of @xmath65 max - pooling , and a fully connected layer @xcite
. there are three places for applying 50% dropout : @xmath66 the test errors for the two dropout methods are similar , see figure [ fig : mnist - test - errors , ] . , scaledwidth=38.0% ] for a first experiment with cifar-10 we used a small convolutional network with small filters .
the network is a scaled down version of the network from @xcite ; there are four places to apply dropout : @xmath67 the input layer is @xmath68 .
we trained the network for 1000 epochs using randomly chosen subsets of the training images , and reflected each image horizontally with probability one half .
for testing we used the centers of the images . in figure
[ fig : cifar-10-results - using ] we show the effect of varying the dropout probability @xmath39 .
the training errors are increasing with @xmath39 , and the training errors are higher for batchwise dropout .
the test - error curves both seem to have local minima around @xmath69 .
the batchwise test error curve seems to be shifted slightly to the left of the independent one , suggesting that for any given value of @xmath39 , batchwise dropout is a slightly stronger form of regularization . .
batchwise dropout produces a slightly lower minimum test error .
[ fig : cifar-10-results - using],scaledwidth=45.0% ] we trained a deep convolutional network on cifar-10 _ without _ data augmentation . using the notation of @xcite , our network has the form @xmath70{2})_{12}-832c2 - 896c1-\mathrm{output},\ ] ]
i.e. it consists of 12 @xmath65 convolutions with @xmath71 filters in the @xmath50-th layer , 12 layers max - pooling , followed by two fully connected layers ; the network has 12.6 million parameters .
we used an increasing amount of dropout per layer , rising linearly from 0% dropout after the third layer to 50% dropout after the 14th . even though the amount of dropout used in the middle layers is small
, batchwise dropout took less than half as long per epoch as independent dropout ; this is because applying small amounts of independent dropout in large hidden - layers creates a bandwidth performance - bottleneck .
as the network s max - pooling operation is stochastic , the test errors can be reduced by repetition .
batchwise dropout resulted in a average test error of 7.70% ( down to 5.78% with 12-fold testing ) .
independent dropout resulted in an average test error of 7.63% ( reduced to 5.67% with 12-fold testing ) .
we have implemented an efficient form of batchwise dropout . all other things being equal
, it seems to learn at roughly the same speed as independent dropout , but each epoch is faster . given a fixed computational budget , it will often allow you to train better networks .
there are other potential uses for batchwise dropout that we have not explored yet : * restricted boltzmann machines can be trained by contrastive divergence @xcite with dropout @xcite . batchwise dropout could be used to increase the speed of training . *
when a fully connected network sits on top of a convolutional network , training the top and bottom of the network can be separated over different computational nodes @xcite .
the fully connected top - parts of the network typically contains 95% of the parameters keeping the nodes synchronized is difficult due to the large size of the matrices . with batchwise dropout
, nodes could communicate @xmath33 instead of @xmath72 and so reducing the bandwidth needed .
* using independent dropout with recurrent neural networks can be too disruptive to allow effective learning ; one solution is to only apply dropout to some parts of the network @xcite .
batchwise dropout may provide a less damaging form of dropout , as each unit will either be on or off for the whole time period . *
dropout is normally only used during training .
it is generally more accurate use the whole network for testing purposes ; this is equivalent to averaging over the ensemble of dropout patterns .
however , in a `` real - time '' setting , such as analyzing successive frames from a video camera , it may be more efficient to use dropout during testing , and then to average the output of the network over time .
* nested dropout @xcite is a variant of regular dropout that extends some of the properties of pca to deep networks .
batchwise nested dropout is particularly easy to implement as the submatrices are regular enough to qualify as matrices in the context of the sgemm function ( using the lda argument ) .
* dropconnect is an alternative form of regularization to dropout @xcite . instead of dropping hidden units ,
individual elements of the weight matrix are dropped out . using a modification similar to the one in section [ sub : fixed - dropout - amounts ] , there are opportunities for speeding up dropconnect training by approximately a factor of two
. 10 d. ciresan , u. meier , and j. schmidhuber .
link : www.idsia.ch/~juergen / cvpr2012.pd[multi - column deep neural networks for image classification ] . in _ computer vision and pattern recognition ( cvpr ) , 2012 ieee conference on _ , pages 36423649 , 2012 .
ben graham .
fractional max - pooling , 2014 .
http://arxiv.org/abs/1412.6071 .
hinton and salakhutdinov .
http://www.cs.toronto.edu/~hinton/science.pdf[reducing the dimensionality of data with neural networks ] .
, 313 , 2006 .
alex krizhevsky .
http://www.cs.toronto.edu/~kriz/cifar.html[learning multiple layers of features from tiny images ] .
technical report , 2009 .
alex krizhevsky .
one weird trick for parallelizing convolutional neural networks , 2014 .
http://arxiv.org/abs/1404.5997 .
y. l. le cun , l. bottou , y. bengio , and p. haffner .
gradient - based learning applied to document recognition .
, 86(11):22782324 , november 1998 .
oren rippel , michael a. gelbart , and ryan p. adams .
learning ordered representations with nested dropout , 2014 .
http://arxiv.org/abs/1402.0915 .
nitish srivastava , geoffrey hinton , alex krizhevsky , ilya sutskever , and ruslan salakhutdinov .
http://jmlr.org/papers/v15/srivastava14a.html[dropout : a simple way to prevent neural networks from overfitting ] . , 15:19291958 , 2014 .
ilya sutskever , james martens , george e. dahl , and geoffrey e. hinton .
http://jmlr.org/proceedings/papers/v28/[on the importance of initialization and momentum in deep learning ] . in
volume 28 of _ jmlr proceedings _
, pages 11391147 .
jmlr.org , 2013 . li wan , matthew zeiler , sixin zhang , yann lecun , and rob fergus .
http://jmlr.org/proceedings/papers/v28/wan13.html[regularization of neural networks using dropconnect ] , 2013 .
jmlr w&cp 28 ( 3 ) : 10581066 , 2013 .
sida wang and christopher manning .
http://jmlr.csail.mit.edu/proceedings/papers/v28/wang13a.html[fast dropout training ] .
, 28(2):118126 , 2013 .
wojciech zaremba , ilya sutskever , and oriol vinyals .
recurrent neural network regularization , 2014 .
we might have called batchwise dropout _ fast dropout _ but that name is already taken _ _
fast dropout is an alternative form of regularization that uses a probabilistic modeling technique to imitate the effect of dropout ; each hidden unit is replaced with a gaussian probability distribution .
the _ fast _ relates to reducing the number of training epochs needed compared to regular dropout ( with reference to results in a preprint of @xcite ) . training a network 784 - 800 - 800 - 10 on the mnist dataset with 20% input dropout and 50% hidden - layer dropout
, fast dropout converges to a test error of 1.29% after 100 epochs of l - bfgs .
this appears to be substantially better than the test error obtained in the preprint after 100 epochs of regular dropout training . however , this is a dangerous comparison to make .
the authors of @xcite used a learning - rate scheme designed to produce optimal accuracy eventually , _ not _ after just one hundred epochs .
we tried using batchwise dropout with minibatches of size 100 and an annealed learning rate of @xmath73 .
we trained a network with two hidden layers of 800 rectified linear units each .
training for 100 epochs resulted in a test error of 1.22% ( s.d .
0.03% ) . after 200 epochs
the test error has reduced further to 1.12% ( s.d .
moreover , per epoch , batchwise - dropout is faster than regular dropout while fast - dropout is slower . assuming we can make comparisons across different programs , the 200 epochs of batchwise dropout training take less time than the 100 epoch of fast dropout training . | dropout is a popular technique for regularizing artificial neural networks .
dropout networks are generally trained by minibatch gradient descent with a dropout mask turning off some of the units a different pattern of dropout is applied to every sample in the minibatch .
we explore a very simple alternative to the dropout mask .
instead of masking dropped out units by setting them to zero , we perform matrix multiplication using a submatrix of the weight matrix unneeded hidden units are never calculated . performing dropout _ batchwise _ , so that one pattern of dropout is used for each sample in a minibatch , we can substantially reduce training times .
batchwise dropout can be used with fully - connected and convolutional neural networks . |
the correspondence between quantum and classical mechanics has turned out to be particularly rich for systems that are classically chaotic @xcite .
this correspondence is usually considered in terms of the asymptotic behavior of quantum properties .
statistical analysis of the quantum spectra of closed chaotic systems display short - range correlations that are universal and can be described by random matrix theory ( rmt ) @xcite .
the long range behavior of spectral fluctuations require semiclassical treatments that incorporate periodic orbits@xcite . in this paper
we establish a new connection between the statistics of quantum spectra and classical dynamical properties of an open chaotic system , utilizing physical and numerical experiments on the well - known @xmath0-disk billiard geometry .
the @xmath0-disk system @xcite is a paradigm of open hyperbolic systems , with an associated strange repeller @xcite whose strange properties arise from the cantor set of the unstable periodic orbits ( po ) .
for hyperbolic repellers , ruelle @xcite and pollicott @xcite showed that the dynamical evolution of repellers toward equilibrium can be expressed in terms of complex poles , leading to the so - called ruelle - pollicott ( rp ) resonances .
the @xmath0-disk billiards have been extensively addressed theoretically , both classically and semiclassically @xcite , and only recently have been studied experimentally @xcite .
our microwave experiments on the @xmath0-disk billiards measure the quantum transmission spectrum @xmath3 , which is essentially the two - point green s function @xmath4 , and from which the spectral autocorrelation @xmath1 is determined .
the small wave vector @xmath5 ( long time ) behavior of the spectral autocorrelation provides a measure of the quantum escape rate , and is in good agreement with the corresponding classical escape rate @xmath6 @xcite , as predicted by rmt analysis of this universal behavior . for large @xmath7 ( short time ) , non - universal oscillations of the autocorrelation @xmath1
are observed , that can be understood completely in terms of classical rp resonances .
this behavior is similar to that of the variance @xmath8of closed chaotic systems , which also displays leading universal behavior for small @xmath9 , followed by non - universal behavior for large @xmath9 that can be described in terms of system specific periodic orbits @xcite .
the present work demonstrates that @xmath1 of open systems displays a similar trend from short - range universality to long - range non - universality , however the non - universal contributions to quantum correlations are described in terms of other classical phase space structures of the associated repeller , viz .
the rp resonances @xcite .
a concise account of the main results was presented in ref .
this paper further explores the details of the experimental aspects , as well as confirming the results via numerical experiments using simulated quantum spectra .
semiclassical arguments that demonstrate this correspondence are discussed in terms of fluctuations of the density of states and correlations of the particle density .
the wider implications of the present results are also discussed .
the quantum dynamics of the 2-d @xmath0-disk system can be realized in a microwave experiment exploiting the mapping between the helmholtz equation and the schrdinger equation in their stationary forms and in the 2-d limit .
this is because under the conditions of the experiment , the maxwell - helmholtz equations reduce to @xmath10 with @xmath11 the microwave electric field .
the experimental geometry consists of thin copper disks sandwiched between two large copper plates of size @xmath12 @xmath13 in area . in order to simulate an infinite system ,
microwave absorber material eccosorb an-77 was placed between the plates at the edges .
microwaves were coupled in and out through antennas inserted in the vicinity of the scatterers .
all measurements were carried out using an hp8510b vector network analyzer which measured the complex transmission amplitude ( @xmath14 ) and reflection amplitude ( @xmath15 ) @xmath16-parameters of the coax + scatterer system .
an example of the measured transmission @xmath17 of a 4-disk system is shown in fig .
1 . for more details of the microwave experiments ,
@xcite .
an open system can be represented by an effective hamiltonian consisting of two parts , @xmath18 , where @xmath19 is the hamiltonian for the closed system and @xmath20 represents the decay to open channels .
since the total hamiltonian is not hermitian , the eigenfunctions @xmath21 of @xmath22 with eigenvalues @xmath23 do not form an orthogonal set . instead ,
let @xmath24 be the eigenfunction of the adjoint operator @xmath25 with eigenvalues @xmath26 , and the @xmath27 s and @xmath28 s form a bi - orthonormal set with @xmath29 the green s function for an open system is @xcite @xmath30 . for the @xmath0-disk system , with the quantum resonances in wave vector space @xmath31 , we have @xmath32[(k - s_{n})^{2}+s_{n}^{\prime 2 } ] } \\
\hskip -.3cm&&\hskip -.3cm+\frac{2m}{\hbar ^{2}}\sum_{n , m}\,^{^{\prime } } \frac{\psi _ { n}^{*}({\vec{r}% } _ { 1})\phi _ { n}({\vec{r}}_{2})\psi _ { m}({\vec{r}}_{1})\phi _ { m}^{*}({\vec{r}}% _ { 2})}{(k^{2}-k_{n}^{2})(k^{2}-k_{m}^{2})^{*}}.\end{aligned}\ ] ] the second sum with the prime includes the off - diagonal terms with @xmath33 .
their contribution can be neglected if all the resonances are well separated .
the transmission amplitude @xmath3 measured between two point antennas is essentially the green s function @xmath4 of the system , so that @xmath34 with @xmath35 a slowly - varying modulating function @xcite . setting @xmath36 $ ] , and ignoring the off - diagonal terms , we get the transmission coefficient @xmath37 expressed as a sum of lorentzian peaks @xmath38 the coupling @xmath39 in eq .
( [ transmission ] ) depend on the location of the probes .
the @xmath40 dependence of @xmath39 is locally weak near a resonance since @xmath41 $ ] is flat around @xmath42 for sharp resonances .
one can treat @xmath39 practically as constants in the neighborhood of a resonance .
this shows that the measured transmission coefficient @xmath37 shown in fig .
1 directly yields the wave - vector @xmath43 and the half - width @xmath44 of the sharper scattering resonances .
we next compare the experimental resonances with semiclassical calculations .
although any chaotic system can be quantized numerically by directly diagonalizing the hamiltonian truncated in certain expansion space , for a large number of systems , one uses the techniques based upon semiclassical periodic orbit theory @xcite , such as the cycle expansion @xcite , fredholm determinant @xcite or harmonic inversion @xcite . except for the lowest eigenvalues ,
the semiclassical quantization gives very accurate results .
using gutzwiller s trace formula @xcite , the semiclassical ruelle zeta - function can be derived as an _ euler _ product over all the prime po which are the po without repetition @xcite @xmath45 with @xmath46 the semiclassical weight @xmath47 here @xmath48 is the number of collisions of the po with the disks , @xmath49 is the length of the po and @xmath50 the eigenvalue of the instability matrix @xmath51 .
@xmath52 comes in from the decomposition of the trace of the green s function into ruelle zeta - functions with @xmath53 .
the quantum resonances are the poles of the ruelle zeta - functions which is directly related to the trace of scattering matrix or the trace of the green s function .
if a symbolic dynamics exists for the system , only a few prime pos will be needed in the zeta - function to give quite accurate eigenvalues because the curvature in the cycle expansion will decay exponentially with increasing po length @xcite . as an example , consider the non - chaotic @xmath54-disk system with disk separation @xmath55 and disk radius @xmath56 .
there is only one unstable prime po between the disks .
the symmetry group of po is @xmath57 , with two one - dimensional irreducible representations , symmetric @xmath58 and anti - symmetric @xmath59 .
the semiclassical resonances in wave vector space are @xmath60/(r-2a)$ ] with @xmath61 the eigenvalue of the instability matrix in the fundamental domain and @xmath62 . here
@xmath0 is odd for @xmath58 representation , @xmath0 even and @xmath63 for @xmath59 representation @xcite . for the chaotic @xmath0-disk system with @xmath64
, there is no analytical expression of the semiclassical quantum resonances . to do the numerical calculations of the quantum resonances ,
use can be made of the symmetry of the system and the cycle expansion in the fundamental domain @xcite .
we have calculated numerically the quantum resonances of the 3-disk and 4-disk systems with different disk separation ratio @xmath65 by searching the poles of the ruelle zeta function ( [ semi - ruelle ] ) .
an example of the results of the calculation for the @xmath66 for a 3-disk system is shown in fig .
the experiments only access a small range of the quantum resonances with @xmath67 and @xmath68 .
the experimental quantum resonances are well matched by the sharp resonances from the semiclassical calculations .
the very broad resonances are not observed because the coupling to them is weak . for details concerning
the comparison of experimental and calculated quantum resonances , see ref .
@xcite .
the spectral autocorrelation function is determined from the experimental spectra as @xmath69 with average carried out over a band of wave vector centered at certain value @xmath70 and of width @xmath71 @xcite .
the spectral autocorrelation can be fitted well by a superposition of lorentzians : @xmath72 with @xmath73 the classical rp resonances in wave vector space .
a semiclassical derivation will be given in sec .
the experimental rp resonances are obtained by fitting the experimental autocorrelation as shown in fig .
3 with eq . ( [ auto - cor ] ) .
similar comparisons between experiment and theory for the @xmath0-disk system with @xmath74 are presented in ref .
the coupling @xmath75 in the above equation which should depend on the location of the probes determine the decay probability of the classical rp resonances .
since we do not have knowledge of them experimentally , they are chosen to optimize the fitting .
the experimental rp resonances ( @xmath76 ) obtained from the lorentzian decomposition of the experimental @xmath77 as shown in fig .
[ fig - exp - cor ] are displayed in table 1 .
they can be compared with theoretically calculated rp resonances , which are the poles of the classical ruelle zeta - function , and whose calculation will be explained in the next subsection . although the position of the peaks of the oscillations in the experimental autocorrelation in fig .
[ fig - exp - cor ] is quite accurately given by the imaginary part @xmath78 of the rp resonances , the experimental half - width of the oscillation is almost always smaller than the calculated real part @xmath79 of the rp resonances .
this is possibly because the absorber used is not ideal , and leads to quantum resonances that are slightly sharper than calculated .
this results in systematically sharper widths of the experimental rp resonances .
the theoretical rp resonances are essential ingredients of a liouvillian description of the classical dynamics , and can be calculated as the poles of the classical ruelle zeta - function .
the classical dynamics of hamiltonian flow in symplectic phase space can be described by the action of a linear evolution operator which is also the perron - frobenius operator @xmath80 here @xmath81 is the trajectory of the initial point @xmath82 in phase space .
the trace of the perron - frobenius operator is @xmath83 .
the first summation is over all the prime classical po @xmath84 with period @xmath85 , the second one is over the repetition @xmath86 of the prime po .
@xmath51 again is the instability matrix .
the evolution operator has a lie group structure @xmath87 with @xmath88 the generator of the hamiltonian flow .
this generator @xmath88 has the classical resonances as its eigenvalues , @xmath89 , which we call the rp resonances @xcite .
we have @xmath90 with @xmath91 the multiplicity of the resonances .
these resonances determine the time evolution of any classical quantity .
the trace formula for classical flows is obtained from the laplace transform of the above expression @xcite @xmath92 in order to calculate the rp resonances , the ruelle zeta - function is introduced as @xmath93 here the product is over all the prime po with @xmath94 the classical weights of the periodic orbits @xmath95 here the integer @xmath96 comes in from the expansion of the determinant @xmath97 .
one can see that the above classical ruelle zeta - function is very similar to the semiclassical one ( eq.([semi - ruelle ] ) ) .
the classical ruelle zeta - function is exact and can be directly derived from the above trace formula ( [ tra - class ] ) of the perron - frobenius operator .
the topological pressure @xmath98 can be defined as the simple pole of the classical ruelle zeta - function @xcite with @xmath99 extrapolated to the entire real space .
all the characteristic quantities of the classical dynamics , such as the kolmogorov - sinai entropy , escape rate , fractal dimensions , can be derived from @xmath98 @xcite .
for the hard - disk system , the classical velocity @xmath100 of particles is constant .
so @xmath101 .
for simplicity , one can set @xmath102 .
the rp resonances are then calculated in wave - vector space .
for example , the classical rp resonances of the 2-disk system are @xmath103/(r-2a)$ ] with @xmath104 given before . here
@xmath0 is even for @xmath58 representation and @xmath0 odd for @xmath59 representation .
similar to the situation of the quantum resonances @xmath105 , there is no analytical expression of the rp resonances @xmath106 for the chaotic @xmath0-disk system .
they can be calculated numerically as the poles of the classical ruelle zeta - function eq .
( [ classical - ruelle ] ) @xcite .
14 prime po up to period 3 were used in our calculation of the rp resonances of the 4-disk system , the results of which are shown in table 1 . to gain further insight into the established experimental correspondence , we have utilized the results of the calculations of the quantum and classical resonances to carry out a numerical simulation of the physical experiment .
we calculate numerically a close approximation to the experimental quantum spectrum using the quantum resonances .
we them determine the spectral autocorrelation of the simulated spectrum , and compare its decomposition with the calculated classical resonances . writing eq .
( [ transmission ] ) explicitly , we have following form of the transmission @xmath107 \left [ ( k - s_{n})^{2}+s_{n}^{\prime 2}\right ] } $ ] . here
we assume the presence of the probes poses a small perturbation and will not change the quantum spectrum . for a closed chaotic system with time - reversal symmetry
, the wave density is expected to follow the porter - thomas ( pt ) distribution in rmt @xcite . for a very open system and when the probes are far away from the scatterers , the wave density will not follow the pt distribution .
but in the vicinity of the scatterers , one may assume the density to follow the pt distribution .
setting @xmath108 and @xmath109 , one has , @xmath110 with @xmath111 being the ensemble averaged density .
the densities at the location of different antennas were found not to be correlated in closed cavities @xcite .
we assume the same is true for an open chaotic system in the vicinity of the scatterer disks . for simplicity
, we consider the ensemble average of the transmission coefficient @xmath112 \left [ ( k - s_{n})^{2}+s_{n}^{\prime 2}\right ] $ ] . from this expression , one can see that the contribution from broad resonances is suppressed .
the main contribution is from sharp resonances .
this has already been observed in our experiments@xcite . for these sharp resonances , if they are far away from the origin , @xmath113 .
thus , for large @xmath114 , the above transmission can be approximated as @xmath115 ^{-1}$ ] .
the function @xmath116 is found to be proportional to @xmath40 @xcite . for convenience , we set @xmath117 and also @xmath118 . we thus get @xmath119 } .
\label{numeric}\ ] ] the @xmath40 dependence of @xmath116 can be understand as follows . in the experimental setup , one of the antenna can be regarded as a diploe radiating electromagnetic waves .
the radiated electrical field of a dipole is proportional to @xmath40 if an alternating current with constant amplitude @xmath120 was maintained on it @xcite .
the voltage picked up on another antenna is just the electrical field at the antenna location times the length @xmath121 of the antenna inside the cavity .
the transmission amplitude @xmath122 measured by the analyzier is the ratio of the output voltage on one antenna to the input current of another antenna , @xmath123 , thus @xmath124 .
the expression ( [ numeric ] ) is thus used to get the averaged transmission and then to calculate the autocorrelation @xmath125 numerically .
we have done that for the @xmath0-disk system with @xmath126 using @xmath127 cm .
here we just present the results for the @xmath128-disk system as shown in fig . 4 and 5 .
note that @xmath129 , @xmath130 is equivalent to @xmath131@xmath132 .
the symmetry group for the chaotic 3-disk system is @xmath133 @xcite , which has two one - dimensional irreducible representations @xmath58 , @xmath59 , and one two - dimensional irreducible representation @xmath134 . in the fundamental domain
, only the resonances of the @xmath59 representation will contribute to the transmission @xcite .
the semiclassical resonances of @xmath59 representation in the range @xmath135 and @xmath136 for the disk separation @xmath137 are obtained from @xmath138 with @xmath139 .
about 200 resonances are obtained as shown in fig .
2 . the corresponding classical resonances shown in fig . 5 are from the @xmath58 representation of the classical ruelle zeta - function ( [ classical - ruelle ] ) with @xmath140 .
8 prime po up to period 4 were used in our calculations of the quantum and rp resonances . the numerical transmission and the autocorrelation are shown in fig . 4 and 5 , respectively
one can see that the oscillations in the autocorrelation are fully determined by the classical rp resonances as predicted by eq.([auto - cor ] ) .
in the above section , we have demonstrated experimentally and numerically that the quantum auto - correlation and hence the statistics of quantum resonances are determined by the rp resonances of the underlying classical dynamics .
here we explore the theoretical justification for this correspondence . for chaotic open scattering systems , the distribution and correlation of the quantum resonances
have been discussed in the framework of rmt @xcite .
however , the rmt results are the ensemble averages of different systems with certain symmetry in common .
consequently , rmt does not make contact with system specific features of the underlying classical dynamics .
thus it is unable to make any concrete connection between the quantum and classical dynamics , for which it is necessary to go beyond rmt .
the appearance of nonuniversality requires the use of structures such as periodic orbits within semiclassical calculations . to proceed ,
we first consider the correlation of the density of states . in the semiclassical periodic orbit theory ,
the fluctuation part of the density of states is @xcite @xmath141 the trace of the green s function is the integration over the configuration space with @xmath142 .
the autocorrelation is @xmath143 substituting the semiclassical expression ( [ density ] ) of @xmath144 into the above equation , one gets @xcite
@xmath145 \label{semi - cor}\end{aligned}\ ] ] with @xmath146 .
since the spectrum of the generator @xmath88 is @xmath147 , one can see that in the semiclassical theory , the autocorrelation of the density of states is determined solely by the classical rp resonances .
the transmission spectrum @xmath148is a projection of the density of states , one expects that the autocorrelation of the transmission is also determined by the rp resonances .
the approximation in eq .
( [ semi - cor ] ) is valid only for hyperbolic systems with finite symbolic dynamics . for these systems as explained in sec .
2.1 , the contribution to the autocorrelation @xmath149 is mainly from a few fundamental prime pos . for those po
, one has @xmath150 with @xmath151 the average period . for weakly open or closed system
, the approximation will break down . the experimental quantity whose correlations we examined in sec .
2 is the green s function , while theoretical arguments usually examine the density of states or the time delay @xcite which are however not directly measurable experimentally . to develop the correlation theory for the transmission we measured ,
consider the stationary green s function @xmath152 which is the solution of the following equation @xmath153 with certain boundary condition .
the quantum mechanical propagator is the fourier transform of the green s function .
one can construct a time evolution of a wave packet @xmath154 with @xmath155 with @xmath156 . here
the integration is performed around @xmath157 , with the range @xmath158 and @xmath159 the group velocity of the wave packet .
the integration in the @xmath160 space can be changed into that in the @xmath40 space .
we get @xmath161 the propagator @xmath162 is just the wave function at point @xmath163 due to a @xmath164-function excitation of the system at point @xmath165 and time @xmath166 .
the particle density @xmath167 is thus @xmath168 with the average given by @xmath169 a probability measure is assumed for the above average to converge @xcite . here
@xmath9 is the size of the whole system .
for an open system , @xmath170 .
here we used the definition of the @xmath164 function @xmath171 the autocorrelation of the particle density is @xmath172 using the diagonal approximation , we get @xmath173 note that we have @xmath174 if one assumes that the correlation of the particle density is classical , one has @xcite
@xmath175 where the coefficients @xmath176 are the coupling , and @xmath177 the rp resonances evaluated in wave - vector space . the fourier transformation of ( [ fourier ] ) and ( [ decay - mode ] ) gives @xmath178 with @xmath179 .
the autocorrelation is usually normalized as @xmath180 so that the system size @xmath9 will be cancelled .
thus the above expression is valid for both closed and open systems .
the above arguments closely parallel agam s @xcite recent application of the diagrammatic techniques for disordered systems to open chaotic systems .
we remark that the above arguments differ somewhat from those of agam , whose results are valid when the two probes are far away from each other which implies @xmath181 .
this is the case in the diffusive limit with a poisson distribution of the transmission coefficient @xmath182 .
the above correspondence between the autocorrelation derived from semiclassical theory and the classical resonances is obviously not exact , but requires corrections due to quantum interference .
these corrections are small for open systems , but are expected to dominate for closed systems .
the results presented here establish an alternative path from quantum to classical achieved by considering the correspondence of quantum correlations and classical decay modes . as fig .
[ fig - exp - cor ] shows , the small @xmath5 behavior of @xmath183is universal , while the large @xmath184 behavior shows non - universal oscillations that are completely described by the classical rp resonances .
this demonstrates that in an open system , quantum correlation functions are intimately related to decay laws of classical survival probabilities or evolution of particle ensembles .
most theoretical calculations in quantum chaos using ensemble averaging are restricted only to the leading order consistent with rmt , which probes only the average properties of the system , such as the average escape rate @xmath185 .
the leading rmt contribution to the correlation @xmath1 is lorentzian @xcite @xmath186 , when the classical decay is exponential , which occurs for hyperbolic systems such as the @xmath0-disk billiards .
our earlier experiments established this result in the open microwave billiards @xcite .
as we have shown , our experiments go well beyond the average information represented by rmt - the spectral data contain further details of the fine structure embodied in the decay modes or classical resonances of the system . in terms of the fractal repeller
, the wave mechanical experiments thus yield not only the leading average property @xmath6 related to the fractal dimension , but also further higher order properties of the repeller in terms of the classical resonances .
while we have demonstrated this for the @xmath0-disk billiard , the above result should be valid generally for hyperbolic open chaotic systems .
the relevant classical limit here is the evolution of classical particle ensembles .
the notion of ensemble averaging that is common in quantum mechanics is also useful in classical treatments of chaotic systems .
this is because while the prediction of individual particles is sensitive to initial conditions , the dynamics of appropriate averages of a perturbed system relaxing towards equilibrium is well defined .
the rp resonances determine the decay modes of the dynamical systems evolving from a nonequilibrium state to equilibrium or steady state .
for an open system , gaspard and nicolis @xcite have derived an escape rate formula to relate the coefficient of diffusion and the escape rate of large open systems .
while for an open system , the escape rate @xmath187 is nonvanishing , for closed system , @xmath188 , and the corresponding eigenstate is just the equilibrium state .
the rp resonances are essential ingredients of a description of irreversibility and thermodynamics based upon microscopic chaos @xcite
. however these resonances have remained mathematical objects .
the present experiments have provided physical reality to them .
it is remarkable that we are able to observe classical resonances in a quantum experiment . because microwaves do not interact ( collide ) with each other , ensemble averaging is easier to achieve in microwaves than in fluids .
thus theories that employ ensemble averaging are ideally applicable to the microwave experiments .
this is true for both open systems as shown here , and for disordered billiards @xcite .
the present work shows that the information concerning the classical resonances is coded into the quantum spectrum .
this is somewhat anticipated in periodic orbit theory , where one can observe the strong similarity between the classical ( eq.([classical - ruelle ] ) and ( [ weight - classical ] ) ) and semiclassical ( eq.([semi - ruelle ] ) and ( [ weight - semi ] ) ) ruelle zeta - functions . a one - to - one correspondence exists between @xmath189 and @xmath190 for the @xmath54-disk system as discussed in sec.i .
another example where an exact relation between the classical and quantum resonances exists is the system of a free particle sliding on a compact surface of constant negative curvature , the rp resonances were found by biswas and sinha @xcite to be @xmath188 , @xmath191 for @xmath192 with real @xmath193 the solution of a certain equation while the quantal excitation eigenvalues are @xmath194 .
one would also expect a relation between the distribution of the classical and semiclassical resonances .
the rp resonances have recently been observed in the distribution of the zeroes of the riemann zeta function @xcite .
our work demonstrates that suitable quantum correlations diffuse just like classical observables in an open system .
an interesting question is what will happen when the number @xmath0 of disks are very large , a situation that is frequently referred to as the lorentz gas . due to the exponential separation of the nearby orbits ,
a particle is forced to move through the bouncing paths , or undergo diffusive motion , for long time .
one may argue that the particle motion occurs in a system of size @xmath9 * , * in which it has a diffusive path and hence the average escape time @xmath195 @xmath196 @xmath197 , with @xmath198 the diffusion coefficient .
thus the escape rate @xmath199 @xmath200 . a more general expression for diffusion coefficient @xmath198
is derived through the escape rate formula in ref .
@xcite . for the lorentz gas with large @xmath9 ,
by taking the general form of the escape rate @xmath201 where @xmath202 is the lyapunov exponent and @xmath203 is the kolmogorov - sinai entropy of the fractal repeller @xmath204 , one has @xmath205 ^{2}[\lambda ( { \cal f}_{l})-h_{ks}({\cal f}_{l})]$ ] .
there is a fundamental connection between the present observation of classical resonances and the diffusion operator approach of the supersymmetry theories @xcite of disordered systems .
the statistics of the spectra and eigenfunction of disordered systems is directly related to the classical spectral determinant @xmath206 , where @xmath160 is the energy ( in unit of @xmath207 , the average level spacing ) , and @xmath208 are the non - zero eigenvalues of the diffusion operator @xcite .
these are just the rp resonances that we have observed in open systems . using the supersymmetric techniques , agam , altshuler , and andreev @xcite calculated the level correlation and showed that it can be fully determined by the classical spectrum of the perron - frobenius operator .
bogomolny and keating @xcite arrived at a similar conclusion using the periodic orbit approach .
the applicability of these theories for closed systems is currently a matter of debate @xcite , but the applicability of the supersymmetry theory to disordered systems is well established , and is strongly supported by microwave experiments on disordered billiards @xcite .
as pointed out by berry @xcite , the two limits @xmath209 and @xmath210 do not commute with each other , thus the long - time quantum evolution is fundamentally different from long time classical evolution . the probability @xmath211 that the quantum chaotic system decays at time @xmath212 after its formation follows a power law @xcite @xmath213 for @xmath212 larger than the equilibrium time of the system . here
@xmath214 is the average decay width , @xmath215 the number of open channels . as the system is opened up with more channels , the decay will be almost exponential since the above equation can be approximated by @xmath216 , $ ] for large @xmath215 and the weight of the algebraic tail becomes negligible .
more precisely , as pointed out in ref .
@xcite , there exists a new quantum relaxation time scale @xmath217 , with @xmath218 the heisenberg time . beyond that time
, the quantum correlation decays algebraically . for shorter times
, the quantum correlation decays exponentially in the same way as the classical correlation does . for an open system ,
since the spectrum is continuous , the heisenberg time is actually infinite , and so is the quantum relaxation time @xmath219 .
this is one reason that the classical resonances are clearly visible in the present experiments , and the lorentzian decomposition of the spectral autocorrelation works so well .
the transmission spectrum measured in the microwave experiment directly corresponds to the conductivity measured in electronic quantum dots . in quantum dot experiments though
, the magnetoconductivity @xmath220 is typically measured , so that the resulting autocorrelation @xmath221 is related to the area spectrum , rather than the wave vector dependence @xmath77 measured here which is related to the length spectrum @xcite .
nevertheless similar arguments should be applicable there also , as well as to a variety of other physical systems which are essentially open and chaotic , such as molecular photo - dissociation @xcite and chemical reactions @xcite .
while the coupling to the quantum resonances is well understood , the statistics and also the physical meaning of the experimental coupling @xmath75 to the classical resonances in eq .
( [ auto - cor ] ) deserves further clarification .
it would also be very interesting to examine the role of classical resonances in mixed phase space systems @xcite , such as the open billiards frequently considered in quantum dots where the classical decay is non - exponential @xcite .
more detailed work should be devoted to these and other interesting questions raised by the present results .
we have benefited greatly from formal and informal discussions at the `` quantum chaos y2k '' nobel symposium . we thank p. cvitanovi and o. agam for important insights into the semiclassical arguments for the correspondence between quantum and classical correlations .
we also thank o. bohigas , b. altshuler , j. keating , e. bogomolny , d. ullmo and p. lebeouf for useful discussions .
cvitanovi , p. and eckhardt , b. , nonlinearity * 6 * , 277 ( 1993 ) ; cvitanovi , p. , artuso , r. , mainieri , r. , and vattay , g. , `` classical and quantum chaos : a cyclist treatise '' _ , _
http://www.nbi.dk / chaosbook/. fyodorov , y. v. , titov , m. and sommers , h. -j .
e * 58 * , r1195 ( 1998 ) ; fyodorov , y. y. and khoruzhenko , b. a. , phys .
lett . * 83 * , 65 ( 1999 ) , fyodorov , y. v. and sommers , h. -j . , j. math .
phys . * 38 * , 1918 ( 1997 ) . | we show that the autocorrelation of quantum spectra of an open chaotic system is well described by the classical ruelle - pollicott resonances of the associated chaotic strange repeller . this correspondence is demonstrated utilizing microwave experiments on 2-d @xmath0-disk billiard geometries , by determination of the wave - vector autocorrelation @xmath1 from the experimental quantum spectra @xmath2 .
the correspondence is also established via numerical experiments " that simulate @xmath2 and @xmath1 using periodic orbit calculations of the quantum and classical resonances .
semiclassical arguments that relate quantum and classical correlation functions in terms of fluctuations of the density of states and correlations of particle density are also examined and support the experimental results .
the results establish a correspondence between quantum spectral correlations and classical decay modes in an open systems .
pacs ref : 05.45.mt , 05.45.ac , 03.65.sq , 84.40.-x |
it is widely accepted that quantum chromodynamics ( qcd ) is the correct theory for strong interactions .
qcd is a renormalizable quantum field theory which is closely modeled after quantum electrodynamics ( qed ) , the most accurate physical theory we have to date .
however , in the low - energy regime , qcd tells us that the interactions between quarks and gluons are strong , so that quark - gluon dynamics becomes non - perturbative in nature .
understanding the structures of hadrons directly from qcd remains an outstanding problem , and there is no indication that it will be solved in the foreseeable future . in 1989 , it was realized that , in low energy situations where the typical gluon momenta are small compared with the heavy quark mass @xmath0 , qcd dynamics becomes independent of the flavor and spin of the heavy quark @xcite . for the heavy flavors ,
this new symmetry called heavy quark symmetry ( hqs ) . of course , even in this infinite heavy quark mass limit , low energy qcd dynamics remains non - perturbative , and what hqs can do for us is to relate otherwise unrelated static and transition properties of heavy hadrons , and hence enormously reduces the complexity of theoretical analysis .
in other words , hqs allows us to factorize the complicated light quark and gluon dynamics from that of the heavy one , and thus provides a clearer physical picture in the study of heavy quark physics . beyond the symmetry limit
, a heavy quark effective theory ( hqet ) can be developed by systematically expanding the qcd lagrangian in powers of @xmath1 , with which hqs breaking effects can be studied order by order @xcite . in the experimental area ,
all masses of @xmath2-wave charmed hadrons and bottomed mesons which containing one heavy quark are found at present .
however , except the particle @xmath3 was already found in the early 1980 s , there has not been significant progress in searching @xmath2-wave bottomed baryons until these months . recently some bottomed baryons were discovered at fermilab .
they are the exotic relatives of the proton and neutron @xmath4 and @xmath5 by cdf collaboration @xcite and the triple - scoop baryon @xmath6 by d0 and cdf collaborations @xcite .
it is reasonable that the remainder particles , which include @xmath7 , @xmath8 , @xmath9 , and @xmath10 , will be observed in the foreseeable future .
all these heavy hadrons provide a testing ground for hqet with the phenomenological models to the low - energy regime of qcd . in this paper
we focus on one static property , that is , the mass spectrum of heavy hadrons and combine hqet with the known experimental data to predict the mass splitting of some heavy baryons .
the phenomenological models are not needed here .
the paper is organized as follows . in sec .
ii brief introductory notes are given for hqet . in sec
iii we formulate the hyperfine mass splitting for heavy mesons and baryons . in sec .
iv we evaluate the numerical results and predict some mass differences between heavy baryons . finally , the conclusion is given in sec . v.
the full qcd lagrangian for a heavy quark ( @xmath11 , @xmath12 , or @xmath13 ) is given by _
q = |q ( i_d^- m_q ) q , [ lag ] where @xmath14 with @xmath15 . inside a hadronic bound state containing a heavy quark
, the heavy quark @xmath16 interacts with the light degrees of freedom by exchanging gluons with momenta of order @xmath17 , which is much smaller than its mass @xmath18 .
consequently , the heavy quark is close to its mass shell , and its velocity does not deviate much from the hadron s four - velocity @xmath19 .
in other words , the heavy quark s momentum @xmath20 is close to the
kinetic " momentum @xmath21 resulting from the hadron s motion p^_q = m_q v^+ k^ , [ pk ] where @xmath22 is the so - called residual " momentum and is of order @xmath17 .
to describe the properties of such a system which contains a very heavy quark , it is appropriate to consider the limit @xmath23 with @xmath19 and @xmath24 being kept fixed . in this limit ,
it is evident that the quantity @xmath21 is frozen out " from the qcd dynamics , so it is appropriate to introduce the large " and small " component fields @xmath25 and @xmath26 , which is related to the original field @xmath27 by h_v ( x ) = e^im_q vx p_+ q(x ) , + h_v ( x ) = e^im_q vx p_- q(x ) , where @xmath28 and @xmath29 are the positive and negative energy projection operators respectively : @xmath30 .
so that q ( x ) = e^-im_q v x .
[ hh ] it is clear that @xmath25 annihilates a heavy quark with velocity @xmath19 , while @xmath26 creates a heavy antiquark with velocity @xmath19 . in the heavy hadron s rest frame @xmath31 , @xmath32
correspond to the upper ( lower ) two components of @xmath33 . in terms of the new fields , the qcd lagrangian for a heavy quark given by eq .
( [ lag ] ) takes the following form _
q = |h_v i vd h_v - |h_v ( i vd + 2 m_q ) h_v + |h_v i d _ h_v + |h_v i d _ h_v [ nlag ] where @xmath34 is orthogonal to the heavy quark velocity , @xmath35 . from eq .
( [ nlag ] ) , we see that @xmath25 describes massless degrees of freedom , whereas @xmath26 corresponds to fluctuations with twice the heavy quark mass .
the heavy degrees of freedom represented by @xmath26 can be eliminated using the equations of motion of qcd .
substituting eq .
( [ hh ] ) into @xmath36 gives i d h_v + ( i d - 2 m_q ) h_v = 0 .
multiplying this equation by @xmath37 , one obtains -i v d h_v = i d_h_v , [ 1 - 14 ] + ( i v d + 2 m_q ) h_v = i d_h_v .
[ 1 - 15 ] the second equation can be solved schematically to give h_v = 1 i d_h_v , which shows that the small component field @xmath26 is indeed of order @xmath1 .
one can insert this solution back into eq .
( [ 1 - 14 ] ) to obtain the equation of motion for @xmath25 .
it is easy to check that the resulting equation follows from the effective lagrangian _ q , eff = |h_v i vd h_v + |h_v i d _ 1 i d_h_v .
[ lm ] @xmath38 is the lagrangian of the heavy quark effective theory ( hqet ) , and the second term of eq .
( [ lm ] ) allows for a systematic expansion in terms of @xmath39 .
taking into account that @xmath40 , and using the identity p_+ i d _ i d_p_+ = p_+ p_+ , where g^ = t_a g^_a = i[d^,d^ ] is the gluon field strength tensor , one finds that _ q , eff = |h_v i vd h_v + 1 |h_v ( id_)^2 h_v + g |h_v _ g^ h_v + o ( 1 ) . [ expand ] the new operators at order @xmath1 are _ 1 = 1 |h_v ( id_)^2 h_v , [ o1 ] + o_2 = g |h_v ^ g _
h_v , [ o2 ] where @xmath41 is the gauge invariant extension of the kinetic energy arising from the off - shell residual motion of the heavy quark , and @xmath42 describes the color magnetic interaction of the heavy quark spin with the gluon field .
it is clear that both @xmath41 and @xmath42 break the flavor symmetry , while @xmath42 breaks the spin symmetry as well .
for instance , @xmath43 would introduce a common shift to the masses of pseudoscalar and vector heavy mesons , and @xmath42 is responsible for the color hyperfine mass splittings @xmath44 .
the full expansion of @xmath38 in @xmath39 can be organized as follow _
q , eff = _ n=0 ( 1)^n l_n [ expansion ] where the first few terms are given by _ 0 & = & |h_v ( i v d ) h_v , + l_1 & = & |h_v ( id_)^2 h_v + g |h_v _ g^ h_v , + l_2 & = & g |h_v _ v_i d^g^ h_v + g |h_v v _
i d_g^ h_v .
we must emphasize that this effective theory comes from first principle directly , and it is in terms of the power of @xmath1 , which is small enough , to calculate the physical quantities which concerning heavy quarks perturbatively , i.e. , order by order . if one chooses the appropriate frame and phenomenological model , then one can handle many physical processes systematically .
we also note that , as shown in eq .
( [ expansion ] ) , the information of heavy quark flavor is only involved in the factor @xmath45 . in other words ,
all @xmath46 s are independent of the heavy quark flavor .
we will use this property to evaluate the mass splittings of some heavy hadrons .
first , we consider the hyperfine mass splitting between pseudoscalar and vector heavy mesons .
the operators that break hqs to order @xmath1 are @xmath41 and @xmath42 given in eqs .
( [ o1 ] ) and ( [ o2 ] ) respectively .
@xmath41 can be separated into a kinetic energy piece and a one - gluon exchange piece : _ 1 = o_1k + o_1 g where _ 1k & & -1 |h_v [ _ ^+ ( v)^2]h_v , + o_1 g & & -g_s |h_v [ ( p+p)_- v(p+p)v_]a^ h_v , also , @xmath42 can be reexpressed as _ 2 = -g_s t^a _ ^a^a . with the @xmath1 corrections
included , the heavy meson masses can be expressed as m_m = m_q + |^q -1(_1^q + d_m ^q_2),[mmm ] where @xmath47 comes from @xmath41 and @xmath48 comes from @xmath42 .
@xmath49 receive two different contributions , one from @xmath50 and the other from @xmath51 , thus _
1^q = _ 1k + _ 1g^q .
the parameter @xmath52 in eq .
( [ mmm ] ) is the residual mass of heavy mesons in the heavy quark limit . in other words
, @xmath52 is independent of heavy quark flavor .
@xmath53 comes from the heavy quark kinetic energy , @xmath54 and @xmath48 are respectively the chromoelectric and chromomagnetic contributions .
@xmath49 parameterizes the common mass shift for the pseudoscalar and vector mesons , and @xmath48 accounts for the hyperfine mass splitting . in both the non - relativistic and relativistic quark models ,
the hyperfine mass splitting comes from a spin - spin interaction of the form h_hf ~s_q j_l . where @xmath55 is the spin operator of the heavy quark and @xmath56 is the angular momentum operator of the light degree of freedom .
thus d_m & = & -m(v ) |4 s_q j_l |m(v ) + & = & -2[s_m(s_m+1 ) - s_q(s_q+1)- j_l(j_l+1 ) ] , where @xmath57 is spin quantum number of the meson @xmath58 .
consequently , @xmath59 for a vector meson , and @xmath60 for a pseudoscalar meson .
therefore , we obtain the hyperfine mass splitting , m_vp m_v - m_p = 2 _ 2^q .
[ masslp ] we next consider the hyperfine mass splitting among the baryons containing one heavy quark ( @xmath16 ) and two light quarks ( @xmath61 , @xmath62 ) .
each light quark is in a triplet @xmath63 of the flavor @xmath64 .
since @xmath65 and the lowest lying light quark state has @xmath66 and @xmath67 ( @xmath68-wave ) , there are two different diquarks : a symmetric sextet ( @xmath69 ) and an antisymmetric antitriplet ( @xmath70 ) .
when the diquark combines with a heavy quark , the sextet contains both spin-@xmath71 ( @xmath72 ) and spin-@xmath73 ( @xmath74 ) baryons , and the antitriplet contains only spin-@xmath71 ( @xmath75 ) baryons .
the multiplets @xmath76 and @xmath77 are illustrated in fig .
1 ( a ) and ( b ) , respectively , and their quantum numbers are listed in table i. 8.5 cm .[tab1 ] the @xmath2-wave heavy baryons and their quantum number , where the subscript @xmath78 stands for the quantum number of the two light quarks . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ]
in this paper , based on hqet , we have presented a formalism to describe the hyperfine mass splittings of the heavy baryons . furthermore ,
through the known experimental data in charmed sector , we predicted the hyperfine mass differences in bottomed sector .
the parameters appearing in this analysis are the ratio @xmath79 and the residual mass of heavy baryons in the heavy quark limit @xmath80 . on the one hand
the ratio @xmath79 is fixed by the experimental values of heavy mesons , and on the other hand the residual mass difference @xmath81 , due to it is independent of heavy flavor , is obtained by the known mass differences of charmed baryons .
the prediction of @xmath82 is in agreement with the experimental values , we expect the deviations of the other predictive mass differences are all small for the future experimental data . in addition , in both heavy meson and baryon systems , we find that the flavor @xmath64 breaking effect of the hyperfine mass splitting is very small .
finally we also estimated the masses of @xmath7 and @xmath8 and used the gell - mann / okubo formula to calculate the masses of @xmath9 and @xmath10 .
the uncertainties of these four heavy baryon masses mainly come from the error of the measured value @xmath83 . to get the more confidence in hqet ,
the more precise experimental data are needed . | we study the hyperfine mass differences of heavy hadrons in the heavy quark effect theory ( hqet ) .
the effects of one - gluon exchange interaction are considered for the heavy mesons and baryons .
base on the known experimental data , we predict the masses of some heavy baryons in a model - independent way . |
thick branes have been proposed as a smooth generalization of the randall - sundrum scenario @xcite . in this model ,
five - dimensional gravity is coupled to scalar fields .
thick brane models consist in a more realistic scenario than the randal - sundrum one , since no singularities appear due to the form of the scalar potential functions . as a matter of fact
, thick brane models have been comprehensively used in the task of localization of physical fields on the brane .
the importance of this subject stays in the fact that the introduction of extra dimensions affects both gravitational interactions and particle physics phenomenology , and leads to modifications in the standard cosmology .
if the extra dimensions indeed exist , it will inevitably change our ideas about the universe .
the quest of field localization can guide us to which kind of brane structure is more acceptable phenomenologically @xcite . in this context
, gravitons and fermions , as well as gauge fields can be localized on the brane in thick brane models .
gauge fields , in particular , are localized only with the help of the dilaton field . the kalb - ramond field localization in this scenario
was also studied by @xcite . there
the use of the dilaton was again necessary in order to localize the kalb - ramond field on the brane . on the other stand point
, scenarios have been proposed where thick brane solutions are extended to spacetimes with dimension more than five @xcite . among these works , we have some where branes are embedded in a bulk with a string - like metric .
the mainly motivation to study branes in the presence of a string - like bulk comes from the fact that most of the standard model fields are localized on a string - like defect .
for example , spin-@xmath0 , spin-@xmath1 , spin-@xmath2 , spin-@xmath3 and spin-@xmath4 fields are all localized on a string - like structure . particularly , the bosonic fields are localized with exponentially decreasing warp factor , and the fermionic fields are localized on the defect with increasing warp factor @xcite .
even more interesting is the fact that spin-@xmath1 vector @xcite , as well as the kalb - ramond field @xcite , which are not localized on a domain wall in randal - sundrum model , can be localized in the string - like defect .
however , most of the thick brane models in six dimensional scenarios , proposed so far , have been suffering from some drawbacks .
the first difficult is related with the introduction of scalar fields as a matter - energy source in the equations . in this case
it is very difficult to find analytical solution to the scalar field and to the warp - factor as well .
koley and kar @xcite have suggested a model where analytical solutions can be found in a six dimensional scenario , however they run into a second difficult .
this difficult is related with the positivity of the components of the energy - momentum tensor and has been found by other authors also @xcite .
finally , field localization are not possible for any field in the thick branes defined in ref .
@xcite , at least for physically acceptable solutions . on the other hand , in a recent work @xcite ,
a model was proposed where a brane is made of a scalar field with bounce - type configurations and embedded in a bulk with a string - like metric .
this model produces a sound ads scenario where none of the important physical quantities is infinite . among these quantities
are the components of the energy momentum tensor , which have its positivity ensured by a suitable choice of the bounce configurations .
another advantage of this model is that the warp factor can be obtained analytically from the equations of motion for the scalar field , obtaining as a result a thick brane configuration , in a six dimensional context .
it has been shown that scalar field localization is suitable in the scenario proposed in ref .
@xcite , paving the way in the sense of localization of other fields .
therefore , in the present work we will study the possibility of localization of vector and fermion fields in these scenario , in order to test its applicability and robustness . this paper is organized as follows . in section [ section2 ]
, we address the model introduced in ref .
@xcite , where a bulk scalar field with bounce - type configurations generates a brane which is embedded in a bulk with a string - like metric . in section [ section3 ]
we address the vector field localization , and in section [ section4 ] the fermion field localization is established . on the other hand , section [ section4b ] cope with qualitatively analysis of the fermionic massive modes .
section [ section5 ] is devoted to remarks and conclusions .
the use of bulk scalar fields to generate branes was introduced by goldberger and wise @xcite , and has been largely studied in the literature @xcite . in the six - dimensional context , koley and kar @xcite
have built a scenario where the brane is made of scalar fields and analytical `` thin brane '' solutions have been found out .
several progress have been obtained in the work by koley and kar in the intend of construct brane solutions in six dimensions , as well as , in the task of localize physical fields . however , some troubles with the energy conditions ( wec , sec , nec ) @xcite were found . in this model
, the energy momentum tensor violates all the energy conditions since its components are not positive defined .
recently , an ads type solution was found in a model which assumes a six dimensional action for a bulk scalar field in a double well @xmath5 potential minimally coupled to gravity in the presence of a cosmological constant @xcite . in this scenario , which will underly the present article ,
it is admitted that the scalar field equation possess bounce - like statics solutions depending only on the radial extra dimension , where the simplest is @xmath6 .
the model is described by the action @xmath7\ ; , \label{action}\ ] ] where @xmath8 is the @xmath9-dimensional gravitational constant , and @xmath10 is the bulk cosmological constant .
the fields live in a string - like scenario with the following metric @xmath11 where @xmath12 denote the @xmath9-dimensional space - time indices , @xmath13 the @xmath14-dimensional brane ones , and @xmath15 denote the @xmath2-extra spatial dimension ones . also @xmath16 , @xmath17 and @xmath18 . in the same way of the model introduced by koley and kar ,
the one introduced in @xcite has the advantages to be analytical
. however , the introduction of the bounce - type configurations to the scalar field that generates the brane supports a way to solve the problems with the energy conditions since the energy density may be positive or negative on the brane depending on the choice of the bounce configurations .
moreover , the finiteness of the relation between the four ( @xmath19 ) and six ( @xmath20 ) dimensional reduced plank scale @xcite is ensured by the form of the warp factor found out by the authors which is given by @xmath21 in the expression above @xmath22 .
moreover , the warp factor found out is equal to 1 at @xmath23 which ensures that on the brane one has a 4d minkowski space - time .
besides , as @xmath24 goes to zero or infinity , the warp factor goes to @xmath0 . in the reference @xcite
, the authors also pointed the interesting possibility of localization of the standard model fields in this scenario . in this way
, the scalar field localization has been implemented paving the way for the localization of other fields .
an interesting result is that any non - gravitational trapping mechanism has not been necessary to localize scalar field in this model , which can been seen as an advantage when compared with results of dzhunushaliev and folomeev @xcite . in the present work
, we will deal with vector and fermionic fields localization .
it is know that it is possible to localize chiral fermions in the `` 5d version of this model '' @xcite .
however , to localize vector field in this set up , in five dimensions , we need to have a dilaton field present in the model forming a bounce - gravity - dilaton system " @xcite .
we expect , in this work , to localize fields in this scenario that is more realistic than the rs model ones , without the necessity of the dilaton field .
[ vetor - local ] in this section we will address the issue of the vector field localization . as we will see , it is possible to localize the vector field in this scenario if one has an exponentially decreasing warp factor , without the necessity of the dilaton field , as in the case of the the scalar field discussed in ref .
@xcite . in order to deal with the vector field localization ,
let us introduce the action @xmath25 where @xmath26 . from the action above ,
one obtains the equation of motion @xmath27 which results in @xmath28 this equation can be written in terms of equations for the vector field components as follows @xmath29 @xmath30 and @xmath31 as has been done in @xcite , if we choose the gauge condition @xmath32 and assume the decomposition @xmath33 and @xmath34 we can show that there exist the s - wave ( l = 0 ) constant solution @xmath35 and @xmath36 .
note that we assume that @xmath37 , where @xmath38 is defined by @xmath39 . by substituting the constant solution in the initial action , the resultant integral in the variable @xmath24 is @xmath40 in order to have zero mode localization for the vector field in this model
, we need that @xmath41 to be finite .
it is clear that in the case for @xmath42 and with @xmath43 given by eq .
the condition above is satisfied * ( as can be seen in figures ( [ fig.ia1 ] ) and ( [ fig.i1]))*. it is interesting to note that , in the domain wall case , this term is not present . that is why in 5d domain wall it is not possible to localize the vector field only by means of the gravitational interaction . for further reference we may say that it is possible to obtain the zero mode localization for the vector field without imposing the gauge condition @xmath32 . indeed , if one consider a @xmath24 dependence of @xmath44 , say , @xmath45 , the system ( - ) will assume the form @xmath46 @xmath47 and @xmath48 if the fields @xmath49 , and @xmath50 are decomposed as in eq .
and eq . it is possible to obtain the constant solution @xmath35 and @xmath36 for the zero mode and s - wave .
but in this case the function @xmath44 has to satisfy the equation @xmath51 whose general solution may be expressed as follows @xmath52 where @xmath53 are integration constants .
obviously this equation admits a more simple solution as @xmath54 . for this specific case
the results obtained for the vector field localization would be the same found above for @xmath32 .
however , as will be seen this constant solution will be relevant for the localization of the fermion field in the next section .
it is worthwhile to mention here that other solutions for @xmath44 are possible , since they are finite for all @xmath24 _ and _ @xmath55 .
however the simplest solution to achieve fermion confinement is @xmath56 . + +
in this section , we will address the issue of fermionic field localization in a string - like defect scenario . at first , we will following the procedure of the reference @xcite , where fermionic fields are localized on a string - like defect .
however , as will be shown , it is not possible to localize the fermionic field in this scenario if one requires an exponentially decreasing warp factor , which is possible in the case of scalar and vector fields as we have seen . on another standpoint there
exist an approach developed by liu and collaborators @xcite , where it is possible to localize the fermionic field in the case of a decreasing warp factor . in this approach ,
one modifies the covariant derivative by adding a minimal coupling between the fields . on other hand , here we propose a mechanism in order to localize the fermion field using the minimal coupling in the bouncing brane keeping an decreasing warp factor in scenarios of codimension two brane . to begin with ,
let us proceed as in the reference @xcite .
we have that the fermionic field action can be written in a scenario with six dimensions as @xmath57 and the equation of motion related with this action is @xmath58 where the matrices @xmath59 represent the dirac matrices in a curved spacetime .
these matrices are related with the dirac matrices in the flat spacetime as @xmath60 where the _ vielbein _ @xmath61 is given by the relation @xmath62 the covariant derivative has the standard form @xmath63 where the spin connection @xmath64 is defined by @xmath65 in order to write explicitly the equation , we have to calculate the matrices @xmath59 as well as the covariant derivative . using the metric and the equation
, we have that the relation between the dirac matrices in a curved space - time and the dirac matrices in the flat space - times is given by @xmath66 the nonvanishing components of the spin connection are @xmath67 moreover , one can explicitly write the covariant derivative components as @xmath68 in order to write the equations of motion to these fields , we have to set the way how the dirac matrices act on the _ spinor _ @xmath69 .
this approach was presented by the references @xcite , which we follow closely .
first , let us assume that the spinor can be written in two parts , the _ right _ part @xmath70 and the _ left _ part @xmath71 as @xmath72 the @xmath73 matrices act on these spinor as @xmath74 or yet in terms of the @xmath75 matrices in the flat spacetime , as @xmath76 naturally , for @xmath77 one has @xmath78 these equations can still be put in the form @xmath79
@xmath80 we require here that @xmath81 must satisfies the dirac equation on the brane , namely @xmath82 .
thus , taking into account the s - wave case , and the equations , , , , and , we will have that the equations of motion can be written as @xmath83 the general solution to the equation above is given by @xmath84 where @xmath85 is a constant of integration .
it is still necessary to verify if the solution is normalizable . in order to do this ,
we need analyze the action with the spinor @xmath86 replaced by this solution and verify if the resultant integral in the variable @xmath24 is finite . from this ,
the interesting integral for the case analyzed here assumes the following form @xmath87 at last , let us replace the expression for @xmath88 in the equation in a way that , for the case where @xmath89 , we have @xmath90 we note that no localization of the fermionic fields is possible in the case of a smooth warp factor given by @xmath91 .
even the possibility of assume @xmath92 is not possible since @xmath93 . in consequence , in our case
it is not possible to appeal to a growing warp factor .
therefore , the only alternative that remains is to use the treatment introduced in the reference @xcite which consists in modify the covariant derivative by adding a term of minimal coupling .
if one does this , the new covariant derivative reads @xmath94 where @xmath95 is a gauge field , @xmath96 the electrical charge , and @xmath97 is the imaginary unit .
such modification does not change the relation between the gamma matrices , neither the nonvanishing components of the spin connection . on the other hand , the covariant derivatives components in this case ,
will be changed to the following forms @xmath98 \psi , \nonumber \\ \\
\nonumber d_{r } \psi & = & \left ( \partial _ { r } - i e
a{r}\right ) \psi , \\ \\
\nonumber d_{\theta } \psi & = & \left [ \partial _ { \theta } - ( q'/8q ) \gamma _ { r } \gamma _ { \theta } - i e a_{\theta } \right ] \psi.\end{aligned}\ ] ] using the conditions , , and , taking into account only the zero mode of the field and for the s - wave , the equation of motion for the _ right mode _ reads @xmath99 where we have assumed that the dirac equation @xmath100 is valid on the brane , and the field @xmath95 can be decomposed in its components @xmath101 , @xmath102 , and @xmath103 . the solution of the equation is given by @xmath104 inserting this solution in the action , the integral in the variable @xmath24 reads @xmath105 the form of field @xmath44 in the above integral is essential to ensure the zero mode localization , i.e , to ensure the finiteness of the integral . in this case
it is sufficient that @xmath106 .
the integral ( ) is naturally convergent for @xmath107 , so we have to study it in the other limit , @xmath108 .
it is easy to see that in this limit the function @xmath109 is linear on @xmath24 and , in this case , it is possible to write @xmath110 . in this situation
we can easily see that ( ) converges .
indeed , it will assumes the simple form @xmath111 now , setting @xmath112 and since the exponential function always assumes values greater than values from a linear function , one sees that the integral above is convergent ( as can be seen in figures ( [ fig.ia12 ] ) and ( [ fig.i12 ] ) ) .
moreover , for left handed fermions we must set @xmath113 in order to converge that integral . + +
in this section we treat the fermionic massive modes using a qualitative analysis . in order to do this , we put the equations of motion of the massive spinor in the form of the schrdinger equation .
first , we consider the massive counterpart of the eq .. indeed , with help of eq .
we arrive at @xmath114\alpha_{r , l}(r)=\pm m p^{-\frac{1}{2}}\alpha_{l , r } ( r),\end{aligned}\ ] ] where @xmath115 .
note that the expression is composed by two equations with coupled quiralities . applying the following change in the independent variable @xmath116 and decoupling the left mode and the right mode we have @xmath117\left [ \partial _
{ z } + h_{l , r}(z)\right]\alpha_{r , l}(z)=-m^2\alpha _ { r , l}(z).\end{aligned}\ ] ] through a new change of variable @xmath118\alpha_{r , l}(z)$ ] , we can put eq . in the form of schrdinger equation , namely @xmath119 \right)\tilde{\alpha}_{r , l}(z)=m^2\tilde{\alpha}_{r , l}(z).\end{aligned}\ ] ]
therefore our potential can be written in term of @xmath24 as @xmath120 .\end{aligned}\ ] ] now we study the choice @xmath121 which it was convenient for confinement of the zero modes of the vectorial and spinorial fields . for this choice ,
we present a sketch of the potentials in fig.([fig.vbounce ] ) .
we note from the fig.([fig.vbounce ] ) that both potentials are gap free when @xmath122 .
also we note that only the left mode potential assumes a volcano type form .
this feature indicates that only left handed fermions are confined in our brane . as a matter of fact ,
the potential for right handed fermions is always attractive .
also it reaches their asymptotic value on the infinity under the axis , which would not guarantee the right mode localization @xcite .
the complete analysis of the massive modes must to include the calculation of the resonant modes .
however this calculation requires the numerical solution of equation .
we defer this numerical analysis for a future work .
this work adds results in studies about thick brane in codimension two spaces .
here we have implemented a mechanism in order to localize vector and fermion fields in the scenario introduced in ref .
@xcite , where a thick brane is generated from a scalar field on a string - like defect .
we have found that both , vector and fermion fields , can be localized in this scenario only with the gravitational interaction , which confirms the applicability and robustness of this model .
it is worthwhile to mention that in order to localize fermion field we show that , for the first time treating a bouncing brane , a previous localization of the gauge field is required .
indeed , we must have a component of the gauge field in the direction of the angular extra coordinate to obtain a convergent result for the integral of localization . in the literature , this component is usually turns to be zero as a gauge choice @xcite . in the case under analysis here
it was not necessary to appeal to a growing warp factor in order to localize fermion fields , since no localization of these fields is possible in the case of the smooth warp factor .
therefore , the only alternative in order to localize fermions in this context it is to modify the covariant derivative by adding a term of minimal coupling @xcite , which was never used for the bouncing brane . in this case
, the fermionic field can be localized on the brane .
finally , we perform a qualitative analysis of the fermionic massive modes .
we conclude that only left handed fermions could be localized in the brane .
a numerical analysis of the massive modes is deferred for a future work .
the authors thank the fundao cearense de apoio ao desenvolvimento cientfico e tecnolgico ( funcap ) , the coordenao de aperfeioamento de pessoal de n ' ivel superior ( capes ) , and the conselho nacional de desenvolvimento cient ' ifico e tecnol ' ogico ( cnpq ) for financial support . | in a recent work , a model has been proposed where a brane is made of a scalar field with bounce - type configurations and embedded in a bulk with a string - like metric .
this model produces an ads scenario where the components of the energy momentum tensor are finite and have its positivity ensured by a suitable choice of the bounce configurations . in the present work
, we study the issue of gauge and fermion field localization in this scenario .
in contrast with the five dimensional case here the gauge field is localized without the dilaton contribution . nevertheless , it is remarkable that the localization of the fermion field depends on the introduction of a minimal coupling with the angular component of the gauge field , which differs clearly from five dimensional scenarios .
furthermore , we perform a qualitative analysis of the fermionic massive modes and conclude that only left handed fermions could be localized in the brane . |
increasing computational power and algorithmic advancements are making many computational materials problems more tractable .
for example , density functional theory ( dft ) is used to assess the stability of potential metal alloys with high accuracy . however ,
dft computational burdens prevent feasible exploration of all possible configurations of a system . in certain cases ,
one can map first - principles results on to a faster hamiltonian , the cluster expansion ( ce ) @xcite . over the past 30 years
, ce has been used in combination with first - principles calculations to predict the stability of metal alloys @xcite , to study the stability of oxides @xcite , and to model interaction and ordering phenomena at metal surfaces @xcite . numerical error and
_ relaxation _ effects decrease the predictive power of ce models .
the aim of this paper is to demonstrate the effects of both and to provide a heuristic for knowing when a reliable ce model can be expected for a particular material system .
ce treats alloys as a purely configurational problem , i.e. , a problem of decorating a fixed lattice with the alloying elements @xcite .
however , ce models are usually constructed with data taken from `` relaxed '' first - principles calculations where the individual atoms assume positions that minimize the total energy , displaced from ideal lattice positions .
unfortunately , cluster expansions of systems with larger lattice relaxation converge more slowly than cluster expansions for unrelaxed systems @xcite .
in fact , ce with increased relaxation may fail to converge altogether . no rigorous description of conditions for when the ce breakdown occurs exists in the literature .
a persistent question in the ce community regards the impact of relaxation on the accuracy of the cluster expansion .
proponents of ce argue that the ce formalism holds even when the training structures are relaxed because there is a one - to - one correspondence in configurational space between relaxed and unrelaxed structures . in this paper
, we demonstrate a relationship between relaxation and loss of sparsity in the ce model .
as sparsity decreases , the accuracy of ce prediction decreases .
in addition to the effects of relaxation , we also examine the impact of numerical error on the reliability of the ce fits .
there are several sources of numerical error : approximations to the physics of the model , the number of @xmath0-points , the smearing method , basis set sizes and types , etc .
most previous studies @xcite only examine the effect of gaussian errors on the ce model , but arnold et al .
@xcite also investigated systematic error ( round - off and saturation error ) .
they showed that , above a certain threshold , the ce model fails to recover the correct answer , that is , the ce model started to incorporate spurious terms ( i.e. , sparsity was reduced ) . a primary question that we seek to answer
is how the shape of the error distribution impacts predictive performance of a ce model . in this study
, we quantify the effects of : 1 ) relaxation , by comparing ce fits for relaxed and unrelaxed data sets , and 2 ) numerical error , by adding different error distributions ( i.e. , gaussian , skewed , etc . )
to ideal ce models .
we study more than one hundred hamiltonians ranging from very simple pair potentials to first - principles dft hamiltonians .
we present a heuristic for judging the quality of the ce fits .
we find that a small mean - squared displacement is indicative of a good ce model . in agreement with past studies ,
we show that the predictive power of ce is lowered when the level of error is increased .
we find that there is no clear correlation between the shape of the error profile and the ce predictive power .
it is possible to decide whether the computational cost of generating ce fitting data is worthwhile by examining the degree of relaxation in a smaller set of 50150 structures .
relaxation is distinct from numerical error
it is not an error but it has a similar negative effect . when relaxations are significant , it is less likely that a reliable ce model exists .
relaxation is a systematic form of distortion , the local adjustment of atomic positions to accommodate atoms of different sizes .
atoms `` relax '' away from ideal lattice sites to reduce the energy , with larger atoms taking up more room , smaller atoms giving up volume .
the type of relaxations ( i.e. , the distortions that are possible ) for a particular unit cell are limited by the symmetry of the initially undistorted case , as shown in fig .
[ fig : symmetry_allowed_relaxations ] . in the rectangular case ( left ) ,
the unit cell aspect ratio may change without changing the initial rectangular symmetry . at the same time , the position of the blue atom is _ not allowed _ to change because doing so would destroy rectangular symmetry .
in contrast , the two blue atoms in the similar structure shown in the right panel of the figure can move horizontally without reducing the symmetry .
( color online ) symmetry- allowed distortions for two different unit cells .
the atomic positions of the cell on the left do not have any symmmetry - allowed degrees of freedom , but the aspect ratio of the unit cell is allowed to change . for the unit cell on the right , the horizontal positions of the atoms in the middle layer may change without destroying the symmetry .
( the unit cell aspect ratio may also change.),scaledwidth=55.0% ] conceptually , the cluster expansion is a technique that describes the local environment around an atom and then sums up all the `` atomic energies '' ( environments in a unit cell ) to determine a total energy for the unit cell . for the cluster expansion model to be sparse to be a predictive model with few parameters it relies on the premise that any specific local neighborhood contributes the same atomic energy to the total energy regardless of the crystal in which it is embedded . for example , the top row of fig .
[ fig : relaxex ] shows the same local environment ( denoted by the hexagon around the central blue atom ) embedded in two distinct crystals .
if the contribution of this local environment to the total energy is the same in both cases , then the cluster expansion of the energy will be sparse .
the effect of relaxation on the sparsity becomes clear in the bottom row of fig .
[ fig : relaxex ] . in the left - hand case [ panel ( a )
] , the crystal relaxes dramatically and the central blue atom is now _ four - fold coordinated _ entirely by red atoms .
by contrast , in the right - hand case [ panel ( b ) ] , a collapse of the layers is not possible and the blue atoms are allowed by symmetry to move closer to each other . from the point of view of the cluster expansion ,
the local environments of the central blue atom are the same for both cases .
this fact , that two different relaxed local environments have identical descriptions in the cluster expansion basis , leads to a slow convergence of cluster expansion models .
the problem is severe when the atomic mismatch is large and relaxations are significant ( i.e. , when atoms move far from the ideal lattice positions . ) ( color online ) relaxation scheme .
the top images show the original unrelaxed configurations , while the bottom figures show the relaxed configuration .
the left images ( a ) shows the relaxation where the hexagon is contracted as shown by the black arrows in bottom left figure .
the relaxation in the right images ( b ) is restricted to displacement of the blue atoms as shown by the black arrows in bottom right figure .
, scaledwidth=55.0% ] we investigated the predictive power of cluster expansions using data from more than one hundred hamiltonians generated from density functional theory ( dft ) , the embedded atom method , lennard - jones potential and stillinger - weber potential . to investigate the effects of relaxation
, we examined different metrics to measure the degree of atomic relaxation in a crystal configuration .
first - principles dft calculations have been used to simulate metal alloys and for building cluster expansion models @xcite . however , dft calculations are too expensive to extensively examine the relaxation in many different systems ( lattice mismatch ) .
thus , we examine other methods such as the embedded atom method ( eam ) which is a multibody potential .
the eam potential is a semi - empirical potential derived from first - principles calculations .
eam potentials of metal alloys such as ni - cu , ni - al , and cu - al have been parameterized from dft calculations and validated to reproduce their experimental properties such as bulk modulus , elastic constants , lattice constants , etc . @xcite .
eam potentials are computationally cheaper , allowing us to explore the effects of relaxation for large training sets ; however , we are limited by the number of eam potentials available . therefore , we also selected two classical potentials , lennard - jones ( lj ) and stillinger - weber ( sw ) , to adequately examine various degrees of relaxation , which can be varied using free parameters in each model .
the lennard - jones potential is a pairwise potential . using the lj potential , we can model a binary ( @xmath1 ) alloy with different lattice mismatch and interaction strength between the a and b atoms by adjusting the @xmath2 parameter in the model .
additionally , we also examined the stillinger - weber potential which has a pair term and an angular ( three - body ) term . in attempting to determine the conditions under which the ce formalism breaks down , we implemented a set of parameters in the sw potential where
the angular dependent term could be turned on / off using the @xmath3 coefficient @xcite .
for example , depending on the strength of @xmath3 , the local atomic environment in 2d could switch between 3- , 4- and 6-fold coordination by changing a single parameter .
thus , when the system relaxes to a different coordination , the ce fits would no longer be valid or at least not sparse .
all first - principles calculations were performed using the vienna ab initio simulation package ( vasp ) @xcite .
we used the projector - augmented - wave ( paw ) @xcite potential and the exchange - correlation functional proposed by perdew , burke , and ernzerhof ( pbe ) @xcite . in all calculations
, we used the default settings implied by the high - precision option of the code .
equivalent @xmath0-point meshes were used for brillioun zone integration to reduce numerical errors @xcite .
we used 1728 ( @xmath4 ) @xmath0-points for the pure element structures and an equivalent mesh for the binary alloy configurations .
each structure was allowed to fully relax ( atomic , cell shape and cell volume ) .
relaxation was carried out using molecular dynamics simulations for eam , lj and sw potentials .
two molecular dynamics packages were used to study the relaxation : gulp @xcite and lammps @xcite .
details for the lj , sw and eam potentials and the dft calculations can be found in the supplementary materials .
the universal cluster expansion ( uncle ) software @xcite was used to generate 1000 derivative superstructures each of face - centered cubic ( fcc ) , body - centered cubic ( bcc ) and hexagonal closed - packed ( hcp ) lattice . for the dft calculations
, we used only 500 structures instead of 1000 due to the computational cost .
we generated a set of 1100 clusters , ranging from 2-body up to 6-body interactions .
100 independent ce fits were performed for each system ( hamiltonian and lattice ) .
we performed cluster expansions using the uncle software @xcite .
we briefly discuss some important details about cluster expansion here , but for a more complete description , see the supplementary materials and past works @xcite .
cluster expansion is a generalized ising model with many - body interactions .
the cluster expansion formalism allows one to map a physical property , such as e , to configuration ( @xmath5 ) : @xmath6 where e is energy , @xmath7 is the correlation matrix ( basis ) , and @xmath8 is coefficient or effective cluster interaction ( eci ) . when constructing a ce model , we are solving for the effective cluster interactions , or @xmath8s .
we used the compressive sensing ( cs ) framework to solve for these coefficients @xcite . the key assumption in compressive
sensing is that the solution vector has few nonzero components , i.e. , the solution is sparse @xcite .
the cs framework guarantees that the sparse solution can be recovered from a limited number of dft energies . using the @xmath8s
, we can build a ce model to interpolate the configuration space .
each ce fit used a random selection of 25% of the data for training and 75% for validation .
results were averaged over the 100 ce fits with error bars computed from the standard deviation .
we defined the percent error as a ratio of the prediction root mean squared error ( rms ) over the standard deviation of the input energies , @xmath9 this definition of percent error allowed us to consistently compare different systems . currently , there is no standard measure to indicate the degree of relaxation .
we evaluated different metrics as a measure of the relaxation : normalized mean- squared displacement , ackland s order parameter @xcite , difference in steinhardt order parameter ( @xmath10 ) @xcite , soap @xcite , and the centro - symmetry parameter @xcite .
we compared the metrics across various hamiltonians to find a criterion that is independent of the potentials and systems .
we found that none of these metrics are descriptive / general enough except for the normalized mean- squared displacement . to measure the relaxation of each structure / configuration
, we used the mean- squared displacement ( msd ) to measure the displacement of an atom from its reference position , i.e. , the unrelaxed atomic position .
the msd metric is implemented in the lammps software @xcite , which also incorporates the periodic boundary conditions to properly account for displacement across a boundary .
the msd is the total squared displacement averaged over all atoms in the crystal : @xmath11 - x[0])^2\ ] ] where @xmath12 is the final relaxed configuration and 0 is the initial unrelaxed configuration .
additionally , we defined a normalized mean - square displacement ( nmsd ) percent : @xmath13 which is the ratio of msd to volume of the system .
this allows for a relaxation comparison parameter that is independent of the overall scale . to explore the effects of relaxation on ce predictability ,
we examine relaxation in various systems from very high accuracy ( dft ) to very simple , tunable systems ( lj and sw potentials ) .
we examine more than one hundred different hamiltonians and we find several common trends among the different systems . in most cases , we find that the relaxed ce fits are worse ( higher prediction error and higher number of coefficients ) than the unrelaxed .
for example , fig .
[ fig : unrel - rel - vasp - eam ] shows the cluster expansion fitting for unrelaxed and relaxed data sets of ni - cu alloy system using dft and eam with two different primitive lattices .
though it seems strange for us to model ni- cu using a bcc primitive lattice when ni - cu is closed - packed , this is a method for us to evaluate the relaxation of ni - cu for a highly relaxed system . as fig .
[ fig : unrel - rel - vasp - eam ] shows , ni - cu alloy fitting for a fcc lattice is below 10% error , while bcc fitting result in more @xmath8s and higher percent error ( above 10% ) .
we find similar results in the relaxation of ni - cu alloy using first - principles dft and eam potential .
the difference between relaxed and unrelaxed ce fits are negligible when relaxations are small .
this is shown in fig .
[ fig : unrel - rel - vasp - eam ] for the relaxation of fcc superstructures using a ni - cu eam potential .
( color online ) cluster expansion fits for ni - cu alloy using dft or eam potential .
each bar represents the average percent error and error bar ( standard deviations ) for 100 independent ce fits .
the blue bars represent the unrelaxed ce fits , while the red bars represent the relaxed ce fits .
the colored number represents the average number of coefficients used in the ce models .
when the configurations are relaxed , we find that the ce fits are often worse ( higher prediction error and higher number of @xmath8s ) than unrelaxed system .
however , we show that in one case ( ni - cu eam ) the unrelaxed and relaxed ce fits are identical ( same error and same number of coefficients ) and this is due to a small relaxation.,scaledwidth=50.0% ] 0.55 [ fig : vasp - fcc - prob ] 0.55 0.180 0.178 0.55 [ fig : vasp - bcc - prob ] 0.55 0.178 0.192 fig . [
fig : unrel - rel - vasp - eam ] shows that increased relaxation is associated with reduced sparsity ( increased cardinality of @xmath8s ) .
one possible implication is that number of coefficients ( @xmath8 ) could be used to evaluate the predictive performance of the ce fits . the number of coefficients used in the fits ( such as in fig .
[ fig : unrel - rel - vasp - eam ] ) is a simple way to determine whether or not a ce fit can be trusted .
[ fig : vasp - fcc - ce ] and [ fig : vasp - bcc - ce ] show similar clusters across the 100 independent ce fittings ; thus , vertical lines indicate the presence of the same cluster across all ce fits . when the fit is good , only a small subset of clusters is needed ( fig .
[ fig : vasp - fcc - ce ] ) . on the other hand ,
[ fig : vasp - bcc - ce ] shows some common clusters in all of the ce fits with several additional clusters .
[ fig : rms - js ] shows the correlation of the percent error with the number of terms in the expansion .
we find that as the number of coefficients increases the percent error increases .
however , this is not a sufficient metric as shown in fig . [
fig : rms - js ] where the number of coefficient varies a lot .
nonetheless , the number of coefficients may be used as a general , quick test .
0.5 displays the ce fitting error vs the number of coefficients , while plot [ fig : msd - js ] highlights the relationship between number of coefficients and relaxation .
the dashed line approximates what we consider as the maximum acceptable error for a ce model ( 10% ) . the dashed line in fig .
[ fig : msd - js ] marks the estimated threshold for acceptable relaxation level .
each symbol represents 100 independent ce fittings for each hamiltonian .
higher error correlates with a higher number of coefficients.,title="fig : " ] 0.5 displays the ce fitting error vs the number of coefficients , while plot [ fig : msd - js ] highlights the relationship between number of coefficients and relaxation .
the dashed line approximates what we consider as the maximum acceptable error for a ce model ( 10% ) . the dashed line in fig .
[ fig : msd - js ] marks the estimated threshold for acceptable relaxation level .
each symbol represents 100 independent ce fittings for each hamiltonian .
higher error correlates with a higher number of coefficients.,title="fig : " ] the degree of relaxation is crucial to define whether or not the ce model is accurate or not . however , there is no standard for _ when _ cluster expansion fails due to relaxation . thus far
, we have made some remarks about relaxation and ce fits .
but the question of how much relaxation is allowed has not been addressed . by examining a few metrics : nmsd , soap @xcite , d6 @xcite , ackland @xcite and centro - symmetry @xcite
, we find that there is a relationship between degree of relaxation and the quality of ce fits .
as shown in the supplementary information , we have used these metrics to investigate over 100 + systems ( different potentials , lattice mismatches , and interaction strengths ) . here , we present a heuristic to measure the degree of relaxation based on the nmsd
. in general , cluster expansion will fail when the relaxation is large .
figure [ fig : msd - js ] shows that a small nmsd weakly correlates with a small number of coefficients .
however , figure [ fig : rms - msd ] highlights the correlation between degree of relaxation and prediction error .
there is a roughly linear relationship between the degree of relaxation and the ce prediction .
we partition the quality of the ce models into three regions : good ( nmsd @xmath14 0.1% ) , maybe ( 0.1% @xmath15 nmsd @xmath15 1% ) and bad ( nmsd @xmath16 1% ) .
the `` maybe '' region is the gray area where the ce fit can be good or bad .
this metric provide a heuristic to evaluate the reliability of the ce models , i.e. , any systems that exhibit high relaxation will fail to provide an accurate ce model .
as we have shown in the previous section , greater relaxation results in worse ce fitting .
in addition to the effects of relaxation , we now investigate the effects of numerical error on reliability of ce models .
numerical error arises from various sources such as the number of @xmath0-points , the smearing method , minimum force tolerance , basis set sizes and types , etc .
these errors are not stochastic error or measurement errors ; they arise from tuning the numerical methods .
we assume that the relaxation - induced change in energy for each structure is an _ error term _ that the ce fitting algorithm must handle .
the collection of these `` errors '' from all structures in the alloy system then form an error profile ( or distribution ) . using the simulated relaxation error profiles from the previous section together with common analytic distributions , we built `` toy '' ce models with known coefficients .
we then examined whether or not the shape of the error distribution affects the ce predictive ability .
the numerical errors in dft calculations are largely understood , but it is difficult to disentangle the effects of different , individual error sources . instead of studying the effects of errors separately , we added different distributions of error to a `` toy '' model in order to imitate the aggregate effects of the numerical error on ce models .
hence , we opt to simplify the problem by creating a `` toy '' problem for which the exact answer is known . to restrict the number of independent variables
, we formulated a `` toy '' cluster expansion model by selecting five non - zero values for a subset of the total clusters . using this toy ce ,
we predicted a set of energies @xmath17 for 2000 known derivative superstructures of an fcc lattice , these @xmath17 values are used as the true energies for all subsequent analysis .
we added error to @xmath17 , chosen from either : 1 ) `` simulated '' distributions obtained by computing the difference between relaxed and unrelaxed energies predicted by either dft , eam , lj or sw models ( fig .
[ fig : experimentaldistrbutions ] ) ; or 2 ) common analytic distributions ( fig .
[ fig : analyticequalwidthdistributions ] ) . to generate the simulated distributions , we chose a set of identical structures and fitted them using a variety of classical and semi - classical potentials , and quantum mechanical calculations using vasp .
for each of the potentials we selected , we calculated an unrelaxed total energy @xmath17 for each structure and then performed relaxation to determine the lowest energy state , @xmath18 .
the difference between these two energies ( @xmath19 ) was considered to be the `` relaxation '' error .
( color online ) distributions from real relaxations using classical and semi - classical potentials , as well as dft calculations .
the distributions are all normalized to fall within 0 and 1 .
the widths , @xmath20 , were calculated by taking the difference between the 25@xmath21 and 75@xmath21 percentiles.,scaledwidth=50.0% ] ( color online ) the analytic , equal width distributions used for adding error to the toy model ce fit.,scaledwidth=50.0% ] certain assumptions are usually made about the error in the signal , namely that it is gaussian .
the original cs paradigm proves that the @xmath22 error for signal recovery obeys @xcite : @xmath23 where @xmath24 bounds the amount of error in the data , @xmath25 is the cs solution , @xmath26 is the true solution , and @xmath27 is the vector @xmath26 with all but the largest @xmath28 components set to zero .
this shows that , _ at worst , the error in the recovery is bounded by a term proportional to the error_. for our plots of this error , we first normalized @xmath29 so that @xmath30 $ ] using @xmath31 not surprisingly , the various potentials produced different error profiles . the expectation value of the distributions was set to be a percentage of the average , unrelaxed energy across all structures
. thus , `` 15% error '' means that each unrelaxed energy was changed by adding a randomly drawn value from a distribution with an expectation value of 15% of the mean energy .
we performed ce fits as a function of the % -error added ( 2,5,10 and 15% ) for each distribution .
although we only present the 15% error results in the next section , all results at different error levels can be found in the supporting information . for each data point
, we performed 100 independent ce fits and used the mean and standard deviation to produce the values and error bars for the plots .
as shown in fig .
[ fig : equalwidthsummary ] , the error is weakly uniform across all ( analytic and simulated ) distributions , implying that there is no correlation between specific distribution and error .
none of the normal quantifying descriptions of distribution shape ( e.g. width , skewness , kurtosis , standard deviation , etc . ) show a correlation with the ce prediction error .
the error increased proportionally with the level of error in each system ( 2 , 5 , 10 and 15% error ) .
we therefore turn to the compressive sensing ( cs ) formalism for insight .
( color online ) comparison of the predictive error in ce fits as the shape of relaxation error changes .
( a ) refers to the analytic distribution while ( s ) refers to simulated distribution .
the fits are ordered from lowest to highest distribution width .
fits were averaged over 100 randomly selected subsets with 500/2000 data points used for training ; the remaining 1500 were used to verify the model s predictions .
the black and red colored symbols represent 2% and 15% error levels , respectively .
the circles and triangles represent the analytical and simulated distributions , respectively .
higher error produces higher prediction errors .
, scaledwidth=50.0% ] the theorems of tao and cands @xcite guarantee that the solution for an underdetermined cs problem can be recovered _ exactly _ with overwhelming probability provided : * the solution is sparse within the chosen representation basis .
* sufficient data points , sampled independent and identically distributed ( i.i.d ) .
* the sensing and representation bases are maximally incoherent .
if all of these conditions are met , we know that cs will provide a solution that is very close to the true answer .
conversely , if cs can not converge to a good solution , it means that one of these conditions has been violated .
we have control over the number of training points , and the incoherence of the sensing- representation bases .
however , we _ can not _ control whether the true physical solution is sparsely represented for relaxed systems .
this suggests a useful connection between the cs framework and the robustness of ce : if cs can not reproduce a good ce fit ( quantified below ) , then sparsity has been lost . in the cs framework
, the foundational assumption is that of sparsity , meaning that the compressed signal ( or cluster expansion ) requires only a few terms to accurately represent the true signal ( physics ) .
thus , the number of terms recovered by cs to produce the ce is a good measure of the quality of the cs fit .
this begs the question : can we use the number of terms within the cs framework to heuristically predict _ in advance _ whether the ce fit will converge well ? in answering the question of predictability for a good ce fit , we define three new quantities : 1 . [
item : xidefinition ]
@xmath32 : total number of unique clusters used over 100 ce fits of the same dataset .
we also call this the model complexity .
[ item : notindefinition ] @xmath33 : number of `` exceptional '' clusters .
these are clusters that show up fewer than 25 times across 100 fits , implying that they are not responsible for representing any real physics in the signal , but are rather included because the ce basis is no longer a sparse representation for the relaxed alloy system .
they are sensitive to the training / fitting structures .
3 . [ item : lambdadefinition ] @xmath34 : number of _ significant _ clusters in the fit ; essentially just the total number of unique clusters minus the number of `` exceptional '' clusters , @xmath35 . in the relaxation section
, we showed that the average number of coefficient is not sufficient to determine the quality of the ce model . here , we decompose the number of @xmath8s into three new quantities to provide additional insights into the reliability of the ce fits . in fig .
[ fig : errorvsmodelcomplexity ] , we plot the ce error , ordered by model complexity and show that it reproduces the trend identified by the number of coefficients ( indeed they are intimately related , @xmath32 being the statistically averaged number of coefficients across many fits ) .
an ordering by the number of exceptional clusters @xmath33 produces an identical trend , showing that it may also serve to quantify a good fit .
( color online ) prediction error over 65% of the structures for the `` toy '' cluster expansion ( at 15% error added ) .
the systems are ordered by @xmath32 , which is the total number of unique clusters used by any of the 100 ce fits for the system .
this ordering shows a definite trend with increasing @xmath32.,scaledwidth=50.0% ] as indicated earlier , all these experiments were performed for a _ known _ ce model that had 5 non - zero terms .
additional insight is gained by plotting the errors , ordered by @xmath34 , the number of significant clusters ( fig .
[ fig : errorvssignificantterms ] ) .
[ fig : errorvssignificantterms ] shows that in almost all cases , once we remove the exceptional clusters @xmath36 , the remaining model is almost _ exactly _ the known ce model that we started with .
the cs framework provides a rigorous mathematical framework for this statement because it guarantees to _ exactly _ recover the original function with high probability as long as we have enough measurements and our representation basis is truly sparse .
once the cluster expansion stops converging , we lose sparsity and cs fails .
this gives us confidence to use the cs framework as a predictive tool for ce robustness .
( color online ) prediction error over 65% of the structures for the `` toy '' ce model ( at 15% error added ) .
the errors are ordered by @xmath34 , the number of significant terms in the expansion . as expected , the values are close to the known model complexity ( 5 terms ) and the ordering once more appears random.,scaledwidth=50.0% ] provided the training structures are independent and identically distributed , we do not necessarily need hundreds of costly dft calculations to tell us that the ce will not converge . using our toy ce model , we discovered that for all error distributions , a training set size of 50 data points was sufficient to recover the actual model complexity ( 5 terms ) .
for actual dft calculations , where relaxation was known to disrupt ce convergence , we saw a similar trend with about 100 data points needed to identify whether the ce would converge with more data or not .
we conclude that ce robustness for relaxed systems can be predicted with a _ much _ smaller number of data points than is typically needed for a good ce fit ( on the order of 5 - 10% from our experience ) .
the proposed heuristic to verify convergence of the relaxed ce , when trained with a limited dataset , is to examine the values of @xmath34 and @xmath32 over a large number of independent fits . if the number of the exceptional clusters @xmath33 is significant compared to @xmath34 , then it is likely that the ce will _ not _ converge on a larger dataset as shown in fig .
[ fig : errorvssingleshow ] .
figure [ fig : norm - training ] highlights the ce fitting as function of training set size .
we observe small relaxation ( black curve ) correlates with a small number of coefficients ; thus the ce can fit using a small number of @xmath8s even with 5% ( 25 ) to 10% ( 50 ) of the structures . on the other hand , red and blue curves which have high relaxation ,
do not converge . by using a small relaxed dataset ( 50 to 100 structures ) , we can predict whether or not the computational cost of relaxing _ many _ structures is fruitful .
( color online ) plot of predictive error over 65% of the structures for the `` toy '' problem ( at 15% error added ) .
the systems are ordered by @xmath33 the number of clusters that were used less than 25 times across all 100 ce fits .
these are considered exceptions to the overall fit for the system . as for figure
[ fig : errorvsmodelcomplexity ] , there is a definite trend toward higher error for systems with more exceptional clusters.,scaledwidth=50.0% ] ( color online ) for a reliable ce model , the number of coefficients converges as a function of the training set size .
a total of 500 structure were available for training .
the number of coefficients in a fit and its error bars give us an indication of the predictive power of ce with only a small training set .
the black curve represents a good ce fit ; only 25 to 50 ( or 5 to 10% ) of training structures were needed . on the other hand ,
the red and blue curves show that ce fails to fit the data due to a slowly converging expansion .
the error bars on the blue points indicate extremely bad fitting.,scaledwidth=50.0% ]
relaxation and error decrease the reliability of the cluster expansion fit because the ce model is no longer sparse .
nevertheless , until now , there has been no measure of relaxation that provides a heuristic as to when the ce fitting data is reliable . using four different hamiltonians ( first - principles , lennard - jones , stillinger - weber and embedded atom method ) ,
we show that the normalized mean - squared displacement of alloy configuration is a good measure of relaxation and ce predictability .
a small displacement percent , e.g. , less than 0.1% , will usually generate a reliable ce model .
the number of cluster terms in the ce models is also a good indicator of how well cluster expansions can perform
. as the number of clusters increases , the predictability of ce model decreases .
ce tends to fail when the number of @xmath8s exceeds 80 . in our error analysis
, we investigated the ability of the compressive sensing framework to obtain fits to a toy , cluster expansion model as the energy of relaxation changes in a predictable way .
we used 16 relaxation error distributions ( both analytic and simulated ) and compared the prediction errors of the resulting ce fits for the relaxed vs. unrelaxed case .
no clear correlation appears between the statistical measures of distribution shape and the predictive errors . however , there are clear correlations between the predictive error , the complexity of the resulting ce model , and the number of significant terms in that model
. we can not use the relaxation distributions alone to determine the viability of a ce fit in advance .
however , the analysis does reveal that the majority of the clusters used by the unrelaxed ce fit will also be present in the relaxed case ( albeit with adjusted @xmath8 values ) if the ce fit is viable .
this suggests that it may be possible to decide whether the computational cost of full ce is worthwhile by making predictions for a few relaxed systems ( 50 - 100 ) and determining whether the error remains small enough .
the authors thank volker blum , lance j. nelson and mark k. transtrum for useful discussions . this work was supported by funding from onr ( muri n00014 - 13 - 1 - 0635 ) .
we thank the fulton supercomputing lab at brigham young university for allocation of computing time .
10 j. m. sanchez , f. ducastelle , and d. gratias . .
, 128(1 - 2):334350 , 1984 .
j. m. sanchez . .
, 48(18):1401314015 , nov 1993 .
j. m. sanchez . .
, 81(22):224202 , jun 2010 .
j. w d connolly and a. r. williams . . ,
27(8):51695172 , apr 1983 .
l. g. ferreira , a a mbaye , and alex zunger . .
, 35(12):64756478 , apr 1987 .
l. g. ferreira , a. a. mbaye , and a. zunger . .
, 37(18):10547 , 1988 .
l. g. ferreira , s .- h .
wei , and a. zunger . .
, 40(5):3197 , 1989 .
z. w. lu , s .- h . h. wei , alex zunger , s. frota - pessoa , and l. g. ferreira . .
, 44(2):512544 , 1991 .
c wolverton and a zunger . .
, 8(1 - 2):107121 , 1997 . a. van de walle and g. ceder .
, 23(4):348359 , aug 2002 .
sergey v. barabash , volker blum , stefan mller , and alex zunger . . ,
74(3):035108 , jul 2006 .
m asato , h takahashi , t inagaki , n fujima , r tamura , and t hoshino . . ,
48(7):17111716 , 2007 .
lance j. nelson , vidvuds ozoli , c. shane reese , fei zhou , and gus l. w. hart . .
, 88(15):155105 , 2013 . p.h .
gargano , p.r .
alonso , and g.h .
rubiolo . .
, 9:239245 , 2015 .
manuel aldegunde , nicholas zabaras , and jesper kristensen . . ,
( july ) , 2016 .
ravi chinnappan , b.k .
panigrahi , and axel van de walle . .
, 54:125133 , 2016 .
g ceder , a van der ven , c marianetti , and d morgan . .
, 8:311321 , 2000 .
atsuto seko , koretaka yuge , fumiyasu oba , akihide kuwabara , and isao tanaka . .
, 73(18):15 , 2006 .
atsuto seko , atsushi togo , fumiyasu oba , and isao tanaka . .
, 100(4):25 , 2008 .
isao tanaka , atsushi togo , atsuto seko , fumiyasu oba , yukinori koyama , and akihide kuwabara . .
, 20(46):10335 , 2010 .
isao tanaka , atsuto seko , atsushi togo , yukinori koyama , and fumiyasu oba .
, 22(38):384207 , 2010 .
stefan mller . .
, 15(34):r1429r1500 , 2003 .
ralf drautz , r singer , and m fhnle . .
, 67(3):035418 , 2003 .
s. mller , m. sthr , and o. wieckhorst . .
, 82(3 spec .
iss.):415419 , 2006 .
laura m. herder , jason m. bray , and william f. schneider . .
, 640:104111 , 2015 .
ryohei tanaka , kazuhito takeuchi , and koretaka yuge . .
56(7):10771080 , 2015 . david b. laks , l. g. ferreira , sverre froyen , and alex zunger . .
, 46(19):1258712605 , nov 1992 .
bjoern arnold , alejandro diaz ortiz , gus l w hart , and helmut dosch . .
, 81(9):111 , 2010 .
alejandro daz - ortiz and helmut dosch . .
, 76(1):012202 , jul 2007 .
alejandro daz - ortiz , helmut dosch , and ralf drautz . .
, 19:406206 , 2007 .
s. m. foiles , m. i. basker , and m. s. daw . .
, 33(12 ) , 1986 .
frank h. stillinger and thomas a. weber . .
, 31(8):52625271 , 1985 .
g. kresse and j. hafner . .
, 47(1):558561 , 1993 . g. kresse and j. hafner . .
, 49(20):1425114269 , 1994 .
g. kresse and j. furthmller . .
, 6(1):1550 , 1996 .
g. kresse and j. furthmller . .
, 54(16):1116911186 , 1996 . p. e. blchl . .
, 50(24):1795317979 , dec 1994 .
john p. perdew , kieron burke , matthias ernzerhof , department of physics , and new orleans louisiana 70118 john quantum theory group tulane university . . ,
77(18):38653868 , 1996 .
sverre froyen . . , 39(5):31683172 ,
feb 1989 .
julian d. gale . .
, 93(4):629637 , 1997 .
julian d. gale and andrew l. rohl . .
, 29(5):291341 , may 2003 .
s. plimpton . .
, 117(june 1994):119 , 1995 . see supplemental materials for details about the different potentials and their parameters .
gus l. w. hart and rodney w. forcade . .
, 77(22):112 , 2008 .
gus l w hart and rodney w. forcade . .
, 80(1):18 , 2009 .
d lerch , o wieckhorst , g l w hart , r w forcade , and s mller . .
, 17(5):055003 , sep 2009 . see supplemental materials for more details on cluster expansion .
ferreira , s .- h .
wei , and alex zunger . .
, 5(1):3456 , mar 1991 .
alex zunger , l g wang , gus l w hart , and mahdi sanati . . ,
10(6):685706 , 2002 .
axel van de walle . . ,
65(11):15231532 , nov 2013 .
lance j. nelson , gus l. w. hart , fei zhou , and vidvuds ozoli . .
, 87(3):035125 , 2013 .
emmanuel j. candes and terence tao .
, 52(12):54065425 , dec 2006 . e.j .
candes and m.b .
, 25(2):2130 , mar 2008 .
g. j. ackland and a. p. jones . . , 73(5):054104 ,
feb 2006 .
paul j. steinhardt , david r. nelson , and marco ronchetti . .
, 28(2):784805 , jul 1983 .
albert p. bartk , risi kondor , and gbor csnyi . .
, 87(18):116 , 2013 .
cynthia l. kelchner , s. j. plimpton , and j. c. hamilton . . ,
58(17):1108511088 , nov 1998 .
see supplemental materials for details about the different relaxation metrics . in our experience
, a percent error above 10% often gives unreliable ce model .
see supplemental materials for additional plots at different levels of error .
the supplementary data has many plots from the hundreds of ce fits that we performed for this analysis . for typical binary ces
we typically need about 300 - 500 structures to get a good fit , which is then verified using an additional 200 + dft calculations .
cluster expansion is a generalized ising model with many - body interactions .
the ce model provides a fast , accurate way to compute physical properties which are function of the configuration .
consider a binary alloy @xmath37 , the alloy is treated as a lattice problem .
each site of a given lattice is assigned a occupation variable , @xmath38 ( i=1,2, ... ,n ) with @xmath38 = 1 or + 1 depending on the site @xmath39 being occupied by an a or a b atom . any atomic configuration on a given lattice
can then be specified using a vector of the occupation variable , @xmath40 $ ] .
a physical quantity such as energy can be expressed as a linear combination of basis function : @xmath41 where the argument to the function is a vector of occupation variable , @xmath42 .
the @xmath43 are the basis function or often referred to as the cluster functions .
each cluster function corresponds to a cluster of lattice sites .
the coefficients @xmath44 are the effective cluster interactions or eci s .
the main task of building a ce model is to find the @xmath8s and their values .
we can solved for the @xmath8s using the structure inversion method @xcite .
however , we use a new approach based on compressive sensing to solve for these coefficients @xcite .
this is a more extensive version of the method present in the paper including the various parameters , forms of the potential and relaxation metrics .
two molecular dynamics packages were used to study the relaxation : gulp and lammps .
gulp ( general utility lattice program ) is written to perform a variety of tasks based on force field methods such molecular dynamics , monte carlo and etc @xcite .
gulp is a general purpose code for the modeling of solids , clusters , embedded defects , surfaces , interfaces and polymers @xcite .
lammps ( large - scale atomic molecular massively parallel simulator ) is a widely used molecular dynamics program @xcite .
we used these two programs to minimize / relaxed each structure .
we computed the energy of each structure ( unrelaxed and relaxed ) .
relaxation of each structure was obtained by minimization of total energy using gulp or lammps via a conjugate gradient scheme .
molecular dynamics simulations were carried for the embedded atom method ( eam ) , lennard - jones ( lj ) and stillinger - weber ( sw ) potentials .
the eam potential is a semi - empirical potential derived from first - principles calculations .
the embedded atom method ( eam ) potential has following form : @xmath45 eam potentials of metal alloys such ni - cu , ni - al , cu - al have been parameterized from first - principle calculations and validated to reproduce experimental properties , bulk modulus , elastic constants , lattice constants , etc @xcite .
compared to first - principles calculations , eam potentials are computationally cheaper .
thus , this allows us to explore the effect of relaxation for large training sets . nonetheless , we are limited by the number of eam potentials available .
we used various eam potentials to study the relaxation ; these binary eam potentials are shown in table [ eam - parameter ] . .eam
potentials used to study the relaxation .
lattice mismatch is shown in percentage .
lattice mismatch = ( @xmath46)/ ( ( @xmath47)/2 ) @xmath48 , where @xmath49 is the lattice constant of the pure element . [ cols="^,^",options="header " , ] in order to distinguish and measure the relaxation of the atoms from their ideal positions , we examined several metrics ( order parameters ) to quantify the relaxation : normalized mean - squared displacement or nmsd ( see the method in the main article ) , ackland s order parameter @xcite , @xmath10 , soap , and centro - symmetry .
we found that some of these order parameters are not descriptive / general enough for all cases ( potentials and crystal lattices ) .
we used the ackland s order parameter to identify the crystal structure after relaxation .
ackland s op identify each atomic local environment and assign it as fcc , bcc , hcp and unknown @xcite .
we used this op to determine which structures remain the same or on lattice and which structures undergo a structural change .
we can use this order parameter to separate / sort those structure that remain the same to examine the robustness of ce due the relaxation of crystal structure . similar to the ackland s order parameter , the centro - symmetry identifies the crystal structure of each atom based on the local arrangement ( neighbors ) . for example , figure [ fig : relaxed - sw - ack - fcc ] shows the mapping of msd and ackland order parameter for each structure at 5% ( top plot ) and 15% ( bottom plot ) .
overall , we can see that the msd increases with higher lattice mismatch . the spread of the ackland s order parameter is also affected .
going from 5% ( system a in table [ sw - parameter ] ) to 15% ( system c in table [ sw - parameter ] ) , the ce fitting error increases from 41.2% ( 110 clusters ) to 63.5% ( 156 clusters ) for the bcc .
when the lattice mismatch increases , the mean - squared displacement also increases .
ackland s order parameter and centro - symmetry are useful since they provide information about individual atoms . however , ackland s order parameter and centro - symmetry is too specific and it does not provide a useful measure of relaxation .
in addition to using the crystallographic information as a measure of relaxation , we used a variant of the steinhardt s bond order parameter @xcite that we called the @xmath10 order parameter or the @xmath10 metric , which is a measure of the difference between the local atomic environment ( relaxed and unrelaxed ) .
we computed the local atomic environment using the @xmath50 ( spherical harmonic with @xmath51 ) for the unrelaxed and relaxed configuration .
we averaged the difference of the two configurations , @xmath52 .
figure [ fig : d6-op ] shows the @xmath10 metric as a measure of relaxation vs the mean - squared displacement ( msd ) .
we observe that @xmath10 metric does not correlate withe msd . as relaxation increases ( higher msd ) ,
we expect that the @xmath10 value also increase .
however , this metric is not robust for all systems , that is , we can not compare the relaxation across all hamiltonians ( potentials and crystal lattices ) .
similar to the @xmath10 metrics , we used another metrics known as the soap ( smooth overlap of atomic position ) similarity kernel .
the soap similarity kernel measures the difference in configuration ( 1 when it is identical and decreasing as the difference increases ) .
the soap kernel is invariant to rotation and translation @xcite ; however , this metric is not applicable for multiple species cases .
[ fig : soap - error ] shows the prediction error vs soap .
similar to @xmath10 , the soap value does not correlate with the prediction error or displacement , that is , these metrics are too broad and vary too much for small displacements .
this problem lies in the normalization of soap and @xmath10 values . as a measure of the relaxation using a lennard- jones potential .
we show that the @xmath10 metric does not correlated with the displacement .
we show the relaxation of three crystal lattice .
lj favor fcc/ hcp ; thus , we should not observe high relaxation ( this is indicated by the displacement which is less than 0.001 @xmath53 .
however , the @xmath10 metrics show a very broad range from 0.0 ( identical configuration ) up to 0.1 .
although we only show this plot for lj , the results of sw and eam potential reveal the same conclusion , that is , @xmath10 is not a sufficient metric to analyze the various crystal lattices and potentials.,scaledwidth=50.0% ] none of the normal quantifying descriptions of distribution shape ( e.g. , width , skewness , kurtosis , standard deviation , etc . )
show a correlation with the ce prediction error .
the error increased proportionally with the level of error in each system ( 2 , 5 , 10 and 15% error ) .
0.47 ( color online ) width , skewness and kurtosis using the relaxation energies .
the relaxation energies are obtained by taking the absolute difference of unrelaxed and relaxed energies .
we show only the 2% and 15% error instead of all four error levels .
this allows us to illustrate the effect of error level on the width , skewness and kurtosis of the distribution.,title="fig : " ] 0.47 ( color online ) width , skewness and kurtosis using the relaxation energies .
the relaxation energies are obtained by taking the absolute difference of unrelaxed and relaxed energies .
we show only the 2% and 15% error instead of all four error levels .
this allows us to illustrate the effect of error level on the width , skewness and kurtosis of the distribution.,title="fig : " ] 0.47 ( color online ) width , skewness and kurtosis using the relaxation energies .
the relaxation energies are obtained by taking the absolute difference of unrelaxed and relaxed energies .
we show only the 2% and 15% error instead of all four error levels .
this allows us to illustrate the effect of error level on the width , skewness and kurtosis of the distribution.,title="fig : " ]
one other possibility is that the presence of outliers has a large impact on the performance of the bcs fit . to rule out that possibility , we performed fits with 0 , 1 , 2 , 10 , 20 , 30 , 40 , 50 and 60 outliers added to the error ( representing between 0 and 3% of the total data ) .
outliers were selected randomly from between 2 and 4 standard deviations from the mean and then appended to the regular list of errors drawn from the distribution ( the total number of values equaling 2000 again to match the number of structures ) . the summary is plotted in figure [ fig : outliersummary ] .
the difference between fits as the number of outliers changes is comparable to the variance in the individual fits .
we conclude then that outliers have no direct effect on the error profile s performance .
+ ( color online ) fitting errors as outliers are added to the error profile .
the difference between fits as the number of outliers changes is comparable to the variance in the individual fits.,scaledwidth=50.0% ] to further elucidate the claims relative to the ce framework s failures , we investigated whether we could measure the change in configuration upon relaxation . since the @xmath8 values selected in the model are backed by geometric clusters , the presence or absence of certain clusters has some correlation to the configuration of the physical system . when the physics is mostly dependent on configuration , the function can be sparsely represented by the ce basis .
thus , we expect that the sparsity will be a good heuristic in determining when the ce breaks down .
when the expansion terms do not decay well in the representation , it shows a misapplication of the ce basis to a problem that is not mostly configurational . 1 .
[ item : xidefinition ] @xmath32 : total number of unique clusters used over 100 ce fits of the same dataset .
we also call this the model complexity as shown in fig .
[ fig : errorvsmodelcomplexity ] .
[ item : notindefinition ] @xmath33 : number of `` exceptional '' clusters .
these are clusters that show up fewer than 25 times across 100 fits , implying that they are not responsible for representing any real physics in the signal , but are rather included because the ce basis is no longer a sparse representation for the relaxed alloy system ( shown in fig .
[ fig : errorvssingleshow ] .
it is sensitive to the training subsets .
[ item : lambdadefinition ] @xmath34 : number of _ significant _ clusters in the fit ; essentially just the total number of unique clusters minus the number of `` exceptional '' clusters , @xmath35 ( see fig .
[ fig : errorvssignificantterms ] ) . | cluster expansion ( ce ) is effective in modeling the stability of metallic alloys , but sometimes cluster expansions fail .
failures are often attributed to atomic relaxation in the dft - calculated data , but there is no metric for quantifying the degree of relaxation .
additionally , numerical errors can also be responsible for slow ce convergence .
we studied over one hundred different hamiltonians and identified a heuristic , based on a normalized mean - squared displacement of atomic positions in a crystal , to determine if the effects of relaxation in ce data are too severe to build a reliable ce model . using this heuristic ,
ce practitioners can determine a priori whether or not an alloy system can be reliably expanded in the cluster basis .
we also examined the error distributions of the fitting data .
we find no clear relationship between the type of error distribution and ce prediction ability , but there are clear correlations between ce formalism reliability , model complexity , and the number of significant terms in the model .
our results show that the _ size _ of the errors is much more important than their distribution . |
since its discovery by huygens when observing coupled pendula in the 17th century @xcite , synchronization has been described in physics , chemistry , biology and even social behavior @xcite , becoming a paradigm for research of collective dynamics .
it has been thoroughly studied in classical nonlinear dynamical systems @xcite , and extended to chaotic ones @xcite . only recently synchronization has been studied in quantum systems , e.g. ,
two coupled quantum harmonic oscillators @xcite , a qubit coupled to a quantum dissipative driven oscillator @xcite , two dissipative spins @xcite , and two coupled cavities @xcite .
last year , connections between quantum entanglement and synchronization have been discussed in continuous variable systems @xcite .
a decade ago , hampton and zanette introduced a new concept termed measure synchronization ( ms ) for coupled hamiltonian systems @xcite .
they found that two coupled hamiltonian systems experience a synchronization transition from a state in which the two subsystems visit different phase space regions to a state in which `` their orbits cover the same region of the phase space with identical invariant measures '' @xcite .
the control parameter is the coupling strength between the two subsystems .
the key difference between ms and conventional synchronization is that ms takes place in absence of dissipation . in
standard synchronization dissipation plays a key role , as it is responsible for the collapse of any trajectory of the system in phase space . for
coupled hamiltonian systems , phase space volume must be conserved following liouville s theorem , thus preventing the collapse of any trajectory in phase space . in the case of ms ,
two coupled hamiltonian systems become synchronized when they cover the same phase space domain , without requiring that the synchronized systems have the same evolution trajectories . in this article
we introduce measure synchronization , a concept up to now only considered in a classical framework , into the quantum many - body regime .
first exploratory studies of measure synchronization in quantum systems have been done in different contexts , i.e. , ms has been discussed in coupled hamiltonian systems associated with nonlinear schrdinger equations @xcite ; also , ms transitions have been revealed on meanfield theories describing condensed bosonic quantum many - body systems @xcite .
however , it is worth stressing that in the above cases the dynamical variables describing these quantum systems are classical , i.e. quantum fluctuations are neglected , a reasonable approximation in bosonic systems which are fully condensed @xcite
. this made the correspondence between the classical ms concept introduced in ref .
@xcite and the ms studies of these quantum systems straightforward .
so conceptually , measure synchronization discussed in these contexts remained classical .
here , we tackle the problem in a fully many - body quantum mechanical way .
a major conceptual difference is that , in the general case we need to identify quantum many - body observables which allow us to characterize ms - like behaviors provided the very definition of the area covered by each subsystem in phase space is absent .
we consider two quantum many - body systems ( qmbs ) which are coupled through a local interaction term .
our main finding is that we characterize a crossover behavior from non - ms to ms in the evolution of the quantum many - body properties of the subsystems .
this implies that two qmbs , which if non coupled would develop different quantum correlations , will , if sufficiently coupled , have similar condensed fractions , particle fluctuations , etc .
this is an effect which will affect the behavior of future qmbs and quantum simulators , and which , if properly controlled , can be employed to share or to induce quantum correlations between different degrees of freedom in the system .
ms is a dynamical feature which we will show to appear in the evolution of qmbs .
it describes how , under certain premises , the dynamics of two weakly coupled quantum subsystems becomes coherent after a short transient time .
ms describes how two subsets of a qmbs will evolve in a collective way , exchanging energy during the full evolution , exploring similar average values of relevant observables and developing similar quantum correlations .
it is worth emphasizing that our ability to understand and utterly control quantum correlations in qmbs is the key to producing powerful technological applications .
a notable recent example is the case of pseudo - spin squeezed states @xcite , which can be produced in bosonic josephson junctions @xcite . in this case , producing fragmented ultracold gases is shown to notably raise the precision achievable in quantum metrology experiments @xcite .
these applications will become a reality in the near future thanks to the miniaturization of ultracold atomic systems @xcite .
as shown here , ms can be used to transfer , or sympathetically produce , fragmentation in one subsytem of the qmbs which can , for instance , improve interferometric signals . the article is organized in the following way . in sec .
[ sec : teo ] we describe the many - body hamiltonian . in sec .
[ sec : res ] we present our results , concerning the onset of ms and how it shows in the many - body properties of the system . in sec .
[ sec : exp ] we sketch an experimental implementation with ultracold atomic gases .
a summary and conclusions are provided in sec .
[ sec : sum ] .
to illustrate the many - body quantum ms we consider the simplest implementation we can think of .
these are two different kinds of bosons , @xmath0 and @xmath1 , populating solely two quantum states , @xmath2 and @xmath3 .
we will consider a linear coupling between the two quantum states and contact interaction for @xmath4 , @xmath5 , and @xmath6 bosons .
the many - body hamiltonian considered for @xmath7 and @xmath8 atoms in two modes is , @xmath9 where @xmath10 -j_{a}(\hat{a}_{l}^{\dag } \hat{a}_{r}+\hat{a}_{r}^{\dag } \hat{a}_{l } ) \nonumber \\ \hat{h}_b&=&\frac{u_{b}}{2 } \left[(\hat{b}_{l}^{\dag}\hat{b}_{l})^{2}+(\hat{b}_{r}^{\dag}\hat{b}_{r})^{2}\right ] -j_{b}(\hat{b}_{l}^{\dag } \hat{b}_{r}+\hat{b}_{r}^{\dag } \hat{b}_{l } ) \nonumber\\ \hat{h}_{ab } & = & u_{ab } ( \hat{a}_{l}^{\dag}\hat{a}_{l}\hat{b}_{l}^{\dag } \hat{b}_{l } + \hat{a}_{r}^{\dag}\hat{a}_{r}\hat{b}_{r}^{\dag } \hat{b}_{r } ) \,.\end{aligned}\ ] ] @xmath11 @xmath12 and @xmath13 @xmath14 are creation ( annihilation ) operators for the single - particle modes @xmath2 or @xmath3 of the two species .
the terms proportional to @xmath15 are the linear coupling terms , which in absence of any interaction would induce periodic rabi oscillations of the populations between the states @xmath2 and @xmath3 .
@xmath16 , @xmath17 , and @xmath18 measure the @xmath4 , @xmath6 , and @xmath5 contact interactions .
the @xmath18 term is the only one coupling the dynamics of the @xmath0 and @xmath1 subsystems , and will be responsible for the ms between both of them .
the hamiltonian can be numerically diagonalized in the @xmath19 dimensional space spanned by the many - body fock basis tensor product of the @xmath0 and @xmath1 fock states , e.g. , for the @xmath0 , @xmath20 , with @xmath21 and @xmath22 .
the most general @xmath23-particle state can be written as @xmath24 the time evolution of any given initial state is governed by the time dependent schrdinger equation , @xmath25 .
once we have computed the many - body state , we can obtain average particle numbers on modes @xmath2 and @xmath3 , @xmath26 , @xmath27 , with @xmath28 .
the imbalance of population for each species is defined as @xmath29 . to characterize the degree of condensation of each subsystem , @xmath0 and @xmath1 , at
any given time we will make use of the one - body density matrix , @xmath30 @xcite . for a state @xmath31
it is defined as , e.g. , for species @xmath0 , @xmath32 , with @xmath33 , and @xmath34 .
the traces of @xmath35 and @xmath36 are normalized to the number of atoms in each subsystem , @xmath37 and @xmath38 .
the two normalized eigenvalues ( divided by the total number of atoms @xmath37 ) are @xmath39 , with @xmath40 .
we always have @xmath41 .
the larger eigenvalue is also called the condensed fraction .
similar definitions are used for species @xmath1 . , @xmath42 , and @xmath43 for species @xmath0 ( red ) and @xmath1 ( green ) .
( a ) @xmath44 ( non - ms ) , ( b ) @xmath45 ( non - ms ) , and ( c ) @xmath46 ( ms ) . in panel
( c ) the two clouds overlap .
@xmath47 , @xmath48 , and @xmath49 . in all panels
the points correspond to periodic moments within one time evolution , we show them as dots , instead of a continuous line , to avoid one of the species hiding the other one . ]
a way to characterize the transition from non - ms to ms dynamics in classical systems is by looking at the time average of the energies of subsystems @xmath0 and @xmath1 @xcite . in ms dynamics , both subsystems cover , with equal density , the same phase space domain , which reflects on equal long - time averages of the energies of the subsystems defined as @xmath50 where the expectation values of the energy for each subsystem @xmath0 ( and @xmath1 ) at time @xmath51 are @xmath52 , and @xmath53 , with @xmath54 the evolved quantum state .
we set both intraspecies interactions to be the same , i.e. , @xmath55 , with @xmath56 , and also choose equal linear couplings , @xmath57 .
we take as a unit of time the rabi time , @xmath58 , and as a unit of energy , @xmath59 .
our initial states will in all cases be coherent states for both the @xmath0 and @xmath1 species , in which all atoms populate the single particle state @xmath60 [ with initial population imbalance , @xmath61 .
these states will evolve under the action of the many - body hamiltonian .
we will look for a transition from non - ms to ms in the collective dynamics of the many - body state as we vary the interspecies interaction strength @xmath18 .
the transition from non - ms to ms dynamics is shown in fig .
we plot the average value of the pseudo - angular momentum operators which can readily be constructed from the creation and annihilation operators of each species @xcite , @xmath62 , @xmath63 , @xmath64 . in our conditions ,
fixed @xmath7 and @xmath8 , these operators build the symmetric representation of @xmath65 of dimension @xmath66 and @xmath67 .
as shown in the three - dimensional ( 3d ) figure , in the non - ms cases , @xmath44 and @xmath68 , the domains of ( @xmath69 , @xmath70 , @xmath71 ) explored by each subsystem are disjointed . in the ms case
, however , both domains completely overlap .
this feature can be regarded as the many - body counterpart of the classical definition of ms , in which the phase space domain covered by both subsystems is the same . in classical systems
ms implies that both subsystems have similar long - time averages of their energies .
importantly , the non - ms to ms transition in classical systems is discontinuous , which allows one to define a critical point to characterize the dynamical phase transition @xcite .
this is seen in fig .
[ fig2 ] , where we depict the average energy @xmath72 and @xmath73 as a function of the interspecies interaction @xmath18 , with @xmath74 characterizing ms .
the classical results @xcite , depicted in green and blue , feature the known discontinuity . in the many - body case
the situation is different ; the dynamical phase transition is replaced by a crossover behavior , therefore no critical point can be unambiguously defined .
there is no criticality which involves logarithmic singularity in the quantum measure synchronization as compared with classical theory of measure synchronization @xcite .
also note that in the many - body case , ms appears at higher values of @xmath18 as compared to the classical transition .
the inset in fig .
[ fig2 ] shows the behavior of @xmath75 and @xmath76 for the two different regions . in the ms case , the two subsystems exchange energy in such a way that their energies oscillate around the same average value maintaining an almost constant sum . in the non - ms dynamics
, the energies of the subsystems are never fully exchanged , and @xmath0 has always more average energy than @xmath1 . for different initial conditions , particle numbers , and parameters
@xmath77 we obtain a similar picture ( see appendix [ app ] ) , the main difference being the size of the ms and non - ms regions . ) , for the two species @xmath78 ( solid ) and @xmath79 ( dashed ) as a function of @xmath18 .
ms dynamics corresponds to equal averages , @xmath80 .
@xmath47 , @xmath48 , and @xmath49 . in the insets
we depict @xmath75 and @xmath76 for two specific values of @xmath81 ( a ) and @xmath82 ( b ) .
the two values are marked with arrows in the main figure . @xmath83 .
the green and blue lines correspond to the classical prediction of ref .
] now we concentrate on the evolution of the many - body properties of both subsystems . in fig .
[ fig3 ] we consider the same initial state and two values of @xmath18 , @xmath84 and @xmath85 , giving rise to non - ms and ms , respectively .
figure [ fig3 ] , panels ( a ) and ( b ) show the population imbalance between the two quantum states ( @xmath2 and @xmath3 ) for each subsystem .
the quantum many - body evolution becomes apparent , with characteristic collapse and revival dynamics @xcite .
this collapse and revival dynamics has been addressed for single component junctions @xcite and it can be understood in finite systems due to the finite number of frequencies entering in any dynamical evolution in the system .
we note that before reaching ms , the dynamics of the two subsystems is different both in the amplitude of the oscillation and on the times for collapses and revivals . after reaching ms , the times for collapses and revivals are the same .
this striking feature provides a way to characterize quantum measure synchronization in the rhythms of the coupled hamiltonian systems .
furthermore , we note that the oscillating amplitude for the two subsystems will be the same once ms is achieved , which is a feature that also shows in classical measure synchronization states @xcite .
( black ) and @xmath1 ( red ) as a function of time , for a non - ms dynamics , @xmath45 ( left panels ) and for ms dynamics , @xmath46 ( right panels ) .
the population imbalances [ ( a ) , ( b ) ] , condensed fractions [ ( c ) , ( d ) ] , and dispersion of the population imbalance @xmath86 [ ( e ) , ( f ) ] .
all magnitudes show a signature of the difference between the non - ms and ms dynamics .
@xmath47 , @xmath48 , and @xmath49 . ]
a crucial feature of quantum many - body bosonic systems is the appearance of correlations stronger than those present in bose - einstein condensed clouds .
an initially condensed system loses condensation during the evolution , and becomes fragmented @xcite [ see panels ( c ) and ( d )
. this fragmentation also takes place if there is no coupling between the subsystems @xcite .
interestingly , in the ms dynamics , the condensed fraction of both subsystems gets clearly correlated after a very short transient time , having the same envelope of the oscillation amplitudes , which is a key feature of ms .
this feature is found with all particle numbers studied @xmath87 ( see appendix [ app ] ) .
this similar behavior is also exhibited in the dispersions of particle differences [ fig .
[ fig3 ] , panels ( e ) and ( f ) ] .
this is of special significance , as this is directly related to the emergence of cat - like many - body states or pseudo - spin squeezed states in the evolution @xcite .
the latter provide a direct application of this physics to improve our precision measurements @xcite .
the aforementioned ms can be studied with state - of - the - art experimental techniques in ultracold atomic physics .
the almost perfect decoupling of ultracold atoms from their environment enables the investigation of the quantum measure synchronization in conservative systems .
we describe a feasible system which can simulate with good precision the many - body hamiltonian in eq .
( [ eq : h ] ) using trapped ultracold atomic gases @xcite .
we consider a two - species ultracold atomic cloud trapped in a symmetric double - well potential . in the weakly interacting regime , assuming the atom - atom interactions are correctly described by a contact interaction , and following similar steps as in ref .
@xcite , one obtains eq .
( [ eq : h ] ) .
the classical predictions of eq .
( [ eq : h ] ) have been studied in refs .
@xcite and some many - body features in ref .
the linear couplings @xmath88 are proportional to the energy splitting between the quasidegenerate ground state of the double - well potential .
the atom - atom interaction terms , are given by @xmath89 , @xmath90 , where @xmath91 refers to atoms @xmath0 or @xmath1 .
@xmath92 are localized single particle states ; the localized single particle states have been normalized as @xmath93 .
@xmath94 is @xmath95-wave scattering length between atoms @xmath91 , with mass @xmath96 .
@xmath97 is the interspecies @xmath95-wave scattering length .
the scattering lengths are varied routinely in ultracold atom experiments by means of feshbach resonances @xcite or confinement induced resonances @xcite .
a possible specific experimental implementation could be an external double - well potential as in ref .
@xcite or the double - well inside the quantum chip used in ref .
it has been shown that two mode becs can be prepared in coherent states experimentally @xcite .
it would be possible to characterize quantum ms by investigating the times of collapses and revivals for the two species .
there are also other experimental options for consideration , i.e. , microcavity exciton - polaritons system @xcite , or coupled micropillars system @xcite . even though ultracold atomic samples are well isolated from the environment , there is one source of decoherence which could affect the onset of ms to non - ms transitions .
this is the presence of losses in the system .
these have been studied in detail for single component josephson junctions , finding a constraint on the maximum attainable correlation which can be produced in the junction @xcite .
a study of the effect of losses on the ms to non - ms transition falls beyond the scope of the present article .
we introduced the concept of measure synchronization in quantum many - body systems . to exemplify the phenomenon we have considered a two - species bosonic josephson junction made of a small number of atoms which can be experimentally studied in a number of different setups .
importantly , the measure synchronization occurs at the many - body quantum level , showing how properties such as the condensed fraction or the fluctuations in particle number of the two species behave coordinately above a certain coupling strength between the two systems .
the findings reported apply to a variety of quantum many - body systems .
an important application which can be envisaged is to profit from the ms described here to build targeted quantum correlations of certain degrees of freedom in the system in a sympathetic way .
that is , when one can experimentally control and prepare the quantum correlations in one part of the system ( e.g. one of the species ) , this method can be used to build similar quantum correlations in other parts of the system which can not be experimentally controlled ( e.g. the other species ) . in this ms regime
, different parts of the system will develop similar quantum correlations and other quantum properties after a short transient time .
the authors thank j. martorell for useful comments on the manuscript .
this work was supported by the national natural science foundation of china ( grant no .
11104217 ) .
we also acknowledge partial financial support from the dgi ( spain ) ( grant no.fis2011-24154 ) and the generalitat de catalunya ( grant no .
2009sgr-1289 ) . b.j .-
d . is supported by the ramn y cajal program .
) , respectively .
( c ) and ( d ) are the full quantum and classical prediction for the ms case ( @xmath98 ) , respectively . [ figa1 ] ] in this appendix we analyze , by considering increasingly larger particle numbers , the relation between the classical and full quantum descriptions .
the main interest of the present article is to extend the concept of measure synchronization to systems which do not accept a full classical description .
thus , we have emphasized the effects on the magnitudes which have no direct classical analog , such as the condensed fractions and the fluctuations of the particle numbers , shown in figs .
[ fig3](c ) and [ fig3](d ) and figs .
[ fig3](e ) and [ fig3](f ) , respectively .
interestingly , particle number fluctuations can be measured experimentally and provide a good way of pinning down correlated states in these systems @xcite . to take the classical limit in a meaningful way
we will perform exact numerical simulations for @xmath87 , keeping both @xmath88 and @xmath99 constant , and compare to the corresponding classical predictions . as occurred in the case of a usual bosonic josephson junction ,
the most remarkable difference between the classical and quantal results is the presence of quantum collapses and revivals in the latter @xcite .
this can be seen already on the evolution of the average values of the particle number imbalances of species a and b. in fig .
[ figa1 ] we depict the comparison between the average particle imbalance of each species reported in fig . 3 ( obtained for @xmath100 ) and the corresponding classical prediction @xcite .
also , this is one of the signatures that shows that ms can be characterized by the rhythms observed in the dynamical evolution of the two coupled subsystems .
in contrast , classical measure synchronization is characterized by a spatial localization in the phase space associated to the conjugate variables describing the imbalance and the phase difference of each subsystem with no need of synchronization in the time evolution of the variables of each subsystem @xcite .
u for non - ms , and @xmath101 u for the ms case .
exact many - body results for @xmath102 and @xmath103 are given panels [ ( a ) , ( d ) ] and [ ( b ) , ( e ) ] , respectively .
panels [ ( c ) , ( f ) ] contain the classical predictions @xcite . @xmath47 and @xmath48 . ]
u for non - ms , and @xmath101 u for the ms case .
exact many - body results for @xmath102 and @xmath103 are given panels [ ( a ) , ( d ) ] and [ ( b ) , ( e ) ] , respectively . panels [ ( c ) , ( f ) ] contain the classical predictions @xcite .
@xmath47 and @xmath48 .
] as expected , increasing the number of particles , the classical predictions better describe the initial behavior of the quantum ones . in fig .
[ figa2 ] we present the average values of the energy of each subsystem as a function of time comparing the classical results to quantum ones at @xmath102 and 80 . in the ms case @xmath104 ,
panels ( a)(c ) , the classical result shows quasiperiodic oscillations for both @xmath75 and @xmath76 .
as expected for the ms case , both subsystems have the same mean energy , when averaged over long times .
this feature is also present in the quantum calculation , as shown in fig . 2 for @xmath49 , already for @xmath102 , fig . [ figa2 ] ( a ) .
in this case the oscillation is clearly damped for times around @xmath105 , departing from the classical results fairly early .
as the total number of particles is increased to @xmath103 , the time of the first collapse increases @xmath106 . in fig .
[ figa2a ] we depict the evolution of the average population imbalance in both situations , which clearly shows how the classical result improves the description of the initial time evolution as @xmath23 is increased . in the non - ms situation
the classical result departs earlier from the quantal predictions [ see fig . [
figa2](d)[figa2](f ) ] . in this case , the classical case clearly shows different long - time - averaged values of the energies for each subsystem , a feature of non - ms . in the quantum results this
is also observed , albeit in this case even for @xmath103 the classical and quantum results differ quantitatively already for times of the order of @xmath107 .
note the collapses and revivals inherent to the quantum description make it difficult to talk about long time averages of the signals . as discussed above ,
we find a synchronization of the fragmentation of the subsystems in the ms case as opposed to the non - ms situations . in the classical description
this is of course absent , as the subsystems are fully condensed during the evolution . in the quantum case even for small number of particles
@xmath102 we find this clear feature [ see fig . [ figa3](a ) and fig .
[ figa3](d ) ] . in the ms case both subsystems clearly fragment in a synchronized way [ figs .
[ figa3](a ) - [ figa3](c ) ] as opposed to the non - ms case [ figs .
[ figa3](d)[figa3](f ) ] . note also that ms produces more overall fragmentation in the system , as it is the less fragmented component a , the one that follows the more fragmented one b. the time scale in which the system fragments is found to be mostly independent of the number of particles , for the particle numbers considered . at @xmath108 the maximum degree of correlation
is already built in the system .
also the amount of fragmentation is found to be almost independent of the particle number considered , although as expected it decreases slowly with particle number .
s. b. papp and c. e. wieman , phys .
lett . * 97 * , 180404 ( 2006 ) ; s. b. papp , j. m. pino , and c. e. wieman , phys . rev . lett .
* 101 * , 040402 ( 2008 ) ; g. thalhammer , g. barontini , l. de sarlo , j. catani , f. minardi , and m. inguscio , phys . rev . lett . * 100 * , 210402 ( 2008 ) ; d. j. mccarron , h. w. cho , d. l. jenkin , m. p. kppinger , and s. l. cornish , phys .
a * 84 * , 011603(r ) ( 2011 ) . | the concept of measure synchronization between two coupled quantum many - body systems is presented . in general terms
we consider two quantum many - body systems whose dynamics gets coupled through the contact particle - particle interaction .
this coupling is shown to produce measure synchronization , a generalization of synchrony to a large class of systems which takes place in absence of dissipation .
we find that in quantum measure synchronization , the many - body quantum properties for the two subsystems , e.g. , condensed fractions and particle fluctuations , behave in a coordinated way . to illustrate the concept we consider a simple case of two species of bosons occupying two distinct quantum states . measure synchronization can be readily explored with state - of - the - art techniques in ultracold atomic gases and , if properly controlled , be employed to build targeted quantum correlations in a sympathetic way . |
the standard universe model successfully explains the observations of cmbr but there are still some unresolved issues .
inflationary cosmology is proved to be a cornerstone to resolve long - standing theoretical issues including horizon problem , flatness , magnetic monopole issue and origin of fluctuations .
scalar field as a primary ingredient of inflation provides the causal interpretation of the origin of lss distribution and observed anisotropy of cmb @xcite .
inflationary standard models are classified into slow - roll and reheating epochs . in slow - roll period ,
potential energy dominates kinetic energy and all interactions between scalar ( inflatons ) and other fields are neglected , hence the universe inflates @xcite .
subsequently , the universe enters into reheating period where the kinetic energy is comparable to potential energy .
thus the inflaton starts an oscillation about minimum of its potential losing their energy to other fields that present in the theory @xcite .
after this epoch , the universe is filled with radiation .
warm inflation @xcite ( opposite to cold inflation ) has an attractive feature of joining the end stage of inflation with the current universe .
it is distinguished from cold inflation in a way that thermal radiation production occurs during inflationary epoch and reheating period is avoided .
the dissipation effects become strong enough due to the production of thermal fluctuations of constant density which play a vital role in the formation of initial fluctuations necessary for lss formation .
the thermalized particles are produced continuously by microscopic processes which must occur at a timescale much faster than hubble scale @xmath0 , hence the decay rates of the particles must be larger than @xmath1 . during this regime ,
density fluctuation arises from thermal rather than quantum fluctuation @xcite .
warm inflationary era ends when the universe stops inflating , then the universe enters into the radiation dominated phase smoothly .
finally , the remaining inflatons or dominant radiation fields create matter components of the universe @xcite .
the feasibility of the warm inflation scenario from various view points is discussed in @xcite .
their results as a whole show that it is extremely difficult to realize the idea of warm inflation .
dissipative effects could lead to a friction term in the equation of motion for an inflaton field during the inflationary era .
the friction term may be linear as well as localized and is described by a dissipation coefficient .
bastero - gil et al .
@xcite made considerable explicit calculations using quantum field theory method that compute all the relevant decay and scattering rates in the warm inflationary models .
berera et al .
@xcite presented particular scenario of low - temperature regimes in the context of dissipation coefficient .
they considered the value of dissipation coefficient in supersymmetric ( susy ) models which have an inflaton together with multiplets of heavy and light fields .
dissipation leads to two important consequences : weak @xmath2 and strong regimes @xmath3 . during weak dissipative regime ,
the primordial density perturbation spectrum is determined by thermal fluctuations rather than vacuum fluctuations while restrictions on the gradient of inflaton potential may be relaxed in strong regime .
inflationary universe has many interesting exact solutions that can be found using an exponential potential often called a power - law inflation . here , the scale factor has a power - law type evolution @xmath4 , where @xmath5 @xcite .
another exact solution is obtained in the de sitter inflationary universe where a constant scalar potential is considered @xcite .
exact solutions of the inflationary cosmology can also be obtained for two particular scenarios , i.e. , intermediate and logamediate with specific growth of the scale factors @xcite .
this type of expansion is slower than de sitter inflation but faster than power - law inflation , so dubbed as intermediate " .
the string or m theory motivates intermediate inflation according to which higher order curvature invariant corrections to the einstein - hilbert action must be proportional to gauss - bonnet terms for the ghost free action .
these terms arise naturally as the leading order in the expansion of inverse string tension @xmath6 to low energy string effective action .
it has been found that 4-dimensional gauss - bonnet interaction with dynamical dilatonic scalar coupling leads to a solution , i.e. , intermediate form of the scale factor @xcite . on the other hand , logamediate inflation ( generalized model of the expanding universe )
is motivated by applying weak general conditions on the indefinite expanding cosmological models @xcite .
these models were originally developed as exact solutions of inflationary cosmology but were best formulated using slow - roll approximation . during slow - roll approximation , it is possible to find a spectral index @xmath7 .
in particular , intermediate inflation leads to @xmath7 for special value @xmath8 ( but this value is not supported by the current observational data @xcite ) that corresponds to the harrizon - zeldovich spectrum @xcite . in both models ,
an important observational quantity is the tensor to scalar ratio @xmath9 , which is significantly non - zero @xcite .
recently , the effects from bicep2 experiment of gravitational waves in the b - mode have been analyzed which predict @xmath10 @xmath11 c.l . ) and take out the value @xmath12 at a significance of @xmath13 @xcite .
therefore , the tensor modes should not be neglected .
setare and kamali @xcite studied warm inflation " with vector and non - abelian gauge fields during intermediate and logamediate scenarios using flat frw background with constant and variable dissipation coefficients .
motivated by this , we have discussed inflation with both types of field and proved that locally rotationally symmetric ( lrs ) bianchi i ( bi ) universe model is also compatible with wmap7 observations @xcite .
del campo and herrera @xcite explored dynamics of warm - chaplygin inflationary universe model and discussed cosmological perturbations in warm inflationary universe model with viscous pressure .
herrera et al .
@xcite studied intermediate generalized chaplygin gas ( cg ) inflationary universe model with standard as well as tachyon scalar fields and checked its compatibility with wmap7 data .
we have studied inflationary dynamics of generalized cosmic cg using standard and tachyon scalar fields ( with and with out viscous pressure ) during intermediate and logamediate scenarios @xcite .
setare and kamali @xcite investigated dynamics of warm inflation with viscous pressure in frw universe model and on the brane with constant as well as variable dissipation and bulk viscosity coefficients . we have extended this work to lrs bi universe model @xcite .
herrera et al .
@xcite analyzed the possible realization of an expanding intermediate and logamediate scale factors within the framework of a warm inflationary frw as well as loop quantum cosmology models .
they checked that how both types of inflation work with a generalized form of dissipative coefficient during weak and strong dissipative regimes .
the frw universe is just an approximation to the universe we see as it ignores all the structure and other observed anisotropies , e.g. , in the cmb temperature @xcite .
one of the great achievements of inflation is having a naturally embedded mechanism to account for these anisotropies .
although the new era of high precision cosmology of cmb radiation improves our knowledge to understand the infant as well as the present day universe . there arises a question about the main assumption of an exact isotropy of the cmb .
there are two pieces of observational evidence demonstrating that there is no exact isotropy .
the first is the existence of small anisotropy deviations from isotropy of the cmb radiation and second the presence of large angle anomalies that are shown as real features by the planck satellite results .
this helps to construct an alternative model to decode effects of the early universe on the present day lss without affecting the processes of nucleosythesis @xcite .
bianchi models can be alternatives to the standard frw models with small deviations from exact isotropy to explain anisotropies and anomalies in the cmb .
martinez - gonzalez and sanz @xcite proved that the small quadrupole component of cmb temperature found by cobe implies that if the universe is homogeneous but anisotropic bi then there must be a small departure from the flat friedmann model .
bianchi spacetimes are geometries with spatially homogeneous ( constant @xmath14 ) surfaces which are invariant under the action of a three dimensional lie group .
bianchi type i , the straight forward generalization of the flat frw metric with symmetry group described by @xmath15@xmath16 corresponds to flat hypersurfaces .
following this idea , we extend this work to lrs bi universe model which is a generalization of our previous paper @xcite in which a particular choice of dissipation coefficient @xmath17 is considered .
the paper is organized as follows .
section * 2 * provides basic formalism of warm inflation for lrs bi universe model . in section
* 3 * , we deal with weak dissipative regime and develop the model in two particular scenarios ( i ) intermediate inflation ( ii ) logamediate inflation .
we evaluate explicit expressions for inflaton , potential and rate of decay as well as perturbation parameters .
the behavior of these physical parameters is checked through graphical analysis by constraining the model parameters with recent observations .
finally , the results are summarized in section * 4*.
in this section , we present basic formalism of warm inflation in the background of lrs bi universe model whose line element is given as @xcite @xmath18 where @xmath19 and @xmath20 denote the expansion measure along @xmath21-axis and @xmath22-axis , respectively . under a linear relationship , @xmath23 ( @xmath24 be the anisotropic parameter ) , the above metric is reduced to @xmath25 the basic ingredients of the universe are assumed to be self - interacting scalar field ( @xmath26 ) and radiation field @xmath27 .
the inflaton possesses following energy density @xmath28 and pressure @xmath29 , respectively @xmath30 where @xmath31 is the effective potential associated with @xmath26 and dot stands for derivative with respect to cosmic time @xmath14 .
the anisotropic warm inflation is described by the first evolution equation and conservation equations of inflaton and radiation given by @xmath32 where @xmath33 is the radiation density , @xmath34 is the directional hubble parameter and dissipation factor @xmath35 is introduced to measure the decay rate .
the second law of thermodynamics suggests @xmath36 implying that @xmath37 dissipates into @xmath33 .
it is found that it can be considered as a constant , function of inflaton @xmath38 , function of temperature @xmath39 , function of both @xmath40 and equivalent to @xmath31 in some papers @xcite . here
, we take a general form of the dissipative coefficient as @xmath41 where @xmath42 be any arbitrary integer and @xmath43 is associated to the dissipative microscopic dynamics @xcite . in this reference , zhang and basero - gil et al .
analyzed different choices of the integer @xmath42 which correspond to different expressions for dissipation coefficient .
in particular , for @xmath44 , @xmath45 , where @xmath46 ( @xmath47 is the multiplicity of the @xmath48 superfield and @xmath49 is the number of decay channels available in @xmath48 s decay ) @xcite .
the value @xmath50 leads to @xmath51 ( represents the high - temperature susy case ) , @xmath52 generates @xmath53 ( corresponds to an exponentially decaying propagator in the susy case ) and @xmath54 leads the decay rate @xmath55 ( corresponds to the non - susy case ) . the case @xmath44 implies the most common form @xmath56 considered for the warm intermediate and logamediate models @xcite . for convenience ,
we only focus on the parameter regime @xmath57 .
in particular , we have paid attention to the cosmological results , i.e. , the potential takes the monomial form and the hybrid - like form when @xmath58 . during inflationary regime ,
stable regime can be obtained by applying an approximation , i.e. , @xmath59 . under this limit
, the evolution equation is reduced to @xmath60 using this equation along with conservation equation of inflaton , we have @xmath61 where decay rate is the ratio of @xmath62 by @xmath34 , given by @xmath63 . in warm inflation ,
the radiation production is assumed to be quasi - stable where @xmath64 . using eq.([5 ] ) and
quasi - stable condition in the last equation of eq.([2 ] ) , it follows that @xmath65 where @xmath66 in which @xmath67 is known as the number of relativistic degrees of freedom .
the temperature of thermal bath can be extracted from second equality of the above equation as @xmath68^{\frac{1}{4}}.\ ] ] substituting the value of @xmath69 in eq.([3 ] ) , we have @xmath70 where @xmath71^{\frac{n}{4}}$ ]
. the effective potential can be obtained from first evolution equation with the help of eqs.([5 ] ) and ( [ 6 ] ) as @xmath72.\ ] ] in the following , we shall develop warm inflationary model during intermediate and logamediate scenarios for weak dissipative regime .
here , we consider that warm anisotropic inflationary model evolves according to weak dissipative regime where @xmath73 . in this era , the scale factor evolution is given by @xcite @xmath74 using this scale factor in eq.([5 ] ) , we find solution of inflaton field as @xmath75 where @xmath76^{\frac{1}{2}}$ ] . substituting the value of @xmath14 from the above solution with @xmath77 , we obtain @xmath34 in terms of @xmath26 as @xmath78 in weak dissipative regime , @xmath31 turns out to be @xmath79 the friction term under weak dissipation is reduced to @xmath80 and can be written in terms of inflaton as @xmath81^{\frac{n}{4-n } } \psi^{\frac{4f^{\ast}(1+n)-2n}{f^{\ast}(4-n)}}.\ ] ] in @xcite , a detailed analysis of inflationary scenario with anisotropy is proposed .
remarkably , they have found that degrees of anisotropy are universally determined by the slow - roll parameter .
since the slow - roll parameter is observationally known to be of the order of a percent , therefore anisotropy during inflation can not be entirely negligible . in order to analyze the slow - roll dynamics , the dimensionless slow - roll parameters are @xmath82 and @xmath83 @xcite , where @xmath82 is the standard slow - roll parameter while @xmath83 can be expressed in terms of @xmath82 . a sensible inflation not only demands @xmath84 but also @xmath83 must be small over a reasonably large time period .
these parameters are represented as a function of @xmath26 @xmath85 the condition for the occurrence of inflation @xmath86 is satisfied when inflaton is restricted to @xmath87^{\frac{1}{2}}\alpha_{0}.\ ] ] the initial inflaton @xmath88 is obtained at the earliest possible inflationary stage where @xmath89 @xmath90^{\frac{1}{2}}\alpha_{0}.\ ] ] the number of e - folds @xmath91 interpolated between two different times @xmath92 ( beginning of inflation ) and @xmath93 ( end of inflation ) is defined as follows @xmath94 the second equality shows @xmath95 in terms of @xmath26 using eq.([12 ] ) .
perturbations are usually characterized in terms of four quantities , i.e. , scalar ( tensor ) power spectra @xmath96 ( @xmath97 be the wave number ) and corresponding scalar ( tensor ) spectral indices @xmath98 . the power spectrum is introduced to measure the variance in the fluctuations produced by inflaton @xcite . for a standard scalar field ,
the density perturbation could be written as @xmath99 @xcite . in weak dissipative regime , @xmath100 @xcite .
the scalar power spectrum is obtained by combining eqs.([7 ] ) , ( [ 12 ] ) and ( [ 14 ] ) as @xmath101^{\frac{1}{4-n}}h_{2}^{\frac{11 - 3n}{4-n } } \\\label{16}&\times&(-\dot{h_{2}})^{\frac{n-3}{4-n } } \psi^{\frac{1-n}{4-n}}=\alpha_{1}\psi^{-\beta_{1}},\end{aligned}\ ] ] where @xmath102^{\frac{1}{4-n}}(a^{\ast}f^{\ast})^{2 } ( 1-f^{\ast})^{-\frac{3-n}{4-n}}\alpha_{0}^{\frac{1-n}{4-n}+\beta_{1}},\\\nonumber \beta_{1}&=&\frac{10 - 2n - f^{\ast}(17 - 5n)}{f^{\ast}(4-n)}.\end{aligned}\ ] ] the scalar spectral index for our model is defined as @xmath103\left[\frac{10 - 2n - f^{\ast}(17 - 5n ) } { f^{\ast}(4-n)}\right]\psi^{-2}.\ ] ] inserting @xmath104 in the expression of @xmath95 , we obtain the value of final inflaton @xmath105^{\frac{1}{2}}.\ ] ] using this equation , @xmath106 becomes @xmath107,\ ] ] which yields the specific value of @xmath108 in terms of @xmath95 and @xmath106 @xmath109 the parameter @xmath110 can be calculated through eq.([16 ] ) as @xmath111^{\frac{f^{\ast}}{2n-10 + 2f^{\ast}(13 - 4n)}}(1+f^{\ast}(n-1 ) ) ^{\frac{10 - 2n - f^{\ast}(17 - 5n)}{2(2n-10+f^{\ast}(13 - 4n))}}\\\label{b}&\times&p_{r } ^{\frac{f^{\ast}(4-n)}{2n-10+f^{\ast}(13 - 4n)}},\end{aligned}\ ] ] where @xmath112 the rate of decay as a function of @xmath106 can be calculated as follows @xmath113^{\frac{2(2-n)+f^{\ast}(3n-2)}{f^{\ast}(4-n)}}\\\label{c } & \times&\frac{\kappa}{(\mu+2)(a^{\ast } f^{\ast})\alpha_{0}^{\frac{2(1-f^{\ast})}{f^{\ast}}}}.\end{aligned}\ ] ] figures * 1 * and * 2 * verify that @xmath114 for specific values of the parameters .
the corresponding tensor perturbations are @xmath115 combining eqs.([16 ] ) and ( [ 17a ] ) , the tensor to scalar ratio ( in terms of @xmath116 ) becomes @xmath117\\\label{d}&=&\alpha_{2 } \left[\frac{12(1 + 2\mu)(1-f^{\ast})(10 - 2n - f^{\ast}(17 - 5n))}{\kappa ( \mu+2)^{2}{f^{\ast}}^{2}(4-n)(1-n_{s})}\right]^{\frac{\beta_{2}}{2}},\end{aligned}\ ] ] where @xmath118 during weak dissipative regime , we have calculated constraint on the parameter @xmath43 using eqs.([b ] ) , ( [ c ] ) and ( [ d ] ) in which @xmath119 .
it is proved that in anisotropic model the dissipation coefficient @xmath120 evolutes during weak dissipative regime for these constraints ( figures * 1 * and * 2 * ) .
the range of parameter @xmath43 for three different values of @xmath42 and corresponding @xmath108 are given in table * 1*. the parameter @xmath108 is obtained through eq.([a ] ) by fixing @xmath121 and @xmath122 .
figures * 3 * and * 4 * show the dependence of @xmath123 on @xmath106 for specific values of the parameters .
the left plot of figure * 3 * ( blue curve ) shows that warm anisotropic model is well fitted with recent observations ( wmap9 , plank , bicep2 ) for @xmath124 during weak intermediate regime .
we have also computed an upper bound , @xmath125 that improves the compatibility of our model with recent observations .
planck data places stronger bound on @xmath126 as compared to wmap9 and bicep2 .
figure * 3 * ( right panel ) is plotted for @xmath52 and three different values of @xmath43 make the model well supported by the wmap9 , planck as well as bicep2 for @xmath127 .
similarly , for @xmath54 ( figure * 4 * ) , our model remains compatible with recent observations during @xmath128 .
we have found that the presence of anisotropic parameter @xmath129 leads to increase the values of @xmath43 as compared to frw @xcite .
it is also noted that the value of @xmath43 decreases with the increase of @xmath42 .
.constraints on @xmath43 for different values of @xmath42 [ cols="^,^,^",options="header " , ] in order to constrain the physical parameters used in warm logamediate model , we numerically solve eqs.([20 ] ) and ( [ f ] ) for three different values of @xmath42 .
the values of the model parameters @xmath110 and @xmath130 for particular @xmath43 are picked up from the defined range are given in table * 3*. figures * 6 * and * 7 * show two - dimensional marginalized constraints on the inflationary parameters @xmath123 and @xmath106 derived from recent data . these graphs are plotted for three different values of @xmath131 and @xmath43 .
the left panel of figure * 6 * proves that our anisotropic warm inflationary model is compatible with recent observations in the range @xmath132 as @xmath122 generates very weak bound @xmath133 .
analogously , figure * 6 * ( right panel ) verifies that the range @xmath134 remains well consistent with recent observations .
our anisotropic model satisfies the bounds of wmap9 , planck and bicep2 for @xmath135 ( figure * 7 * ) .
one of the characteristics of warm inflation is @xmath136 , where @xmath69 is the temperature of thermal background of the radiation .
further , we assume that @xmath137 , it is not a free parameter , we can restrict @xmath138 using recent planck data @xmath139 and an upper bound for @xmath140 at the pivot point @xmath141 .
we have checked the condition of warm inflation ( @xmath142 ) for our model .
figure * 2 * shows a restriction on @xmath143 , i.e. , @xmath144 .
the idea of warm inflation is influenced by the friction term in the inflaton equation of motion .
the magnitude of the damping term suggests the possibility that it could be the dominant effect prolonging inflation . in this paper
, we study the possible realization of an expanding intermediate and logamediate scale factors in weak dissipative regime and analyze how these two types of inflation work with generalized form of the dissipation coefficient @xmath145 . to this end
, we have used the framework of homogeneous but anisotropic lrs bi universe model which is asymptotically equivalent to the standard frw universe .
recent data from the planck satellite prove that large angle anomalies represent real feature of the cmb map of the universe .
this result has a key importance since the small temperature anisotropies and large angle anomalies may be caused by some unknown mechanism or an anisotropic phase during the evolution of the universe .
this statement is particularly interesting because it helps to modify the present model or to construct an alternative model to decode the effects of the early universe on the present day lss without affecting the processes in the nucleosynthesis @xcite .
scalar field leads to isotropisation , necessary for compatibility with standard cosmological models at late times , as well as inflation which accounts for structure formation .
anisotropic model provides bouncing solutions that corresponds to the late - time evolution which is driven to isotropy and spatial flatness .
inflationary models has ability to make a transition form early anisotropic phase to late - time isotropic evolution .
we have assumed that the universe is composed of standard scalar field and radiation . under slow - roll approximation , we have found solutions of the first evolution equation ( field equation ) in weak dissipative regime . during intermediate and logamediate eras , the explicit expressions for inflaton @xmath146 , corresponding effective potential @xmath147 and rate of dissipation @xmath148 are calculated .
moreover , we have evaluated perturbation parameters including slow - roll parameters @xmath149 to find the more general conditions on the starting and ending conditions for the occurrence of inflationary era , scalar and tensor power spectra @xmath150 , scalar spectral index @xmath151 and finally observational parameter of interest , i.e. , tensor to scalar ratio @xmath9 . in each case , we have constrained the model parameters @xmath152 by wmap9 , planck and bicep2 data for three particular values of @xmath153 . in both regimes
, we have proved that the values of the model parameters given in tables * 1 * -*3 * lead to @xmath114 as shown in figures * 1 * , * 2 * and * 5*. we conclude that theses constraints make our warm anisotropic inflationary model with generalized form of @xmath62 well supported by recent observations .
the trajectories of @xmath154 plotted in both regimes are the verification of our results .
the results of this paper deviate from frw universe in the following way . during weak intermediate era
, it is observed that due to the presence of anisotropic parameter , @xmath129 , constraints on the important model parameter @xmath43 increase as compared to frw universe .
for example , in anisotropic universe ( lrs bi ) , @xmath50 leads to @xmath155 while for isotropic universe ( frw ) , this is @xmath156 .
further , it is observed that the value of @xmath43 decreases with the increase of @xmath42 due to anisotropic parameter .
since @xmath43 acts as coupling parameter , so its decreasing nature is well consistent with observations , i.e. , coupling between two constituents of the universe must be decreasing with the cosmic evolution .
this result has opposite effect as compared to isotropic universe .
similarly , in logamediate regime , we are successful in constraining @xmath135 that leads to compatibility of the model with wmap9 , planck as well as bicep2 .
it is observed that the compatibility range of @xmath43 is less than frw universe .
it is also found that compatibility of the model disturbs for too large values of the anisotropic parameter @xmath157 .
the case @xmath44 where ( @xmath158 ) is discussed for strong dissipation regime during intermediate and logamediate regimes @xcite .
it is worth mentioning here that all the results reduce to the isotropic universe for @xmath159 @xcite and @xmath44 leads to @xcite .
the work in strong dissipative regime and interpolation between weak and strong regimes is under process .
barrow , j.d . : phys .
b * 235*(1990)40 ; barrow , j.d . and saich , p. : phys .
b * 249*(1990)406 ; muslimov , a. : class . quantum grav . * 7*(1990)231 ; rendall , a.d . : class .
quantum grav .
* 22*(2005)1655 .
barrow , j.d . and
liddle , a.r .
d * 47*(1993)r5219 ; starobinsky , a.a .
. lett . * 82*(2005)169 ; del campo , s. , herrera , r. , saavedra , j. , campuzano , c. and rojas , e. : phys . rev .
d * 80*(2009)123531 ; herrera , r. and videla , n. : eur .
j. c * 67*(2010)499 ; herrera , r. , olivares , m. and videla , n. : eur .
j. c * 73*(2013)2295 .
moss , g. and xiong , c. : arxiv : hep - ph/0603266 ; berera , a. , moss , i.g . and ramos , r.o .
. phys . * 72*(2009)026901 ; bastero - gil , m. , berera , a. and ramos , r.o . : j. cosmol .
. phys . * 09*(2011)033 ; cerezo , r. and rosa , j.g . :
j. high energy phys . * 01*(2013)024 . | this paper explores the dynamics of warm intermediate and logamediate inflationary models during weak dissipative regime with a general form of dissipative coefficient .
we analyze these models within the framework of locally rotationally symmetric bianchi type i universe . in both cases ,
we evaluate solution of inflaton , effective scalar potential , dissipative coefficient , slow - roll parameters , scalar and tensor power spectra , scalar spectral index and tensor to scalar ratio under slow - roll approximation .
we constrain the model parameters using recent data and conclude that anisotropic inflationary universe model with generalized dissipation coefficient remains compatible with wmap9 , planck and bicep2 data .
* keywords : * warm inflation ; slow - roll approximation . + * pacs : * 98.80.cq ; 05.40.+j . |
topology has played an increasingly important role in condensed matter physics . an early application involved using topological chern - simons theory in quantum hall systems @xcite . more recently
, topological ideas applied to brillouin zone ( bz ) have played a significant role in classifying topological phases of matter ( for example , see @xcite and references therein ) . whereas the former application involved a dynamical gauge field in physical spacetime , the latter cases were formulated in the momentum space and involved topological aspects of static gauge fields such as the berry connection .
it is thus natural to ask : can dynamical gauge fields naturally occur in a condensed matter system ?
the aim of this paper is to argue that this can be done at least in the context of three - dimensional ( 3d ) time reversal invariant ( tri ) superconductors , where ( a slightly modified version of ) berry s gauge field can be viewed as dynamical in the brillouin zone and where the action of the theory is the chern - simons action . in this class of theories ,
the fluctuations of the berry connection are induced from quantum fluctuations of the superconductor s pairing amplitude .
more interestingly , 3d tri superconductors are known to be capable of having gapless nodal lines in the bz @xcite , and we will show that they play exactly the role of wilson loop observables in the chern - simons theory . in nature , a large class of materials belongs to 3d tri superconductors , which includes most of the conventional s - wave superconductors described by the bardeen - cooper - schrieffer theory , and a large number of unconventional superconductors such as many heavy fermion superconductors , cuprates , iron - based superconductors and the 3d topological superconductors . in particular , unconventional superconductors often exhibit a highly anisotropic pairing amplitude such as p - wave , d - wave or their hybridizations with s - wave , and have much richer phenomena in experiments @xcite .
previous studies have shown from different perspectives the significance of topology in these tri unconventional superconductors .
gapped 3d tri superconductors are shown to have a @xmath0 topological classification @xcite , which fit in the generic k - theory classification framework of gapped topological phases @xcite .
topological superconductors are defined as such superconductors with a nonzero topological number ( e.g. the @xmath1he - b phase ) , and are shown to support gapless topological majorana fermions on the surfaces @xcite . on the other hand ,
the gapless 3d tri superconductors contain gapless nodal lines in the bz that are allowed by the time - reversal symmetry , with known examples such as the heavy fermion superconductor cept@xmath2si and the cuprates @xcite .
several recent studies show that topological numbers can also be defined for these nodal - line superconductors in terms of k - theory @xcite , which give rise to various types of topological majorana surface states @xcite .
in addition , in analogy to 3d weyl semimetals which can be viewed as intermediate phases between 3d topological insulators and trivial insulators @xcite , nodal - line superconductors can also arise as gapless intermediate phases between conventional tri and topological tri superconductor phases @xcite .
these different while related facts strongly indicate the existence of a unified topological field theory that describes both gapped and gapless 3d tri topological superconductors .
this also motivates us to consider the dynamical chern - simons theory in the 3d bz for these superconductors .
there is already a hint that a topological field theory in bz can be physically relevant .
in particular , it has been shown @xcite that for an insulator in odd spatial dimensions , the value of the chern - simons ( cs ) action for berry connection of the filled bands in the bz computes the effective theta angle of the corresponding @xmath3 term ( or a gravitational analog proportional to @xmath4 in the 3d case @xcite ) , where @xmath5 is the electromagnetic field in the physical spacetime .
in other words , there is a coupling of the form @xmath6\times \left[\int_{\mathbb r^{2n } } f^n\right]\ .\]]we will be specializing to the case of @xmath7 , i.e. , 3d space in this paper .
( a similar term can be used to compute the gravitational response involving @xmath8 ) .
note that the cs action , which is an angular quantity and has a shift ambiguity , has the correct structure to be the coefficient in front of @xmath9 , which needs to be defined only up to shift symmetry . in this context , it is very natural to ask whether the berry connection @xmath10 can fluctuate .
in particular , we can imagine having in physical spacetime a pulse where @xmath11 , which leads to an effective level of @xmath12 for the cs theory in the bz .
can the berry connection behave as if it were dynamical ? for the answer to be yes , the classical background for @xmath10 must be flat , as is demanded by the cs equations of motion .
this is certainly not the case in general .
however , as we shall see , it is the case for tri superconductors . with this encouraging result
, one may then ask whether there are natural objects in the bz corresponding to wilson loops of the cs theory . indeed , for tri superconductors the symmetries allow gapless nodal lines in the bz , which we will see end up playing the role of wilson lines for the cs theory . for superconductors ,
the relevant term to compute is the gravitational @xmath13 term .
but with the gapless modes , as would be the case if we have nodal lines , the theta angle is ambiguous .
it turns out the choices of resolving this ambiguity by introducing infinitesimal time reversal breaking perturbations to get rid of gapless modes are in 1 - 1 correspondence with allowed basic charges for the cs theory ! we thus find that dressing up the nodal lines with this data gives an unambiguous theta angle for the coefficient of @xmath4 term , which is identified with the free energy of cs theory in the presence of wilson loops and leads to physically measurable effects .
in particular , as we shall see , when the nodal lines change from linked to unlinked , the theta angle changes in units of @xmath14 .
our main discussion is in the context of the abelian chern - simons theory , but we also indicate briefly how the theory gets extended to non - abelian case , and how in particular the @xmath15 case can be potentially realized in experiments
. we will also connect aspects of our discussions with topological strings @xcite which is formulated in a 6-dimensional symplectic manifold ( which is typically taken to be a calabi - yau 3-fold ) . in our context
, the phase space which is @xmath16 turns out to play the role of this 6d symplectic space . in the topological string setup , cs theory lives on 3d lagrangian defects ( ` d - branes ' ) . if they are oriented along the bz @xmath17 , they give rise to the cs theory we find in the bz . in this context ,
wilson loop observables arise from a pair of such 3d defects intersecting along a loop @xcite .
we will show that line defects in physical space lead to such 3d lagrangian branes and can also give rise to wilson loop observables ( nodal lines ) in the bz .
the organization of this paper is as follows . in section
ii , we review the hamiltonian of tri superconductors and define the modified berry connection . in section iii , we relate the topological @xmath18 angle to a chern - simons term .
section iv incorporates the nodal lines as wilson loops , first in minimal model and then in multi - band system . in section
v , we consider non - abelian nodal lines , first in a @xmath15 example and then generalize to the @xmath19 example .
in section vi we discuss connections with topological strings and show how line defects in physical space also lead to effective wilson loops in the bz .
we present our conclusions in section vii , and in the appendix are some computational details .
at the single - particle level , superconductors are described by the bogoliubov - de gennes ( bdg ) hamiltonian which has an inherent charge conjugation symmetry @xmath20 . for a superconductor with @xmath21 electron bands , the bdg hamiltonian can be written in terms of the nambu basis @xmath22 as @xmath23 , where @xmath24 is a @xmath25 matrix , @xmath26 is the momentum , and @xmath27 is the @xmath21-component electron basis of the system . both @xmath28 and @xmath29 are @xmath30 matrices .
@xmath28 represents the single - particle hamiltonian of the system before superconductivity arises , while @xmath29 is the pairing amplitude satisfying @xmath31 as required by the fermion statistics .
the charge conjugation is defined as @xmath32 , or equivalently @xmath33 , where @xmath34 is a @xmath25 matrix , with @xmath35 denoting the pauli matrices ( for particle - hole basis ) and @xmath36 the @xmath30 identity matrix .
the charge conjugation symmetry of the bdg hamiltonian can be seen via the relation @xmath37 .
tri superconductors are a large class of superconductors that are quite robust in nature . in the presence of time reversal symmetry
@xmath38 , electrons which are time - reversal kramers pairs have the maximal spatial wave - function overlap , thus fall into cooper pairs most easily @xcite . for fermions ,
the time reversal symmetry @xmath38 is anti - unitary and satisfies @xmath39 .
it acts on the fermions as @xcite @xmath40 , where @xmath41 is an @xmath30 matrix satisfying @xmath42 , @xmath43 .
the bdg hamiltonian of a tri superconductor can then be shown to satisfy @xmath44 and @xmath45 , namely , @xmath46 is hermitian .
these conditions can be more compactly written as @xmath47 , where @xmath48 is the @xmath25 time - reversal transformation matrix of the nambu basis .
it is useful to define the chiral transformation @xmath49 , which is a unitary hermitian matrix that anti - commutes with the bdg hamiltonian , @xmath50 . upon diagonalizing @xmath51 to @xmath52 under a new basis @xmath53
, the bdg hamiltonian becomes @xmath54 where @xmath55 .
since both @xmath28 and @xmath46 are hermitian , @xmath56 is a general @xmath30 complex matrix .
if @xmath56 is non - singular everywhere , the superconductor is in a fully gapped phase . on the other hand ,
the superconductor becomes gapless ( nodal ) at momentum @xmath26 if @xmath57 . in the absence of additional symmetries other than the time reversal symmetry ,
the gapless submanifolds in the momentum space are nodal points for two - dimensional ( 2d ) superconductors , and are 1d nodal lines for 3d superconductors @xcite .
when the superconductor is centrosymmetric , the nodal lines ( points ) in 3d ( 2d ) become doubly degenerate , as we shall show in sec .
[ secu2 ] . in this paper
we focus on 3d superconductors with nodal lines , which are widely found in noncentrosymmetric superconductors @xcite and quasi-2d centrosymmetric superconductors such as cuprates and iron - based superconductors @xcite .
in general , the matrix @xmath56 can be singular - value - decomposed ( svd ) into @xmath58 , where @xmath59 , @xmath60 are unitary matrices , and @xmath61 is a diagonal matrix with all elements real and nonnegative .
the bdg hamiltonian can then be diagonalized as @xmath62 in literature @xcite , it is conventional to define a @xmath63-form u(@xmath21 ) berry connection for the @xmath21 occupied bands with negative energy as @xmath64 , where @xmath65 is the exterior derivative in the momentum space , and @xmath66 is the eigenstate wave function of the @xmath67-th occupied band at momentum @xmath26 .
this can be rewritten as @xmath68 where we have used @xmath69 , @xmath70 short for @xmath59 , @xmath60 .
however , such a berry connection @xmath71 can not be regarded as a u(@xmath21 ) gauge field for @xmath72 , since the eigenstates @xmath66 are fixed by the hamiltonian and the gauge freedom of @xmath71 is u(@xmath63)@xmath73 instead of u(@xmath21 ) .
( see appendix [ appberry ] ) . in this paper
, we define a modified u(@xmath21 ) berry connection as @xmath74 , where @xmath75 is the vacuum state with no electrons , and @xmath67 runs over all the @xmath76 eigenstates . in the tri case here , such a berry connection can be rewritten as follows : @xmath77 physically , @xmath78 is the berry connection felt by an electron when its wave function is projected onto the bdg eigenstates and evolves adiabatically . for superconductors ,
this definition of @xmath79 is closely related to the conventional berry connection @xmath80 , as we shall see in sec .
[ seccsa ] .
in particular , in the case @xmath81 , we have @xmath82 , and the two definitions are identical up to a sign . in the generic @xmath21 band case , even though the eigenstates are fixed by the hamiltonian , the u(@xmath21 ) gauge transformation for @xmath79 can be realized via a unitary transformation of the basis @xmath83 ( instead of transformation of eigenstates ) , where @xmath84 is a u(@xmath21 ) matrix . by the definition given above , the berry connection @xmath79 transforms as @xmath85 , which is exactly a u(@xmath21 ) gauge transformation .
this is however not the case for @xmath80 ( see appendix [ appberry ] ) . in this sense ,
the modified berry connection @xmath79 is more naturally defined than the conventional @xmath80 as a u(@xmath21 ) gauge field .
accordingly , a slow external potential @xmath86 added to the system will induce an effective hamiltonian perturbation @xmath87 in the momentum space .
the topological nature of most weakly interacting condensed matter systems is reflected in the berry connections @xcite .
in particular , a 3d gapped superconductor is shown to be described by an effective gravitational topological action @xcite @xmath88 where @xmath89 is the riemann tensor of the @xmath90d spacetime , and @xmath18 is the topological theta angle determined by the superconductor up to multiples of @xmath91 . for spacetimes which are spin manifolds ,
the topological action is quantized to @xmath92 with @xmath93 being the gravitational instanton number @xcite . in the weak field limit
, the einstein gravitation can be reformulated in the gravitoelectromagnetism framework @xcite , and the above action can be rewritten as @xcite @xmath94 where @xmath95 is the effective coupling constant , while @xmath96 and @xmath97 are the gravitoelectric field and the gravitomagnetic field , respectively . when a 3d gapped superconductor is tri , @xmath98 is required and the theta angle @xmath18 must take @xmath99 or @xmath14 . in general , a tri topological superconductor characterized by topological number @xmath100 will have a theta angle @xmath101 @xcite . in particular , a superconductor in the @xmath1he - b phase has @xmath102 .
this form of gravitational topological action is in direct analogy with the topological electromagnetic action for 3d topological insulators and axion insulators @xcite .
these gravitational fields are closely related to the thermal transport of the superconductor @xcite .
the gravitoelectric field @xmath96 can be produced by a gradient of temperature @xmath103 as @xmath104 , while the gravitomagnetic field @xmath97 characterizes the rotational energy flow and is proportional to the effective angular velocity of the matter in the system . as a result ,
the superconductor s response to gravito - magneto - electric effects is affected by this @xmath18 angle . in the band theories of gapped systems ,
the topological theta angle is given by the momentum - space chern - simons action of the non - abelian berry connection as follows @xcite : @xmath105=\frac{1}{4\pi}\int\text{tr}\left(a'\wedge da'-i\frac{2}{3}a'\wedge a'\wedge a'\right),\ ] ] where @xmath80 is as defined in eq .
( [ aprime ] ) for superconductors and is defined similarly for occupied bands of insulators , and
the integration is in the whole brillouin zone . when the gravitational instanton number @xmath106 , the total action @xmath107 will become a chern - simons action at level @xmath12
however , @xmath108 $ ] can not be regarded as defining a non - abelian chern - simons theory .
this is because the berry connection @xmath80 has only u(@xmath63)@xmath73 instead of u(@xmath21 ) gauge freedom , and is generically non - flat for @xmath72 : namely , the non - abelian berry curvature @xmath109 for generic band structures , which is contrary to the classical equation of motion @xmath110 of the chern - simons theory . to obtain @xmath18 for gapped tri superconductors from a well - defined momentum - space chern - simons theory , we consider the modified berry connection @xmath79 , and define @xmath111=-\frac{1}{4\pi}\int\text{tr}\left(a\wedge da - i\frac{2}{3}a\wedge a\wedge a\right).\ ] ] for 3d gapped tri superconductors
, it can be shown that @xmath112=-\mathcal{a}_{cs}[a']=n\pi$ ] with @xmath113 an integer , and the two definitions of @xmath18 coincide ( see appendix [ appberry ] ) .
formula ( [ csa ] ) has the advantage that @xmath79 is a well - defined gauge field with u(@xmath21 ) gauge freedom as we have shown .
more importantly , the equation of motion of chern - simons theory , i.e. , the flatness of @xmath79 , is satisfied by a natural physical condition which is generically true in mean - field theories of tri superconductors : two electrons will form a cooper pair only when they are time - reversal partners and have the same kinetic energy .
this condition is equivalent to stating that the two hermitian matrices @xmath28 and @xmath46 commute with each other . as a result
, they can be simultaneously diagonalized via a unitary transformation as @xmath114 and @xmath115 . therefore
, @xmath59 and @xmath60 in the svd of @xmath56 are related by @xmath116 , where @xmath117 is a diagonal unitary matrix , and @xmath118 is the complex phase of @xmath119 @xmath120 .
following eq .
( [ a ] ) , one can readily show the berry connection is in the following form : @xmath121 which is just a gauge transformation of @xmath122 .
it is easy to see that @xmath123 as a diagonal matrix of exact @xmath63-forms is flat , namely , @xmath124 .
therefore , we reach the conclusion that @xmath79 is also flat , namely , the non - abelian field strength ( modified berry curvature ) @xmath125 at the mean field level . unlike insulators where the berry connection @xmath80 is fixed
, superconductors always have quantum fluctuations of the pairing amplitude @xmath29 around the mean - field solution , which lead to fluctuations in both @xmath80 and @xmath79 .
the pairing fluctuations @xmath126 generically need not respect the time - reversal symmetry or preserve the commutation of @xmath28 and @xmath46 , therefore leading to non - flat fluctuations in the berry connection @xmath79 . to the lowest order , the fluctuation of @xmath79 is linear to @xmath126 .
explicitly , for the simplest @xmath81 case where @xmath79 is a u(1 ) gauge field in 3d , the fluctuation of the phase of @xmath29 breaks the time - reversal symmetry and gives rise to a @xmath127 term in the @xmath128 hamiltonian @xmath129 , which makes the berry connection @xmath79 generically non - flat . in calculating the partition function of the tri superconductors , all the quantum fluctuations in @xmath29 around the mean - field solution should be included in the path integral .
correspondingly , any fluctuations in @xmath79 induced by the pairing fluctuation should be taken into account in evaluating the @xmath18 angle . these fluctuation in @xmath79
can then be naturally regarded as gaussian fluctuations in the vicinity of the flat connection which is a classical solution . since the partition function of the gravitational topological action @xmath107 generically contains contributions from topological sectors of all gravitational instanton numbers @xmath93 , the action @xmath92 in each topological sector should be regarded as a physical action . for this reason
, we shall view the topological @xmath18 angle as a functional @xmath130 $ ] of the gauge field @xmath79 in the rest of the paper , i.e. , as the free energy of the chern - simons theory .
more specifically , we will focus on the case where the gravitational instanton pulse has integer unit @xmath131 , which leads to level 1 chern - simons theory using eq .
( [ grav ] ) , and @xmath18 will be identified with the free energy of this level 1 cs theory .
when a tri superconductor has nodal lines , the system becomes gapless and dissipative , and the gravitational topological action @xmath107 does not seem to have a clear physical meaning . on the other hand ,
nodal lines look like they can behave as wilson line observables of the chern - simons theory . but what picks the charges of these wilson lines ?
it turns out that these two problems cancel out , as we will show below : to make @xmath107 well defined , we need to gap the system , which can be done by choosing small time - reversal breaking perturbations . the choices for these perturbation turn out to pick the charges that the wilson lines carry .
in other words , we will argue that fixing the charges on the wilson line , which leads to well - defined amplitudes for the cs theory , _ is equivalent _ to choosing small perturbations which gap the system , leading to well defined @xmath18 angle , which will be identified with the free energy of the cs theory in the presence of wilson loop observables . in this section
, we will focus on tri superconductors without additional symmetries , such as the noncentrosymmetric superconductors , while tri nodal - line superconductors with inversion symmetry will be discussed in the next section . in a tri superconductor with no other symmetry ,
the bdg band structure is in general nondegenerate except for some zero - measure submanifolds which generically consists of nodal lines .
therefore , each nodal line is associated with a definite pair of bands which are related by the chiral transformation @xmath51 .
it is helpful to consider the minimal toy model with @xmath81 before dealing with the general case , though it is not quite realistic as the spin degrees of freedom ( or equivalently , @xmath39 ) requires the total number of bands @xmath21 to be even .
the bdg hamiltonian is then a @xmath128 matrix @xmath132 where the matrix @xmath56 from eq .
( [ jk ] ) reduces to a complex number @xmath133 , with @xmath134 being its complex phase . accordingly , in the svd @xmath135 , one can choose @xmath136 and @xmath137 where @xmath138 is an arbitrary real function corresponding to the u(@xmath63 ) gauge freedom , and the berry connection @xmath79 reduces to a u(@xmath63 ) abelian gauge field @xmath139 the nodal lines are characterized by the equation @xmath140 , which are closed loops in the momentum space .
therefore , the berry connection @xmath79 becomes ill - defined on each nodal line @xmath141 .
as one winds around a nodal line @xmath141 once , the phase @xmath134 will change by @xmath91 as long as @xmath142 is a generic continuous function . besides
, the gauge function @xmath138 must be chosen to change by @xmath143 ( @xmath144 ) per winding around the nodal line @xmath141 , so that @xmath59 and @xmath60 are single - valued .
the gauge field @xmath79 then satisfies @xmath145 where @xmath146 is a loop that winds around the nodal line @xmath141 once , and @xmath113 is an arbitrary integer .
therefore , a nodal line can be viewed as a vortex line in the momentum space that carries a @xmath147 berry flux @xmath148 .
such a u(@xmath63 ) berry connection can be derived from the equation of motion of the chern - simons theory by regarding the nodal lines as wilson loops . by modifying the topological theta angle to @xmath149=-\frac{1}{4\pi}\int a\wedge da+\sum_bq_b\oint_{\mathcal{l}_b}a\ ] ] and regarding it as a physical action
, one finds the equation of motion @xmath150 where @xmath151 stands for the delta function at the nodal line @xmath141 in the two perpendicular directions @xmath152 and @xmath153 of @xmath141 .
the flux formula in eq . ( [ flux ] ) is then satisfied by assigning each wilson loop a half - integral charge @xmath154 . to better understand such a modified theta angle
, we can consider a weak time reversal symmetry breaking in the system .
more explicitly , we introduce a small tri breaking mass term @xmath155 to the system , so that the hamiltonian becomes @xmath156 and a nodal line will be lifted into a gap of size @xmath157 ( @xmath155 is real ) .
this mass term may emerge effectively from magnetic disorders , or could be induced via uniform tuning of the parameters of the system ( e.g. , the phase of @xmath29 ) .
if @xmath155 is nonzero on all nodal lines @xcite , the superconductor becomes fully gapped , and the berry connections @xmath79 and @xmath158 become well - defined everywhere .
in particular , in the limit @xmath159 , the berry phase along a loop @xmath146 that winds around a nodal line @xmath141 once becomes definite : @xmath160 where @xmath161 denotes the sign of @xmath155 .
this is analogous to the berry phase around a dirac point in the 2d graphene , which falls into @xmath162 depending on the sign of the dirac mass added .
accordingly , the wilson loop charge becomes @xmath163 and is no longer indefinite , either . in principle
, different nodal lines can have different signs of mass , and therefore different wilson loop charges .
therefore , the ambiguity of wilson loop charges @xmath154 disappears once the time reversal symmetry of the superconductor is explicitly weakly broken , and are restricted to @xmath164 .
let us clarify the meaning of half - integral @xmath165 charge . recall that we are dealing with a @xmath165 chern - simons theory at level @xmath63 .
the hilbert space of this theory consists of conformal blocks of the 2d cft associated with a complex free fermion.@xcite this theory also admits an extra field , the spin field , which creates branch cuts for the fermion correlations . using bosonization
, we can view the fermion as @xmath166 and the spin field as @xmath167 .
the insertion of the @xmath168 fields corresponds to charge @xmath169 states in the @xmath165 chern - simons theory at level 1 .
here we have restricted to gravitational instanton number @xmath131 . for a more general @xmath12 ,
the action is @xmath170 $ ] , which we argue emerges from cs theory at level @xmath12 coupled with wilson loops : for eq .
( [ mono ] ) to still hold , we also need to rescale the charges on the wilson loops so that the charge carried by the wilson loops are now @xmath171 . the gravitational topological action @xmath107 with coefficient @xmath130 $ ] defined in eq .
( [ u1theta ] ) also acquires a physical meaning under such a weak time reversal symmetry breaking .
since the mass term @xmath155 gaps out the superconductor , @xmath107 is now well - defined as discussed in sec .
[ seccsa ] , and one has to go back to eq .
( [ theta ] ) for the definition of @xmath18 in the absence of time reversal symmetry . on the other hand , making use of the equation of motion in eq .
( [ eom ] ) , one can show @xmath172 and thus the classical value of @xmath130 $ ] defined in eq .
( [ u1theta ] ) is @xmath173=\frac{1}{4\pi}\int a\wedge da=\frac{1}{4\pi}\int a'\wedge da'=\mathcal{a}_{cs}(a')\ , \ ] ] which agrees exactly with the definition of @xmath18 in eq .
( [ theta ] ) , and we have used the fact that @xmath82 for @xmath81 .
this verifies the validity of the topological theta angle @xmath130 $ ] defined in the presence of nodal lines . as a topological quantum field theory
, the chern - simons theory has an intrinsic connection with the knot invariants as first revealed by witten @xcite , which plays an important role in the description of @xmath174d topological many - body states such as the quantum hall states , and quasiparticle statistics therein @xcite .
indeed , we will now argue that the partition function of the theory as measured by the @xmath18-angle , in the presence of nodal lines , behaves exactly as expected for a quantum chern - simons theory with the corresponding wilson loops .
in the u(@xmath63 ) case at level @xmath63 , the partition function of cs theory with wilson loops is given by @xcite @xmath175=\pi\sum_{b , c}q_bq_c\phi(\mathcal{l}_b,\mathcal{l}_c)\ , \ ] ] where @xmath176 is the gauss linking number between nodal lines @xmath141 and @xmath177 .
in the case of u(@xmath63 ) chern - simons theory , this is also the classical valuation of the action at the critical point ( because the action is quadratic ) . since the wilson loop charges @xmath154 are half integral , changing the linking number between two nodal lines by @xmath63 yields a @xmath178 shift in @xmath130 $ ] . besides , the calculation of eq .
( [ thetalinking ] ) needs a proper regularization of the self - linking number for @xmath179 , which is related to the framing of the wilson loop .
change in the @xmath130 $ ] angle .
( b ) this shows how a nodal line @xmath141 can cross itself , after which its self - linking number changes by @xmath180 .
( c ) the generic skein relation for knot polynomials , which gives the rules for the local surgery of a knot configuration.,width=316 ] now we would like to derive this same result using the definition of the @xmath18-angle for our system .
we will show that as we change the linking number between nodal lines , the @xmath18-angle changes exactly as expected for @xmath165 cs theory at level 1 .
consider the local crossing of two nodal lines @xmath141 and @xmath177 as shown in fig .
[ fig1](a ) .
@xmath141 and @xmath177 can be either different nodal lines or different parts of the same nodal line . the two configurations in fig .
[ fig1](a ) can be adiabatically connected by a parameter @xmath181 $ ] , where @xmath182 and @xmath183 corresponds to the left and right configurations , respectively .
for example , this can be realized by assuming the off - diagonal element in eq .
( [ hu1 ] ) to be @xmath184 locally , where @xmath141 and @xmath177 are at @xmath185 and @xmath186 , respectively .
accordingly , the field strength in the 4d parameter space @xmath187 is @xmath188 $ ] . besides , the wilson loop integrals @xmath189 and @xmath190 change by @xmath191 and @xmath192 , respectively .
the resulting total variation in the expectation value of @xmath130 $ ] is thus @xmath193\big|_{t=-1}^{t=1}=-\int_{-1}^{1}d\mathcal{a}_{cs}[a]+2\pi q_bq_c+2\pi q_bq_c\\ & = -\frac{1}{4\pi}\int_{4d}\text{tr}(f\wedge f)+4\pi q_bq_c=-2\pi q_bq_c+4\pi q_bq_c\\ & = 2\pi q_bq_c=\pm\frac{\pi}{2}\ \mod 2\pi\ , \end{split}\ ] ] where we have used the fact that @xmath194 are @xmath169 .
if we define @xmath195=e^{i\theta_{cl}[a]}=\langle e^{i\theta[a]}\rangle$ ] as a knot invariant , the above result yields a simple skein relation @xmath196-e^{i2\pi q_bq_c}w_2[a]=0\ , \ ] ] where @xmath197 $ ] and @xmath198 $ ] correspond to two configurations which are only different locally as shown in fig . [ fig1](a ) .
this is a special case of the general skein relation as shown in fig .
[ fig1](c ) , with coefficients @xmath199 , @xmath200 and @xmath201 @xcite .
it is convenient to assign each nodal line @xmath141 a direction along which the berry flux is @xmath202 , i.e. , @xmath203 , so that skein relation is unambiguous with @xmath204 . in particular , when a nodal line @xmath141 crosses a distinct nodal line @xmath177 the @xmath18-angle changes the contribution to @xmath18-angle changes by @xmath178 exactly as is expected from eq .
( [ thetalinking ] ) .
this is easy to explain : both computations can be reduced to finding classical solutions of the cs theory with a given nodal line and evaluating the action at the classical value .
however , we also need to discuss the nodal line crossing itself as shown in fig . [ fig1](b ) . here
$ ] should also change by @xmath178 due to the skein relation , which indicates the ` self - linking number ' @xmath206 should change by @xmath180 .
the self - linking number is also known as the framing of the knot @xcite , which we have not yet discussed in the context of the nodal lines . as we will argue below , due to the doubling of the nodal lines due to time reversal invariance , such self - linking , or framing contributions , cancel pairwise as a nodal line crosses itself .
in other words , we can delete the @xmath179 terms in the sum in eq .
( [ thetalinking ] ) as far as its contribution to the total @xmath18-angle is concerned .
@xcite in practice , in tri superconductors , nodal lines usually occur in pairs at momenta related by the time - reversal symmetry . therefore , the nodal line crossings are always doubled , and the change in @xmath130 $ ] will be doubled to @xmath99 or @xmath14 , depending on the signs of masses of each nodal line in the weak breaking of time - reversal symmetry .
for instance , if one assign @xmath141 and its time - reversal partner the same mass , but assigns @xmath177 and its time - reversal partner opposite masses , the crossing of @xmath141 and @xmath177 ( and their time - reversal ) will induce a @xmath14 shift in @xmath130 $ ] .
however , if a nodal line @xmath141 crosses itself once and recovers the original configuration , due to the time - reversal doubling , time - reversal related nodal lines always have opposite self - crossing phases , and the total change in @xmath205 $ ] is always zero .
this justifies why we may neglect the contribution of self - linking ( or framing number ) in finding the contribution of linked nodal lines to @xmath18 .
now we proceed to consider nodal lines in general @xmath21-band tri superconductors . following the discussion above eq .
( [ flata ] ) , we shall assume @xmath28 and @xmath46 are simultaneously diagonalized by @xmath207 , where @xmath208 is the @xmath67-th normalized eigenvector of @xmath28 ( @xmath209 ) , or the wave function of the @xmath67-th electron band before superconducting . in the absence of additional symmetries , each nodal line @xmath141
is associated with a definite electron band .
the modified topological theta angle can then be naturally generalized as @xmath210=-\mathcal{a}_{cs}[a]+\sum_bq_b\oint_{\mathcal{l}_b}u_{\alpha_b}^\dag au_{\alpha_b}\ , \ ] ] where @xmath79 is now a u(@xmath21 ) gauge field , @xmath112 $ ] is the non - abelian chern - simons action , and @xmath211 is the @xmath21-component wave function of the electron band @xmath212 associated with nodal line ( or wilson loop ) @xmath141 .
similarly , the wilson loop charges @xmath154 are half - integral .
the wilson loops @xmath141 are coupled to the projected u(@xmath63 ) berry connection @xmath213 , and thus breaks the u(@xmath21 ) gauge symmetry .
if the 3d momentum space is viewed as a @xmath214d `` spacetime '' , the wilson loops can be interpreted as world lines of particles , while @xmath211 are the u(@xmath21 ) charges carried by the particles .
due to such a coupling , the chern - simons theory governed by @xmath130 $ ] is enriched with more structures .
the equation of motion then becomes @xmath215 where @xmath216 is now the u(@xmath21 ) berry curvature .
when two nodal lines @xmath141 and @xmath177 cross each other as shown in fig .
[ fig1](a ) , a calculation similar to eq .
( [ calcrossing ] ) yields the jump in theta angle @xmath217=2\pi q_bq_c|u_{\alpha_b}^\dag u_{\alpha_c}|^2=2\pi q_bq_c\delta_{\alpha_b\alpha_c}\ , \ ] ] provided the eigenvectors @xmath211 and @xmath218 are nonsingular during the crossing .
therefore , the crossing produces a change in @xmath205 $ ] only if the two nodal lines are in the same band . in metals which become superconductors at low temperatures
, nodal lines always live on the 2d fermi surfaces of the metals .
one may therefore wonder what kind of fermi surface can give rise to linked nodal lines . fig .
[ fig1](a ) shows how two nodal lines of linking number @xmath63 can be drawn on a @xmath180-torus fermi surface .
we note that they are not time - reversal invariant by themselves .
therefore , their time - reversal partners should coexist on another @xmath180-torus fermi surface in the bz . to unlink the two nodal lines @xmath219 and @xmath220
, one must imagine a process during which the poloidal radius of a certain part of the torus shrinks to zero and then expands back .
the two nodal lines can then cross each other at the zero poloidal radius point . and
annihilated at point @xmath221 , whose trajectory is a closed surface @xmath222 .
( c)-(d ) the sign of pairing @xmath223 on a fermi surface of a gapped tri superconductor can be reversed by creating nodal lines and making them sweep around the fermi surface , which leads to a phase transition between topological superconductors if the fermi surface chern number @xmath224 is nonzero.,width=316 ] the theta angle @xmath130 $ ] defined in eq .
( [ untheta ] ) also explains in a different way the fact that @xmath101 for a tri gapped topological superconductor with topological number @xmath225 .
this fact comes from the following observation : consider the process in which a nodal line @xmath141 is created at point @xmath226 and then annihilated at point @xmath221 as shown in fig .
[ fig2](b ) as a function of adiabatic parameter @xmath181 $ ] , during which the system remains tri .
the trajectory of @xmath141 is thus a closed 2d manifold @xmath222 in the 3d momentum space .
accordingly , the change in @xmath205 $ ] is @xmath227\big|_{t=-1}^{t=1}=-\int_{-1}^{1}d\mathcal{a}_{cs}[a]+q_b\int_{-1}^1d\oint_{\mathcal{l}_b}a^{(\alpha_b)}\\ & = -\frac{1}{4\pi}\int_{4d}\text{tr}(f\wedge f)+q_b\oint_\mathcal{m}da^{(\alpha_b)}\\ & = 0+q_b\oint_\mathcal{m}f^{(\alpha_b)}=2\pi q_b c_1(\alpha_b)\ , \end{split}\ ] ] where we have used the fact that @xmath228 when there is no crossing of nodal lines , while @xmath229 is the projected u(@xmath63 ) berry curvature of electron band @xmath212 , and @xmath230 is the first chern number of band @xmath212 on the 2d manifold @xmath222 .
we note that even if the u(@xmath21 ) berry curvature @xmath231 , the projected u(@xmath63 ) berry curvature @xmath232 and thus @xmath230 can be generically nonzero .
therefore , @xmath130 $ ] of a tri gapped superconductor can be changed by creation and annihilation of nodal lines .
physically , this can be done by adiabatically changing the pairing amplitude @xmath29 of the bdg hamiltonian .
we shall still assume @xmath114 and @xmath115 , so a nodal line in band @xmath67 is given by @xmath233 .
we note that @xmath234 defines the fermi surface @xmath235 of electron band @xmath67 , while @xmath236 is a real function of @xmath26 .
when the superconductor is gapped , @xmath236 is nonzero everywhere on the fermi surface @xmath235 , and thus has a definite sign @xmath237 there .
qi , hughes and zhang @xcite has shown that the topological number @xmath225 of a gapped tri superconductor is given by @xmath238 where @xmath239 is the fermi - surface first chern number of electron band @xmath67 . now consider a fermi surface @xmath235 with @xmath240 and @xmath241 as shown in fig .
[ fig2](c ) at the beginning .
one can then create two nodal lines @xmath219 and @xmath220 ( which are time - reversal partners ) from the north and south poles of the fermi surface as shown in fig .
[ fig2](d ) , and make them annihilate on the equator . since @xmath223 on @xmath235 changes sign across a nodal line , the sign @xmath242 is reversed at the end of the process , and @xmath225 decreases by @xmath63 . on the other hand , eq .
( [ c1 ] ) tells us that the change in the theta angle @xmath18 is @xmath243 , which therefore agrees with the fact @xmath244 for tri topological superconductors .
this also verifies that the wilson loop charges @xmath154 have to be half - integral .
this also indicates that @xmath130 $ ] for @xmath72 could acquire a geometric contribution solely from the deformation of a nodal line , without involving any nodal - line crossings .
this is different from the u(@xmath63 ) chern - simons theory discussed in sec .
[ secu1a ] which is purely topological .
such an additional phase change is due to the berry curvature of the wave functions @xmath245 . in practice
, @xmath39 always requires the total number of bands @xmath21 to be even , so such a geometric contribution is in general present and nonzero .
so far , we have discussed tri superconductors with nondegenerate bands , where nodal lines behave as wilson loops coupled to projected u(@xmath63 ) berry connections . to construct wilson loops with a non - abelian coupling to the berry connections
, one needs the band structure of the superconductor to have an @xmath21-fold degeneracy @xcite .
theoretically , we can discuss all @xmath72 cases , but it turns out that the @xmath246 case is particularly well motivated in certain superconductors : in particular , adding inversion symmetry to the superconductor yields a @xmath180-fold degeneracy and leads to u(@xmath180 ) nodal lines , as we shall show below . in subsection
a , we discuss the @xmath15 case and in the subsection b , we turn to the more general @xmath19 case .
many tri superconductors are also centrosymmetric , namely , have an inversion symmetry @xmath247 . in this case , the combined symmetry @xmath248 is an anti - unitary symmetry with @xmath249 .
since the @xmath248 symmetry keeps the momentum @xmath26 invariant , such a symmetry ensures the @xmath21 electron bands ( @xmath21 is even ) to be doubly degenerate according to the kramers theorem .
if one denotes the action of @xmath247 on the electron basis as @xmath250 where @xmath251 , one can show @xmath252 and @xmath253 , namely , the matrix @xmath254 has the same properties as @xmath41 .
the inversion symmetry of the superconductor requires @xmath255 , and @xmath256 , where @xmath257 and @xmath258 signs indicate even and odd parity pairings , respectively .
note that @xmath31 , so this parity requirement can be rewritten as @xmath259 , i.e. , the matrix @xmath260 is anti - symmetric for even parity , and symmetric for odd parity .
when the superconductor has a pairing amplitude @xmath29 of even parity , nodal lines are allowed to arise and are stable against perturbations .
this can be seen as follows .
assume @xmath261 , where @xmath262 with @xmath263 being a diagonal matrix , while the basis of each @xmath264 ( @xmath128 identity matrix ) consists of two eigenstates related by the @xmath248 transformation .
effectively , one can regard the two states as carrying spin up and down , respectively .
the matrix @xmath254 can be written as @xmath265 , where @xmath266 denote the pauli matrices in the doublet basis of @xmath248 related eigenstates .
we note that @xmath59 is not unique . since @xmath28 commute with the hermitian matrix @xmath46 , the same matrix @xmath59 diagonalizes @xmath46 into @xmath267 blocks . together with the condition that @xmath268 is anti - symmetric ( symmetric ) for even ( odd ) parity , we find @xmath269 for even parity pairing , where @xmath270 are real functions even in @xmath26 , and @xmath271 for odd parity pairing , where @xmath272 are real vector functions odd in @xmath26 . when the parity is odd , the gapless condition for the bdg hamiltonian is @xmath273 , which is overdetermined in 3d and generically has no solution . on the contrary ,
when the parity is even , the bdg band structure becomes gapless @xmath274 , which defines the nodal lines in the doubly - degenerate electron band @xmath275 . in practice ,
such nodal lines widely exist in cuprates with @xmath276-wave pairings @xcite and iron - based superconductors with modulated @xmath277-wave pairings @xcite . due to the double degeneracy of the bdg band structure ,
the nodal lines are coupled to the berry connection as non - abelian u(@xmath180 ) wilson loops . in the @xmath246 case , the berry connection @xmath79 is itself a u(@xmath180 ) gauge field , and the @xmath18-angle
is simply given by the conventional u(@xmath180 ) non - abelian chern - simons theory .
these wilson loops reside in the fundamental representation @xmath278 of u(@xmath180 ) where the generators are given by @xmath279 and @xmath280 ( @xmath281 ) @xcite , and accordingly the berry connection can be decomposed into @xmath282 .
the matrix @xmath56 in this case is simply @xmath283 , where @xmath142 is a complex number function .
a nodal line is defined by @xmath140 , while winding around the nodal line increases the phase @xmath134 of @xmath142 by @xmath91 . to compute the classical value of gauge field @xmath79
, one can choose @xmath284 and @xmath285 in the vicinity of the nodal line , in which case the field strength is @xmath286 alternatively , one can choose @xmath287 and @xmath288 with a unit vector @xmath289 , which yields a field strength @xmath290 both choices ensure @xmath59 and @xmath60 to be single - valued .
this choice of classical solutions exactly results from the u(@xmath180 ) gauge freedom of the nodal line .
similar to the u(@xmath63 ) case , the field strength of a nodal line becomes well defined when the time reversal symmetry or the inversion symmetry of the system is weakly broken .
for instance , when the hamiltonian acquires a small mass term and becomes @xmath291 explicitly , there are four degrees of freedom , which are @xmath292 and @xmath293 ( @xmath294 ) , which is the expected number for arbitrary fluctuations of a u(@xmath180 ) gauge field in @xmath295 because @xmath296 , where here @xmath297 and @xmath295 .
it is important to note that the field strength @xmath216 will depend on the signs of the eigenvalues of the upper @xmath128 block .
if the signs of the mass eigenvalues of the upper @xmath128 block is @xmath298 , this leads to the curvature sitting in the abelian subalgebra of @xmath15 given by @xmath299\ \delta^2(\mathbf{k}_\perp-\mathbf{k}_{b})dk_\perp^{1}\wedge dk_\perp^{2}.\ ] ] this can be interpreted ( when the gravitational instanton number @xmath131 ) as the response of the @xmath15 system to a charged object with charges @xmath300 $ ] in the cartan of @xmath15 .
if both signs are the same , the wilson loop will carry only a @xmath301 charge .
if the signs are opposite , the two possibilities correspond to the fundamental of @xmath302 together with some overall @xmath165 charge .
note that when the inversion symmetry is broken , the u(@xmath180 ) nodal lines naturally breaks down to abelian nodal lines with wilson loop charges @xmath164 , in agreement with our observation in sec .
[ secinco ] .
the wilson loop observables etc . will work exactly as in the abelian case we already discussed , because the level of the @xmath15 cs theory is 1 .
if we change the gravitational instanton number so that the level of the @xmath15 cs theory is changed to @xmath12 , we would need to choose different weights in the @xmath15 weight lattice in order to get the same @xmath216 we computed in eq .
( [ fsigns ] ) .
in particular , the cartan weights will now be @xmath303 .
apart from the @xmath165 part , this belongs to the @xmath12-fold symmetric representation of the fundamental of @xmath304 , which is known to generate a free current and thus means that its correlations behave as in the abelian case , leading exactly to the same result expected , where @xmath305 gets multiplied by a factor @xmath12 . though uncommon in nature , condensed systems with @xmath113-fold degenerate ( @xmath306 ) band structures are theoretically possible when the symmetries of the systems are high enough . for instance , systems where electrons carry both a spin @xmath307 and a pseudospin @xmath307 may have a @xmath308-fold degenerate band structure when there are no spin - pseudospin couplings
. when such a system is tri and develops a superconductivity , nodal lines with a u(@xmath113 ) symmetry may arise . for simplicity
, here we only consider the case when the total number of electron bands is @xmath309 , which is the minimal system that realizes u(@xmath113 ) nodal lines .
the matrix @xmath56 in eq .
( [ jk ] ) is then of the form @xmath310 with @xmath133 being a complex function .
a nodal line is defined by @xmath140 ; it occupies the fundamental representation of u(@xmath21 ) group , of which the generators are @xmath311 and the @xmath312 generators @xmath313 of su(@xmath21 ) group @xcite .
the matrices @xmath69 and @xmath70 around the nodal line can be generically chosen as @xmath314 where @xmath315 ( @xmath316 ) , and @xmath317 ( @xmath318 ) generates the element @xmath319 of the center of su(@xmath21 ) .
the resulting field strength @xmath216 is @xmath320 which is a classical solution of the u(@xmath21 ) chern - simons theory with u(@xmath21 ) wilson loops . in paricular , just as in the @xmath15 case this will lead to curvature which is proportional to @xmath321 $ ] , which is in the fundamental representation of @xmath19 at level 1 , leading to similar formulae as before .
in many cases it is desirable to consider the effect of defects in the physical space of a condensed matter system . in the physical space of 3d superconductors , vortex lines can be naturally created as line defects , while crystallographic line defects ( dislocations and disclinations ) may also be present , which lead to a breaking of the translational symmetry .
interestingly , as we shall show below , such line defects in physical space give rise to effective nodal lines in the bz in a semiclassical approximation . for simplicity ,
let us consider a line defect along @xmath322-direction located at @xmath323 in the physical space .
the line defect will generically produce a potential energy @xmath324 for the bdg quasi - particles that is centered at @xmath323 .
an quasi - particle with kinetic energy @xmath325 will then have a hamiltonian : @xmath326 where we have used the fact that @xmath327 is still a good quantum number .
assume the hamiltonian has a bound state @xmath328 with a bound state energy @xmath329 .
such a bound state is in general a wave packet of size @xmath330 in the momentum space with the center located at @xmath331 and @xmath332 , where @xmath333 is the localization length of the bound state ( which we assume is much larger than the lattice constants ) , and the functions @xmath334 and @xmath335 are determined by the line defect hamiltonian @xmath336 .
therefore , all the bound states on the physical space line defect live on a `` smeared out '' line @xmath337 in the bz given by @xmath338 , which is a closed loop .
now we assume a quasi - particle is bounded on the line defect .
we then introduce an additional slow potential @xmath339 which slightly breaks the translational symmetry along the line defect .
such a slow potential could result from inhomogeneity of the system or slight distortions of the defect line .
consider a local minimum where @xmath340 .
one can write down a @xmath63d hamiltonian for the quasi - particle : @xmath341 where @xmath26 is restricted on the loop @xmath337 , the bound state energy @xmath342 , and @xmath343 is the berry connection in the bz defined in sec .
ii . for generality
, we have also included a new berry gauge field @xmath344 in the lagrangian space @xmath345 @xcite , which may in general arise when the translational symmetry of the system along @xmath346 and @xmath347 directions is broken @xcite .
the corresponding lagrangian of the particle in the momentum space is then @xmath348 where @xmath349 is the time derivative of @xmath26 .
note that @xmath350 in this lagrangian can be viewed as a periodic potential along the loop @xmath337 .
accordingly , the quasi - particle will have a tunnelling amplitude of winding @xmath351 times along the loop , which can be calculated via the wkb approximation as @xmath352 .
the path integral of the bounded quasi - particle is thus @xmath353 therefore , one sees the loop @xmath337 plays exactly the role of a wilson loop ( nodal line ) that couples to both @xmath79 and @xmath354 , where the wilson loop charge is now an integer winding number @xmath351 .
we note that the slow potential @xmath355 only enters into the tunnelling action @xmath356 , while does not affect the wilson loop integral . in this case , the wilson loop charge @xmath351 and the level of cs action of @xmath79 , i.e. , the gravitational instanton number @xmath12 , are independent of each other . in principle , the gauge field @xmath354 may also have a physical cs action at some level @xmath357 in the lagrangian space @xmath345 . in this case , one will obtain a double cs theory which is more interesting .
in general , the number of quasi - particles bounded on the line defect is not limited to one . to take all number of quasi - particles into account , we can second quantize the above action , rewriting the path integral as @xmath358\psi_j}\ , \ ] ] where @xmath359 is the fermion mode associated with a particular closed path @xmath360 on the loop @xmath337 , and @xmath361 is the period in the time direction .
we note that the berry connection @xmath79 here should be treated as defined at the specific point @xmath362 of the physical space where the quasi - particles are trapped , i.e. , the local minimum of the line defect .
in fact , the strict definition of the berry connection @xmath79 does require specifying a point @xmath363 in the physical space as an origin @xcite , namely , @xmath79 is defined via the bloch wave functions in the unit cell at @xmath363 .
in particular , since the bloch wave functions have a shift ambiguity @xmath364 under a translation @xmath365 ( where @xmath366 is a lattice vector ) , the berry connections @xmath79 defined at @xmath363 and @xmath367 are related by @xmath368 .
this shift ambiguity is closely related to the 3d quantum hall effect @xcite , but does not affect the calculation of @xmath130 $ ] of tri systems here .
similarly , the gauge field @xmath354 and the associated lagrangian space @xmath345 should be viewed as located at the specific position @xmath369 ( local minimum of @xmath355 ) and @xmath338 @xcite .
it turns out that the above description is very natural in the context of topological strings , which was introduced in @xcite and gives a realization of chern - simons theory as a string theory .
this corresponds to strings propagating on a six dimensional symplectic manifold which is typically taken to be a calabi - yau 3-fold .
@xmath21 d - branes wrapping lagrangian 3-cycles lead to @xmath19 chern - simons theory living on it .
moreover , intersecting d - branes along a loop leads to insertion of wilson loop observables as in eq .
( [ sum ] ) on the intersection @xcite . to connect this to the present discussion
, we note that the phase space @xmath370 is a sympletic manifold @xmath16 .
we take two lagrangians to be one wrapped around @xmath371 giving us the cs theory with gauge field @xmath343 , and the other to represent the defect , given by the lagrangian @xmath372 with gauge field @xmath373 living on it .
these lagrangians intersect along @xmath327 , leading to a fermion field which couples to the two gauge fields exactly as in eq .
( [ mohem ] ) @xcite .
it is remarkable that phase space in 3d can naturally give a realization of topological strings !
though our discussion here is based on a straight line defect in physical space , the result holds for generic smooth line defects . in the semiclassical approximation , one can define a momentum along the defect line which plays the role of @xmath327 here , and all the rest of the derivation will follow .
we have seen how a modified berry s connection in the bz of a tri superconductor can behave as a dynamical gauge field in 3d with cs action whose level is given by the gravitational instanton number in the physical spacetime .
moreover , we have seen that gapless nodal lines play the role of wilson loop observables for the chern - simons theory .
the free energy of this cs theory computes the @xmath18-angle of the topological gravitational response @xmath4 of this system .
changing of the topological mutual linking numbers of nodal lines leads to a shift of the @xmath18-angle in units of @xmath14 .
whereas the linking of degenerate manifolds is also studied for the 5d weyl semimetals @xcite , it is reasonable to expect linking and knot invariants to play an important role in more generic gapless topological states of matter .
the examples we have discussed exhibit mostly an abelian structure . even in the multi - band case where we got @xmath19 cs theory ,
the computations reduce to @xmath374 abelian cs theory
. it would be nice to find ways where the linked nodal amplitudes are genuinely non - abelian .
this paper provides an example of how the berry connection in the bz can become dynamical .
in addition , in the discussion of physical space defects in sec . [ secdefect ]
, we see possibility of realizing topological string theory in a condensed matter set - up . we have seen an intriguing realization of topological strings on the six dimensional phase space where defects play the role of d - branes .
it would be very interesting to develop this connection further , for instance , to see how the gauge field @xmath354 in the lagrangian manifold @xmath345 can also be made dynamical and governed by a cs action .
it would also be useful to find more examples along these lines in different condensed matter systems in different dimensions , which need not be limited to topological theories .
we would like to thank c. liu and c. xu for discussions .
we gratefully acknowledge support from the simons center for geometry and physics , stony brook university , where some of the research for this paper was performed during the 2016 simons summer workshop .
cv would also like to thank stanford institute for theoretical physics for hospitality during part of this work .
the research of bl and scz is supported by the nsf grant dmr-1305677 .
the research of cv is supported by the nsf grant phy-1067976 .
the research of fv is supported by the nsf grant dmr-1151208 .
the berry connection @xmath64 defined in the literature @xcite is usually seen as a natural non - abelian generalization of the u(@xmath63 ) berry connection proposed by berry @xcite , where @xmath66 are the eigenstates of the hamiltonian @xmath375 and @xmath376 runs over the @xmath21 occupied bands . however , when the eigenstates are nondegenerate , the only degrees of gauge freedom in this definition are the phase rotations of the eigenstates @xmath377 , which leads to a gauge transformation @xmath378 with @xmath379 being the kronecker symbol .
this just gives @xmath80 a gauge freedom u(@xmath63)@xmath73 .
one may wonder what happens if one makes a unitary transformation of the basis used for writing the hamiltonian as @xmath83 , which is also the basis for writing the eigenstates @xmath66 .
such a transformation leads to @xmath380 and @xmath381 ( as we will show later ) , and according to eq .
( [ aprime ] ) , leads to a transformation @xmath382\ ] ] which is not a gauge transformation .
therefore , the basis transformation does not transform @xmath80 legally .
we conclude @xmath80 only has a u(@xmath63)@xmath73 gauge freedom instead of a u(@xmath21 ) gauge freedom , and can not be regarded as a u(@xmath21 ) gauge field . on the other hand
, we shall show the modified berry connection @xmath74 is a well - defined u(@xmath21 ) gauge field , which is uniquely defined for superconductors . to see this
, we note that under a unitary u(@xmath21 ) basis transformation @xmath83 , one has @xmath383 , therefore @xmath380 and @xmath381 , and @xmath79 given in eq .
( [ a ] ) transforms as @xmath384 which is exactly the gauge transformation of the u(@xmath21 ) gauge field .
therefore , we can view the basis transformation @xmath83 as the u(@xmath21 ) gauge transformation of @xmath79 .
therefore , for gapped superconductors where @xmath69 and @xmath70 can be well defined everywhere in the brillouin zone , we have @xmath385=-\mathcal{a}[a']$ ] . besides
, note that the integral in the 3d brillouin zone @xmath389 is an integer for unitary matrix @xmath69 , we conclude that @xmath385=-\mathcal{a}[a']=n\pi$ ] with @xmath144 , in agreement with the requirement of the time reversal symmetry . in the presence of nodal lines
, however , @xmath69 and @xmath70 become ill - defined on the nodal lines , and this conclusion no longer holds . | berry connection is conventionally defined as a static gauge field in the brillouin zone . here
we show that for three - dimensional ( 3d ) time - reversal invariant superconductors , a generalized berry gauge field behaves as a dynamical fluctuating field of a chern - simons gauge theory .
the gapless nodal lines in the momentum space play the role of wilson loop observables , while their linking and knot invariants modify the gravitational theta angle .
this angle induces a topological gravitomagnetoelectric effect where a temperature gradient induces a rotational energy flow .
we also show how topological strings may be realized in the 6 dimensional phase space , where the physical space defects play the role of topological d - branes . |
the observation of bose einstein condensation ( bec ) in ultracold alkali gases has led to intensive experimental and theoretical investigations of the phenomenon @xcite .
since the atomic gases of the experiments are dilute , and hence weakly interacting , their physical properties may be explored theoretically in terms of basic interatomic interactions and as a first approximation the interactions may be neglected @xcite . this is strongly opposed to the physical properties of related phenomena such as superconductivity in metals and superfluidity in liquid he , which are fundamentally modified by strong interparticle interactions . in the remarkable pioneering experiment @xcite @xmath0rb vapour
was magnetically trapped and evaporatively cooled to approximately @xmath1 nk , where a finite fraction of the particles was observed to occupy the ground state .
when further lowering the temperature this fraction increased abruptly , signalling a bose - einstein condensation .
the confining magnetic field was a time - averaged , orbiting potential @xcite whose effective potential to first order is a three - dimensional anisotropic harmonic oscillator .
the potential was not anisotropic in all three directions , but was cylindrically symmetric .
shortly after , a similar experiment was performed with @xmath2na gas @xcite ; in this case the potential was completely anisotropic . with these experiments in mind ,
we will in the present paper extend our previous work on the thermodynamics of an ideal isotropic gas @xcite to the anisotropic case .
although the interactions among the atoms are weak , they are quite important in the condensate . both experiments and theory show that due to the interaction the actual ground state is considerably larger than the single particle ground state .
however , close to the transition temperature just a few particles occupy the ground state and the gas is much more dilute . at these temperatures
one would believe that the ideal gas is a good approximation . in ref .
@xcite it has been proposed that highly anisotropic traps may freeze out one or two dimensions , so that the system to a good approximation is lower dimensional .
this implies that it is of interest to study these systems also , and we present a preliminary analysis of them in this work .
the plan of the article is as follows . in the next section we study noninteracting bosons in a confining anisotropic harmonic potential in three dimensions ( 3@xmath3 ) . in the framework of the grand canonical ensemble we calculate the condensate fraction ,
the internal energy and the heat capacity analytically using euler - maclaurin summation and compare with exact numerical summation . in section three
we discuss the same system in one and two spatial dimensions . in particular , we study the effect of using different ensembles in one - dimensional systems . in section four
we summarize and conclude .
in this section we will discuss bose - einstein condensation in three - dimensional harmonic traps .
firstly , we consider the isotropic potential and then we move on to the anisotropic trap .
however , we would first like to discuss bose - einstein condensation and phase transitions in a more general setting , and comment upon earlier work .
let us begin our discussion by considering free bosons in 3@xmath4 in the thermodynamic limit ( @xmath5 , @xmath6 finite )
. in this case one may show ( see any standard textbook in statistical mechanics ) that the fraction of particles in the ground state is given by @xmath7 where @xmath8 is defined through @xmath9 as can be seen from the above equation , @xmath8 defines the onset of the condensation of particles into the ground state .
moreover , it can be demonstrated that the chemical potential vanishes for @xmath10 and that @xmath11 for @xmath12 .
thus , one can _ define _ the critical temperature to be the highest temperature for which @xmath13 .
the derivative of @xmath14 with respect to temperature is discontinuous at @xmath15 , and so there is a first order phase transition .
+ + generally , the chemical potential must satisfy @xmath16 , where @xmath17 is the ground state energy of the system , in order to have positive occupation numbers . in @xcite
, kirsten and toms propose that a necessary criterion for bose - einstein condensation is that @xmath14 must reach the ground state energy @xmath17 at finite temperature . from this , they show that bec only occurs for free bosons when @xmath18 , when @xmath19 . ] and not for free bosons in a finite volume in three dimensions . in a system with a constant magnetic field
it only happens in five or more dimensions .
this latter result has also been verified independently by elmfors _
+ + from theoretical work as well as from an experimental point of view we will argue that this criterion is somewhat too strict . in ref .
@xcite it has been demonstrated numerically that the chemical potential never reaches its critical value in the case of bosons in a confining harmonic potential . according to the criterion mentioned above , this system does not exhibit bec .
however , the occupation number of the ground state changes considerably in a narrow temperature range , and attains a macroscopic value .
more specifically , it has been shown ( see e.g. @xcite ) that the condensate near @xmath20 varies as @xmath21 this abrupt change in number of particles in the ground state takes place near the temperature at which the heat capacity has a pronounced maximum ( which is @xmath22 ) .
we would not hesitate to call this a condensation , although there is no critical behaviour and so it is not obvious how to define a critical temperature .
however , in ref .
@xcite it has been proposed that the critical temperature is defined by the maximum of the specific heat . according to this criterion
there is a phase transition in two and three dimensions , but not in one ( see also section three ) .
we think is it natural to call the temperature @xmath22 the _ condensation temperature _ , since this temperature determines the onset of condensation of particles into the ground state .
this view is supported by experiment , namely the fact that one has measured a large increase in the number of particles in the ground state over a narrow temperature region .
the eigenfunctions of the three - dimensional harmonic oscillator hamiltonian can always be written as products of three eigenfunctions of the one - dimensional oscillator , which are hermite polynomials .
hence , the eigenfunctions can be labeled by three integers @xmath23 , @xmath24 and @xmath25 and the energy is @xmath26 in the isotropic case ( we have set the ground state energy to zero ) . each energy level can thus be characterized by a single quantum number @xmath27 and the corresponding degeneracy is @xmath28 , which simply is the number of ways of writing @xmath27 as the sum of the three positive integers @xmath23 , @xmath24 and @xmath25 .
the quantum number @xmath27 then takes the values @xmath29 . due to the high degree of symmetry of the isotropic harmonic
well , there is another convenient basis set .
the hamiltonian commutes with @xmath30 and @xmath31 and hence the eigenfunctions can be labeled by @xmath32 , @xmath33 and the energy .
the eigenfunctions are laguerre polynomials times a spherical harmonic .
+ + at temperature @xmath34 and chemical potential @xmath14 the particle number @xmath35 is given by @xmath36 where @xmath37 and @xmath38 is the usual bose - einstein distribution function . introducing the fugacity @xmath39
, the ground state particle number may be written as @xmath40 the number of particles in excited states is then @xmath41 which is a function of the _ effective fugacity _
@xmath42 where @xmath43 .
+ since the particle number is given , one must express the chemical potential as a function of temperature and particle number by inverting eq .
( [ free:1 ] ) in order to derive thermodynamic quantities describing the system . of particular interest to us is the particle number in the ground state as well as the specific heat of the system as a function of temperature .
now let us move on to actual calculations and consider different ways of handling eq .
( [ free:1 ] ) . at high temperature @xmath44 , the difference between the terms in the sum is small , and one can approximate the sum by an integral .
the dominating contribution at high @xmath34 is found by setting @xmath45 , leading to the semiclassical limit @xcite .
the transition temperature , as defined by the temperature at which @xmath46 , is in this approximation , found to be @xcite @xmath47 in the same approximation the specific heat @xmath48 exhibits a discontinuity at this temperature with the shape of a @xmath49 .
the discontinuity is however an artifact of the approximation .
an exact numerical calculation @xcite reveals that @xmath48 is a continuous function of @xmath34 .
this is a consequence of the finiteness of the particle number . as discussed in @xcite , an efficient way of performing summations
is given by the euler - maclaurin summation method @xcite .
the sum is then written as an integral plus an infinite series of correction terms @xmath50 + \frac{1}{12}[f'(b ) - f'(a ) ] + \cdots\end{aligned}\ ] ] applying this formula to the sum over excited states , the contributions from the upper limit vanish ( it is natural to single out the ground state from the sum and use euler - maclaurin on the rest . moreover
, the series converges much faster doing so ) .
higher order derivatives mainly bring down higher powers of @xmath51 , although differentiation of the degeneracy factor yields small corrections to lower order . at high temperature
we may therefore safely truncate the series . including only the first derivative
, the number of excited particles becomes @xmath52 we have here introduced the functions @xmath53 these functions satisfy a simple recursion relation @xmath54 they can be expressed in terms of the more familiar polylogarithmic functions @xmath55 which are generalizations of the riemann @xmath56-function . for @xmath57 and @xmath58
we have @xmath59 the @xmath60-functions are obtained when the lower limit of the integral in eq .
( [ lower ] ) is set to zero .
for @xmath61 they may be written in terms of more familiar functions , e.g. @xmath62 .
the polylogarithms satisfy the functional relation @xmath63 in terms of these functions , the total number of particles reads @xmath64 the internal energy as well as the heat capacity have been derived in a similar way in ref .
@xcite .
the potentials used in the experiments @xcite are either partially or completely asymmetric .
we therefore need expressions for the cases where the frequencies are different .
let us first consider the case @xmath65 . as for the isotropic potential we use the euler - maclaurin formula for each of the sums .
the calculations are similar and we skip the details .
the result is @xmath66 we have here introduced the notation @xmath67 etc . where @xmath68 and so on .
we also need the internal energy and the result using the euler - maclaurin formula is @xmath69 the specific heat is now found by differentiating @xmath70 with respect to @xmath34 . in doing
so we need @xmath71 which is found by differentiation of @xmath35 . + + in the @xmath2na experiment all three frequencies were different @xcite .
thus , we need the expression for the particle number in this case also . by applying the euler - maclaurin approximation
one finds that @xmath35 reads @xmath72\mbox{li}_1(\lambda_{xyz } ) \nonumber\\ & & + \frac{1}{8}\left[1 + \frac{1}{3}\left(\frac{b_x}{b_y } + \mbox{sym.}\right ) \right]\mbox{li}_0(\lambda_{xyz } ) + \frac{1}{b_xb_y}\mbox{li}_2(\lambda_{xy } ) \nonumber\\ & & + \frac{1}{2}\left(\frac{1}{b_x } + \frac{1}{b_y}\right)\mbox{li}_1(\lambda_{xy } ) + \frac{1}{4 } \left[1 + \frac{1}{3}\left(\frac{b_x}{b_y } + \frac{b_y}{b_x}\right)\right ] \mbox{li}_0(\lambda_{xy } ) + \mbox{sym.}\nonumber\\ & & + \frac{1}{b_x}\mbox{li}_1(\lambda_x ) + \frac{1}{2}\mbox{li}_0(\lambda_x ) + \mbox{sym . }
+ { \cal o}(\mbox{li}_{-1}),\end{aligned}\ ] ] here , `` sym . '' means symmetrization , e.g. @xmath73 the two leading terms at high temperature are @xmath74 in @xcite grossmann and holthaus considered a potential where @xmath75 and @xmath76 .
they constructed a continuous density of states based upon the two leading terms in the degeneracy . for the @xmath77-term they found a coefficient @xmath78 where @xmath79 .
a numerical summation gave @xmath80 .
this numerical factor is now easily found from eq .
( [ free : xx ] ) : @xmath81 which agrees quite well with their approximate result .
the same value for @xmath82 has been obtained independently in @xcite . + + for the internal energy one finds @xmath83 \mbox{li}_2(\lambda_{xyz})\nonumber\\ & & + \left[\frac{3}{4 } + \frac{b_x}{3b_y } + \frac{b_x^2}{12b_yb_z } + \mbox{sym.}\right]\mbox{li}_1(\lambda_{xyz})\nonumber\\ & & + \frac{2}{b_xb_y}\mbox{li}_3(\lambda_{xy } ) + \frac{3}{2}\left(\frac{1}{b_x } + \frac{1}{b_y}\right)\mbox{li}_2(\lambda_{xy } ) \nonumber\\ & & + \left(1 + \frac{b_x}{2b_y } + \frac{b_y}{2b_x}\right)\mbox{li}_1(\lambda_{xy } ) + \mbox{sym . }
+ \frac{1}{b_x}\mbox{li}_2(\lambda_x ) \nonumber\\ & & + \mbox{li}_1(\lambda_x ) + \mbox{sym . } + { \cal o}(\mbox{li}_0).\end{aligned}\ ] ] again the specific heat can be found by differentiation of this expression . with the formulas derived above at hand
, we can make predictions for the actual physical systems within the ideal gas approximation . in doing
so we have used the values for the oscillator frequencies and particle numbers as given in @xcite . for the completely anisotropic system ( @xmath2na ) the frequencies are 410 , 235 and 745 hz , and the particle number
@xmath35 is 700 000 .
the frequencies for the @xmath0rb experiment are @xmath84 hz and @xmath85 , and the particle number is 20 000 .
for these systems we have calculated both the condensate fraction and the specific heat using the euler - maclaurin approximation .
the results are compared with those of direct numerical summation and the semiclassical approximation .
the temperature is measured in units of the semiclassical transition temperature @xmath86 .
it is 1666 nk for the sodium gas and 127 nk for the rubidium gas .
1 we plot the condensate fraction as function of temperature .
the numerical and analytical results agree well . for @xmath87 700 000
the transition is quite sharp , and the condensate nearly vanishes about 1% below @xmath86 . for the smaller particle number ,
the transition is smoother . the specific heat is shown in fig .
we again find good agreement between numerical and analytical results .
the discontinuity in the semiclassical approximation is smoothened by the corrections .
still , for @xmath87 700 000 , the specific heat falls rapidly over a temperature range of less than 10 nk . for this case
it has a maximum value of @xmath88 , quite close to the semiclassical value @xmath89 .
all bec - experiments on trapped bose gases reported so far are performed on three - dimensional systems .
however , by the use of optical dipole traps or highly anisotropic magnetic traps one might , effectively , freeze out oscillations in one or two of the dimensions at low temperatures . in ref .
@xcite the authors suggest the use of a field configuration with a radial energy level spacing of 200 nk , and a much smaller axial spacing . at temperatures below 100
nk this system should effectively be one - dimensional . in the previous section we discussed in some detail different criteria for bec .
the same arguments apply here and we make a few comments . for a free gas in the thermodynamic limit , the chemical potential vanishes first at @xmath90 .
this implies that the critical temperature in one and two dimensions is zero and that the condensate vanishes at finite temperature . in a finite volume there
is again no critical behaviour and there is a condensate even at finite temperature .
this is also the case for an ideal gas trapped in an external harmonic potential , as we shall see below . in this section
we consider the trapped bose gas in both one and two dimensions using the same methods as for three dimensions .
we start by examining the case of a highly anisotropic three - dimensional system , and demonstrate the effective reduction to one dimension when two of the oscillator frequencies are much higher than the third . after this
, we briefly discuss the use of different ensembles in one dimension . as opposed to two and three dimensions it is here quite straightforward to use the canonical ensemble . for completeness
, we also give some results for the two - dimensional case .
excitations in a given direction of the trap will be suppressed when the temperature is below the energy scale set by the corresponding oscillator strength .
the highly anisotropic trap with @xmath91 should therefore effectively be one - dimensional at temperatures @xmath92 . in fig .
3 we have plotted the condensate fraction and specific heat @xmath48 for the case @xmath93 . in the same figure
we also give the results for a one - dimensional system with frequency @xmath94 .
+ the three - dimensional condensate fraction is seen to agree very well with the one - dimensional one at low @xmath34 , but at higher temperatures the former is somewhat reduced due to radial excitations .
these excitations have a more dramatic effect on the specific heat .
the two curves have high temperature limits of three and one , respectively .
the expressions for the particle number and the energy density in the one - dimensional case are given by : @xmath95 @xmath96.\end{aligned}\ ] ] this one - dimensional system has earlier been studied by ketterle and van druten @xcite . for @xmath97
, they obtained the equation @xmath98 the contribution from the excited states is similar to the first term in eq .
( [ our ] ) , but has a different effective fugacity . from eq .
( [ nn0 ] ) , ketterle and van druten have defined a transition temperature @xmath22 by setting @xmath99 and @xmath100 .
we then have @xmath101 . in fig .
4 we compare various approximations for the density , plotting the condensate fraction as function of temperature for @xmath102 . including all three terms in eq .
( [ eq : onedimn ] ) , the analytical and numerical results agree to three decimal places .
the heat capacity is a smooth function of temperature in the limit @xmath103 , and so there is no phase transition in one dimension .
the trapped bose gas has so far been described in terms of the grand canonical ensemble .
this implies an open system exchanging energy and particles with a reservoir , where the average particle number is given by the chemical potential .
the experiments are performed at fixed energy and particle number , and should therefore be described in terms of the microcanonical ensemble .
the derivation of the corresponding partition function is in general a difficult task , which involves complicated combinatorial problems .
using the canonical ensemble simplifies the calculations considerably , and still satisfies the constraint of fixed particle number .
the different ensembles are known to give the same predictions for thermodynamic quantities in the large @xmath35 limit @xcite .
below we compare the results for the grand canonical and canonical ensembles .
the results for the specific heat turn out to agree well .
however , for the condensate fraction we find significant deviations .
+ assuring that each configuration is counted only once , the partition function in the canonical ensemble can be written as @xmath104 here , @xmath105 .
the internal energy is now easily found to be @xmath106 and the specific heat becomes @xmath107 the derivation of the condensate fraction is a little more involved : it can be written as @xmath108 where @xmath109 is the probability of finding @xmath110 particles in the lowest energy level .
this probability is found summing all contributing configurations : @xmath111 in fig .
5 we have plotted the condensate fraction and the specific heat for @xmath112 and @xmath113 , as given by the two ensembles . the difference in specific heat is hardly visible for the lower particle number , and for @xmath113 the results of the two different ensembles agree to six decimal places .
the situation for the condensate fraction is very different .
one can show @xcite that the error made calculating @xmath114 in the grand canonical ensemble is , at most , of the order @xmath115 .
this is seen from the numerical results given in fig . 5 , where a significant deviation is found at intermediate temperatures even for @xmath116 .
this result is in contrast to the case of a free gas in three dimensions , where the deviation is known to go like @xmath117 @xcite .
a more complete discussion of the finite @xmath35 effects for this system in the canonical ensemble will be given elsewhere @xcite .
for completeness we also include results for an isotropic two - dimensional trapped gas .
the equation for the particle number here takes the form @xmath118 the expression for the internal energy reads @xmath119.\end{aligned}\ ] ] + + for large @xmath35 one may define an approximate condensation temperature setting @xmath120 and @xmath46 .
the leading term at high temperature then gives @xcite @xmath121 the limiting value at large @xmath35 for the condensate fraction is @xmath122 at temperatures just below @xmath86 and large @xmath35 the energy takes the value @xmath123 .
this gives a specific heat @xmath124 above @xmath86 the leading terms for @xmath48 are , again in the large @xmath35 limit @xmath125 setting @xmath126 , we see that the second term vanishes , and the specific heat approaches the same value as found above .
there is thus no jump in the specific heat even in the large @xmath35 limit in two dimensions . in fig .
6 we have plotted the condensate fraction for different values of @xmath35 .
the transition sharpens as @xmath35 grows .
the specific heat displayed in fig .
7 is seen to develop a sharp , though continuous peak with increasing @xmath35 .
+ this system behaves similarly to the one in three dimensions : the specific heat has a maximum , and the condensate has a powerlike fall off in the large @xmath35 limit ( @xmath127 in 2@xmath3 and @xmath128 in 3@xmath3 ) . thus , as opposed to the free case , a two - dimensional trapped bose gas undergoes a phase transition to a condensed state at _ finite _ temperature .
in the present paper we have discussed the thermodynamics of an ideal bose gas trapped in an anisotropic harmonic well in some detail .
we have investigated the condensate fraction and the heat capacity for realistic values of the frequencies by numerical as well as analytical methods .
there is an excellent agreement between the two methods in all dimensions .
moreover , we have studied the highly anisotropic trap where one of the frequencies is much smaller than the others .
this system should be realizable in future experiments and it is expected to be effectively one - dimensional for low temperatures .
we have therefore compared the condensate and the heat capacity for this system with the corresponding quantities in the one - dimensional trap .
the overall agreement for low temperatures confirms the above expectations .
we have also considered thermodynamic quantities in different ensembles at finite @xmath35 .
the most remarkable result is the deviation of the number of particles in the ground state .
finally , we would like to mention several possible improvements of the present treatment of bec .
the most obvious is the inclusion of interactions between the atoms in the gas .
this could either be based on perturbation theory or some nonperturbative method .
there are already a few articles on the subject @xcite , but more work needs to be done .
secondly , one should use the microcanocial ensemble in order to compare with experiment .
it is essentially a combinatorial challenge to do so .
more insight into finite @xmath35 effects of the ideal system is also of interest , and a detailed discussion will be presented in a follow up paper by one of the present authors @xcite .
davis , m .- o .
mewes , m.r .
andrews , n.j .
van druten , d.s .
durfee , d.m .
kurn , and w. ketterle , phys .
lett . * 75 * , 3969 ( 1995 ) .
m. h. anderson , j. r. ensher , m. r. matthews , c. e. wieman and e. a. cornell , science * 269 * , 198 ( 1995 ) . | we study the thermodynamic behaviour of an ideal gas of bosons trapped in a three - dimensional anisotropic harmonic oscillator potential . the condensate fraction as well as
the specific heat is calculated using the euler - maclaurin approximation . for a finite number of particles
there is no phase transition , but there is a well defined temperature at which the condensation starts .
we also consider condensation in lower dimensions , and for one - dimensional systems we discuss the dependence of the condensate fraction and heat capacity on the ensemble used .
= -15truemm = 165truemm = 220truemm * bose - einstein condensation in * * anisotropic harmonic traps * t. haugset , h. haugerud and j. o. andersen _ institute of physics _ _ university of oslo _ _
n-0316 oslo , norway _ |
the resistive plate chamber ( rpc ) @xcite muon detector of the compact muon solenoid ( cms ) experiment@xcite utilizes a gas recirculation system called closed loop ( cl ) @xcite , @xcite to cope with large gas mixture volumes and costs .
a systematic study of closed loop gas purifiers has been carried out in 2008 and 2009 at the isr experimental area of cern with the use of rpc chambers exposed to cosmic rays with currents monitoring and gas analysis sampling points .
goals of the study @xcite were to observe the release of contaminants in correlation with the dark current increase in rpc detectors , to measure the purifier lifetime , to observe the presence of pollutants and to study the regeneration procedure .
previous results had shown the presence of metallic contaminants , and an incomplete regeneration of purifiers @xcite,@xcite .
the basic function of the cms cl system is to mix and purify the gas components in the appropriate proportions and to distribute the mixture to the individual chambers .
the gas mixture used is 95.2% of c@xmath0h@xmath0f@xmath1 in its environmental - friendly version r137a , 4.5% of @xmath2c@xmath1h@xmath3 , and 0.3% sf@xmath4 to suppress streamers and operate in saturated avalanche mode .
gas mixture is humidified at the 45% rh ( relative humidity ) level typically to balance ambient humidity , which affects the resistivity of highly hygroscopic bakelite , and to improve efficiency at lower operating voltage .
the cl is operated with a fraction of fresh mixture continuously injected into the system .
baseline design fresh mixture fraction for cms is 2% , the test cl system was operated at 10% fresh mixture .
the fresh mixture fraction is the fraction of the total gas content continuously replaced in the cl system with fresh mixture .
the filter configuration is identical to the cms experiment .
in the cl system gas purity is guaranteed by a multistage purifier system : * the purifier-1 consisting of a cartridge filled with 5 ( 10% ) and 3 ( 90% ) type linde@xcite molecular sieve@xcite based on zeolite manufactured by zeochem@xcite ; * the purifier-2 consisting of a cartridge filled with 50% cu - zn filter type r12 manufactured by basf@xcite and 50% cu filter type r3 - 11 g manufactured by basf ; * the purifier-3 consisting of a cartridge filled with ni alo@xmath5 filter type 6525 manufactured by leuna@xcite . the experimental setup ( fig . [ fig : setup ] )
is composed of a cl system and an open mode gas system .
a detailed description of the cl , the experimental setup , and the filters studied can be found in @xcite .
the cl is composed of mixer , purifiers ( in the subunit called
filters in the fig .
[ fig : setup ] ) , recirculation pump and distribution to the rpc detectors .
eleven double - gap rpc detectors are installed , nine in cl and two in open mode .
each rpc detector has two gaps ( upstream and downstream ) whose gas lines are serially connected .
the the gas flows first in the upstream gap and then in the downstream gap .
the detectors are operated at a 9.2 kv power supply .
the anode dark current drawn because of the high bakelite resistivity is approximately 1 - 2 @xmath6a .
gas sampling points before and after each filter in the closed loop allow gas sampling for chemical and gaschromatograph analysis .
the system is located in a temperature and humidity controlled hut , with online monitoring of environmental parameters .
chemical analyses have been performed in order to study the dynamical behaviour of dark currents increase in the double - gap experimental setup and correlate to the presence of contaminants , measure lifetime of unused purifiers , and identify contaminant(s ) in correlation with the increase of currents . in the chemical analysis set - up ( fig .
[ fig : chem_setup ] ) the gas is sampled before and after each cl purifier , and bubbled into a set of pvc flasks .
the first flask is empty and acts as a buffer , the second and third flasks contain 250 ml solution of lioh ( 0.001 mole / l corresponding to 0.024
g / l , optimized to keep the ph of the solution at 11 ) .
the bubbling of gas mixture into the two flasks allows one to capture a wide range of elements that are likely to be released by the system , such as ca , na , k , cu , zn , cu , ni , f. at the end of each sampling line the flow is measured in order to have the total gas amount for the whole period of sampling .
( @xmath7 m ) have been installed upstream the flasks .
the sampling points ( fig .
[ fig : sampling_point ] ) are located before the whole filters unit at position hv61 , after purifier-1 ( zeolite ) at hv62 , after purifier-2 ( cu / zn filter ) at hv64 and after the ni filter at position hv66 .
rpc are very sensitive to environmental parameters ( atmospheric pressure , humidity , temperature ) , this study has been performed in environmentally controlled hut with pressure , temperature and relative humidity online monitoring . the comparison of temperature and humidity inside and outside the hut is displayed in fig .
[ fig : isrtemp ] and fig .
[ fig : isrrh ] , respectively , over the whole time range of the test .
the inside temperature shows a variation of less than @xmath8 ; the inside humidity still reveals seasonal structures between 35% and 50% , it is , however , much smaller than the variation outside .
gas mixture composition was monitored twice a day by gaschromatography , which also provided the amount of air contamination , stable over the entire data taking run and below 300 ( 100 ) ppm in closed ( open ) loop .
purifiers were operated with unused filter material .
the data - taking run was divided into cycles where different phenomena were expected .
we have four cycles ( fig .
[ fig : closedloop ] ) , i.e. , initial stable currents ( cycles 1 and 2 ) , at the onset of the raise of currents ( cycle 3 ) , in the full increase of currents ( cycle 4 ) .
cycle 4 was terminated in order not to damage permanently the rpc detectors .
the currents of all rpc detectors in open loop were found stable over the four cycles .
[ fig : closedloop ] shows the typical behaviour of one rpc detector in cl .
while the current of the downstream gap is stable throughout the run , the current of the upstream gap starts increasing after about seven months .
such behaviour is suggestive of the formation of contaminants in the cl which are retained in the upstream gap , thus causing its current to increase , and leaving the downstream gap undisturbed . while the production of f@xmath9 is constant during the run period , significant excess of k and ca is found in the gas mixture in cycles 3 and 4 .
the production of f@xmath9 is efficiently depressed by the zeolite purifier ( fig .
[ fig : fmeno ] ) . the observed excess of k and ca could be explained by a damaging effect of hf ( continuously produced by the system ) on the zeolite filter whose structure contains such elements .
further studies are in progress to confirm this model .
preliminary results show that the lifetime of purifiers using unused material is approximately seven months . contaminants ( k , ca ) are released in the gas in correlation with the dark currents increase .
the currents increase is observed only in the upstream gap .
the study suggests that contaminants produced in the system stop in the upstream gap and affect its noise behaviour , leaving the downstream gap undisturbed .
the presence of an excess production of k and ca in coincidence with the currents increase suggests a damaging effect of hf produced in the system on the framework of zeolites which is based on k and ca .
further studies are in progress to fully characterize the system over the four cycles from the physical and the chemical point of view .
the main goal is to better schedule the operation and maintenance of filters for the cms experiment , where for a safe and reliable operation the filter regeneration is presently performed several times per week .
a second run is being started with regenerated filter materials to measure their lifetime and confirm the observation of contaminants .
finally , studies in high - radiation environment at the cern gamma irradiation facility are being planned . the technical support of the cern gas group is gratefully acknowledged .
thanks are due to f. hahn for discussions , and to nadeesha m. wickramage , yasser assran for help in data taking shifts .
this research was supported in part by the italian istituto nazionale di fisica nucleare and ministero dell istruzione , universit e ricerca .
9 r. santonico and r. cardarelli , `` development of resistive plate counters , '' nucl .
instrum .
meth . * 187 * ( 1981 ) 377 .
cms collaboration , `` the cms experiment at the cern lhc '' , jinst * 3 * ( 2008 ) s08004 .
m. bosteels et al .
, `` cms gas system proposal '' , cms note 1999/018 . l. besset et al . , `` experimental tests with a standard closed loop gas circulation system '' , cms note 2000/040 .
m. abbrescia _ et al .
_ , `` proposal for a systematic study of the cern closed loop gas system used by the rpc muon detectors in cms '' , frascati preprint lnf-06/27(ir ) , available at http://www.lnf.infn.it/sis/preprint/
. g. saviano _ et al .
_ , `` materials studies for the rpc detector in cms '' , presented at the rpc07 conference , mumbai ( india ) , january 2008 . s.bianco _
, `` chemical analyses of materials used in the cms rpc muon detector '' , cms note 2010/006 . manufactured by zeochem , 8708 uetikon ( switzerland ) .
basf technical bulletin .
leuna data sheet september 9 , 2003 , catalyst kl6526-t .
grace davison molecular sieves data sheet .
linde technical bullettin .
l. benussi _ et al .
_ , `` sensitivity and environmental response of the cms rpc gas gain monitoring system , '' jinst * 4 * ( 2009 ) doi:10.1088 1748 - 0221 4 08 p08006 [ arxiv:0812.1710 [ physics.ins-det ] ]
. m. abbrescia _ et al .
_ , `` hf production in cms - resistive plate chambers , '' nucl .
* 158 * ( 2006 ) 30 .
nuphz,158,30 ; g. aielli _ et al .
_ , `` fluoride production in rpcs operated with f - compound gases '' , 8th workshop on resistive plate chambers and related detectors , seoul , korea , 10 - 12 oct 2005 . published in nucl.phys.proc.suppl .
* 158 * ( 2006 ) 143 . | the cms rpc muon detector utilizes a gas recirculation system called closed loop ( cl ) to cope with large gas mixture volumes and costs .
a systematic study of cl gas purifiers has been carried out over 400 days between july 2008 and august 2009 at cern in a low - radiation test area , with the use of rpc chambers with currents monitoring , and gas analysis sampling points .
the study aimed to fully clarify the presence of pollutants , the chemistry of purifiers used in the cl , and the regeneration procedure .
preliminary results on contaminants release and purifier characterization are reported .
rpc , cms , gas , purifier detectors hep muon |
the distribution of galaxies in the universe on large scales , as revealed by sky surveys such as the cfa redshift survey ( geller & huchra 1989 ) , the 2df galaxy redshift survey ( colless et al .
2001 ) or the sloan digital sky survey ( gott et al . 2005 ) is structured into a filamentary web in which the intersections correspond to the most massive objects ( i.e. galaxy clusters , bond et al .
this specific distribution is now commonly and easily reproduced by cosmological hydrodynamical n - body simulations of the @xmath0cdm cosmology which also indicate that the distribution of dark matter , the dominant mass component of the universe ( komatsu et al .
2011 ) , follows the same trend .
indeed , since in the traditional picture of galaxy formation , galaxies are supposed to form when baryonic gas falls into the gravitational potential of their host dark matter halo , the distribution of dark matter is therefore expected to faithfully trace that of the baryons .
however on galactic scales , observations tend to suggest that the spatial distributions of dark matter and baryons ( especially in the form of gas ) may display some substantial differences . in particular , it has been shown that galaxies are missing most of their baryons , most galaxies are severely baryons - depleted relative to the cosmological fraction ( see for instance bell et al .
2003 ; hoekstra et al . 2005 ; mcgaugh 2010 ) .
this so - called missing baryons problem " ( see bregman 2007 for a complete review ) , if _
real _ , calls for two alternative scenarios .
either a significant part of the gas never collapsed into the gravitational potential wells of protogalaxies in the first place , or some of the gas has been expelled by galaxy formation feedback processes such as supernova winds .
hence solving the missing baryon problem may prove to be central in order to constrain galaxy formation models .
past studies based on hydrodynamical simulations indicate that most of the `` missing '' baryons might lie in a gaseous phase ( the so - called warm - hot intergalactic medium , hereafter whim ) in the temperature range @xmath5 @xmath6 and at moderate over - density ( cen & ostriker 1999 ; dav et al .
2001 ; cen & ostriker 2006 ) .
according to these studies , the whim is primarily shock - heated during the formation of large - scale structures , while feedback mechanisms associated with star formation should also have an additional impact on this phase during galaxy formation .
for instance , for low mass systems , the cosmological uv - background is supposed to reduce both the star formation and baryon content ( thoul & weinberg 1996 ; bullock , kravtsov & weinberg 2000 ; gnedin 2000 ; somerville 2002 ; ricotti , gnedin & shull 2002 ; benson et al .
2002 ; read , pontzen & viel 2006 ; hoeft et al . 2006 ; okamoto , gao & theuns 2008 ; peirani 2010 ; nickerson et al .
2011 , hambrick et al .
supernova feedback is also expected to expel some gas , especially for low mass systems , as suggested by various numerical investigations ( scannapieco et al .
2008 , 2009 ; faucher - gigure , kere & ma 2011 and references therein ) .
other mechanisms have been proposed , such as turbulence in the baryonic intergalactic medium ( zhu , feng & fang 2011 ) or pre - heating by pre - virialisation ( mo et al . 2005 ) , although crain et al .
( 2007 ) claimed that this process is unable to prevent the collapse of gas ( instead , according to these authors , non - gravitational feedback is required in order to reduce the efficiency of gas cooling and star formation in dwarf galaxies ) .
another less investigated potential mechanism , which may also affect the evolution of the distribution of baryons on small scales , is the effect of accretion and merger events ( and in particular their associated tidal disruption ) that are expected to be frequent in the framework of the hierarchical model .
for instance , in an early paper , navarro & white ( 1993 ) have shown that during a merger involving dark matter and adiabatic gas , there is a transfer of energy between these two components that leads to a situation where the gas is less tightly bound .
more recently , sinha & holley - bockelmann ( 2010 , 2011 ) found that a few percent of gas can be driven into the intergalactic medium ( igm ) by galaxy mergers , using either merger trees calculated from the press - schechter formalism or idealized simulations of mergers of galaxy clusters . using hydrodynamical simulations , nickerson et al .
( 2011 ) have also concluded that tidal forces may cause significant mass loss from satellites of all masses . in the present work , we make use of cosmological `` zoom '' simulations with an extended treatment of the physics of baryons to study the formation of milky way - like galaxies .
our aim is to characterise the relative role of supernova feedback to accretion and mergers in the evolution of the distribution of baryons for objects of such masses .
we will also test scenarios in which a significant fraction of gas in progenitors is expelled at high redshift by more powerful sources of feedback , such as agn associated with massive black holes .
this will allow us to quantify two distinct processes which may allow us to address the so - called missing baryon problem , should it persist .
this paper is organised as follows . in section 2 , we summarize the numerical modelling , and section 3 presents our main results on the evolution of the distribution of baryons in our simulated local group universes .
we give our main conclusions in the last section .
the numerical methodology used in the present paper is described in detail in peirani ( 2010 ) to which we refer the reader for more information . for the sake of clarity ,
we summarize the main steps below .
we analyse three cosmological zoom simulations for a @xmath0cdm universe using wmap5 parameters ( komatsu et al .
2009 ) , namely @xmath7 , @xmath8 , @xmath9 , @xmath10 km / s / mpc , @xmath11 and @xmath12 .
each simulation was performed in a periodic box of side @xmath13 mpc with @xmath14 effective dark matter and gas particles in the highest resolution region ( a sphere of @xmath15 mpc of radius ) . in this region of interest , the mass resolution are @xmath16 and @xmath17 .
the plummer - equivalent force softening adopted for the high mass resolution particles were 1 and 0.5 @xmath18 kpc for dark matter and gas particles respectively and were kept constant in comoving units .
initial conditions have been generated from the mpgrafic code ( prunet et al . 2008 ) and the simulations were performed using gadget2 ( springel 2005 ) , with added prescriptions for metal - dependent cooling , star formation ( in this work we remind that the star formation efficiency is @xmath19=0.02 ) , feedback from type ia and ii supernovae ( sn ) , uv background ( starting at @xmath20 ) and metal enrichment .
the three simulations have common initial conditions whose phases are consistent with the local group but essentially differ in the quantity of energy released by sn derived from star particles @xmath21 ( @xmath22 ) .
as mentioned in peirani ( 2010 ) , we consider that a fraction @xmath23 of this energy is deposited in the @xmath24 neighbour gas particle by applying a radial kick to its velocity with a magnitude @xmath25 , where @xmath26 is the weighting based on the smoothing kernel and @xmath27 is the mass of gas particle @xmath28 .
the first simulation @xmath29 uses the standard value of @xmath30 while in the second one , @xmath31 , we have considered a higher efficiency @xmath32 in order to investigate how our results would be affected . in the third simulation ( @xmath33 )
, our aim is to study the effects of earlier high energy ejection to the ism induced either by intermediate mass black holes or other high energy processes such as hypernovae events .
intermediate mass black holes are the likely missing link between population iii and the supermassive black holes in quasars and plausibly as part of the hierarchy of structure formation .
hypernovae are likely to be more prevalent than normal supernovae in the earliest phases of star formation : for example , the population iii imf is most likely top - heavy , and this may also be the norm for the precursors of spheroids , as evidenced for example by the radial distribution of sne in disturbed ( including recently merged ) galaxies ( habergham , anderson & james 2010 ) and the frequency of low mass x - ray binaries in ultracompact dwarf galaxies ( dabringhausen et al .
. a higher frequency of hypernovae could plausibly provide an order - of - magnitude higher feedback efficiency than supernovae .
the case of agn is intriguing , as even higher feedback is required in deep potential wells , where supernovae are relatively inefficient , in order to account for the correlation between black hole mass and spheroid velocity dispersion ( silk & nusser 2010 ; debuhr et al .
moreover , the recent detection of a massive gas outflow in a distant quasar ( @xmath34 ) strongly suggests that a strong quasar activity is already in place at very early times ( maiolino et al .
2012 ) . for this purpose
, a simple modelling was used in which a much higher efficiency ( @xmath35 ) was considered during a very short ( @xmath36myr ) at earlier times ( @xmath37 ) and @xmath30 otherwise .
we also justify the choice of this high efficiency by simple energetics and momentum comparisons between supernovae and agn .
for instance , for a @xmath38m@xmath39 black hole along with its @xmath40m@xmath39 in stars , the number of type ii sne produced in star formation phase ( @xmath41m@xmath39 per sn for a chabrier imf ) is @xmath42 .
thus , the total sne energy injected is @xmath43ergs over a dynamical time , say @xmath44yr , or @xmath45ergs / s . since the eddington luminosity is @xmath46ergs / s energetics favour agn .
furthermore , for supernovae momentum conservation starts at shell velocity @xmath47km / s . since the ejection velocity is @xmath48km / s , the momentum injected is @xmath49 of initial momentum namely @xmath50gmcm / s or @xmath51dynes . for agn , momentum injected is @xmath52 or @xmath53dynes .
thus agn injects 100 times more momentum flux than supernovae .
finally , there is another factor of 10 - 100 that favours agn , the so - called mechanical advantage factor due to ram pressure on the expanding bubble as suggested by recent hydrodynamical simulations ( wagner , bicknell & umemura ) .
therefore the value of @xmath35 seems to be reasonable in order to account for the accumulation of the different potential sources of feedback at high redshifts .
the feedback parameters are summarized in table [ table1 ] .
.fractions @xmath23 of the energy released by supernovae [ cols="<,<,<,<",options="header " , ] [ table2 ] the time evolution of the virial , dark matter ( dm ) , gas and stellar masses of each object at the virial radius and for the three simulations is shown in figure[mass ] .
first , we notice that there is no particular difference between the evolution of dark matter halo masses ( and hence viral masses ) between @xmath29 and @xmath31 .
however , the baryonic compositions are quite different .
indeed , stellar masses are significantly reduced in @xmath31 due to higher amounts of energy released by the sn .
the mass of the gas component is therefore higher in @xmath31 , as expected . for @xmath33 ,
the stellar masses are even more reduced .
this is mainly due to the fact that a significant fraction of the gas has been expelled at high redshifts . indeed , while the dm masses are quite similar between the 3 runs , the virial masses for objects in @xmath33 tend to be lower at high redshift , which indicate that a fraction of gas expelled at high z does not recollapse at some later stage .
interestingly , @xmath54 undergoes a minor merger event at @xmath55 , which corresponds for instance to a sudden variation in the evolution of its stellar mass .
this same accretion event takes place at @xmath56 in @xmath33 .
moreover , we can also identify clearly in @xmath33 another minor merger event at @xmath57 that has not occurred ( yet ? ) in @xmath29 and @xmath31 .
this explains why the dark matter mass ( and therefore the virial mass ) of @xmath54 is higher in @xmath33 than @xmath29 or @xmath31 .
this also suggests that small perturbations in the properties of haloes at earlier times can have some significant impact on the whole accretion history ( see for instance , thiebaut et al .
it is also worth mentioning that although the two main central objects have the same virial mass at @xmath1 , they have undergone different mass accretion histories . indeed , while @xmath54 grows regularly through smooth mass accretion , the evolution of the mass of @xmath58 is more sporadic thanks to several major merger events .
it is therefore instructive to see to which extent these two different mass evolutions could affect the evolution of the distribution of baryons in the vicinity of these two objects .
it is also worth studying how the stellar masses found in the different simulations are related to the mass of their host dark matter haloes . from table
[ table2 ] , the mass of dark matter haloes varies from @xmath59 to @xmath60 while according to observational and numerical analysis , the expected stellar mass range inside those haloes is @xmath61\times10^{10 } m_\odot$ ] ( guo et al .
2010 ; moster et al . 2010 ; brook et al .
in order to make a suitable comparison , instead of considering the stellar mass inside the virial radius @xmath3 , we define the galaxy radius as @xmath62 and all stars inside this radius contribute to the galaxy stellar mass @xmath63 . by doing this ,
we only select stars which belong to the central main galaxy while excluding stars of accreted satellites inside @xmath3 . in figure[mass ] , the evolution of each stellar mass inside @xmath64 is shown in dashed line .
one can see that at @xmath1 , galaxy stellar mass ranges are @xmath65\times10^{11 } m_\odot$ ] , @xmath66\times10^{11 } m_\odot$ ] and @xmath67\times10^{11 } m_\odot$ ] for @xmath29 , @xmath31 and @xmath33 respectively .
thus , each simulation tends to produce too many stars in the central region of haloes and the trend is more severe in @xmath29 and @xmath31 relative to @xmath33 .
this problem was already pointed out from other hydrodynamical simulations which indicate that low feedbacks may result in more than an order - of - magnitude too many stars ( see for instance piontek & steinmetz 2011 ; brook et al .
2012b ) . as a first conclusion , stronger and earlier feedbacks seem to be required in order to improve the stellar mass - halo mass relation (
also recently suggested by stinson et al .
2012 ) and , in the present study , one probably also needs the action of additional sources of feedback in the late phase of galaxy formation such as agn to get even closer to the expected galaxy stellar masses .
moreover the fact that our model leads to too high central stellar masses suggests that the effect of all different sources of feedback has been underestimated .
thus in the following , baryonic fraction values found in our simulations may be slightly overestimated .
finally , figure[fig_sfr ] shows the evolution of the star formation rate ( sfr ) for each object and simulation . at @xmath1 ,
sfr values are around @xmath68 $ ] @xmath69yr which is consistent with observational values derived by robitaille & whitney ( 2010 ) using spitzer data .
note that these cosmic evolutions are similar for the two haloes : at high redshifts ( @xmath70 ) the lower the sn efficiency the higher the stellar mass . and the higher the mass of formed stars at high redshift , the higher the subsequent sn feedback .
this explains why , at low redshifts , a higher fraction of stars is produced in @xmath33 relative to @xmath29 and @xmath31 .
fig.[distribution1 ] shows the projected distribution of baryons in our lg type universes derived from @xmath31 and @xmath33 and at four specific epochs ( i.e. @xmath71 and @xmath72 ) .
we clearly see that the evolution of the distributions of baryons and dark matter do not follow the same trend . indeed , the regions of the universe in red and dark blue correspond to regions where the baryonic fraction is lower than the universal value . and
those specific regions are mainly located around the forming protogalaxies or galaxies at each redshift , while in the very inner part of dark matter haloes , @xmath2 is higher .
this is an expected result : due to cooling , the gas collapses to the center of haloes where stars can be formed .
however , note that the size of these `` red '' cavities increases over time which suggests that the fraction of gas that has collapsed to the center of the halo is not immediately replaced by some fresh gas from its vicinity . note also , at high and low redshift , the existence of relative high baryon content regions which are located either in the filaments or beyond the virial radii .
such anisotropic distribution seems to be more pronounced in @xmath33 relative to @xmath31 . if the high baryon fractions in filaments can be understood by the dissipative nature of gas , allowing it to cool to the dense filaments , it is crucial to characterise the mechanisms that drive high baryon content regions beyond the virial radii
. 0.5 cm let us first only consider the role of supernovae feedback in the evolution of the mass budget of milky way - type galaxies .
we therefore focus on both @xmath29 and @xmath31 in this sub - section .
the cosmic evolution of the baryonic fraction @xmath2 estimated at the virial radius for the two main haloes is shown in figure[fb_evolution ] . in each of these two simulations
, they follow the same trend . at high redshift
, @xmath2 is close to the universal value and sometimes is slightly higher .
cold flows provides gas to form stars .
note that the virial radius is estimated according to the local density , which is dominated by the newly formed stars .
since the resolution in these simulations is limited , especially for low mass systems , the virial radius may be underestimated , and subsequently this baryonic fraction may be in fact overestimated at high redshifts . from @xmath73
, @xmath2 is decreasing with cosmic time until it reaches the values of @xmath74 and @xmath75 at @xmath1 for @xmath58 and @xmath54 respectively .
these latter values are quite close to the universal value which is not surprising since similar trends have been already found from other hydrodynamical simulations using weak or no supernovae feedback , for objects with a mass of @xmath76 at @xmath1 ( crain et al . 2007 ;
faucher - gigure , kere & ma 2011 ) .
more interestingly , no particular differences are seen in the evolution of @xmath2 between @xmath29 and @xmath31 .
this suggests that higher sn feedback can reduce the star formation rate ( and therefore the final stellar mass ) but is rather inefficient in expelling the gas outside the virial radius at high and low redshifts for massive galaxies .
this results are in agreement with stinson et al .
( 2011 ) who show that increased feedback in haloes of such mass scale affects the star formation more than baryon content in the circumgalactic medium ( cgm ) .
moreover , they show that higher feedback models can account for sufficient ovi and hi gas in the cgm compatible with the observed distributions . therefore ,
in the present scenario , the relatively small differences in the baryon fraction between @xmath77 and @xmath78 could make supernovae feedback prescriptions difficult to distinguish and one probably needs to focus on stellar mass and/or properties of the intergalactic medium in order to have better constraints ( see also dav , oppenheimer & sivanandam 2008 ; shen , wadsley & stinson 2010 ; stinson et al . 2011,2012 ; scannapieco et al .
2012 ; brook et al .
in particular , it seems that higher feedback of @xmath78 compared to @xmath77 may be necessary in reproducing values closer to the correct stellar masses and properties of intergalactic medium as we will see in section 3.3 .
we have also studied the distribution of the integral @xmath79 and differential @xmath80 as respect to the radius @xmath81 at three different redshifts @xmath82 , @xmath83 and @xmath72 in figure[fb_dis ] . only results from @xmath54 are shown since similar trends are obtained for @xmath58 . at the center of each halo
, @xmath2 is high due to the presence of the galaxy . at larger radii , @xmath84
, @xmath79 tends towards the universal value as expected .
interestingly , for high redshifts ( @xmath85 ) , @xmath79 converges to the universal value from above , while this is not the case at lower redshifts , which suggest , as mentioned above , that the gas infall is not recurrently replaced by some fresh gas from the outskirt of the haloes .
also , an excess of baryons at @xmath86 can be clearly seen in the variations of the differential @xmath80 for different cosmic times , which correspond to regions in yellow in fig.[distribution1 ] around haloes .
similar plots were recently derived from observations using chandra and suzaku facilities for an isolated elliptical galaxy with a @xmath87 milky way mass ( humphrey et al .
2011 ) and a fossil group ( humphrey et al .
2012 ) . in @xmath33
however , due to earlier and important eviction of gas from other potential sources of feedback ( agn , hypernovae , ... ) @xmath2 reaches its lowest value @xmath88 at @xmath89 , then increases until @xmath90 and becomes nearly constant ( @xmath4 ) until the present time .
this strongly suggests that sources of feedback acting at high redshift , even for a very short period , can have a stronger impact on the final mass budget of massive galaxies than those acting in the later and longer phases of galaxy formation .
one can also notice from figure[fig_sfr ] that the gas outflow taking place at high redshift lead to the quenching of the star formation during a few million years ( @xmath91 ) .
this is in nice agreement with conclusions of valiante , schneider & maiolino ( 2012 ) using semi - analytical models of the formation and evolution of high - redshift quasars . indeed , according to their model , the observed outflow of a distant quasar at @xmath34 is dominated by quasar feedback which _ `` has considerably depleted the gas content of the host galaxy , leading to a down - turn in the star formation rate at @xmath92''_. nevertheless , future observational proofs of star formation quenching at these redshifts are required in order to confirm such statements .
the time evolution of the baryonic fraction inside the virial radius depends on the competition between the evolution of the mass that arrives within @xmath3 , and the mass that leaves @xmath3 .
it raises the following question : are the low values of @xmath2 found at @xmath1 due to the fact that more baryons have been expelled out of the virial radius or because more dark matter has been accreted inside @xmath3 ? to answer this question , instead of studying the evolution of the mass that enter or leave the virial radius at each subsequent step , we focused on the location at a given redshift @xmath93 of all particles that have been accreted up to this specific redshift . to do this
, we have first identified all the particles that have been accreted at each snapshot of the simulation from @xmath94 to `` @xmath93 '' .
then , we have computed the fraction of those particles that are inside or outside @xmath3 at the considered redshift . by doing this
, the interpretation of our findings will be easier since each accreted particle contributes only once in the final result .
indeed , certain particles can be accreted , ejected and accreted again during the whole evolution of these haloes .
figure[mass_acc2 ] shows the evolution of the total accreted dm and baryon mass from @xmath94 to a given redshift , @xmath93 , for both haloes , and for both @xmath31 and @xmath33 ( results derived from @xmath29 are not shown since they are similar to those from @xmath31 ) .
we also plot the fraction of these accreted particles that are located inside or outside @xmath3 at this given redshift @xmath93 .
first , we look at the evolution relative to @xmath58 .
one can see that a higher fraction of dark matter mass has been accreted from @xmath95 to @xmath1 relative to the baryon component ( solid lines ) .
moreover , a similar fraction of dark matter and baryons are located outside the virial radius at @xmath1 .
that means that the fraction of baryons released during the formation of @xmath58 is not higher than that of dark matter , which demonstrate that the low value of @xmath2 at @xmath1 follows from the fact that more dark matter particles has been accreted over cosmic time .
this can be understood given the difference in nature between the dissipative gas and the non - dissipative dark matter .
in particular , the gas component can be shock - heated at the virial radius which considerably affects and slows down its further accretion on to the halo as already shown in hydrodynamical simulations using no feedback ( see for instance oorbe et al .
in contrast , especially after an important episode of accretion or merger event , the dark matter component can `` damp out '' its orbital motion until it completely relaxes .
for @xmath54 the same trends are observed but the mass of baryons outside the virial radius turns out to be higher that that of dark matter .
however this does not necessarily mean that a higher fraction of baryons has been expelled through physical mechanisms such as feedback or merger events .
indeed , if we represent the fraction of baryons and dm released from @xmath96 to @xmath1 ( see figure[mass_acc2 ] ) , one can see that same amounts of mass have been expelled at @xmath1 by the two components .
it implies in fact that a significant fraction of dark matter which has been expelled between @xmath95 and @xmath96 has been re - accreted between @xmath96 and @xmath1 .
this is not the case for baryons which are generally heated through the expelling process and therefore remain outside the virial radius afterwards .
in order to illustrate this , figure[illustration ] presents the evolution of the distribution of baryons and dark matter for 1 . from @xmath97 ,
the main progenitor undergoes two mergers with mass ratios of 1:1.8 and 1:5 .
here , we have selected all particles inside the virial radius of each of the 3 objects .
note that the orbits of the two mergers are quite radial .
after the two merger events , the spatial distributions of baryons and dark matter are quite different . most of the dark matter particles are located inside the virial radius , though a small part has been expelled in the same orbital direction .
the baryons tends to either collapse toward the center or are expelled all around .
this phenomenon is observed at @xmath1 where a non - negligible part of baryons have been expelled and remain outside the virial radius .
figure[illustration ] also shows the evolution of the mass of dm and baryons located outside the virial radius . in the case of the dark matter component , there are 3 distinct phases .
first , from the beginning of the experiment until @xmath98 gyr , both satellites are accreted inside @xmath3 . then in the second phase ( from @xmath98 to @xmath99 gyr ) some of the dark matter is expelled via tidal disruption .
finally from @xmath99 gyr some of the dark matter particles are re - accreted .
note that evolution from @xmath100 gyr is probably due to the effect of another merger event .
however , the baryon component follows the two first phases but not the third one : once baryons are expelled , most of them remain outside the virial radius . in this specific example
, we found that 2.1% and 5.8% of dark matter and baryon masses respectively are released at @xmath1 from the three main objects .
this is in good agreement with results obtained from idealized simulations ( sinha & holley - bockelmann 2009 ) .
it is also interesting to see that due to mergers , unbound gas diffuses far away from the galaxy ; thus mergers represent a potential mechanism for enriching the igm with metals .
note that this enrichment must occur early , as in our model , to avoid disruption of the cold and weakly enriched lyman alpha forest .
it is also worth mentioning that during a merger event , gas can be expelled via both tidal disruption and outflows from feedback due to the merger driven starburst .
however , since there is no significant difference in the evolution of @xmath2 within the virial radius between @xmath77 and @xmath78 , this suggests that supernovae feedbacks have a lower contribution relative to tidal disruption for milky way mass galaxies . as observed in the evolution of the dark matter component
, tidal disruption should guarantee escape whereas a starburst driven outflow is much more problematic ( see for instance powell , slyz and devriendt 2011 ) .
finally , it is interesting to point out that @xmath58 undergoes major merger events between @xmath101 and @xmath102 , leading to important mass outflows from @xmath102 and @xmath1 .
in particular , figure[mass_acc2 ] indicates that more dark matter particles are released than baryons after the merger events .
most of this dark matter is not yet re - accreted , which explains why @xmath58 tends to have values of @xmath2 higher than @xmath54 at @xmath1 .
this also explains why in figure[distribution1 ] the high @xmath2 value regions around @xmath58 seem to be less pronounced in comparison to those around @xmath54 .
thus , the evolution of the mean baryonic fraction value @xmath2 at the virial radius is essentially governed by the relative efficiency at which the dark matter and baryons are accreted .
indeed , the evolution of the accreted mass of both dark matter and baryons follows two different regimes : a rapid growth at high redshifts ( @xmath103 ) and a slower one at lower redshifts .
these evolutions are directly related to the total mass evolutions in figure[mass_acc2 ] which also presents these two regimes .
this behaviour compares well to the evolution of individual dark matter haloes ( see for instance wechsler et al .
2002 ; vitvitska et al .
2002 , peirani , mohayaee & pacheco 2004 , mcbride et al .
2009 and tillson , miller & devriendt 2011 ) , indicating a fast growth of the mass for @xmath104 , followed by a phase where the accretion rate is slower .
hydrodynamical simulations have also suggested that galaxy formation presents two phases : a rapid early phase ( @xmath103 ) during which stars mainly form within the galaxy from infalling cold gas ( katz et al .
2003 ; kere et al . 2005 ; ocvirk , pichon & teyssier 2008 ; dekel et al .
2009 ) and a second one ( @xmath105 ) during which stars are accreted ( oser et al . 2010 ; hirschmann et al .
this latter phase is dominated by the accretion of diffuse matter and small satellites ( fakhouri & ma 2010 ; genel et al . 2010 ; wang et al . 2011
; van de voort et al . 2011 ) .
hence the high @xmath2 values at high redshift derived from @xmath29 and @xmath31 can be explained by the accretion of high gas - to - dark matter ratio from dense region such as filament via cold flows .
but at lower redshift , @xmath2 is decreasing due to the accretion of low gas - to - dark matter ratio material , in particular from the diffuse region .
indeed , we found that the baryonic fraction of the diffuse accreted matter ( namely @xmath106 ) is on the average much lower than the universal value : as the haloes become more massive , the temperature at the virial radius increases and the gas is shock - heated and this process slows down its accretion onto the halo . for @xmath33
, @xmath2 reaches its lowest value ( @xmath88 ) at @xmath107 right after some significant expulsion of gas .
after that , there is a short phase where @xmath2 increases because the accreted matter has a higher baryonic content ( but not necessarily higher than the universal value ) .
then , in the late phase of evolution of each halo , @xmath2 tends to be constant ( @xmath4 ) because the average accreted matter tends to have the same baryonic fraction value as the mean value inside virial radii .
the whim has proven to be difficult to detect and represents a serious candidate for the missing baryons .
it is therefore instructive to estimate the fraction of the whim in our simulated local universes . for this purpose
, we have divided the inter - galactic medium ( igm ) into four components : ( 1 ) stars , ( 2 ) cold gas ( @xmath108 ) , ( 3 ) whim ( @xmath109 ) and hot gas ( @xmath110 ) .
figure[whim ] shows the time evolution of three baryon components ( stars , cold gas and whim ) within a sphere centred on the local universe - type haloes and of comoving radius 1.5 mpc / h .
the evolution of the hot gas ( @xmath111k ) is not shown in figure[whim ] since this component represents less than @xmath112 of the total baryon mass budget in our selected region , as expected since we are not in a cluster environment .
it shows that the whim represents @xmath113% of the baryon budget at @xmath1 , in good agreement with past work based on hydrodynamical simulations ( cen & ostriker 1999 ; dav et al .
2001 ; cen & ostriker 2006 ; rasera & teyssier 2006 ; tornatore et al .
2010 ) when agn feedback is not taken into account .
we also found that the warm and hot gas becomes the dominant baryonic component at large radii ( @xmath114 kpc ) from each main halo , in good agreement with previous observational analysis of massive galaxies ( humphrey et al .
2011 , 2012 ) or with previous numerical simulations ( van de voort et al . 2011
; van de voort & schaye 2012 ) .
however , when the contribution of the whim is not taken into account in the estimation of the baryonic fraction at the virial radius , namely when only the cold gas ( @xmath115k ) and stars are considered , @xmath2 values obtained at @xmath1 only decrease by @xmath116% .
this follows from the fact that stellar masses obtained in our experiments have a large contribution in the final @xmath2 values and thus we probably need the additional effect of agn feedback to regulate the sfr and also increase the mass of whim inside the virial radius .
nevertheless , the distribution of baryonic fraction seen in figure[distribution1 ] is significantly affected when the whim is not taken into account .
indeed , figure[whim2 ] shows that in this latter case , the baryonic fraction is lower than @xmath117 in the direct vicinity of the virial radius . outside the virial radii , our analysis in the previous section has suggested that the igm is composed of gas that either never collapses onto haloes or is released by mechanisms such as feedback or merger events .
but what are the relative proportions ? from figure[mass_acc2 ] we can estimate that @xmath118 of the total accreted baryons is expelled from @xmath95 to @xmath1 for @xmath58 and @xmath54 , which is quite significant .
these results are also in good agreement with sinha & holley - bockelmann ( 2010 ) who found from semi - analytic techniques that up to @xmath87 25% of gas can be unbind over the course of galaxy assembly . from our analysis
, we found that the gas that has been released after being accreted on to haloes is mainly located between @xmath3 and @xmath119 at @xmath1 .
if we focus on @xmath58 which is quite isolated , we found that the gas component between @xmath3 and @xmath119 consists of 55.9% , 58.6% and 60.5% of gas that has been expelled from @xmath58 from @xmath29 , @xmath31 and @xmath33 respectively .
in other words @xmath87 40% of this gas has never collapsed onto @xmath58 .
it is also worth mentioning that @xmath120 of the gas that has been released through merger events or feedback is in the form of whim at @xmath1 but only @xmath87 65% in the case the gas that has never collapsed into the halo .
hence , the gas that has been ejected through different mechanisms during the galaxy formation and evolution significantly increases the mass of the whim .
in this paper , we have analysed hydrodynamical simulations with an extensive treatment of the physics of baryons in order to study the evolution of the distribution of baryons in a local group - type object . in particular , we have investigated the impact of i ) supernova feedback , ii ) more powerful sources of feedback at high redshift such as agn from intermediate mass black holes or hypernovae and iii ) accretion / ejection and merger events in the evolution of the mass content of milky - way type galaxies .
first , when no high early feedback is considered as is the case in @xmath77 and @xmath78 , we found that the evolution of the mean baryonic fraction value @xmath2 estimated at the virial radius of the two main central objects is essentially governed by the relative efficiency at which dark matter and baryons are accreted . indeed , in the first and early phase of galaxy formation ( the cold mode of accretion @xmath103 ) during which stars mainly form within the galaxy from infalling cold gas ( katz et al .
2003 ; kere et al .
2005 ; ocvirk , pichon & teyssier 2008 ; dekel et al . 2009 ) @xmath2 is in general higher than the universal value .
but in the second phase ( the hot mode of accretion ) , dominated by the accretion of diffuse matter or small satellites , @xmath2 tends to decrease and is 10 - 15% lower than the universal value , 0.166 , at @xmath1 .
this decrease is not due to the fact that more baryons ( relative to dm ) are expelled beyond the virial radius by physical mechanisms such as feedback or merger events .
rather this essentially follows from the relative nature of dissipative gas and non - dissipative dark matter .
specifically , the gas component can be shock - heated at the virial radius , which considerably slows down its further accretion into the halo ( see for instance oorbe et al .
however , shock - heating will presumably not be a large factor in low mass galaxies , but as mentioned in the introduction , feedbacks such as uv - background and supernovae winds are supposed to play a major role in the effective reduction of the baryon content of these galaxies . additionally , we found that accretion and particularly merger events can release a fraction of both dark matter and baryons . but again , dark matter that has been expelled can be re - accreted more efficiently than the corresponding baryons , hereby enhancing the decrease of @xmath2 inside the virial radius .
consequently , our simulations suggest that the gas located beyond the virial radius is mainly in the form of whim that comes from either gas that never collapsed prior to the formation of the proto - galaxy ( @xmath121 ) or gas expelled by feedback and merger processes ( @xmath122 ) .
our comparison of 1a and 1b also suggest that globally increasing somewhat the efficiency of supernovae does not significantly affect the evolution of @xmath2 inside the virial radius . in other words ,
sn feedback is not powerful enough to expel the baryons out of the gravitational potential of dark matter haloes of such mass scales .
this also means that , if future observations indicate that @xmath2 inside the virial radius of milky way mass galaxy at @xmath1 is relatively close to the universal value , then in order to make clear a distinction between the different feedback models , the stellar mass and the properties of the intergalactic medium may provide better constraints than baryon fractions , as suggested by recent numerical works ( dav , oppenheimer & sivanandam 2008 ; shen , wadsley & stinson 2010 ; stinson et al . 2011,2012
; scannapieco et al . 2012 ; brook et al .
but at high redshift the situation is quite different since haloes are less massive and thus feedback mechanisms have a more effective contribution .
indeed for our other scenario ( for which a higher fraction of gas is expelled at high redshift ( @xmath94 ) by more powerful source of feedback such as agn from intermediate mass black holes or hypernovae implemented in 2 ) the accretion of gas is strongly affected at high redshift , and drives the mean baryonic value inside the virial radius to a lower values at @xmath1 .
this is one important conclusion of this paper : to get a low value of @xmath2 at present time , physical mechanisms able to expel the gas _ at high redshift _ will have a stronger impact on the deficit of baryons in the mass budget of milky way type - galaxy today than those that expel the gas in the longer , late phases of galaxy formation .
thus , if future observations favour a relatively low value of @xmath2 at @xmath1 , then this high redshift feedback " scenario is the most probable .
however , it this worth mentioning that the effective contribution of non - gravitational heating such as galactic winds from supernovae in the evolution of @xmath2 within the virial radius of milky way type galaxy is still a controversial and a debated topic in the recent literature
. from our modelling and experiments , we found that increasing the efficiency of supernova feedback leads to the reduction of the stellar mass but has no particular impact on the evolution of @xmath123 for system of such mass scales ( cf powell , slyz & devriendt 2011 ) .
such statements were also advocated by anderson & bregman ( 2010 ) from observational constraints on the density of hot gas around the milky way . moreover , using cosmological simulations , faucher - gigure , kere & ma ( 2011 ) have investigated the net baryonic accretion rates through @xmath124 .
although the net baryonic accretion rate is sensitive to galactic outflows , especially for low mass systems , the baryon mass fractions at @xmath1 for objects with a mass higher than @xmath125 and for different sn feedback prescriptions have similar values to those derived in the present study . however , other hydrodynamical simulations from scannapieco et al .
( 2008 , 2009 ) using higher sfr efficiency ( @xmath126 ) found that at @xmath1 , the baryon fractions inside the virial radius are @xmath127 within the virial radius _
`` indicating that a significant amount of baryons has been lost through winds''_. agn feedback is supposed to play an important role in the formation of massive objects such as giant ellipticals in galaxy groups ( tabor & binney 1993 ; ciotti & ostriker 1997 ; silk & rees 1998 ) . from previous numerical studies , the inclusion of agn
is expected to reduce the stellar mass ( see for instance sijacki et al . 2007 ; dubois et al . 2010 , 2012 ;
hambrick et al .
2011b ) and also increase the mass of whim ( cen & ostriker 2005 ; tornatore et al .
2010 ; durier & pacheco 2011 ; roncarelli et al .
our study does not include the effect of agn feedback at low redshift , but we think that such sporadic events will not have a major contribution in the baryonic fraction value inside the virial radius at present time .
indeed , from observationally motivated constraints , agn and sn at low to moderate redshift seem not to produce the expected correlations with the baryonic tully - fisher relationship ( anderson & bregman 2010 ) .
moreover , as mentioned above , observations of massive objects reported by humphrey et al .
( 2011 , 2012 ) seem to suggest that agn feedback has no particular impact on the baryonic fraction values within the virial radius .
also , recent hydrodynamical simulations show that agn and sn feedback may mutually weaken one another s effect by up to an order of magnitude in haloes in the mass range of @xmath128 ( booth & schaye 2012 ) . at high redshift ,
the effects of galactic winds on the igm are also controversial . in the present paper
, our simple minded modelling indicates that galactic winds can expel some gas outside the virial radius and affect significantly the evolution of the baryon content of massive galaxies .
similar conclusions have been recently drawn by roncarelli et al .
( 2012 ) from hydrodynamical simulations suggesting that galactic winds acting at earlier epochs can prevent the igm to collapse in dense structures .
also , as mentioned above , semi - analytical model from valiante , schneider & maiolino ( 2012 ) indicates that the observed outflow of a distant quasar at @xmath34 ( maiolino et al .
2012 ) is expected to have considerably depleted the gas content of the host galaxy .
however , simulations conducted by di matteo et al .
( 2011 ) or shen et al .
( 2012 ) found that galactic outflows at high redshift may not be very effective at stopping the cold gas from penetrating the central regions . in summary
, our study indicates that in order to reach lower @xmath2 values at @xmath1 for milky way type galaxies , the eviction of cold gas by feedbacks during the first phase of galaxy formation at high redshifts proves to be crucial .
our study also suggests that these high redshift feedback mechanisms more efficiently reduce galaxy stellar mass in the central part of haloes and thus show better agreement to galaxy observable such as the stellar mass - halo mass relation .
moreover , efficient high redshift feedback processes seem also to be required in order to slow down the growth of galaxies at high redshift and thus to reproduce the observed number density evolution as recently pointed out by weinmann et al .
the presence of strong early feedbacks is also motivated in disk galaxy formation theory .
indeed , it has been suggested that early outflows could eject low angular momentum gas allowing disk galaxies to form ( binney , gerhard and silk ( 2001 ) .
such hypothesis was confirmed by hydrodynamical simulations which show that outflows can indeed remove preferentially low angular momentum material which can be re - accreted latter with higher angular momentum and forms a disk ( brook et al .
2011 , 2012a ) .
if such statements are correct , numerical and observational efforts have to be focused towards characterising the respective role of each feedback process on the igm at high redshift .
in particular , are the baryons expelled from protogalaxies or do they never collapse into the gravitational potential well of these objects ? in the present paper , we only focus on the first mechanism , but it is not excluded that processes such as photoionization play a major role especially at high redshifts for low mass systems . in this regard ,
it is interesting to compare our results with those of guedes et al .
( 2011 ) who have studied the formation of late - type spirals from the eris simulation .
specifically , they claimed that the collapse of baryons is heavily suppressed at high redshift by the uv background . and
it is particularly instructive to see that from their highest mass resolution simulation , the evolution of the baryonic fraction ( see their figure 6 in guedes et al .
2011 ) is very similar to those we obtained from @xmath33 .
they also found a similar value at @xmath1 namely @xmath129 .
our simulations do include uv background but our mass resolution may be too small to account for this process properly .
more detailed and well - resolved hydrodynamical simulations are required to investigate the evolution of the mass content of galaxies at high and low redshifts .
but in order to achieve this goal , one must first improve the constraints of the expected energy injection to the igm from sn , agn , uv background ( etc ... ) via observations .
. acknowledges support from the global internship program by the national research foundation of korea ( nrf ) .
we warmly thank the referee for an insightful review that considerably improved the quality of the original manuscript .
we warmly thank d.le borgne , j.devriendt , y.dubois , r.gavazzi , t.kimm , g.mamon , j.a.de freitas pacheco and t.sousbie for interesting discussions .
we also thank d.munro for freely distributing his yorick programming language ( available at ` http://yorick.sourceforge.net/ ` ) which was used during the course of this work .
this work was carried within the framework of the horizon project ( ` http://www.projet-horizon.fr ` ) .
anderson , m. e. , & bregman , j. n. 2010 , apj , 714 , 320 bell , e. f. , mcintosh , d. h. , katz , n. , & weinberg , m. d. 2003 , apjs , 149 , 289 benson , a. j. , frenk , c. s. , lacey , c. g. , baugh , c. m. , & cole , s. 2002 , mnras , 333 , 177 bregman , j. n. 2007 , ara&a , 45 , 221 brook , c. b. , governato , f. , rokar , r. , et al .
2011 , mnras , 415 , 1051 brook , c. b. , stinson , g. , gibson , b. k. , et al .
2011 , mnras , 419 , 771 brook , c. b. , stinson , g. , gibson , b. k. , wadsley , j. , & quinn , t. 2012b , mnras , 424 , 1275 bond , j. r. , kofman , l. , pogosyan , d. , 1996 , nature , 380 , 603 booth , c. m. , & schaye , j. 2012 , arxiv:1203.3802 bullock , j. s. , kravtsov , a. v. , & weinberg , d. h. 2000 , apj , 539 , 517 cen , r. , & ostriker , j. p. 1999 ,
apj , 514 , 1 cen , r. , & ostriker , j. p. 2006 , apj , 650 , 560 ciotti , l. , & ostriker , j. p. 1997 ,
apjl , 487 , l105 colless , m. , dalton , g. , maddox , s. , et al .
2001 , mnras , 328 , 1039 crain , r. a. , eke , v. r. , frenk , c. s. , et al .
2007 , mnras , 377 , 41 dabringhausen , j. , kroupa , p. , pflamm - altenburg , j. , & mieske , s. 2012 , apj , 747 , 72 dav , r. , cen , r. , ostriker , j. p. , et al .
2001 , apj , 552 , 473 debuhr , j. , quataert , e. , & ma , c .- p . 2012 , mnras , 420 , 2221 dav , r. , oppenheimer , b. d. , & sivanandam , s. 2008 , mnras , 391 , 110 dekel , a. , birnboim , y. , engel , g. , et al .
2009 , nature , 457 , 451 di matteo , t. , khandai , n. , degraf , c. , et al .
2012 , apjl , 745 , l29 dubois , y. , devriendt , j. , slyz , a. , & teyssier , r. 2010 , mnras , 409 , 985 dubois , y. , devriendt , j. , slyz , a. , & teyssier , r. 2012 , mnras , 420 , 2662 durier , f. , & de freitas pacheco , j. a. 2011 , international journal of modern physics e , 20 , 44 fakhouri , o. , & ma , c .- p .
2010 , mnras , 401 , 2245 faucher - gigure , c .- a . , kere , d. , & ma , c .- p .
2011 , mnras , 417 , 2982 geller , m. j. , & huchra , j. p. 1989 , science , 246 , 897 genel , s. , bouch , n. , naab , t. , sternberg , a. , & genzel , r. 2010 , apj , 719 , 229 gnedin , n. y. 2000 , apj , 542 , 535 gott , j. r. , iii , juri , m. , schlegel , d. , et al .
2005 , apj , 624 , 463 guedes , j. , callegari , s. , madau , p. , & mayer , l. 2011 , apj , 742 , 76 guo , q. , white , s. , li , c. , & boylan - kolchin , m. 2010 , mnras , 404 , 1111 habergham , s. m. , anderson , j. p. , & james , p. a. 2010 , apj , 717 , 342 hambrick , d. c. , ostriker , j. p. , johansson , p. h. , & naab , t. 2011a , mnras , 413 , 2421 hambrick , d. c. , ostriker , j. p. , naab , t. , & johansson , p. h. 2011b , apj , 738 , 16 hirschmann , m. , naab , t. , somerville , r. s. , burkert , a. , & oser , l. 2012 , mnras , 419 , 3200 hoeft , m. , yepes , g. , gottlber , s. , & springel , v. 2006 , mnras , 371 , 401 hoekstra , h. , hsieh , b. c. , yee , h. k. c. , lin , h. , & gladders , m. d. 2005 , apj , 635 , 73 humphrey , p. j. , buote , d. a. , canizares , c. r. , fabian , a. c. , & miller , j. m. 2011 , apj , 729 , 53 humphrey , p. j. , buote , d. a. , brighenti , f. , et al .
2012 , apj , 748 , 11 katz , n. , keres , d. , dave , r. , & weinberg , d. h. 2003 , the igm / galaxy connection . the distribution of baryons at z=0 , 281 , 185 kere , d. , katz , n. , weinberg , d. h. , & dav , r. 2005 , mnras , 363 , 2 komatsu , e. , dunkley , j. , nolta , m. r. , et al .
2009 , apjs , 180 , 330 komatsu , e. , smith , k. m. , dunkley , j. , et al .
2011 , apjs , 192 , 18 maiolino , r. , gallerani , s. , neri , r. , et al .
2012 , mnras , 425 , l66 mcbride , j. , fakhouri , o. , & ma , c .- p .
2009 , mnras , 398 , 1858 mcgaugh , s. s. , schombert , j. m. , de blok , w. j. g. , & zagursky , m. j. 2010 , apjl , 708 , l14 mo , h. j. , yang , x. , van den bosch , f. c. , & katz , n. 2005 , mnras , 363 , 1155 moster , b. p. , somerville , r. s. , maulbetsch , c. , et al .
2010 , apj , 710 , 903 navarro , j. f. , & white , s. d. m. 1993 , mnras , 265 , 271 nickerson , s. , stinson , g. , couchman , h. m. p. , bailin , j. , & wadsley , j. 2011 , mnras , 415 , 257 ocvirk , p. , pichon , c. , & teyssier , r. 2008 , mnras , 390 , 1326 okamoto , t. , gao , l. , & theuns , t. 2008 , mnras , 390 , 920 oorbe , j. , domnguez - tenreiro , r. , siz , a. , & serna , a. 2007 , mnras , 376 , 39 oser , l. , ostriker , j. p. , naab , t. , johansson , p.
h. , & burkert , a. 2010 , apj , 725 , 2312 peirani , s. , mohayaee , r. , & de freitas pacheco , j. a. 2004 , mnras , 348 , 921 peirani , s. 2010 , mnras , 407 , 1487 piontek , f. , & steinmetz , m. 2011 , mnras , 410 , 2625 powell , l. c. , slyz , a. , & devriendt , j. 2011 , mnras , 660 prunet , s. , pichon , c. , aubert , d. , pogosyan , d. , teyssier , r. , & gottloeber , s. 2008 , apjs , 178 , 179 rasera , y. , & teyssier , r. 2006 , a&a , 445 , 1 read , j. i. , pontzen , a. p. , & viel , m. 2006 , mnras , 371 , 885 ricotti , m. , gnedin , n. y. , & shull , j. m. 2002 , apj , 575 , 49 robitaille , t. p. , & whitney , b. a. 2010 , apjl , 710 , l11 roncarelli , m. , cappelluti , n. , borgani , s. , branchini , e. , & moscardini , l. 2012 , mnras , 424 , 1012 scannapieco , c. , tissera , p. b. , white , s. d. m. , & springel , v. 2008 , mnras , 389 , 1137 scannapieco , c. , white , s. d. m. , springel , v. , & tissera , p. b. 2009 , mnras , 396 , 696 scannapieco , c. , wadepuhl , m. , parry , o. h. , et al .
2012 , mnras , 423 , 1726 shen , s. , wadsley , j. , & stinson , g. 2010 , mnras , 407 , 1581 shen , s. , madau , p. , guedes , j. , mayer , l. , & prochaska , j. x. 2012 , arxiv:1205.0270 sijacki , d. , springel , v. , di matteo , t. , & hernquist , l. 2007 , mnras , 380 , 877 silk , j. , & rees , m. j. 1998 , a&a , 331 , l1 binney , j. , gerhard , o. , & silk , j. 2001 , mnras , 321 , 471 silk , j. , & nusser , a. 2010 , apj , 725 , 556 sinha , m. , & holley - bockelmann , k. 2010 , mnras , 405 , l31 sinha , m. , & holley - bockelmann , k. 2009 , mnras , 397 , 190 somerville , r. s. 2002 , apjl , 572 , l23 springel , v. 2005 , mnras , 364 , 1105 stinson , g. s. , brook , c. , prochaska , j. x. , et al .
2012 , mnras , 425 , 1270 stinson , g. , brook , c. , macci , a. v. , et al .
2012 , arxiv:1208.0002 tabor , g. , & binney , j. 1993 , mnras , 263 , 323 thibaut , j. , pichon , c. , sousbie , t. , prunet , s. , & pogosyan , d. 2008 , mnras , 387 , 397 thoul , a. a. , & weinberg , d. h. 1996 , apj , 465 , 608 tillson , h. , miller , l. , & devriendt , j. 2011 , mnras , 417 , 666 tornatore , l. , borgani , s. , viel , m. , & springel , v. 2010 , mnras , 402 , 1911 valiante , r. , schneider , r. , & maiolino , r. 2012 , arxiv:1205.3488 van de voort , f. , schaye , j. , booth , c. m. , haas , m. r. , & dalla vecchia , c. 2011 , mnras , 414 , 2458 van de voort , f. , & schaye , j. 2012 , mnras , 2882 vitvitska m. , klypin a. , kravtsov a. v. , wechsler r. h. , primack j.r . & bullock j. s. , 2002 , apj 581 , 799 wagner , a. y. , bicknell , g. v. , & umemura , m. 2012 , arxiv:1205.0542 wang , j. , navarro , j. f. , frenk , c. s. , et al .
2011 , mnras , 413 , 1373 wechsler , r. h. , bullock , j. s. , primack , j. r. , kravtsov , a. v. , & dekel , a. 2002 , apj , 568 , 52 weinmann , s. m. , pasquali , a. , oppenheimer , b. d. , et al .
2012 , arxiv:1204.4184 zhu , w. , feng , l .- l . , & fang , l .- z .
2011 , mnras , 415 , 1093 | using hydrodynamical zoom simulations in the standard @xmath0cdm cosmology , we investigate the evolution of the distribution of baryons ( gas and stars ) in a local group - type universe .
+ first , with standard star formation and supernova feedback prescriptions , we find that the mean baryonic fraction value estimated at the virial radius of the two main central objects ( i.e. the milky way and andromeda ) is decreasing over time , and is 10 - 15% lower than the universal value , 0.166 , at @xmath1 . this decrease is mainly due to the fact that the amount of accretion of dissipative gas onto the halo , especially at low redshift , is in general much lower than that of the dissipationless dark matter . indeed , a significant part of the baryons does not collapse onto the haloes and remains in their outskirts , mainly in the form of warm - hot intergalactic medium ( whim ) .
moreover , during the formation of each object , some dark matter and baryons are also expelled through merger events via tidal disruption .
in contrast to baryons , expelled dark matter can be more efficiently re - accreted onto the halo , enhancing both the reduction of @xmath2 inside @xmath3 , and the increase of the mass of whim outside @xmath3 .
varying the efficiency of supernovae feedback at low redshift does not seem to significantly affect these trends .
+ alternatively , when a significant fraction of the initial gas in the main objects is released at high redshifts by more powerful sources of feedback , such as agn from intermediate mass black holes in lower mass galaxies , the baryonic fraction at the virial radius can have a lower value ( @xmath4 ) at low redshift .
hence physical mechanisms able to drive the gas out of the virial radius at high redshifts will have a stronger impact on the deficit of baryons in the mass budget of milky way type - galaxies at present times than those that expel the gas in the longer , late phases of galaxy formation .
galaxies : local group galaxies : haloes dark matter methods : n - body simulations |
quantum mechanics provides some interesting ways for communicating information securely between remote parties @xcite
. however , in practice the quantum channels such as optical fibers are noisy and lossy @xcite .
the transmission loss and the decoherence of photon systems increase exponentially with the distance , which makes it extremely hard to perform a long - distance quantum communication directly . to overcome this limitation , briegel _
et al_. @xcite proposed a noise - tolerant quantum repeater protocol in 1998 . the channel between the two remote parties _ a _ and _ b _ is divided into smaller segments by several nodes , the neighboring nodes can be entangled efficiently by the indirect interaction through flying qubits , and the entanglement between non - neighboring nodes is implemented by quantum entanglement swapping , which can be cascaded to create the entanglement between the terminate nodes _ a _ and _
b_. the implementation of quantum repeaters is compatible with different physical setups assisted by cavity quantum electrodynamics , such as nitrogen vacancy centers in diamonds @xcite , spins in quantum dots @xcite , single trapped ions or atoms @xcite . however , the most widely known approach for quantum repeaters is based on atomic ensembles @xcite due to the collective enhancement effect @xcite . in a seminal paper by duan
_ et al . _
@xcite , the atomic ensemble is utilized to act as a local memory node .
the heralded collective spin - wave entanglement between the neighboring nodes is established by the detection of a single stokes photon , emitted indistinguishably from either of the two memory nodes via a raman scattering process . however , due to the low probability of stokes photon emission required in the duan - lukin - cirac - zoller ( dlcz ) proposal @xcite , the parties can hardly establish the entanglement efficiently for quantum entanglement swapping . in order to improve the success probability ,
photon - pair sources and multimode memories are used to construct a temporal multi - mode modification @xcite , and then the schemes based on the single - photon sources @xcite and spatial multiple modes @xcite are developed .
besides these protocols based on mach - zehnder - type interference , zhao _ et al . _
@xcite proposed a robust quantum repeater protocol based on two - photon hong - ou - mandel - type interference , which relaxes the long - distance stability requirements and suppresses the vacuum component to a constant item .
subsequently , the single - photon sources are embedded to improve the performance of robust quantum repeaters @xcite .
in addition , rydberg blockade effect @xcite is used to perform controlled - not gate between the two atomic ensembles in the middle node @xcite , which makes the quantum entanglement swapping operation be performed deterministically .
since the two - photon interference is performed with the polarization degree of freedom ( dof ) of the photons @xcite , which is incident to be influenced by the thermal fluctuation , vibration , and the imperfection of the fiber @xcite , the fidelity of the entanglement created between the neighboring nodes will be decreased when the photons are transmitted directly @xcite . in other words ,
the more the overlap of the initial photon state used in the two - photon interference is , the higher the fidelity of the entanglement created is .
following the idea of zhao s protocol @xcite , quantum repeaters immune to the rotational polarization noise are proposed with the time - bin photonic state @xcite and the antisymmetric bell state @xcite @xmath0 , respectively .
when the noise on the two orthogonal polarized photon states is independent , zhang _ et al . _
@xcite utilized the faithful transmission of polarization photons @xcite to surmount the collective noise . in the ideal case ,
the two - fold coincidence detection in the central node can successfully get the stationary qubits entangled maximally in a heralded way . apart from this type of entanglement distribution , kalamidas @xcite proposed an error - free entanglement distribution protocol in the linear optical repeater .
an entangled photon source is placed at the center node , and the entangled photons transmitted to neighboring nodes are encoded with their time - bin dof . with two fast pockels cells ( pcs ) ,
the entanglement distribution can be performed with a high efficiency when the polarization - flip - error noise is relatively small . in a recent work ,
mei _ et al . _
@xcite built a controlled - phase - flip ( cpf ) gate between a flying photon and an atomic ensemble embedded in an optical cavity , and constructed a quantum repeater protocol , following some ideas in the original dlcz scheme @xcite . in 2012 , brion _
@xcite constituted a quantum repeater protocol with rydberg blocked atomic ensembles in fiber - coupled cavities via collective laser manipulations of the ensembles and photon transmission .
besides , wang _ et al . _
@xcite proposed a one - step hyperentanglement distillation and amplification proposal , and zhou and sheng @xcite designed a recyclable protocol for the single - photon entanglement amplification , which are quite useful to the high dimensional or multiple dofs optical quantum repeater . in this paper
, we give a general interface between a polarized photon and an atomic ensemble trapped in a single - sided optical cavity .
besides , we show that a deterministic faithful entanglement distribution in a quantum repeater can be implemented with the time - bin photonic state when two identical fibers act as the channels of different spatial dofs of the photons .
interestingly , it does not require fast pcs and the time - slot discriminator @xcite is not needed anymore . by using the input - output process of a single photon based on our general interface , the entanglement between the neighboring atom ensembles can be created in a heralded way , without any classical communication after the clicks of the photon detectors , and the quantum swapping can be implemented with almost unitary success probability by a simplified parity - check gate ( pcg ) between two ensembles , other than @xmath1 with linear optics .
we analyze the performance of our high - efficiency quantum repeater protocol with current experimental parameters and show its robustness to the imperfections involved in the reflection process .
these good features will make it a useful building block in long - distance quantum communication in future .
the elementary node in our quantum repeater protocol includes an ensemble with @xmath2 cold atoms trapped in a single - sided optical cavity @xcite .
the atom has a four - level internal structure and its relevant levels are shown in fig .
the two hyperfine ground states are denoted as @xmath3 and @xmath4 .
the excited state @xmath5 and the rydberg state @xmath6 are two auxiliary states .
the @xmath7 polarized cavity mode @xmath8 couples to the transition between @xmath4 and @xmath5 .
initially , all of the atoms are pumped to the state @xmath3 . with the help of the rydberg state @xmath6
, one can efficiently perform an arbitrary operation between the ground state @xmath9 and the single collective spin - wave excitation state @xcite @xmath10 via collective laser manipulations of the ensembles @xcite . the single collective excited state @xmath11 .
when the rydberg blockade shift is of the scale @xmath12 , the transition between @xmath13 and @xmath14 can be completed with an effective coupling strength @xmath15 and the probability of nonexcited and doubly excited errors @xcite is about @xmath16-@xmath17 .
recently , rotations along axes @xmath18 , @xmath19 , and @xmath20 of a spin - wave excitation with an average fidelity of @xmath21 are achieved in @xmath22 atomic ensembles and they are implemented by making use of stimulated raman transition and controlled larmor procession @xcite . in other words , the high - efficiency single qubit rotations of the atomic ensemble can be implemented faithfully . ,
width=8 ] let us consider an @xmath7 polarized input photon with the frequency @xmath23 , which is nearly resonant to the cavity mode @xmath24 with the frequency @xmath25 .
the coupling rate between the cavity and the input photon can be taken to be a real constant @xmath26 when the detuning @xmath27 is far less than the cavity decay rate @xmath28 ( @xmath29 ) @xcite .
the hamiltonian of the whole system , in the frame rotating with respect to the cavity frequency @xmath25 , is ( @xmath30 ) @xcite @xmath31 \nonumber\\ & & + i\sqrt{\frac{\kappa}{2\pi}}\int{}\!d\delta'\!\left[\hat{b}^{\dagger}(\delta')\hat{a}_{_h}\!-\!\hat{b}(\delta')\hat{a}_{_h}^{\dagger}\right]\!+\!\ ! \int\ !
d\delta'\hat{b}^{\dagger}(\delta')\hat{b}(\delta'),\nonumber\\ \label{hami } % \end{split}\end{aligned}\ ] ] where @xmath32 and @xmath33 are the operators of the cavity mode and the input photon with the properties @xmath34=1 $ ] and @xmath35=\delta(\delta'-\delta'')$ ] , respectively .
@xmath36 is the detuning between the cavity mode frequency @xmath25 and the dipole transition frequency @xmath37 , @xmath38 , and @xmath39 .
@xmath40 represents the spontaneous emission rate of the excited state @xmath41 , while @xmath42 denotes the coupling strength between the _ j - th _ atom transition and the cavity mode @xmath43 . here and after
, we assume @xmath44 and @xmath45 for simplicity . with the hamiltonian @xmath46 shown in eq.([hami ] ) , the heisenberg - langevin equations of motion for cavity @xmath47 and the atomic operator @xmath48 taking into account the atomic excited state decay @xmath49 can be detailed as @xcite @xmath50 here the pauli operator @xmath51 , while @xmath52 is corresponding to the vacuum noise field that helps to preserve the desired commutation relations for the atomic operator .
along with the standard cavity input - output relation @xmath53 , one can obtain the reflection and noise coefficients @xmath54 and @xmath55 in the weak excitation approximation where the ensemble is hardly in the state @xmath56 but predominantly in @xmath14 , that is , @xmath57 where @xmath58 represents the frequency detuning between the input photon and the dipole transition .
@xmath59 means that when the noise field is considered , the energy is conserved during the input - output process of the single - sided cavity . ,
@xmath60 and @xmath61 vs the scaled detuning @xmath62 , with the scaled coupling rate @xmath63 and @xmath64 @xcite .
( b ) @xmath65 vs the scaled coupling rate @xmath66 with detuning @xmath67 and @xmath68.,width=8 ] if the atomic ensemble in the cavity is initialized to be the state @xmath13 , it does not interact with the cavity mode ( i.e. , @xmath69 ) .
the input @xmath7 polarized probe photon feels an empty cavity and will be reflected by the cavity directly .
now , the reflection coefficient can be simplified to be @xcite @xmath70 note that the detuning is small @xmath71 , the pulse bandwidth is much less than the cavity decay rate @xmath28 .
if the strong coupling condition @xmath72 is achieved , one can get the input probe photon totally reflected with @xmath73 or @xmath74 , shown in fig .
the absolute phase shifts versus the scaled detuning are shown in fig .
3 . and @xmath75 that the reflected photon gets , with the ensemble in @xmath13 and @xmath14 , respectively .
the solid line represents the absolute value of the phase shifts difference @xmath76 = @xmath77 .
the inset shows the phase shifts vs the scaled detuning @xmath78 and @xmath79 that the reflected photon gets , with the ensemble in @xmath13 and @xmath14 , respectively.,width=8 ] the principle of our cpf gate on a hybrid quantum system composed of a photon @xmath80 and an atomic ensemble @xmath81 is shown in fig .
4 , following some ideas in previous works @xcite . suppose that the photon @xmath80 is in the state @xmath82 ( @xmath83 ) and the ensemble @xmath84 is in the state @xmath85 ( latexmath:[$ of the photon @xmath80 transmits the polarization beam splitter ( pbs ) and then be reflected by the cavity , while the @xmath87 polarized component is reflected by the mirror @xmath88 .
the optical pathes of the @xmath7 and @xmath87 components are adjusted to be equal and they will be combined again at the pbs with an extra @xmath89 phase shift on the @xmath7 component if the ensemble is in the state @xmath13 .
this process can be described as @xmath90 that is to say , the setup in fig .
4(a ) can be used to accomplish a cpf gate on the atomic ensemble @xmath84 and the photon @xmath80 .
stands for a mirror and the pbs transmits the @xmath7 polarized photon and reflects the @xmath87 component . hwp@xmath91 and hwp@xmath92 are half wave plates performing the bit - flip operation while h represent a hadamard rotation.,width=8 ] the schematic diagram of our pcg on two atomic ensembles @xmath84 and @xmath93 is shown in fig .
let us assume that @xmath81 and @xmath94 are initially in the states @xmath95 ( @xmath96 and @xmath97 , @xmath98 ) .
one can input a polarized photon _
p _ in the state @xmath99 into the import of the setup .
( hwp@xmath92 ) is used to perform the bit - flip operation @xmath100 on the photon @xmath80 by using a half - wave plate ( hwp ) with its axis at @xmath101 with respect to the horizontal direction .
after the two components of @xmath80 are reflected by the two cavities , they combine with each other at pbs@xmath92 .
the state of the system composed of the two atom ensembles and the photon evolves to be @xmath102 and then , another hwp names @xmath103 whose axis is placed at @xmath104 is used to perform a hadamard rotations @xmath105 and @xmath106 on the photon .
the state of the system becomes @xmath107 after the photon is measured with pbs@xmath108 and two single - photon detectors , the parity of @xmath81 and @xmath94 can be determined . in detail , if the photon is in the state @xmath7 , the two ensembles @xmath81 and @xmath94 have an even parity .
if the photon is in @xmath87 , @xmath81 and @xmath94 have an odd parity . with an effective input - output process of a single photon
, one can efficiently complete the pcg on two atomic ensembles .
is a @xmath89 phase shifter.,width=8 ] suppose that there is an entanglement source which is placed at a central station between two neighboring nodes , say alice and bob .
the source produces a two - photon polarization - entangled bell state @xmath109 . here
the subscripts @xmath110 and @xmath111 denote the photons sent to alice and bob , respectively . as shown in fig .
5 ( a ) , the photons _ a _ and _ b _ will pass through an encoder in each side before they enter the noisy channels .
the encoder is made up of a pbs , an hwp , and a beam splitter ( bs ) .
here bs is used for a hadamard rotation on the spatial dof of the photon , i.e. , @xmath112 and @xmath113 , where @xmath114 and @xmath115 represent the upper and the down ports of the bs , respectively . with our faithful single - photon transmission method
( see method ) , alice and bob can share photon pairs in a maximally entangled state , shown in fig .
5 . in detail , after a photon pair from the source passes through the two encoders , its state becomes @xmath116 .
% \end{split}\end{aligned}\ ] ] as the two photons _ a _ and _ b _ suffer from independent collective noises from the two channels , the influence of the channels on the two photons can be described with two unitary rotations @xmath117 and @xmath118 as follows : @xmath119{noise } & \delta_a|h\rangle+\eta_a|v\rangle,\\ u^b_{c}|h\rangle & \xrightarrow[]{noise } & \delta_b|h\rangle+\eta_b|v\rangle,\end{aligned}\ ] ] where @xmath120 ( @xmath121 ) .
the influence on the polarization of the photons arising from the channel noises can be totally converted into that on the spatial dof .
the state of the photons _ a _ and _ b _ arriving at alice and bob becomes @xmath122 this is a two - photon bell state @xmath123 in the polarization dof of the photon pair @xmath124 .
simultaneously , it is a separable superposition state @xmath125 in the spatial dof . to entangle the stationary atomic ensembles @xmath81 and @xmath94 , which are initialized to be @xmath126 and @xmath127 , only two cpf gates are required if alice and bob have shared some photon pairs in the bell state @xmath128 .
let us take the case that the photons @xmath110 and @xmath111 come from the spatial modes @xmath129 and @xmath130 as an example to detail the entanglement creation process .
as for the other cases , the same entanglement between @xmath81 and @xmath94 can be obtained by a similar procedure with or without some single - qubit operations .
first , the photon _ a _ suffers a hadamard operation by passing through a half - wave plate _
, it is reflected by the cavity or the mirror @xmath88 , which is used to complete the cpf gate on the photon _ a _ and the ensemble @xmath81 .
third , alice performs another hadamard operation on the photon _
a_. now , the state of the composite system composed of the photons _ a _ and _ b _ and the ensembles @xmath81 and @xmath94 evolves into @xmath131 , @xmath132\otimes|\varphi\rangle_{_{b}}.\;\;\;\ ; % \end{split}\end{aligned}\ ] ] fourth , alice measures the polarization state of the photon @xmath110 with a setup composed of pbs and single - photon detectors @xmath133 and @xmath134 .
if an @xmath7 polarized photon is detected , the hybrid system composed of _ b _ , @xmath81 , and @xmath94 will be projected into latexmath:[\ ] ] where the normalized coefficient @xmath231 $ ] .
one can obtain the fidelity of the entanglement distribution process @xmath232 and @xmath233 for the cases that @xmath234 and @xmath235 at the bob s node are clicked , respectively .
@xmath236 if one defines the efficiency @xmath237 as the probability that alice detects an @xmath7 polarized photon while bob detects a photon in either @xmath7 or @xmath87 polarization , one has @xmath238 in the above discussion , we detail the performance of our entanglement distribution conditioned on the detection of an @xmath7 polarization photon at alice s node . considering the symmetric property of the system ,
one can easily obtain the performance of the entanglement distribution upon the detection of a @xmath87 polarization photon at alice s node .
now , the fidelities @xmath239 and @xmath240 for the cases that @xmath234 and @xmath235 are clicked at bob s node , have the following relations to that for the cases that an @xmath7 polarized photon is detected by alice , @xmath241 and @xmath242 , see eq .
( [ fmhfmv ] ) for detail . meanwhile , the efficiency @xmath243 of the entanglement distribution process when alice detects a photon in @xmath87 polarization is identical to @xmath237 .
the total efficiency @xmath244 of the entanglement distribution can be written as @xmath245 in our entanglement swapping process , two pcgs are applied on the two ensembles @xmath145 and @xmath146 at the middle node .
in fact , only one pcg is enough if a single - atomic - ensemble measurement on each of the two ensembles @xmath246 and @xmath247 is utilized after the local hadamard operations .
after these measurements , the system composed of the two remote ensembles @xmath81 and @xmath144 is in the state @xmath248 with or without a local unitary operation .
when the fluorescent measurement @xcite or field - ionizing the atoms @xcite with the help of rydberg excitation are used , the state detection on atomic ensembles could be performed with a near - unity efficiency . in other words ,
the fidelity of the quantum entanglement swapping process can equal to that of the pcg operation .
the fidelities of both the entanglement distribution and the entanglement swapping in our repeater scheme are shown in fig .
one can see that all @xmath232 , @xmath233 , and @xmath249 are larger than @xmath250 with the parameters ( @xmath251 , @xmath28 , @xmath49)@xmath252(@xmath253 , @xmath254 , @xmath255 ) mhz achieved in experiment @xcite . meanwhile , all efficiencies involved in our quantum repeater protocol , shown in fig . 10 , can be larger than @xmath256 when the effective coupling @xmath257 with @xmath206 . in a recent experiment with a fiber - based fabry__-__perot cavity constituted by co@xmath92 laser - machined mirrors @xcite , the maximal coupling strength as high as @xmath258 ghz
is achieved for single rb atoms and the cavity decay rate is @xmath259 ghz @xmath260 . in this time , @xmath261 is achieved , and a better performance of our scheme is attainable .
tl was supported by the china postdoctoral science foundation under grant no .
fg was supported by the national natural science foundation of china under grant nos .
11174039 and 11474026 , and the open foundation of state key laboratory of networking and switching technology ( beijing university of posts and telecommunications ) under grant no .
sklnst-2013 - 1 - 13 . | quantum repeater is one of the important building blocks for long distance quantum communication network .
the previous quantum repeaters based on atomic ensembles and linear optical elements can only be performed with a maximal success probability of 1/2 during the entanglement creation and entanglement swapping procedures . meanwhile , the polarization noise during the entanglement distribution process is harmful to the entangled channel created . here
we introduce a general interface between a polarized photon and an atomic ensemble trapped in a single - sided optical cavity , and with which we propose a high - efficiency quantum repeater protocol in which the robust entanglement distribution is accomplished by the stable spatial - temporal entanglement and it can in principle create the deterministic entanglement between neighboring atomic ensembles in a heralded way as a result of cavity quantum electrodynamics .
meanwhile , the simplified parity check gate makes the entanglement swapping be completed with unity efficiency , other than 1/2 with linear optics .
we detail the performance of our protocol with current experimental parameters and show its robustness to the imperfections , i.e. , detuning and coupling variation , involved in the reflection process .
these good features make it a useful building block in long distance quantum communication . |
recently a wealth of x - ray observations has provided more detailed information about spectra of x - ray sources and short - time variability .
this concerns soft x - ray transients as good examples of stellar accreting black holes , high - mass x - ray binaries with a black hole primary star and also active galactic nuclei .
the new observations put more stringent constraints on the modelling of accretion processes . in this connection the change from accretion in a standard cool disk to a hot coronal flow is an important feature .
different solutions of accretion on to compact objects had been investigated in the past .
an account can be found in the review by narayan et al .
the description of an optically thick , geometrically thin disk , now often called `` standard thin disk '' , goes back to work of shakura & sunyaev ( 1973 ) , novikov & thorne ( 1973 ) , lynden - bell & pringle ( 1974 ) and others .
a much hotter self - consistent solution of optically thin gas was discovered by shapiro , lightman & eardley ( 1976 ) , but found to be thermally unstable by piran ( 1978 ) . a solution for even super - eddington accretion rates in which the large optical depth of the inflowing gas prevents radiation to escape
was suggested by katz ( 1977 ) , begelman ( 1978 ) , abramowicz et al .
( 1988 ) and others .
this is an optically thick advection - dominated accretion flow . the fourth solution concerns accretion at low sub - eddington rate ( proposed by shapiro et al . 1976 ,
ichimaru 1977 and rees et al .
detailed theoretical work on this solution by several authors has proven that this optically thin regime of solutions for low accretion rates is relevant for accretion in many cases ( narayan & yi 1994 , 1995a , b , abramowicz , chen & kato et al .
1995 , chen 1995 ; chen et al .
1995 , 1997 ; honma 1996 ; narayan , kato & homna 1997a ; nakamura et al .
1996 , 1997 ) .
the model of an optically thin advection - dominated flow ( adaf ) allows to explain the accretion in soft x - ray transients during quiescence . assuming the old picture of a standard thin accretion disk reaching inward towards the black hole the amount of x - ray flux is clearly inconsistent with the accretion rate estimated from the outer optically thick disk ( mcclintock et al .
an interior adaf allows to understand the observations but requires the change from a standard thin disk to the hot flow at a certain transition radius . the black hole x - ray transients are a laboratory for accretion on to black holes .
the physics of the accretion processes around supermassive black holes in agn is similar , which makes the investigation of adaf even more important and fruitful .
the change from a geometrically thin disk to the hot flow is an essential feature in this overall picture of accretion .
the `` strong adaf principle '' , formulated by narayan & yi ( 1995b ) , suggests that , whenever the accreting gas has a choice between a standard thin disk and an adaf , the adaf configuration is chosen .
this rule provides a transition radius @xmath0 for a given accretion rate ( see fig .
6b in esin et al . 1997 ) . in a different approach honma ( 1996 ) considered the interface of a hot optically thin disk with a cool standard thin disk applying radial diffusive transport of heat . here
we present a model of a corona above a thin disk in order to determine the inner edge of the thin disk , inside of which only a hot optically thin flow exists .
the co - existence of a cool disk and a corona had been investigated for disks in dwarf nova systems , where features otherwise not understood as for example the x - rays observed in quiescence can then be explained ( meyer & meyer - hofmeister 1994 , liu et al .
the situation is the same in disks around black holes ( meyer 1999 ) , both stellar and supermassive .
the physical two - dimensional situation would demand the solution of a set of partial differential equations in radial distance @xmath1 and vertical height @xmath2 .
in particular a sonic transition would require treatment of a free boundary condition on an extended surface . in sect .
2 we show how we derive a simplified set of four ordinary differential equations which describe the physics of the corona above the thin disk . the interaction of corona and thin disk results in evaporation of matter from the thin disk . in sect .
3 we discuss the general features of this evaporation process . the results of our computations are presented in sect .
the main result of this model is the efficiency of evaporation and with this the location where the thin disk has completely changed to a hot coronal flow ( sect .
5 ) . in sect .
6 we point out the invariance of our results for different black hole masses . sects . 7 and 8 concern the validity of the assumptions made in our computations and the importance of synchrotron and compton processes . in sect .
9 the comparison with observations follows , in sect . 10 a discussion of further effects . in sect .
11 we compare with the model of honma ( 1996 ) .
12 contains our conclusions .
in this section we derive the equations that describe the corona above the inner rim of the accretion disk . for this
we will average over an inner ring area and thereby reduce the set of partial differential equations to a system of ordinary differential equations with respect to height .
the solutions then become a function of the radial coordinate of the inner rim .
we start with the five equations of viscous hydrodynamics , conservation of mass @xmath3 the equations of motion @xmath4 = - \nabla p + \nabla \cdot { \bf s}- \rho { \bf \nabla } \mit\phi,\ ] ] and the first law of thermodynamics @xmath5 = - p({\bf \nabla } \cdot{\bf v } ) + ( { \bf s } \cdot \nabla ) \cdot{\bf v } + \dot{q},\ ] ] supplemented by the equation of state for an ideal gas @xmath6 @xmath7 @xmath8 here @xmath9 and @xmath10 are density , pressure and temperature , @xmath11 is the velocity vector , @xmath12 the viscous stress tensor , @xmath13 the gravitational potential , @xmath14 the internal energy , @xmath15 the gas constant , @xmath16 the molecular weight .
we take @xmath17 for a fully ionized gas of cosmic abundances .
@xmath18 is the isothermal speed of sound , @xmath19 the ratio of the specific heats , = @xmath20 , for a mono - atomic gas .
for @xmath12 we take the form appropriate for a mono - atomic gas with cartesian components @xmath21\ ] ] ( sommerfeld 1949 ) . for the ( non - molecular ) viscosity we use an @xmath22-parameterization ( shakura & sunyaev 1973 , novikov & thorne 1973 ) @xmath23 @xmath24 rotational frequency .
@xmath25 is the rate of heating per unit volume .
we write it in two parts , @xmath26 where @xmath27 is the gain or loss of heat by radiation and @xmath28 that by heat conduction @xmath29 for @xmath27 we take here optically thin thermal radiation loss according to fig . 1 of raymond et al .
@xmath30 where @xmath31 and @xmath32 are electron and ion particle densities . for the heat flux
@xmath33 we take @xmath34 ( spitzer 1962 ) for a fully ionized plasma and use the value @xmath35neglecting a weak logarithmic dependence on electron density and temperature .
taking the scalar product of eq.(2 ) with @xmath36 , adding to eq.(3 ) , and using eq.(1 ) we obtain the equation of conservation of energy @xmath37 + { \bf\nabla } \cdot \bigg[{\bf { v}}\bigg(\rho u + \rho \frac{v^2}{2 } + \rho \mit\phi+p\bigg ) } \nonumber \\ & & -({\bf s\cdot\bf{v}})+{\bf { f_{\rm c } } } \bigg]= \dot{q}_{rad}\end{aligned}\ ] ] we introduce cylindrical coordinates , @xmath38 with the @xmath2-axis perpendicular to the midplane of the disk and through the central accretor , and use the vector correspondence to the cartesian advection term @xmath39 where @xmath40 .
the ( newtonian ) potential is then @xmath41 with @xmath42 the gravitational constant and @xmath43 the central mass .
we consider from now on stationary , azimuthally symmetric flows . in the following subsections
we derive values for the rotational velocity @xmath44 and the radial diffusive velocity @xmath45 , and the three remaining ordinary differential equations for the conservation of mass , the vertical dynamic equilibrium , and the energy flow in the corona above the inner zone of the accretion disk . for this
we choose a region between inner disk radius @xmath46 and radius @xmath47 such that @xmath48 where @xmath1 is an area weighted mean radius between @xmath46 and @xmath47 .
we apply therefore , at each radius r , the integration bounds @xmath46=0.72@xmath1 and @xmath47=1.23@xmath1 . the coronal solution depends on the value of @xmath1 as a measure of the distance from the accretor .
we estimate the relative order of magnitude of the various terms in our equations and drop non - dominant terms . for these estimates
we use relations that result from our solutions which are therefore self - consistent .
all our solutions have @xmath49 see fig .
3 . they are subvirial and therefore like `` slim '' accretion disks ( abramowicz et al . 1988 ) with cylindrical symmetry in their ( main ) part in contrast to the more spherical symmetry of adaf solutions . as a result @xmath50 where we have replaced the azimuthal velocity by its kepler value which is a good approximation in the main body of our solution ,
see the remark follwing eq .
a wind from such a corona reaches escape speed at large height leading to @xmath51 see fig .
2 . where in the following we neglect radial derivatives of velocity components we use the order of magnitude relation
@xmath52 this equation yields the rotational velocity .
the radial component of the equation of motion becomes @xmath53 in this equation the pressure term @xmath54 is small compared to the gravitational term in regions @xmath55 dominant with respect to dynamics and energetics of the solutions ( see fig .
2 ) . taking @xmath56 = @xmath57 we estimate @xmath58 this shows a relative unimportance of radial pressure support in these coronae ( in contrast to virialized adaf solutions , narayan & yi ( 1995b ) ) . at a large height @xmath59 this ratio has increased to 0.35 .
but we note that with the topology of fig .
1 that at that height the radial pressure gradient must have significantly decreased because of the axial geometry . therefore to first approximation we can neglect the pressure term .
the viscous terms @xmath60 contribute derivatives of @xmath1- and @xmath2-components of the velocity .
their contributions , evaluated with eqs . ( 16 ) to ( 19 ) , are small compared to @xmath61 by at least a factor @xmath62 .
they as well as the first two terms on the left hand side of eq .
( 20 ) are negligible .
this leaves the balance between centrifugal and radial gravitational force @xmath63 @xmath64 in the main body of the coronal solution the kepler values ( i.e. at @xmath2=0 ) are a good approximation to @xmath44 and @xmath24 . in our calculation
we take the kepler values throughout as a computational convenience .
it overestimates the frictional work and underestimates the release of rotational energy at large height .
these two effects partially compensate each other . at @xmath65
the sum of these effects deviates from the accurate value by less than @xmath66 . at @xmath59 it
is 20@xmath67 .
but at this height the total further contribution of both energy terms already is negligibly small ( due to the exponential decrease of pressure and density with height , see fig .
2 ) and isothermality of the solution completely dominates .
this simplification has thus only a small effect on the accuracy of the solution . the conservation of the @xmath2-component of angular momentum yields the @xmath1-component of the velocity .
multiplication of the equation of motion by @xmath69 unit vectors in @xmath1 , @xmath68 and @xmath2 direction ) yields @xmath70 using ( 1 ) ( for @xmath71 = 0 ) @xmath72 and @xmath73 ( the right side understood as the dyadic product ) and @xmath74 this results in the equation of conservation of @xmath2-angular momentum @xmath75 we integrate over a cylinder with radius @xmath46 and obtain @xmath76 here it is assumed that no significant angular momentum is accreted at the center of the compact object or lost through the boundary @xmath77 . for accretion and jet angular momentum loss from the most interior parts of the accretion region , such losses would be small by order of @xmath78 where @xmath79 is the radius of the inner accretion region , e.g. @xmath80 for the schwarzschild radius @xmath81 of a central black hole ( @xmath82 ) . we replace the global balance of angular momentum flow of eq .
( 28 ) by its local equivalent and obtain the radial diffusive velocity @xmath83 with @xmath84 .
this is the standard bulk radial velocity formula for accretion disks .
it neglects the modification by @xmath2-dependent secondary flows superimposed on the main drift of material ( see kley & lin 1992 ) .
such secondary flows may slightly modify the local vertical structure but will hardly affect the general structure of the solutions
. the vertical component of the equation of motion eq .
( 2 ) yields the vertical dynamic equilibrium .
the vertical component becomes @xmath85 using approximate isothermal pressure layering ( see fig .
2 ) @xmath86 with @xmath87 ( see fig .
3 ) and the estimates of eqs . ( 16 ) to ( 19 ) we estimate @xmath88 to be negligible . further , with @xmath89 the largest contribution of all viscous terms in eq . ( 30 ) appears in @xmath90 from the @xmath2-derivative of @xmath91 ( that is of @xmath92 , see eq .
( 8) times the derivative @xmath93 .
this term is of order @[email protected] and negligible . in the remaining dominant terms
only @xmath2-derivatives appear .
averaging over the inner zone we obtain @xmath96 the equation of mass conservation we integrate over the radial zone between @xmath46 and @xmath47 , of area @xmath97 and obtain @xmath98 when , as we assume , @xmath99 drops off with radius strongly ( e.g. @xmath100 in the scaled solutions , liu et al .
1995 ) the difference on the right hand side is dominated by the first term , and its value will be slightly smaller than this term . we approximately take this into account by replacing the term at @xmath46 by its value at the somewhat larger mean radius @xmath1 .
we replace the left hand side by the area times a mean value .
this leads to the approximation @xmath101 where @xmath45 is the radial flow velocity ( eq .
29 ) . at large height
, however , we have to take into account the flaring of the channel as the ascending gas changes from a cylindrical rise to spherical expansion ( see fig .
1 ) . how can we account for the additional attenuation of the vertical flux density by this change of geometry ?
we proceed in the following way .
approximately the cross section of the vertical flow channel can be taken as proportional to @xmath102 . without sidewise diffusive loss , mass conservation
would then yield @xmath103 = 0\ ] ] if we take the center of the expanding flux tube in the vertical direction . per unit cross section
the resulting conservation equation is then @xmath104 = % \frac{d}{dz}\left(\rho{v}_z\right ) + \frac{2z}{r^2+z^2 } \rho{v}_z % \nonumber } \\ % & & = 0 \frac{1}{1+\frac{z^2}{r^2 } } \frac{d}{dz}\left[\left(1 + \frac{z^2}{r^2}\right ) \rho{v}_z \right ] & = & \frac{d}{dz}\bigg(\rho{v}_z\bigg ) + \frac{2z}{r^2+z^2 } \rho{v}_z \nonumber \\ & = & 0\end{aligned}\ ] ] the gas in a cylindrical column thus experiences an additional geometry - related radial divergence @xmath105 adding the two radial flow components we obtain the final form @xmath106 we note that the diffusive part in this equation becomes unimportant at large height where the diffusive description for the flaring flux tube becomes problematic .
we first derive the frictional flux terms . with eqs .
( 22 ) and ( 13 ) we obtain @xmath107\ ] ] and its @xmath1-component @xmath108\end{aligned}\ ] ] here all terms are negligible except those containing @xmath44 , by use of eqs .
( 16 ) to ( 19 ) .
making use of eqs .
( 8) and ( 19 ) we obtain @xmath109 likewise , only keeping non - negligible @xmath110 terms , the @xmath2-component is @xmath111 = \frac{2}{3 } \alpha p r \frac{\partial v_\varphi}{\partial z}.\end{aligned}\ ] ] the radial component of the divergence can be written , using ( 19 ) as @xmath112= -\alpha p \mit\omega \bigg(\frac{1}{2}+\frac{\partial \rm{ln}p}{\partial \rm{ln}r}\bigg ) = \frac{3}{2}\alpha p \mit\omega .
\end{aligned}\ ] ] for @xmath113 we have here taken its estimated value -2 at the characteristic height @xmath114 ( see fig .
2 ) , which we obtain from eq . ( 31 ) with the scaling @xmath115 that holds when we compare solutions with different distances @xmath1 ( see liu et al .
the resulting term represents the dominant heating mechanism of the corona by release of gravitational energy .
we now integrate over the area between @xmath46 and @xmath47 as in the derivation of conservation of mass @xmath116 -\frac{3}{2 } \alpha p { \mit\omega}-\dot{q}_{rad}\bigg\ } \\ & = & \bigg\{2\pi r \bigg[{v}_r\bigg(\rho\frac{v^2}{2 } + \frac{\gamma}{\gamma-1 } p+\rho{\mit\phi } \bigg ) + ( f_{\rm c})_r \bigg ] \bigg\}_{r_1 } \nonumber \\ & & -\bigg\ { \ldots \bigg\}_{r_2}\nonumber\end{aligned}\ ] ] on the right hand side of this equation we again replace the difference of the terms taken at @xmath46 and @xmath47 by the value taken at the intermediate value @xmath1 , because the value at @xmath47 will only be a fraction of the value at @xmath46 due to the radial drop off of all the quantities . if we take @xmath45 as the diffusive velocity ( eq .
( 29 ) we must still include the divergence of the vertical flow due to the flaring geometry .
we use the same procedure as in the preceding sect .
2.4 to derive the corresponding term and add it on the right hand side .
again , on the left hand side of the equation we write the integral as area times a mean value and obtain the energy equation in its final form @xmath117= } \nonumber \\ & & \frac{3}{2}\alpha p{\mit\omega } + \dot{q}_{rad } \nonumber \\ & & + \frac{2}{r } \bigg[{v}_r \bigg(\rho\frac{{v}^2}{2 } + \frac{\gamma}{\gamma-1 } p+\rho\mit\phi\bigg ) + ( f_{\rm c})_r\bigg ] \nonumber \\ & & - \frac{2z}{r^2+z^2}\bigg[{v}_z\bigg(\rho\frac{{v}^2}{2 } + \frac{\gamma}{\gamma-1 } p + \rho\mit\phi\bigg ) \nonumber \\ & & + \frac{2}{3}\alpha pr\frac{dv_\varphi}{dz}+ ( f_{\rm c})_z \bigg].\end{aligned}\ ] ] had we chosen to integrate the radial divergence @xmath118 $ ] underneath the integral of eq .
( 45 ) instead of evaluating this term as in eq .
( 44 ) the last equation would have the factor 2 instead of @xmath119 .
this difference arises from the different assumptions about radial variation of the quantities , implicit in the two approaches .
it shows a natural limit of numerical accuracy of any such approach .
the lateral inflow of heat by @xmath120 adds to the frictional release of energy in the coronal column .
we show that the relative size of the contribution is small and we will take @xmath121=0 in our computations .
we assume that the temperature at low @xmath122 varies like the virial temperature , @xmath123 , and that at @xmath124 the radial temperature gradient becomes small as the topology changes and the corona fills the full cylindrical region , with @xmath125 on the axis of symmetry . with these assumptions we take @xmath126 the ratio of total radial conductive heat input @xmath127 to the total frictional energy input @xmath128 is then evaluated as @xmath129 dz \bigg/ \int^{z_0}_{0}\frac{3}{2 } \alpha p { \mit\omega}dz=0.22 , \end{aligned}\ ] ] it is independent of the value of @xmath1 as long as we are in the `` advection - dominated '' range @xmath130 ( see fig.3 ) .
this follows when one uses the scaling of @xmath92 , @xmath10 and @xmath24 derived by liu et al .
( 1995 ) ( see also values @xmath131 in table 1 ) . in
the `` conduction - dominated '' corona @xmath132 the temperature becomes constant and the ratio becomes zero . the contribution of @xmath120 can therefore be neglected .
the remaining @xmath133component of eq .
( 11 ) we write as @xmath134 from the five partial differential equations ( 1 ) to ( 3 ) we have now derived three ordinary differential equations ( 34 ) , ( 39 ) and ( 46 ) ( where we take @xmath135 ) which describe vertical dynamical equilibrium and conservation of mass and energy , plus the two expressions for @xmath136 and @xmath45 ( eqs . ( 24 ) and ( 29 ) ) .
our fourth ordinary differential equation is the dependence of the heat flux on the temperature gradient , eq .
( 49 ) . as the 4 dependent variables we have temperature @xmath10 , pressure @xmath92 , mass flow @xmath137 and heat flux @xmath138 .
the 4 boundary conditions are : \(1 ) no pressure at infinity ( i.e. no artificial confinement ) .
this requires sound transition at some height @xmath139 ( free boundary problem ) @xmath140 \(2 ) no influx of heat from infinity .
we neglect a small outward conductive heat flux induced by the temperature gradient in the expanding wind and require @xmath141 \(3 ) chromosphere temperature at the bottom @xmath142
@xmath143 \(4 ) no heat inflow at the bottom @xmath144 this approximates the condition that the chromospheric temperature is kept up by other processes and can not conduct any sizeable thermal heat flow .
the consequence of conditions ( 3 ) and ( 4 ) is that the conductive flux that comes down at a certain level above the boundary must be radiated away .
since the advective cooling by the rising mass flow is small ( in our solutions at @xmath145 only @xmath146 ) this reduces to the balance between thermal conductive flux and radiation and establishes the temperature profile as the one studied by shmeleva & syrovatskii ( 1973 ) for the solar corona . in these lowest layers
the pressure is nearly constant and the temperature rises steeply .
liu et al . ( 1995 ) recalculated the temperature profile using the more recent cooling function of raymond et al .
( 1976 ) and we used the improved version .
this allows to replace the conditions ( 3 ) and ( 4 ) by @xmath147 the relation between @xmath92 and @xmath148 is practically independent of the value of the chromospheric temperature as long as it is small compared to @xmath149k .
these equations and boundary conditions describe a model of the corona - disk interaction for the innermost region of the thin accretion disk where evaporation is most efficient .
the equations derived in the last section describe the physics of a hot corona in equilibrium with a thin cool disk underneath .
aspects of this model had already been investigated for accretion on to a white dwarf ( meyer & meyer - hofmeister 1994 , liu et al .
1995 , for application to wz sge stars and x - ray transients see also mineshige et al .
the picture that arises from the processes involved is the following .
heat released by friction in the corona flows down into cooler and denser transition layers .
there it is radiated away if the density is sufficiently high .
if the density is too low , cool matter is heated up and evaporated into the corona until an equilibrium density is established .
mass drained from the corona by an inward drift is steadily replaced by mass evaporating from the thin disk as the system reaches a stationary state .
this evaporation rate increases steeply with decreasing distance from the central compact star . when the evaporation rate exceeds the mass flow rate in the cool disk , a hole forms up to the distance where both rates are equal .
inside this transition radius @xmath150 only a hot coronal flow exists .
the mass flow in the corona is similar to that in the thin disk .
the matter has still its angular momentum , and differential ( kepler - like ) rotation causes friction .
the main part of the coronal gas flows towards the center directly . for conservation of angular momentum
a small part flows outward in the coronal layer and condensates in the outer cool disk .
since the temperature is very high the corona is geometrically much thicker than the disk underneath .
therefore processes as sidewise energy transport and downward heat conduction are important for the equilibrium between corona and thin disk , together with radiation and wind loss .
we model the equilibrium between corona and thin disk in a simplified way using a one - zone model as described in sect .
this is possible since the evaporation process is concentrated near the inner edge of the thin disk and we can choose a representative radial region from @xmath46 to @xmath47 such that the main interaction between cool disk and corona occurs there and evaporation further outward is not important . in fig .
1 we show the flow pattern as derived from analysis and simplified modelling for the case that a hole in the thin disk exists .
there are three regimes .
( 1 ) near the inner edge of the thin disk the gas flows towards the black hole .
( 2 ) at larger @xmath1 wind loss is important taking away about 20% of the total matter inflow . ( 3 ) at even larger distances some matter flows outward in the corona as a consequence of conservation of angular momentum .
( one might compare this with the flow in a `` free '' thin disk without the tidal forces acting in a binary .
in such a disk matter flows inward in the inner region and outward in the outer region , with conservation of the total mass and angular momentum ( pringle 1981 ) ) .
we solve the differential equations ( 34 ) , ( 39 ) , ( 46 ) , and ( 49 ) using the runge - kutta method integrating from the lower boundary at an assumed height @xmath151 above the midplane to the upper boundary at @xmath152 .
the value @xmath152 is determined by our computations as the height where sound transition occurs ( free boundary problem ) .
the boundary conditions are given in eqs .
( 50 ) , ( 51 ) and ( 54 ) .
we require that the energy flux at the upper boundary is zero but are content in our computations if it is very small , @xmath153 .
we have to vary @xmath131 and @xmath154 , trial values for @xmath92 and @xmath155 ( @xmath156 ) at @xmath142 , to find the coronal structure which fulfills the upper boundary conditions .
the conditions , sound transition and zero heat flux , are met only for a unique pair of values for pressure and mass flow at the bottom . for other values
sound transition is not reached and/or the heat flux is not negligible ( this is the same situation as in coronae of cataclysmic variable disks , compare fig . 2 of meyer & meyer - hofmeister 1994 ) .
numerically , the smaller the distance @xmath1 from the compact object the more accurately the pressure value has to be determined to obtain a consistent solution . for our computations we put the lower boundary at about @xmath157 cm above the midplane , but for @xmath158 cm at correspondingly smaller values @xmath159 . in most cases
the true height of the thin disk is less than the chosen value @xmath159 .
we carried out test computations to check the influence of assuming different values @xmath159 on the coronal structure . as expected the changes
were found to be very small .
@xmath131 and @xmath160 are nearly independent of the value of the chromospheric temperature as long as it is small compared to @xmath149k .
since our results also concern accretion in agn we have confirmed the validity of our assumption for the location of the lower boundary for this case also .
we took 6@xmath161 as our standard parameter value to model the evaporation in accretion disks around stellar black holes .
we compute the structure of the corona for distances of the outer cool disk from the black hole ranging from large values to small values of a few hundred schwarzschild radii . for different distances @xmath1
the structure in @xmath2 direction is similar if considered as a function of @xmath122 . as will be discussed in sect .
6 the coronal structure scales with schwarzschild radius and eddington accretion rate . cm from the 6@xmath161 black hole .
temperature @xmath10 is given in units of 0.2@xmath162virial temperature , pressure p and vertical mass flow @xmath163 ( @xmath164 vertical velocity ) are scaled to their values at the lower boundary .
-@xmath138 is the downward conductive heat flux , measured in units of the frictional energy generation rate in a column of unit cross section .
@xmath18 is the sound velocity.,width=283 ] in fig .
2 we show the structure at @xmath165 cm close to where the evaporation efficiency is highest for a 6@xmath161 black hole . pressure and density decrease significantly within a narrow range above the cool disk .
the downward conductive heat flux -@xmath138 therefore also has its maximum at a low height @xmath2 , then rapidly decreases with height and is negliglibly small where sound transition @xmath166 is reached , our boundary conditions .
this happens at 2.3@xmath167 cm .
the temperature is practically constant in the higher layers of the corona , a consequence of the high thermal conductivity .
the computed coronal structure at distance @xmath1 corresponds to the situation that the cool disk underneath extends from this distance @xmath1 ( inner edge ) outwards .
the derivation in sect .
2 is based on the dominant role of the innermost corona and our model can not be used to describe the coronal structure above the cool thin disk farther out which will be affected by the dominant process farther in .
these results can be compared to those of our earlier investigation of evaporation in cataclysmic variable disks . the coronal structure was considered there at distances @xmath168 to @xmath169 cm from the 1@xmath161 white dwarf adequate for a disk around a white dwarf . due to the scaling with the mass of the compact star the structure above the thin disk around a 1@xmath161 white dwarf at distances logr 9.0 and 9.7 ( shown in liu et al .
1995 , fig .
1a and 1b ) corresponds to the structure at distances @xmath170 and 4.23 from a 6@xmath161 black hole .
cccccccc @xmath171&@xmath172 & @xmath173&@xmath174(k)&@xmath175&@xmath176&@xmath177 & @xmath178 + 10&-4.644&4.140&@xmath179&@xmath180&0.211 & 3.75 & @xmath181 + 9.7&-3.646&5.302&@xmath182&@xmath183&0.185 & 3.45 & @xmath184 + 9.3&-2.399&6.791&@xmath185&@xmath186&0.146 & 3.05 & @xmath187 + 9.0&-1.586&7.825&@xmath188&@xmath189&0.091 & 2.75 & @xmath190 + 8.9&-1.351&8.144&@xmath191&@xmath192&0.066 & 2.65 & @xmath193 + 8.8&-1.143&8.445&@xmath194&@xmath195&0.039 & 2.55 & @xmath196 + 8.7&-0.975&8.717&@xmath197&@xmath198&0.016 & 2.45 & @xmath199 + 8.5&-0.824&9.112&@xmath200&@xmath201&@xmath202 & 2.25 & @xmath203 + 8.3&-0.746&9.421&@xmath204&@xmath205&@xmath206&2.05 & @xmath207 + note : @xmath208 and @xmath131 vertical mass flow rate and pressure at the lower boundary @xmath159 , @xmath10 temperature reached at the upper boundary ( sound transition ) , @xmath209 rate of inward mass flow in the corona ( = evaporation rate ) ; @xmath176 fraction of mass carried away by the wind ; quantities @xmath210 and @xmath211 scaled to schwarzschild radius and eddington accretion rate .
@xmath212 in cgs units .
= @xmath150 of the thin disk for various distances @xmath1 ( in cm ) .
rate of inward mass flow @xmath209(in g / s ) in the corona (= evaporation minus wind loss ) , maximum temperature in the corona and @xmath213 ( h pressure scaleheight ) , solid lines . dashed line : virial temperature .
values of @xmath209 and @xmath1 are for a central mass of 6@xmath161 , scaling with @xmath214 and @xmath81 is indicated.,width=283 ] with distance from the black hole the coronal structure changes .
the main result concerns the evaporation rate , the shape of the function @xmath209 . for small values of @xmath1 the evaporation rate
@xmath209 increases with @xmath1 , then reaches a maximum value and decreases further outward . in fig .
3 we show the evaporation rate , the temperature reached at sound transition and the ratio of the pressure scaleheight @xmath215 to @xmath1 ( maximal temperature and pressure scaleheight from the computed vertical structure of the corona ) . in table 1
we summarize values of characteristic quantities . note the large variation in the values @xmath208 and @xmath131 , the values of @xmath92 and @xmath216 at the lower boundary .
the evaporation rate is given in @xmath217 and scaled to the eddington accretion rate ( @xmath218 ) , the radius in units of schwarzschild radius @xmath210 is added . for different black hole masses
, the evaporation rate can be scaled accordingly ( sect .
6 ) . to include the evaporation process in the evolution of the accretion disk , e.g. for black hole x - ray transients , the @xmath209-@xmath1 relation for values @xmath1 larger than the values in fig .
3 might be needed .
for such distances much larger than the hump location ( fig .
3 , log@xmath219 ) the function @xmath209 can be approximated as found in the earlier investigation by liu et al .
( 1995 ) for accretion in cataclysmic variables . in table 1
we also give the fraction of mass carried away by the wind .
note that this fraction is of the order 20% at large distances @xmath1 and decreases to very small values inside the maximum of the function @xmath209 ( fig .
the location of sound transition is about at the height 3@xmath1 for a distance @xmath165 cm , the hump location , and at increasingly larger heights further inward .
but their uppermost layers are unimportant for the energy balance in the corona because of their very low density
. disks around black holes can reach inward even to the last stable orbit , and a new feature appears , a maximum efficiency of evaporation . in dwarf nova systems
the inner disk edge disk lies much further out measured in schwarzschild radii .
the maximum of the evaporation efficiency has important consequences for the inward extension of the thin disk ( see sect .
this maximum is caused by the change in the physical process that removes the heat released by friction . for larger radii
the frictional coronal heating is balanced by inward advection and wind loss .
this fixes the coronal temperature @xmath10 at about 1/8 of the virial temperature @xmath220 ( @xmath221 ) .
downward heat conduction and subsequent radiation in the denser lower region play a minor role for the energy loss though they always establish the equilibrium density in the corona above the disk .
the ratio of the pressure scaleheight @xmath215 to @xmath1 is almost constant for larger @xmath1 and becomes small closer to the compact object .
black hole at log@xmath222=8.8 .
solid lines : gain by friction and vertical advective flow , dashed lines : loss by radiation , vertical heat flow and sidewise advective outflow ( flow out of the unit volume),width=294 ] . to illustrate the local energy balance we show in fig . 4 all contributions separately as functions of height @xmath2 .
there are two terms of energy gain , by friction and by vertical advective flow .
friction decreases with height with the pressure ( compare the coronal structure in fig .
the gain by vertical advection results from the fact that more mass enters from the layer below than leaves into the layer above .
this is due to the net sidewise outflow out of the unit volume .
this overcompensates the increase of enthalpy with height . concerning the losses , the radiation loss is the essential part in lower layers .
it is proportional to the square of density and decreases steeply towards higher @xmath2 .
the loss by vertical heat flow describes the effect that the heat , released by friction and advection in higher layers is conducted down to layers where it is radiated away .
when the temperature has dropped to @xmath223k at very low height the downward heat flow is then 10 times larger than the upward advective flow .
the balance is then between conduction and radiation .
this leads to the temperature profile calculated by shmeleva & syrovatskii ( 1973 ) and liu et al .
( 1995 ) and fixes the relation between @xmath138 and @xmath92 at this temperature used as boundary condition for our calculations . for log@xmath222=8.8 ( fig .
4 ) the height - integrated losses by advective outflow and by radiation are about equal .
this is a particular case . with distance @xmath1
the relative size of these two losses varies in a characteristic way . to show this we have integrated the two losses over the height of the ( flaring ) column and normalized by the total release of energy by friction .
5 shows the result .
the advective loss term includes the ernergy carried advectively away with the wind .
this latter contribution is also shown separately .
the maximal coronal mass flow rate is reached at that radius where the character of the solution changes from an advection - dominated flow at large radii to a radiation - dominated flow at small radii ( see fig .
( this is an intrinsic feature of the solutions as a dimensional analysis of the global energy balance equation can show ) .
this behaviour contrasts with that of optically thin accretion flows of constant mass flow rate @xmath209 .
these are advection - dominated ( adaf ) at small radii but become radiation - dominated ( shapiro , lightman and eardley 1976 ) at large radii ( narayan & yi 1995b ) . the coronal evaporation , on the other hand , allows for mass exchange between disk and corona and results in a @xmath209 varying with @xmath1 which results in an inversion of that behaviour .
this mass exchange also provides thermal stability to the coronal evaporation solution .
the system responds to a slight increase in temperature by an increased thermal flux that leads to an increase of density due to increased evaporation and thereby to a decrease of temperature by increased radiative cooling .
black hole),width=294 ]
if the mass flow rate in the disk is low , the evaporation rate in the inner disk region may exceed the mass flow rate .
the inner edge of the cool disk then is located where both rates are equal , that is where all matter brought in by the thin disk is evaporated . inside this location only a coronal flow exists .
this coronal flow provides the supply for the adaf towards the compact object .
the higher the mass flow rate in the cool disk the further inward is the transition to a coronal flow / adaf .
this is the situation in transient black hole x - ray binaries during quiescence .
if the mass flow rate exceeds the maximal evaporation rate , @xmath224 ( the peak coronal mass flow rate ) , the inner cool disk can not be completely evaporated anymore , instead it reaches inward to the last stable orbit . in high - mass x - ray binaries fluctuations of the mass flow rate
have the same effect . in fig .
6 we show how the inward extent of the standard thin disk depends on the mass flow rate in the thin disk .
as indicated there the corresponding spectra are hard , if the hot coronal flow / adaf dominates , or , soft , if the standard thin disk radiation dominates near the center .
this picture from our modelling is the same as the description of the accretion flow for different spectral states presented in the investigation of advection - dominated accretion by esin et al .
( 1997 , fig .
1 , except the very high state ) . as illustrated in fig .
5 the transition between dominant advective losses further out and dominant radiative losses further in causes the change in the slope of the evaporation rate at @xmath225= @xmath226 . except for the difference between the sub - virial temperature of the corona and the closer - to - virial temperature of an adaf of the same mass flow rate , a similar critical radius
is then predicted by the strong adaf proposal " ( narayan & yi 1995b ) . in general , however , the strong adaf proposal results in an adaf region larger than that which the evaporation model yields . a detailed discussion of these spectral transitions was given separately ( meyer et al .
2000 ) . as in fig .
extension of the standard thin disk for 3 different mass flow rates @xmath225 in the thin disk ( schematic ) .
note that the standard thin disk reaches inward towards the black hole if @xmath224 .
shown is also the type of spectrum , soft or hard , related to @xmath225,width=313 ] .
the main results from our model are the following .
the transition from a thin disk to an adaf should occur at about the same luminosity for different systems if their black hole masses are not too different .
the inner edge location should then either be at a distance of some schwarzschild radii ( the distance for which @xmath227 ) or the thin disk should reach inward to the last stable orbit .
several high - mass binaries and black hole transients show such a systematic behaviour ( tanaka & shibazaki 1996 , tanaka 1999 ) .
but in some cases the transition seems to be farther inward . in sect .
9 we compare our results with observations .
our equations and boundary conditions are invariant against scale transformations that leave temperature and velocities constant .
an invariant kepler velocity implies that @xmath228 is invariant as is @xmath229 and @xmath230 .
one thus obtains solutions for another central mass @xmath231 from those for @xmath232 by multiplying @xmath233 and @xmath209 values by @xmath234 , and by dividing @xmath235 and @xmath216 values by the same factor while keeping velocities and temperatures the same .
since @xmath236 are both proportional to mass , the relation between @xmath209 and @xmath1 is invariant if plotted in units of eddington accretion rate and schwarzschild radius . in these units
3 is universal and applies to both stellar size and supermassive black holes .
the properties of the thin disk underneath do not scale similarly .
this has no effect on the coronal structure , but shows up in the spectra belonging to a thin disk + corona for a given mass accretion rate .
in our modelling classical heat conduction is assumed . it is only valid if the conductive flux remains sufficiently small against the maximum transport by free streaming electrons .
the relevant ratio is estimated here as 0.16 at the radius of maximal mass flow , constant for smaller radii , and decreasing as @xmath237 for large radii .
all estimates are independent of the central mass .
smaller @xmath22-values increase the validity of the approximations .
our model assumes a one - temperature plasma .
temperature equilibrium between electrons and ions requires that the relevant collision timescale between both species remain small compared to the heating timescale .
their ratio is estimated as 0.27 where the coronal mass flow rate reaches its maximum value , and scales as @xmath238 .
thus at smaller radii this approximation brakes down and electrons and ions will develop different temperatures .
we have performed estimates on the size of these neglected effects .
synchrotron losses are proportional to the number density of electrons , their kinetic energy , and the square of the magnetic field strength .
it is often assumed that the magnetic field is dynamo produced and that its energy density is a fraction @xmath239 of the gas pressure .
the ratio of synchrotron losses to thermal plasma radiation then becomes a pure function of temperature rising with @xmath240 .
the main cooling in our solutions occurs in lower layers at about 1/2 of the final coronal temperature . under such circumstances ,
synchrotron radiation remains less than bremsstrahlung for @xmath241 4 .
values for @xmath242 in the literature include @xmath243 from energy considerations for adafs ( quataert & narayan 1999b ) and @xmath244 from direct numerical mhd calculations ( matsumoto 1999 , hawley 2000 ) .
compton cooling by photons from the underlying disk depends on the mass flow rate in the disk . for a rate equal to the maximal coronal mass flow rate ,
this compton cooling approaches about 1/3 of the frictional heating .
compton cooling ( and heating ) of the corona by radiation from the accretion center depends on the state . in the hard state , taking the example of cyg x-1 ( gilfanov et al .
2000 ) , the spectrum yields a compton equilibrium temperature of @xmath245 k. from the flux one estimates a compton energy exchange rate comparable with the frictional heating rate in the corona .
this indicates that the true coronal temperature at the peak of the coronal mass flow rate of fig .
3 could be somewhat smaller .
a similar result is obtained for compton cooling by the central flux during the soft state .
the kev - photons efficiently cool the corona heated by friction when the temperature obtained with the pure bremsstrahlung assumption is larger than an estimated @xmath245k .
all these cooling mechanism tend to reduce the peak temperature and mass flow rate in fig . 3 and shift the minimal transition radius somewhat outwards .
our estimates are only rough and a more quantitative evaluation of these effects has to be done later .
our model predicts a relation between the evaporation rate and the inner disk location ( function @xmath209 in fig . 3 and 6 ) .
this fact can be compared with observations in two ways : [ 1 ] for low mass flow rates in the thin disk a hole should form ; the location of the change from accretion via the thin disk to a coronal flow should be related to the mass flow rate as predicted ; [ 2 ] the spectral changes hard / soft and soft / hard should occur for a predicted mass flow rate and the location of the inner edge at this time should agree with that from the model . to perform test [ 1 ] we need to know two quantities , the mass flow rate in the thin disk and the inner disk edge location .
fitting of the spectrum based on the adaf model ( for a review see narayan et al .
1998 ) gives values of the mass accretion rate on to the black hole and therefore also the evaporation rate (= mass flow rate in the disk at the inner edge ) .
for the location of the inner edge constraints come from the @xmath246 emission line .
the maximum velocity indicates the distance of the inner disk edge from the black hole . for test [ 2 ]
the luminosity for which the spectral transition happens has to be considered , as well as signatures of the inner disk edge . for the x - ray transients
a0620 - 00 and v404 cyg narayan et al .
( 1996,1997a ) derived mass accretion rates based on spectral fitting of the adaf model .
these rates were used to perform test [ 1 ] . for
a0620 - 00 the resulting inner edge location agrees reasonably with the value from the @xmath246 emission line . for v404 cyg
a discrepancy was found ( meyer 1999 ) , but this disappears with the recent results for v404 cyg ( quataert & narayan 1999b ) and we find agreement also for this system . the comparison for gro j1655 - 40 using the rate from hameury et al .
( 1997 ) and outer - disk - stability arguments is satisfactory . for details of these comparisons
see liu et al .
( 1999 ) .
the agreement of results derived from the disk evaporation on one side and the adaf spectral fitting , including different physics for the description of the innermost region , on the other side , supports both models .
further agreement ( related to test [ 1 ] ) between observation and theoretical modelling of evaporation comes from the computed disk evolution of a0620 - 00 ( meyer - hofmeister & meyer 1999 ) .
the long lasting outburst makes plausible that only little matter is left over in the disk .
if this is the case the evolution during quiescence can be considered independent of the detailed outburst behaviour .
we took into account the evolution of the thin disk together with the corona .
the evaporation determines the inner edge of the thin disk at each time .
the mass flow rate there is important for the disk evolution and therefore for the outburst cycle ( without careful treatment of the inner edge boundary the evolution can not be determined adequately ) .
the evolution based on the same frictional parameter as taken for dwarf nova modelling ( @xmath247=0.05 ) .
leads to the onset of an outburst after the right time and the amount of matter stored in the disk in agreement with the energy estimated from the lightcurve .
concerning test [ 2 ] spectral transitions observed for a few transients can be considered .
the best observed source is nova muscae 1991 ( cui et al .
the transition occurs around @xmath248 erg / s ( ebisawa et al . 1994 ) .
these spectral transitions were modelled based on an adaf by esin et al .
( 1997 ) . our value for the critical mass accretion rate for a 6 @xmath161 black hole
, @xmath249g / s , corresponds to a standard accretion disk luminosity of about @xmath250 erg / s , in agreement with observations .
the accretion rates observed for high - mass x - ray binaries are higher than those for transients , fluctuations are common .
therefore hard / soft and soft / hard spectral transitions occur much more often than for transients where this phenomenon is only related to outburst onset or decline . in our investigation of spectral transitions ( meyer et al .
2000 ) we discuss the behaviour of the three persistent ( high - mass ) black hole x - ray sources lmc x-1 , lmc x-3 and cyg x-1 .
the transition occurs around @xmath248 erg / s as also for transients ( tanaka 1999 ) . for the spectral transitions of cyg x-1 again the adaf model was successfully used to describe the observations ( esin et al .
1998 ) .
our model predicts the inner edge of the thin disk at some 100 schwarzschild radii for the accretion rate related to the hump in fig .
6 , that is at the moment of spectral transition .
we have estimated the timescale on which a standard accretion disk at accretion rates near the critical value @xmath251 and at the related radius can evaporate or form newly and find a value of the order of days .
this agrees with the timescale for the observed spectral transition of a few days for cyg x-1 ( zhang et al .
frequency resolved spectroscopy of cyg x-1 ( revnivtsev et al .
1999 ) also leads to an inner edge of the thin disk at 150 schwarzschild radii .
but the existence of a reflecting component indicates a disk further inward ( gilfanov et al . 1998 , zycki et al . 1999 ) . for a sample of x - ray transients and high - mass x - ray binaries di matteo & psaltis ( 1999 ) found , if they relate the rapid variability properties to those of neutron stars , that the disk reaches down to about 10 schwarzschild radii .
this would be in contrast to the prediction from the adaf based modelling and our evaporation model .
further work is needed to understand these signatures of a thin disk so far inward during the hard state .
liu et al . ( 1999 ) used the results derived in the present paper to predict the location of the inner edge of the thin disk around m87 ( test [ 1 ] ) . using the accretion rate derived by ( reynolds et al .
1996 ) and an inner edge of the thin disk from emission lines broadening ( harms et al .
1994 ) they found good agreement between theory and observations . in connection with new observations
further applications of the evaporation model can be discussed ( liu , meyer and meyer - hofmeister , in preparation ) .
recently the concept of advection - dominated accretion was also applied to the low - luminosity active galactic nuclei m81 and ngc 4579 ( quataert et al .
1999 ) , ngc 4258 ( gammie et al .
1999 ) and low - luminosity elliptical galaxies ( di matteo et al .
1999 , 2000 ) .
it is not clear why for accretion rates which are different by about less than a factor of ten ( scaled to eddington accretion ) the thin disks are truncated at @xmath252 @xmath253 schwarzschild radii or reach inward to @xmath254 100 @xmath255 .
quataert et al .
( 1999 ) argued that the difference might be caused by a different way of mass supply for the adaf .
one might ask whether spectra computed according to our modelling could be used to test our predictions . such spectra would include radiation from an inner adaf region and an outer standard thin disk with the transition at a radius according to our @xmath256 relation .
the spectra determined according to the adaf model ( see e.g. narayan et al .
1998 ) use the observational constraints for the inner edge of the cool disk and fit the observed spectra .
if the derived rate @xmath209 and the inner disk location agree with our @xmath256 relation these already computed spectra are the spectra that our model will yield .
the magnetic nature of friction in accretion disks is now generally accepted , supported by numerical experiments that show the generation and sustainment of magnetic fields in accretion disks of sufficient electrical conductivity ( for a recent study see hawley 2000 ) .
this raises the question whether the neglect of magnetic fields in modelling the corona - disk interaction may not miss important effects .
such effects are twofold : ( a ) where gyro - frequency is large compared to collision frequency electron thermal conductivity is only high along the magnetic field
. this can significantly reduce effective thermal conductivity if the magnetic field is tangled .
this can be taken into account by reducing the conductivity coefficient @xmath257 by the average of @xmath258 over horizontal surfaces , where @xmath259 is the angle of the magnetic field with respect to the vertical .
( b ) magnetic pressure contributes to the pressure support .
this can be taken into account by increasing the value @xmath260 by a factor 1+@xmath239 in our equations , where the @xmath242 is the plasma parameter ( which may also include turbulent pressure ) .
we have performed an order of magnitude approximation to the individual terms in the global energy balance .
as we will we show in separate work this allows to model our numerical results and suggests the following dependence on the parameters @xmath261 and @xmath242 .
@xmath262 @xmath263 @xmath264 where @xmath265 is the saturation temperature reached at small radii and @xmath266 is the subvirial temperature attained at large radii .
the maximum mass flow rate occurs where the two temperatures @xmath266 and @xmath265 become about equal .
this yields the further proportionalities @xmath267 @xmath268 this may serve as a guide on how the results depend on the choice of these parameter values . for values of @xmath242 discussed in the literature see sect .
8 . for our numerical computations we took @xmath22=0.3 .
values discussed in the literature include 0.1 to 0.3 from spectral fits of adaf models to soft x - ray transients ( narayan et al .
1996 ) , @xmath22=0.25 obtained from applying detailed calculations to luminous black hole x - ray binaries ( esin et al .
1997 ) , and @xmath22=0.1 obtained from global magnetohydrodynamical simulations of accretion tori ( hawley 2000 ) .
the above scaling relation shows how our results change if other @xmath22-values are chosen .
x - ray irradiation can cause a corona above the thin disk ( e.g. london et al . 1981 ) .
this can interfere with the assumptions underlying the present model .
we note that optically thick scattering layers in the inner corona can prevent such radiation to reach the disk surface ( e.g. see schandl & meyer 1994 ) .
this can become a complex situation which requires a detailed analysis to clarify under what circumstances this affects the applicability of our model . in general , models like ours should be good for black hole systems with adaf type internal accretion
. aspects of formation of a disk corona have been investigated by de kool & wickramasinghe ( 1999 ) .
these calculations are more detailed with respect to ionization equilibrium than ours , but do not include thermal conduction , sidewise advection and wind loss in the energy balance .
it is therefore difficult to compare with our results .
here we have considered the spatial co - existence of disk and corona and its gradual fading into a corona / adaf only region at its inner boundary .
honma ( 1996 ) in an apparently very different approach considered the effect of a turbulent diffusive heat flow outwards from a hot and mainly non - radiative advection - dominated inner region to an outercool accretion disk . in his simplified modelling
both regions are treated as one disk with radially varying thickness , and optical depth .
he found a steep transition in the disk thickness and temperature at a radius that depends on the mass flow rate . in this radially thin transition layer
most of the systems luminosity is radiated .
though at first sight very different , the two approaches have a basic physical similarity , though in an inverted geometry : what is vertical in the disk+corona system is radial in honma s approach .
heat generated by friction is conducted outward ( downward ) from a hot inner ( upper ) region to a cool outer ( lower ) region where it is radiated away . in both cases
this establishes the condition where the cool disk ends and the hot , adaf type flow takes over .
we note that honma s diffusive heat conductivity bears resemblance to the thermal electron heat conductivity .
the rapid increase of the cooling with density leads to a similarly steep temperature profile in honma s model as at the base of the corona in the disk+corona model .
we have calculated the ratio of honma s effective turbulent heat conductivity to that of electron thermal heat conductivity and obtain the value 1/6 for pressure and temperature at the peak of the coronal mass flow rate in table 1 .
we think that the evaporation model is a more realistic description of the transition from the cool disk to an corona / adaf only flow . in sect .
2.5 we estimated the radial heat inflow and showed that it is a minor contribution . generally , the turbulent outward diffusion of specific energy reduces the inward advective transport ( which is also diffusive ) of energy .
thus the inward flow of energy is somewhat less than the inward flow of mass times the specific energy .
our approximate treatment of the inward flows of mass and energy in the derivations of eqs .
( 39 ) and ( 46 ) in effect leads to such a distinction , for radial power laws as discussed by a factor of 2 . in spite of the very different geometry and simplification honma s model
captures the same physical effect .
his conclusion that his results shed light on the soft - hard transitions of black hole accretors is thus substantiated by the disk+corona model , though quantative differences exist . in sect .
5 we have already remarked on where the `` strong adaf proposal '' agrees with and where it deviates from the results of the disk+corona model .
this brings the `` strong adaf proposal '' , honma s diffusive heat transport model , and the disk+corona model under a common physical aspect and makes a surprising similarity of some of their results results understandable .
our model describes the interaction between the geometrically thin accretion disk and the corona above .
we have derived simplified equations , which include the relevant processes determining the accretion via the cool disk together with the hot coronal flow which yields the matter supply for the adaf towards the black hole .
we determine the structure of the corona which is in equilibrium with the cool disk .
the equilibrium demands the evaporation of a certain amount of matter from the cool disk and we compute the evaporation rate as a function of the distance @xmath1 from the central black hole .
this gives the rate of accretion of matter towards the inner region ( part of the matter is lost by a wind ) . the evaporation efficiency becomes higher with decreasing @xmath1 , but reaches a maximum at some hundred schwarzschild radii . for low accretion rates in the cool disk
the evaporation uses up all matter at the distance where the evaporation rate is equal to the mass flow rate in the cool disk .
the standard thin disk emits a near black body spectrum and , for high enough accretion rates the disk extends inward to the last stable orbit , the total radiation is soft .
but if , for lower accretion rates the thin disk is truncated , the coronal flow / adaf in the inner region provides a hard spectrum . these features of disk evaporation apply to galactic black hole x - ray binaries ( x - ray transients , high - mass x - ray binaries ) and also to agn ( truncated disks were found in low - luminosity agn and elliptical galaxies ) .
the mass accretion rates derived by spectral fits based on the adaf model and the location of the inner disk ( deduced from observations or determined by the fitting ) allow to check the predictions of the model for the location of transition to the coronal flow .
these tests show good agreement for most cases of x - ray transients and the features connected with the maximum evaporation efficiency ( see sect . 9 ) .
we want to point out that our model gives qualitative results for relations between mass accretion rate , location of disk truncation and spectral transitions rather than exact values .
the division in an outer advection - dominated regime and an inner radiation - dominated regime of the coronal structure gives insight to basic features .
computed values for the edge location might differ by a factor of two to three from the " true values .
an investigation of the detailed effect of simplifications in the equations as well as the effect of the @xmath22-value and of magnetic fields will be carried out later .
a truncation of the thin disk farther in than 100 schwarzschild radii is in contrast to the predictions of our model .
the reflecting component observed for cyg x-1 seems to indicate such an inward extent of the thin disk .
but , also for cyg x-1 , another investigation of observations leads to an inner edge location near 100 schwarzschild radii ( see sect .
this situation needs further clarification . but also from the theoretical side the possible existence of an `` interior '' thin disk below a corona in the innermost regions ( closer in than the location of the evaporation efficiency maximum , compare fig .
3 ) needs further investigation . for agn it was found , based on adaf spectral modelling that the truncation of the thin disk can be at very different distances from the supermassive black hole for mass accretion rates which differ much less ( see sect . 9 ) .
our model only predicts the large distances . also here
additional investigations are needed to understand the physical reason for such divergent results .
_ note added after submission of the manuscript : in a parallel investigation , very similar to our approach , raska & czerny ( astro - ph/0004158 , a&a in press ) have also studied the coronal structure and the transition to an adaf .
they take into account the possibility of different temperatures of electrons and ions and include compton cooling by the soft radiation from the underlying disk . in our computed models for the parameters @xmath269 , that means log @xmath270 ( compare table 1 ) , the energy equilibration time between electrons and ions is shorter than the heating time , thus establishing temperature equilibrium between electrons and ions . for the same parameters also
the rate of compton cooling by the radiation from the underlying disk is small compared to the frictional heating rate and thus negligible . _
abramowicz m.a . , czerny b. , lasota j.p .
, 1988 , apj 332 , 646 abramowicz m.a .
, chen x. , kato s. et al . , 1995 ,
apj letters 438 , l37 begelman m.c . , 1978 , mnras 243 , 610 chen x. , 1995 , mnras 275 , 641 chen x. , abramowicz m.a .
, lasota j.p . , narayan r. , yi i. , 1995 , apj 443 , l61 chen x. , abramowicz m.a . , lasota j.p . , 1997 , apj 476 , l61 cui w. , zhang s.n . , focke w. et al . , 1997 , apj 484 , 383 de kool m. , wickramasinghe d. , 1999 , mnras 307 , 449 di matteo t. , psaltis d. , 1999 , apj 526 , 101 di matteo t. , fabian a.c .
, rees m.j .
et al . , 1999 ,
mnras 305 , 492 di matteo t. , quataert e. , allen s.w .
et al , 2000 , mnras 311 , 507 ebisawa k. , ogawa , m. , aoki t. et al . , 1994 , pasj 46 , 375 esin a.a . ,
mcclintock j.e , narayan r. , 1997 , apj 489 , 865 esin a.a . ,
narayan r. , cui w. et al .
, 1998 , apj 505 , 854 gammie c.f . , narayan r. , blandford r. , 1999 , apj 516 , 177 gilfanov m. , churazov e. , sunyaev r. , 1998 , in : 18th texas symposium on relativistic astrophysics and cosmology , eds .
olinto et al .
; world scientific , p.735 gilfanov m.r . ,
churazov e.m . , revnivtsev m. , 2000 , mnras in press hameury j .-
, lasota j .-
, mcclintock et al . , 1997 , apj 489 , 234 harms r.j . ,
ford h.c . ,
tsvetanov z.i .
et al . , 1994 , apj 435 , l35 hawley , j. , 2000 , apj 528 , 462 honma f. , 1996 , pasj 48 , 77 ichimaru s. , 1977 , apj 214 , 840 katz j. , 1977 , apj 215 , 265 kley w. , lin d. , 1992 , apj 397 , 600 liu f.k . , meyer f. , meyer - hofmeister e. , 1995 , a&a 300 , 823 liu b.f .
, yuan w. , meyer f. et al .
, 1999 , apj 527 , l17 london r. , mccray r. , auer l.a . , 1981 , apj 243 , 970 lynden - bell d. , pringle j.e . , 1974 , mnras 168 , 603 matsumoto r. , 1999 , in : numerical astrophysics , eds
. s. miyama , k. tomisaka , t. hanawa , kluwer , dordrecht , p. 195
mcclintock j.e . ,
horne k. , remillard r.a . , 1995 , apj 442 358 meyer f. , 1999 , in : proc . of disk instabilities in close binary systems , eds .
s. mineshige and j.c .
wheeler , universal academic press , kyoto , p.209 meyer f. , meyer - hofmeister e. , 1994 , a&a 288 , 175 meyer f. , liu b.f . , meyer - hofmeister e. , 2000 , a&a 354 , l67 meyer - hofmeister e. , meyer f. , 1999 , a&a 348 , 154 mineshige s. , liu b.f . , meyer f. et al .
, 1998 , pasj 50 , l5 nakamura k.e . , matsumoto r. , kusunose m. , kato s. , 1996 , pasj 48 , 76 nakamura k.e . , kusunose m. , matsumoto r. , kato s. , 1997 , pasj 49 , 503 narayan r. , yi i. , 1994 , apj 428 , l13 narayan r. , yi i. , 1995a , apj 444 , 231 narayan r. , yi i. , 1995b , apj 452 , 710 narayan r. , mcclintock j.e .
, yi i. , 1996 , apj 457 , 82 narayan r. , barret d. , mcclintock j.e . , 1997a , 482 , 448 narayan r. , kato s. , homna f. , 1997b , apj 476 , 49 narayan r. , mahadevan r. , quataert e. , 1998 , in : the theory of black hole accretion discs , eds .
abramowicz et al . , cambridge university press , p.148 novikov i.d . , thorne k.s . ,
1973 in : black holes , eds . c. dewitt and b. dewitt , gordon & breach , p.343 piran t. , 1978 , apj 221 , 652 pringle j.,1981 , ann .
137 quataert e. , narayan r. , 1999a , apj 516 , 399 quataert e. , narayan r. , 1999b , apj 520 , 298 quataert e. , di matteo t. , narayan r. , 1999 , apj 525 , l89 raymond j.c . ,
cox d.p . , smith b.w . , 1976 ,
apj 204 , 290 rees m.j . ,
begelman m.c .
, blandford r.d .
et al . , 1982 ,
nat 295 , 17 revnivtsev m. , gilfanov m. , churazoc e. , 1999 , a&a 347 , l26 reynolds c.s . , di matteo t. , fabian a.c .
et al . , 1996 , mnras 283
, l111 schandl s. , meyer f. , 1994 , a&a 289 , 149 shakura n.i .
, sunyaev r.a .
, 1973 , a&a 24 , 337 , 1995 , apj 446,741 shapiro s.l . , lightman a.p . , eardly d.m . , 1976 , apj 204 , 187 shmeleva d.p . , syrovatskii s.i . , 1973 , solar physics 33 , 341 sommerfeld a. , 1949 , mechanik der deformierbaren medien , akademische verlagsgesellschaft geest & portig , leipzig spitzer l. , 1962 , physics of fully ionized gases , 2nd edition , interscience publ
. , new york , london tanaka y.,1999 , in : proc . of disk instablities in close binary systems , eds .
s. mineshige and j.c .
wheeler , universal academic press , kyoto , p.21 tanaka y. , shibazaki n. , 1996 , ara&a , 34 , 607 zhang s.n . , cui w. , harmon b.a .
et al . , 1997 ,
apj 477 , l95 zycki t. , done c. , smith d.a . , 1999 , mnras 305 , 231 | we present a model for a corona above a geometrically thin standard disk around a black hole .
this corona is fed by matter from the thin disk which evaporates from the cool layers underneath .
an equilibrium establishes between the cool accretion stream and the hot flow . in the inner region evaporation
becomes so efficient that at low accretion rates all matter flows via the corona and proceeds towards the black hole as an advection - dominated accretion flow ( adaf ) .
we investigate this transition for accretion disks around stellar black holes .
we derive a set of four ordinary differential equations which describe the relevant processes and solve them numerically .
the evaporation efficiency has a maximum value where the advection - dominated structure of the corona ( at large distances from the black hole ) changes to a radiation - dominated structure ( at small distances ) .
this maximum can be related to the observed spectral transitions in black hole x - ray binaries .
our solutions of coronal structure are valid for both , accretion in galactic x - ray transients and in agn . |
30 buisson o , gandit p , rammal r , wang y y and pannetier b 1990 a * 150 * 36 ; benoist r and zwerger w 1997 b * 103 * 377 ; schweigert v a and peeters f m 1998 b * 57 * 13817 ; jaddalah h t , rubinstein j and sternberg p 1999 2935 ; yampolskii s v and peeters f m 2000 b * 62 * 9663 van bael m j , temst k , moshchalkov v v and bruynseraede y 1999 b * 59 * 14674 ; van bael m j , van look l , lange m , bekaert j , bending s j , grigorenko a n , moshchalkov v v and bruynseraede y 2002 _ physica _ c * 369 * 97 bespyatykh yu i and wasilevski w 2001 _ fiz .
tela _ ( leningrad ) * 43 * 215 [ 2001 _ sov .
phys.solid state _ * 43 * , 224 ] ; bespyatykh yu i , wasilevski w , gajdek m , nikitin i p and nikitov s a 2001 _ fiz .
( leningrad ) * 43 * 1754 [ 2001 _ sov .
phys.solid state _ * 43 * 1827 ] | within the phenomenological ginzburg - landau theory we investigate the phase diagram of a thin superconducting film with ferromagnetic nanoparticles .
we study the oscillatory dependence of the critical temperature on an external magnetic field similar to the little - parks effect and formation of multiquantum vortex structures .
the structure of a superconducting state is studied both analytically and numerically . little - parks effect @xcite , i.e. oscillations of the critical temperature @xmath0 of multiply connected superconducting samples in an applied magnetic field @xmath1 ,
is one of the striking phenomenon demonstrating coherent nature of superconducting state .
such oscillatory behaviour of @xmath2 is not a distinctive feature of a superconducting thin
wall cylinder and can be observed also in superconductors with columnar defects and holes ( see references @xcite ) and mesoscopic simply
connected samples of the size of several coherence lengths @xcite .
generally , the oscillations of @xmath0 with a change in the external magnetic flux are caused by the transitions between the states with different vorticities ( winding numbers ) characterizing the circulation of the phase of the order parameter . for a system with cylindrical symmetry
the vorticity parameter just coincides with the angular momentum of the cooper pair wave function .
the states with a certain angular momentum @xmath3 can be considered as @xmath3-quanta vortices .
experimental and theoretical investigations of these exotic vortex structures ( multiquanta vortices and vortex molecules ) in mesoscopic superconductors have attracted a great deal of attention .
as we change an external homogeneous magnetic field multiquanta vortices and vortex molecules can transform one into another via first or second order phase transitions .
in this paper we focus on another possibility to create multiquantum vortex states : nucleation of superconducting order parameter in a hybrid system consisting of a thin superconducting film and an array of magnetic nanoparticles .
the interest to such structures is stimulated by their large potential for applications ( e.g. , as switches or systems with a controlled artificial pinning ) .
the enhancement of the depinning critical current density @xmath4 has been observed experimentally for superconducting films with arrays of submicron magnetic dots @xcite , antidots @xcite , and for superconductor ferromagnet ( s / f ) bilayers with domain structure in ferromagnetic films @xcite .
the matching effects observed for magnetic and transport characteristics were explained in terms of commensurability between the flux lattice and the lattice of magnetic particles .
vortex structures and pinning in the s / f systems at rather low magnetic fields ( in the london approximation ) have been analysed in papers @xcite .
provided the thickness of a superconducting film is rather small as compared with the coherence length , the critical temperature of the superconducting transition as well as the structure of superconducting nuclei should be determined by a two dimensional distribution of a magnetic field component @xmath5 ( perpendicular to the superconducting film plane ) induced by the ferromagnetic particles .
obviously , the highest critical temperature corresponds to the nuclei which appear near the lines of zeros of @xmath6 due to the mechanism analogous to the one responsible for the surface superconductivity ( see , e.g. , reference @xcite ) and domain wall superconductivity @xcite .
provided these lines of zeros have the shape of closed loops , the winding number of a superconducting nucleus will be determined by the magnetic flux through the loop .
thus , changing this flux ( e.g. , increasing an external @xmath1 field applied along the @xmath7 axis ) we can control the winding number .
the resulting phase transitions between the multiquantum states with different @xmath3 can cause the oscillations of @xmath0 .
such oscillatory behaviour has been , in fact , observed in reference @xcite for a nb film with an array of gdco particles .
note that a change in a slope of the phase transition curve @xmath2 ( which is probably a signature of the transition discussed above ) has been also found in reference @xcite for a pb film with copd particles . provided the dimensions of the sample in the @xmath8 plane are compared with the coherence length , we can expect a rather complicated picture which is influenced both by the sample edges and by the distribution of an inhomogeneous magnetic field . for a model step - like profile of the magnetic field the resulting phase transitions between different types of exotic vortex states in a mesoscopic disc have been studied numerically in reference @xcite .
the interplay between the boundary effects and magnetic field inhomogeneity influences also the formation of multiquantum vortex states around a finite size magnetic dot embedded in a large area superconducting film @xcite .
the transitions between different multiquantum vortex states with a change in magnetic field and magnetic dot parameters were studied in references @xcite for certain temperature values .
these effects are closely related to the ones observed in mesoscopic and multiply - connected samples , and consequently , we can expect that the oscillations of @xmath2 ( analysed below ) should also be a common feature of multiply - connected superconductors and thin - film systems with magnetic dots . in this work
we do not consider the magnetic phase transitions in the mixed state for @xmath9 and focus on the oscillatory behaviour of @xmath2 in a large area superconducting film caused only by the quantization associated with the characteristics of the inhomogeneous magnetic field produced by ferromagnetic particles .
we neglect the influence of the edge and proximity effects in the s / f system and consider a nanoparticle only as a source of a small - scale magnetic field .
our further consideration is based on the linearized ginzburg - landau model : @xmath10 here @xmath11 is the order parameter , @xmath12 is the vector potential , @xmath13 , @xmath14 is the magnetic flux quantum , @xmath15 is the coherence length , and @xmath16 is the critical temperature of the bulk superconductor at @xmath17 . for the sake of simplicity we neglect the interference effects between the superconducting nuclei appearing near different nanoparticles ( i.e. assume the interparticle distance to be rather large as compared with the superconducting nucleus size ) and consider a single magnetic particle with a fixed magnetic moment chosen perpendicular to the film plane @xmath18 . for a rather thin film ( of the thickness less than the coherence length ) we can neglect the influence of the field components @xmath19 in the film plane and consider an axially symmetrical two dimensional problem ( [ gl - eq ] ) in the field @xmath20 , where @xmath21 is the @xmath7component of the field induced by the ferromagnetic particle and @xmath22 is a cylindrical coordinate system . choosing the gauge @xmath23 one can find the solution of the equation ( [ gl - eq ] ) in the form @xmath24 , where @xmath3 is the vorticity , and @xmath25 for @xmath26 .
the function @xmath27 should be determined from the equation : @xmath28g_m = \frac{1}{\xi^2(t)}g_m \ .\ ] ] here @xmath29 is the total flux through the circle of radius @xmath30 .
the lowest eigenvalue @xmath31 of the schrdinger - like equation ( [ schr ] ) defines the critical temperature @xmath0 of the phase transition into a superconducting state .
obviously , for rather small fields @xmath1 the superconducting order parameter can nucleate either far from the magnetic particle ( @xmath32 ) where the critical temperature @xmath33 is defined by the homogeneous field @xmath34 or in the region close to the circle of the radius @xmath35 where @xmath36 and @xmath0 is controlled by the slope of @xmath37 at @xmath38 and by the flux through the area of the radius @xmath35 .
in the first case we obtain @xmath39 . for the second case we can analyse the behaviour of @xmath2 assuming that the characteristic length scale @xmath40 of the order parameter nucleus is much less than the characteristic scale of the magnetic field distribution . within such local approximation ( similar to the one used in reference @xcite for the description of domain wall superconductivity ) we can expand the flux in powers of the distance from @xmath35 : @xmath41
this local approximation is valid under the following conditions : @xmath42 introducing a new coordinate @xmath43 we obtain the dimensionless equation @xmath44 where the parameters @xmath45 and @xmath46 are given by the expressions : @xmath47{\frac{\phi_0}{\pi \left|b_{\rm z}^\prime(r_0)\right| } } \ , \quad q = -\left(\frac{\phi(r_0)}{\phi_0}+m\right ) \sqrt[3]{\frac{\phi_0}{\pi r_0 ^ 3 b_{\rm z}^\prime(r_0 ) } } \ .\ ] ] we obtain @xmath48 when @xmath49 , and @xmath50 when @xmath51 .
the minimal value of @xmath52 is @xmath53 at @xmath54 .
the final expression for the critical temperature reads : @xmath55{\frac{\phi_0}{\pi r_0 ^ 3 b_{\rm z}^\prime(r_0)}}\right ) + o\left(\frac{\ell^2}{r_0 ^ 2}\right ) \right ] \ .\ ] ] the superconducting nuclei are localized near the ferromagnetic particle at a distance @xmath35 .
the states with different energetically favorable winding numbers @xmath3 correspond to the multiquantum vortex structures very similar to the ones observed in a mesoscopic disc .
as we change an external field @xmath1 , we change the flux @xmath56 and , thus , change the energetically favorable vorticity number and position of the nucleus . to investigate the details of the oscillatory behaviour discussed above we consider a particular case of a small ferromagnetic particle which can be described as a point magnetic dipole with a magnetic moment @xmath57 placed at a height @xmath58 over the superconducting film .
the corresponding expressions for the field and the vector potential are : @xmath59 introducing @xmath60 and a dimensionless coordinate @xmath61 we obtain the equation ( [ schr ] ) in the form : @xmath62 + \frac{m}{\rho}\right)^2 f_m = \frac{h^2}{\xi_0 ^ 2}\left(1-\frac{t_{\rm c}}{t_{\rm c0}}\right)f_{m } \ , \ ] ] where @xmath63 is the dimensionless flux through the area with the positive field @xmath21 and @xmath64 . in the limit of small fields @xmath65
the nucleation of superconductivity occurs at large distances @xmath66 and critical temperatures for different winding numbers @xmath3 are very close .
thus , in this limit the critical temperature is equal to @xmath33 and is not sensitive to the presence of the dipole .
below this temperature we obtain a lattice of singly quantized vortices ( with the concentration determined by @xmath1 ) which is surely disturbed under the dipole .
note , that the behaviour of @xmath67 in this low field regime should modify provided we take account of a finite distance between the magnetic particles . for large absolute values
@xmath68 ( much larger than the maximum field induced by the dipole ) we obtain the following asymptotical behaviour of @xmath0 : @xmath69 for negative @xmath1 and @xmath70 for positive @xmath1 values ( here @xmath71 is the minimum of the dipole field ) .
the superconductivity nucleates near the minima of the total field @xmath72 and , thus , is localized near the dipole . in the intermediate field region ( @xmath73 )
we should expect the oscillatory behaviour of @xmath0 discussed above .
the number of oscillations is controlled by the parameter @xmath74 .
we have carried out the numerical calculations of equation ( [ numer ] ) for the various @xmath74 values .
for the numerical analysis of the localized states of equation ( [ numer ] ) we approximated it on a equidistant grid and obtained the eigenfunctions @xmath75 and eigenvalues by the diagonalization method of the tridiagonal difference scheme .
the results of these calculations as well as the analytical dependence of @xmath0 given by the expression ( [ ana ] ) are shown in figure [ fig1ab ] .
we observed a remarkable asymmetry of the phase transition curve ( @xmath76 ) which is caused by the difference in the distributions of positive and negative parts of the dipole field @xmath21 : the maximum positive field ( @xmath77 ) is much larger than the absolute value of the minimum negative field ( @xmath78 ) . as a result ,
the @xmath0 oscillations appear to be most pronounced for negative @xmath1 which compensate the positive part of the dipole field .
taking @xmath79 ( for a ferromagnetic particle with dimensions @xmath80 and magnetization @xmath81 ) , @xmath82 we obtain @xmath83 , @xmath84 and the characteristic scales of @xmath0 oscillations @xmath85 , @xmath86 for a nb film with @xmath87 and @xmath88 .
we expect that the oscillatory behaviour of @xmath0 can be observable , e.g. , in magnetoresistance measurements of thin superconducting films with arrays of ferromagnetic particles .
the superconducting nuclei localized near the particles should result in the partial decrease in the resistance below the oscillating @xmath2 .
as we decrease the temperature below @xmath2 the supeconducting order parameter around a single particle becomes a mixture of angular harmonics with different @xmath3 values and we can expect the appearance the phase transitions similar to the ones discussed in papers @xcite . with the further decrease in temperature the whole film becomes superconducting and resistivity turns to zero .
we would like to thank a. i. buzdin , a. a. fraerman , yu . n. nozdrin , and i. a. shereshevskii for stimulating discussions .
this work was supported , in part , by the russian foundation for basic research , grant no . 03 - 02 - 16774 , russian academy of sciences under the program quantum macrophysics , russian state fellowship for young doctors of sciences ( md-141.2003.02 ) , university of nizhny novgorod under the program brhe and physics of solid state nanostructures. |
the research on quantum many - body system is an actual topic in physics .
experiments on mesoscopic quantum systems together with updated numerical techniques for quantum many - body theory has produced and puts light on questions related to fundaments of quantum many - body physics and the relationship between microphysics and thermodynamics @xcite . from the theoretical point of view
, we still need analytical theory to clarify questions which answers can not be given with a non analytical theories . in this way
the algebraic formulation of bethe ansatz together with quantum inverse scattering method has been an important tool to propose new theoretical integrable models @xcite .
they can be used to explain some experimental features and also to propose new experiments to investigate new physical phenomena in a mesoscopic scale . for bosonic systems
an important theoretical and integrable model is the two site bose - hubbard model ( _ canonical josephson hamiltonian _ ) @xcite .
the model has been used to understand tunneling phenomena and it explains the qualitative experimental behaviour of bose - einstein condensates in a double well potential system @xcite . in recent literature models with more number of modes have been proposed .
non integrable models with three modes @xcite , four modes @xcite and five modes @xcite have appeared in the literature in the last decades .
however , integrable versions for the above models and even with more number of modes are welcomed because they could give new insights to broad our understanding in the quantum integrability area and consequently they could yield the possibilities to investigate new and opened questions related to the fundamental physics where thermodynamics effects are not taking into account .
for example , in the reference @xcite was speculated that coherent backscattering in ultracold bosonic atoms might act as some sort of precursor to many - body localization .
however , the authors pointed out that more research appears to be useful in order to elucidate those speculative questions . in reference
@xcite was proposed and solved a simple double - well model which incorporates many key ingredients of the disordered bose - hubbard model .
it is believed that the minimal model could be a simple theoretical paradigm for understanding the details of the full disordered interacting quantum phase diagram . that way ,
going beyond the two modes models , it allows us to take into account new effects , like the long - range effect , opening the possibility to do new theoretical investigations needed for the full understanding of the disordered bose - hubbard model . on the other hand side , in order to helping to elucidate the above questions , in the last year it was presented an integrable version for a model with four modes @xcite and more recently a generalized integrable version for a multi - mode models were presented in the references @xcite opening a great opportunity to do new theoretical investigations . in this work
we are interested to presenting the simplest and integrable model with three modes which can be used as a toy model for understanding quantum control devices operation .
the model we show here is a particular case of a more general integrable model @xcite belonging to the family of integrable models presented in the references @xcite .
we believe the model we are presenting here can be useful to put light on subjects related to _ atomtronic _ @xcite and also in quantum manipulation via external fields @xcite .
the rest of the paper is organized as follows . in section
@xmath0 we present the integrable model and its physical meaning .
section @xmath1 is devoted to the classical analysis for the equivalent quantum model where the hamilton s equations , the threshold of discontinuous phase transition and the classical dynamics are presented . in section @xmath2 ,
the quantum dynamics is carried out . in section @xmath3
is devoted for our discussions and overview and finally in section @xmath4 is reserved for our conclusions .
recently the integrable hamiltonian for three aligned well was presented in the reference @xcite and it takes the following hamiltonian @xmath5 where the coupling @xmath6 is the intra - well and inter - well interaction between bosons , @xmath7 is the external potential and @xmath8 , @xmath9 , is the constant couplings for the tunneling between the wells .
there in the reference @xcite it was shown that there are three independent conserved quantities @xmath10 , @xmath11 , the energy @xmath12 and also an important dependent conserved quantity @xmath13 , the total number of bosons in the system . for each fixed value of @xmath14 ,
the dimension @xmath15 of hilbert space associated to the model is given by @xmath16 .
in this paper we will focus on a particular case of the above hamiltonian . setting @xmath17
then the hamiltonian takes the simple form @xmath18 up to an irrelevant constant @xmath19 .
we observe that the independent conserved quantities now take the simple form @xmath20 and @xmath21 .
the above model is basically a model with two wells ( well @xmath22 and well @xmath0 ) interacting with a third well ( well @xmath1 ) . in fig .
[ fig0 ] we show a schematic representation for the model . as the population in the wells @xmath22 and @xmath0
must be conserved and there is tunneling between them , we expect that flow of bosons between wells @xmath22 and @xmath0 can be controlled by the population of the well @xmath1 which is also a conserved quantity . in order to do a parallel with the well - known two site bose - hubbard model @xcite , we first performe the coupling mapping @xmath23 and defining @xmath24 where @xmath7 is the habitual external potential .
so we have now the hamiltonian @xmath25 up to irrelevant constant @xmath26 .
now we observe that @xmath27 works like an _ effective external potential _ that depends on the population in the well @xmath1 which turns on the interwell interactions between wells @xmath22-@xmath1 and wells @xmath0-@xmath1 .
physically the above model , as the apparent similarity with the two well bose - hubbard model , has a different physical interpretation .
here we have a three well model where a two well bose - hubbard model ( well @xmath22 and well @xmath0 with the tunneling possibility between them ) interacting with a third one ( well @xmath1 ) without the possibility of tunneling between them , that is no tunneling between wells @xmath22-@xmath1 and wells @xmath0-@xmath1 .
later we will see that the presence of these interwell interactions will always _ induce _ the localization in the dynamics . before go ahead in the quantum analysis
, firstly we take a look in the classical analogue of the model ( [ h2n2 ] ) .
doing @xmath28 and also the change of variables @xmath29 where @xmath30 is the _ population imbalance _ between wells @xmath22 and @xmath0 and @xmath31 is the phase difference .
so the classical analogue of the model above is @xmath32 with the identification @xmath33 where @xmath34 and @xmath35 is the _ relative population _ between well @xmath1 and wells @xmath22-@xmath0 .
the classical dynamics can be computed through _
hamilton s equations _ @xmath36 defining @xmath37 and @xmath38 and using the conservation of energy , then for any initial condition we have @xmath39 in what follows we consider only the initial state @xmath40 . now
the equation ( [ ehg ] ) @xmath41 \label{thr}\ ] ] must be holded . from this equation
is possible to obtain the threshold couplings for the population imbalance dynamics between wells @xmath22-@xmath0 .
firstly , we will analyse the behavior of the function @xmath42 in the allowed physical interval . in fig.[fig1 ] clearly the dynamics of the system can be localized ( @xmath43 $ ] ) or delocalized ( @xmath44 $ ] ) , with @xmath45 . to find the relationship between the parameters for the threshold curve
, we consider @xmath46 , so the above equation becomes @xmath47 as the straight line @xmath48 is tangent to the curve @xmath49 in some point @xmath50 , then we have that @xmath51 .
now doing @xmath52 we find the threshold coupling point , with the coordinates @xmath53 and @xmath54 , as a parametric equations of @xmath50 : @xmath55 the above equations gives @xmath56 .
clearly there is a minimum point for @xmath57 ( maximum point for @xmath58 ) in the threshold curve .
the minimal ( maximal ) point has coordinates given by @xmath59 by using eq .
( [ tt ] ) we can trace out the threshold curve . these results together with the some solutions of eq .
( [ zaa ] ) are shown in fig .
[ fig12 ] .
as there is a symmetry in the diagram of parameter , hereafter we just consider the case @xmath57 .
for @xmath60 , the threshold curve separates the diagram of parameters in two regions : * region i : for @xmath61 , the dynamics between wells @xmath22-@xmath0 is localized . *
region ii : for @xmath62 : the dynanics between wells @xmath22-@xmath0 is delocalized ; it is worth to highlight for the case @xmath60 that the passing through the region i to the region ii is abrupt configuring the threshold in the classical dynamics . on the other hand side , for @xmath63
, there is not a threshold and the classical dynamics changes smoothly .
these results are shown in fig .
[ fig - abrupt ] . as a function of @xmath64 , for @xmath65 ( solid black line ) , @xmath66 ( dashed red line ) , @xmath67 ( dotted - dashed blue line ) and @xmath68 ( dotted green line ) .
here @xmath69 , where @xmath70 is the set of solutions of equations @xmath71 .
the dashed line is @xmath72.,width=377,height=188 ] integrating the above hamilton s equations we can see how the classical system is evolving .
recall that @xmath73 .
thus , if the system is initially in the delocalized dynamics regime ( region ii ) with @xmath74 and @xmath75 , by increasing the relative population @xmath76 , the system will eventually cross the threshold curve to get into the region i and then the dynamics between wells 1 - 2 will be localized . in the fig . [ evol1 ]
we plot the time evolution for @xmath30 showing the dependence of the parameter @xmath76 . for two threshold points : above is for the point @xmath77 and below is for the point @xmath78 . on the left side
is for @xmath79 ( delocalized region ) and on the right is @xmath80 ( localized region ) . on top
the solid line is for @xmath81 and the dashed line is for @xmath82 ( @xmath83 ) .
on botton the solid line is for @xmath84 and the dashed line is for @xmath85 ( @xmath86 ) . for the both cases we are using the initial state @xmath87,@xmath88 , @xmath89 , @xmath90 and @xmath91.,width=377,height=188 ] clearly from fig .
[ evol1 ] there is a threshold in the classical dynamics when we move from one region to another by varying the parameter @xmath92 ( solid black line ) .
it is worth noting from fig .
[ evol1 ] that the increasing of @xmath76 induces always localization on the classical dynamics .
the important issue here is to understand that if the system is in a delocalized dynamics regime ( region ii ) there is a threshold for relative population ( @xmath93 ) which separates again the dynamics in delocalized and localized regimes .
however if the system is in localized dynamics regime ( region i ) the dynamics smoothly becomes more localized by increasing the relative population in well @xmath1 ( dashed red line ) .
next we look if these behaviors are also present in the quantum dynamics of the model .
using the standard procedure now we compute the time evolution of the expectation value of the population imbalance between wells @xmath22-@xmath0 using the expression @xmath94 where @xmath95 , @xmath96 is the temporal evolution operator and @xmath97 represents the initial state in fock space .
recall that for fixed value of total number of bosons @xmath14 , the hilbert space associated to the hamiltonian ( [ h2n2 ] ) has dimension @xmath16 . as @xmath21 is conserved , the hamiltonian can be written in block - diagonal form for appropriated choice of the basis , where each eigenvalue of @xmath98 gives the dimension of each block by @xmath99 .
the total number of blocks is @xmath100 , such that @xmath101 . if one boson is added in the well @xmath1 , then the dimension of hamiltonian is given by @xmath102 and the dimension of block associated to the eigenvalue @xmath103 is given by @xmath104 .
therefore , for initial state @xmath105 the quantum dynamics is just governed by the block with fixed dimension @xmath106 for any @xmath107 . to start our discussion , firstly
, we choose @xmath108 and @xmath109 to go from one regime to the another . in fig .
[ qdt ] it is shown the time evolution of expectation value of population imbalance @xmath110 using four values of @xmath98 . for two threshold points using different values of @xmath76 ( @xmath98 ) : above is for the point @xmath77 and below is for the point @xmath78 .
on the left hand side @xmath111 and on the right hand side @xmath112 . on top
: the black solid line if for @xmath113 , the dashed red line is for @xmath114 , the dotted - dashed blue line is for @xmath115 and the dotted green line is for @xmath116 .
on botton : the black solid line if for @xmath113 , the dashed red line is for @xmath117 , the dotted - dashed blue line is for @xmath118 and the dotted green line is for @xmath116 . in both cases
the initial state is @xmath119 .
here we are using @xmath120.,title="fig:",width=377,height=151 ] for two threshold points using different values of @xmath76 ( @xmath98 ) : above is for the point @xmath77 and below is for the point @xmath78 . on the left hand side @xmath111 and on the right hand side @xmath112 . on top : the black solid line if for @xmath113 , the dashed red line is for @xmath114 , the dotted - dashed blue line is for @xmath115 and the dotted green line is for @xmath116 . on botton : the black solid line if for @xmath113 , the dashed red line is for @xmath117 , the dotted - dashed blue line is for @xmath118 and the dotted green line is for @xmath116 . in both cases
the initial state is @xmath119 .
here we are using @xmath120.,title="fig:",width=377,height=151 ] in fig .
[ qdt ] we choose the initial state @xmath121 and we observe the time evolution in the localized ( right side ) and delocalized ( left side ) regime by changing the population in well @xmath1 . clearly for both regimes
the increasing of population in well @xmath1 always induces localization of population imbalance between wells @xmath22-@xmath0 .
however in the localized regime ( right side ) the localization of population imbalance is smooth by increasing the population in well @xmath1 even around the threshold of the relative population @xmath93 ( see curves for @xmath122 ( @xmath123 ) on the right hand side of fig .
[ qdt ] ) . in the delocalized regime ,
as predicted classically the dynamics of system will change around @xmath124 , and a new scenario emerges .
a small fraction of bosons putting into well @xmath1 induces a large localization of population imbalance . in fig .
[ a2 ] it is shown the time average @xmath125 , @xmath126 , as function of @xmath127 , for the interval of time @xmath128 .
clearly we can see the localization effect is more prominent when the relative population is below to @xmath129 .
( solid line ) and @xmath130 ( dashed line ) , for @xmath131 against the relative population @xmath132 .
here we are using @xmath133 , @xmath134 and @xmath135 , @xmath136 and @xmath137.,width=377,height=188 ] here it is important to observe that around the threshold for the relative population the evolution changes and it has the same pattern that in the localized regime ( compare , for example , the dotted - dashed blue line on the left hand side of fig .
[ qdt ] with the solid black line on the right hand side of fig .
[ qdt ] ) .
we highlight that if the system is in the delocalized regime and around the threshold , the dynamics can be driven abruptly from one regime to the another just increasing or decreasing the population in well @xmath1 .
this opens the possibility to induce localization or delocalization in the dynamics between wells 1 - 2 just controlling the population in well @xmath1 ( external control ) .
we believe that this property can be useful to explain quantum devices where the switching scheme is needed .
the model presented above is the simplest model we can build using three modes and moreover it is integrable in the sense of bethe ansatz . however , as the apparent simplicity , we believe that this model may be feasible experimentally . taking into account the extent of the interwell interaction is greater than the range required for tunneling , the results obtained from the model could be verifyed in a proposed experiment using two subtly different possible experimental realization leading to the same physical interpretation : * spatially bringing closer the well @xmath22 and well @xmath0 allowing tunneling between them and then leaving well @xmath1 at a fixed distance such the tunneling is not allowed but the interwell interactions among them are turned on . by increasing or decreasing the population in well @xmath1 will change the intensity of the interwell interactions between wells , so in this way we can have control of population imbalance between the wells @xmath22-@xmath0 at some distance . *
keeping the population in well @xmath1 fixed and spatially bringing it closer at the well @xmath22 and well @xmath0 and separating apart after an interval of time . clearly from our previous theoretical results the variation of the population in well @xmath1 or analogously the approximation of the well @xmath1 keeping fixed its population close to the wells @xmath22-@xmath0 can be seen as a variation of the intensity of an external field . by the way quantum manipulation via external field
is a long - standing research area in physics and chemistry @xcite .
a study of coherent control of tunneling for an ultracold atoms can be found in the references @xcite . on the other hand side ,
the proposed theoretical model can be seen as the simplest and useful model for modeling quantum devices which needs some control at a distance by using an external field / potential . in this way
the model presented here can be useful in _ atomtronic _ area @xcite since it can mimic a _ field effect transistor_. taking into account the analogy with the electronic transistor , the well @xmath22 behaves like the _ source _ , the well @xmath0 behaves like the _ drain _ and the well @xmath1 behaves like the _ gate_. we have demonstrated that tunneling of a large fraction of bosons from well @xmath22 to well @xmath0 can be controlled by small fraction of bosons put into the well @xmath1 , without passing through it . in the absence or very small population in well @xmath1 the device is switched on resulting into the strong flux of bosons from well @xmath22 ( the source ) into the well @xmath0 ( the drain ) . increasing the population in the well @xmath1 the flux of bosons from well @xmath22 to
well @xmath0 becomes smaller and the device starts to be switched off .
these results for the model presented in this work is shown in fig .
here we observe that the transistor - like behavior presented here has a different operation from that transistor - like model presented in reference @xcite .
( solid black line ) and @xmath138 ( dashed red line ) for the initial state @xmath139 with different population in well @xmath1 : on the left , top is for @xmath113 and botton , for @xmath140 ; on the right , top is for @xmath141 and botton , for @xmath142 .
here we are using @xmath134 and @xmath135 , @xmath136 , @xmath137 and @xmath143.,width=377,height=188 ] moreover , the proposed model allows us also go further and we can envisioning another interesting scenario .
we can induce a change of quantum phase transition by using an external field by controlling the distance between two parallel bidimensional chains .
for example , if we are imagining bosons weakly interacting loaded in a bidimensional chain where bosons are in a delocalized dynamics regime with tunneling allowed just on the plane .
if another similar bidimensional chain is putting parallel nearby the former one , then taking in account that tunneling between plane is prohibited the change of quantum phase can be induced by changing the interplane distance .
that way , we can go from the superfluid - like phase to insulator - like phase just manipulating the interplane separation between bidimensional planes .
our studies are just in the beginning of the investigations and new results will be published somewhere .
in summary , in this work we presented the simplest model we can build using three modes and moreover it is integrable in the sense of bethe ansatz .
after , the semiclassical analysis for the model was carried out .
the hamilton s equations , the diagram of coupling parameters and the classical dynamics were presented .
posteriorly the quantum dynamics of the model was presented and we show that the flow of bosons between the wells @xmath22-@xmath0 can be controlled by increasing or decreasing the population in the well @xmath1 leading to the conclusions that a small population put into well @xmath1 , relative with population in well @xmath22 and well @xmath0 , generates an effective control of population imbalance between wells @xmath22-@xmath0 .
later we discussed the experimental feasibility of model suggesting an experiment without taking into account the technicality experimental . in the end , we conclude that the proposed model can be very useful to understand quantum devices which needs a quantum control in its operation , for example , in _ atomtronic _ @xcite and also in quantum manipulation via external field @xcite .
anyway our studies are just beginning and more research in this direction appears to be useful in order to elucidate points and thereby contributing to our understanding of the many - body physics in ultracold bosonic atoms systems . 99 t.kinoshita , t. wenger and d. s. weiss , _ nature _ * 440 * , 900 ( 2006 ) m. rigol , v. dunjko ans m. olshanii,_nature _ * 452 * , 845 ( 2008 ) j.m .
deutsch , _ phys .
a _ * 43 * , 2046 ( 1991 ) m. eckstein , m. kollar and p. werner , _ phys .
lett . _ * 103 * 056403 ( 2009 ) m.p .
strzys and j.r .
anglin , _ phys .
a _ * 81 * , 043616 ( 2010 ) d. manzano and e. kyoseva , 2016 an atomic symmetry - controlled thermal switch , arxiv:1508.05691v3 l.d .
faddeev , e.k .
sklyanin and l.a .
takhtajan , _ theor .
_ * 40 * 194 ( 1979 ) p.p.kulish and e.k .
sklyanin , _ lec .
notes phys . _
* 151 * 61(1982 ) l.a .
takhtajan , _ lec .
notes phys . _
* 370 * 3 ( 1990 ) l.d .
faddeev , _ int . j. mod
. phys _ * 10 * 1845 ( 1995 ) a. j. leggett , _ rev .
* 73 * ( 2001 ) , 307 g. j. milburn , j. corney , e. m. wright and d. f. walls , _ phys .
a _ * 55 * 4318 ( 1997 ) j.links , a. foerster , a. tonel and g. santos , _ ann .
henri poincar _ * 7 * 1591 ( 2006 ) a.p .
tonel , j. links , a. foerster , _ j. phys .
_ * 38 * 1235 ( 2005 ) a.p .
tonel , j. links , a. foerster , _ j. phys . a : math .
gen . _ * 38 * 6879 ( 2005 ) j. williams , r. walser , j. cooper , e. a. cornell and m. holland _ phys .
a _ * 61 * 033612 ( 2000 ) m. albiez , r. gati , j. folling , s. hunsmann , m. cristiani and m. k. oberthaler , _ phys .
lett . _ * 95 * 010402 ( 2005 ) k. nemoto , c. a. holmes , g.j .
milburn and w.j .
munro , _ phys .
a _ * 63 * 013604 ( 2000 ) t. lahaye , t. pfau and l. santos,_phys.rev.lett . _
* 104 * 170404 ( 2010 ) a. gallemi , m. guilleumas , r. mayol and a. sanpera , _ phys .
a _ * 88 * 063645 ( 2013 ) d. peter , k. pawlowski , t. pfau and k. rzazewski,_j .
b : at . mol .
phys . _ * 45 * 225302 ( 2012 ) j - m .
liu and y - z .
wang , _ int .
b _ * 20 * 277 ( 2006 ) f. wang , x. yan and d. wang , _ int .
b _ * 23 * 383 ( 2009 ) j.a . stickney , d. z. anderson and a. a. zozulya , _ phys .
rev . a _ * 75 * 013608 ( 2007 ) c.v .
chianca and m.k .
phys.rev . a _ * 83 * , 043607 ( 2011 ) s. de liberato and c. j. foot , _ phys.rev .
a _ * 73 * , 035602 ( 2006 ) i. zerec , v. keppens , m.a .
mcguire , d. mandrus , b.c .
sales and p. thalmeier _ phys .
lett . _ * 92 * , 185502 ( 2004 ) g. lu , w. hai and m. zou , _ j. phys .
b : at . mol .
phys . _ * 45 * , 185504 ( 2012 ) g. mazzarella and v. penna , _
b : at . mol .
phys . _ * 48 * , 065001 ( 2015 ) p. schlagheck and j. dujardin , coherent backscattering in the fock space of ultracold bosonic atoms , arxiv:1610.04350v1 q. zhou and s. das sarma , _ phys .
a _ * 82 * 041601 ( 2010 ) a - x .
zhang , s - l .
tian , r - a .
tang and j - k .
xue , _ chin .
lett . _ * 25 * , 3566 ( 2008 ) a.p .
tonel , l.h .
ymai , a. foerster and j. links , _ j. phys . a : math .
theor . _ * 48 * , 494001(2015 ) l.h .
ymai , a.p .
tonel , a. foerster and j. links , 2016 quantum integrable multi - well tunneling models , arxiv:1606.00816 g. santos , a. foerster , i. roditi , _
j. phys . a : math .
theor . _ * 46 * , 265206 ( 2013 ) g. santos , the exact solution of a generalized bose hubbard model , arxiv:1605.08452v1 a.p .
tonel , l.h .
ymai , k. wilsmann , j. links and a. foerster , in preparation ( 2016 ) b.t .
seaman , m. kramer , d.z .
anderson and m.j .
holland , _ phys .
a _ * 75 * 023615 ( 2007 ) j.a .
stickney , d.z .
anderson and a.a .
zozulya , _ phys .
a _ * 75 * 013608 ( 2007 ) s. moulder , s. beattie , r.p .
smith , n. tammuz and z. hadzibabic , _ phys .
a _ * 86 * 013629 ( 2012 ) h. rabitz , r. de vivie - riedle , m. motzkus and k. kompa _ science _ * 288 * 824 ( 2000 ) e. kierig , u. schnorrberger , a. schietinger , j. tomkovic , and m. k. oberthaler _ phys . rev .
lett . _ * 100 * 190405 ( 2008 ) g. lu , w. hai and h. zhong , _ phys .
a _ * 80 * 013411 ( 2009 ) | in this work we show the simplest and integrable model for bose - einstein condensates loaded in a triple well potential where the tunneling between two wells can be controlled by the other well showing a behavior similar to an electronic field effect transistor . using a classical analysis , the hamilton s equation
are obtained , a threshold indicating a discontinuous phase transition is presented and the classical dynamics is computed .
then , the quantum dynamics is investigated using direct diagonalization .
we find well agreement in both these analysis .
based on our results , the switching scheme for tunneling is shown and the experimental feasibility is discussed . |
fine powders , such as aerosols , volcanic ashes , flours , and toner particles are commonly observed in daily life .
the attractive interaction between fine powders plays major roles @xcite , while there are various studies discussing the effects of cohesive forces between macroscopic powders @xcite . for example
, the johnson - kendall - roberts ( jkr ) theory describes microscopic surface energy for the contact of cohesive grains @xcite .
the others study the attractive force caused by liquid bridge for wet granular particles @xcite .
it should be noted that the cohesive force can not be ignored for small fine powders .
indeed , the intermolecular attractive force always exists .
moreover , the inelasticity plays important roles when powders collide , because there are some excitations of internal vibrations , radiation of sounds , and deformations @xcite .
+ let us consider cohesive powders under a plane shear .
so far there exist many studies for one or two effects of the shear , an attractive force , and an inelastic collision @xcite , but we only know one example for the study of the jamming transition to include all three effects @xcite . on the other hand , when the lennard - jones ( lj ) molecules are quenched below the coexistence curve of gas - liquid phases @xcite , a phase ordering process proceeds after the nucleation takes place @xcite .
it is well known that clusters always appear in freely cooling processes of granular gases @xcite .
such clustering processes may be understood by a set of hydrodynamic equations of granular gases @xcite .
when we apply a shear to the granular gas , there exist various types of clusters such as 2d plug , 2d wave , or 3d wave for three dimensional systems @xcite .
+ in this paper , we try to characterize nonequilibrium pattern formation of cohesive fine powders under the plane shear by the three dimensional molecular dynamics ( md ) simulations of the dissipative lj molecules under the lees - edwards boundary condition @xcite . in our previous paper @xcite , we have mainly focused on the effect of dissipation on the pattern formation in sllod dynamics @xcite . in this study , we systematically study it by scanning a large area of parameters space to draw the phase diagrams with respect to the density , the dimensionless shear rate , and the dissipation rate without the influence of sllod dynamics . +
the organization of this paper is as follows . in the next section ,
we introduce our model and setup for this study .
section [ sec : results ] , the main part of this paper , is devoted to exhibit the results of our simulation . in sec .
[ sec : phase ] , we show the phase diagrams for several densities , each of which has various distinct steady phases .
we find that the system has a quasi particle - hole symmetry .
we also find that the steady states depend on the initial condition of velocities of particles when the dissipation is large . in sec .
[ sec : vdf ] , we analyze the velocity distribution function , and try to reproduce it by solving the kramers equation with coulombic friction under the shear . in the last section , we discuss and summarize our results . in appendix
[ sec : lebc ] , we study the pattern formation of dissipative lj system under the physical boundary condition . in appendix
[ sec : coulombic ] , we illustrate the existence of coulombic friction near the interface of the plate - gases coexistence phase . in appendix
[ sec : av_pypz ] , we demonstrate that the viscous heating term near the interface is always positive . in appendix
[ sec : perturbation ] , we present a perturbative solution of the kramers equation . in appendix
[ sec : moment ] , we show the detailed calculations for each moment . in appendix
[ sec : rotation_f ] , we show the detailed calculations of the velocity distribution function .
in this section , we explain our model and setup of md for cohesive fine powders under a plane shear . we introduce our model of cohesive fine powders in sec .
[ sec : model ] and explain our numerical setup in sec .
[ sec : setup ] . and the coefficient of restitution @xmath0 when the pre - collisional relative velocities ( solid and dashed lines ) are given by @xmath1 and @xmath2 , respectively.,width=207 ] we assume that the interaction between two cohesive fine powders can be described by lj potential , and an inelastic force caused by collisions with finite relative speeds
. the explicit expression of lj potential is given by @xmath3\label{eq : lj}\ ] ] with a step function @xmath4 and @xmath5 for @xmath6 and @xmath7 , respectively , where @xmath8 , @xmath9 , and @xmath10 are the well depth , the diameter of the repulsive core , and the distance between the particles @xmath11 and @xmath12 , respectively .
here , we have introduced the cutoff length @xmath13 to save the computational cost , i.e. @xmath14 for @xmath15 . to model the inelastic interaction ,
we introduce a viscous force between colliding two particles as @xmath16 where @xmath17 , @xmath18 , and @xmath19 are the dissipation rate , a unit vector parallel to @xmath20 , and the relative velocity between the particles , respectively . here ,
@xmath21 and @xmath22 ( @xmath23 ) are , respectively , the position and velocity of the particle .
it should be noted that the range of inelastic interaction is only limited within the distance @xmath9 . from eqs .
( [ eq : lj ] ) and ( [ eq : dashpot ] ) , the force acting on the @xmath11-th particle is given by @xmath24 our lj model has an advantage to know the detailed properties in equilibrium @xcite . the normal restitution coefficient @xmath0 , defined as a ratio of post - collisional speed to pre - collisional speed , depends on both the dissipation rate @xmath17 and incident speed .
for instance , the particles are nearly elastic , i.e. the restitution coefficient , @xmath25 for the case of @xmath26 and the incident speed @xmath27 , where @xmath28 is the mass of each colliding particle .
figure [ fig : cor ] plots the restitution coefficient against the dimensionless dissipation rate @xmath29 , where the incident speeds are given by @xmath1 and @xmath2 , respectively .
we restrict the dissipation rate to small values in the range @xmath30 .
note that small and not too large inelasticity is necessary to reproduce a steady coexistence phase between a dense and a dilute region , which will be analyzed in details in this paper .
indeed , the system can not reach a steady state without inelasticity , while all particles are absorbed in a big cluster when inelasticity is large . in this paper
, we use three dimensionless parameters to characterize a system : the dimensionless density @xmath31 , the shear rate @xmath32 , and the dissipation rate @xmath29 .
it should be noted that the well depth @xmath8 is absorbed in the dimensionless shear rate and the dissipation rate .
thus , we may regard the control of two independent parameters as the change of the well depth .
plane , that is , we choose @xmath33-axis as the shear direction and @xmath34-axis as the velocity gradient direction.,width=226 ] figure [ fig : setup ] is a snapshot of our md for a uniformly sheared state , where we randomly distribute @xmath35 particles in a cubic periodic box and control the number density @xmath36 by adjusting the linear system size @xmath37 . at first , we equilibrate the system by performing the md with the weeks - chandler - andersen potential @xcite during a time interval @xmath38 .
we set the instance of the end of the initial equilibration process as the origin of the time for later discussion .
then , we replace the interaction between the particles by the truncated lj potential ( [ eq : lj ] ) with the dissipation force , eq .
( [ eq : dashpot ] ) under the lees - edwards boundary condition . as shown in appendix [ sec : lebc ] ,
the results under the lees - edwards boundary condition are almost equivalent to those under the flat boundary .
the time evolution of position @xmath39 is given by newton s equation of motion @xmath40 .
in this section , we present the results of our md . in sec .
[ sec : phase ] , we draw phase diagrams of the spatial structures of cohesive fine powders . in sec . [
sec : vdf ] , we present the results of velocity distribution functions and reproduce it by solving a phenomenological model .
.the dimensionless parameters used in fig .
[ fig : steady ] .
[ tab : phase ] [ cols="^,^,^,^",options="header " , ] , @xmath41 , and @xmath42 for ( a)(i ) are listed in table [ tab : phase ] .
we note that gas particles in ( b ) , ( c ) and ( d ) are drawn smaller than the real size for visibility.,width=317 ] , ( b ) @xmath43 , ( c ) @xmath44 , ( d ) @xmath44 for @xmath45 , and ( e ) @xmath46 , respectively . the spatial patterns corresponding to fig .
[ fig : steady](a)(i ) are represented by red filled circles ( fig .
[ fig : steady](a ) ) , blue open circles ( fig .
[ fig : steady](b ) ) , blue filled upper triangles ( fig . [
fig : steady](c ) ) , blue open squares ( fig . [
fig : steady](d ) ) , black open diamond ( fig .
[ fig : steady](e ) ) , black open upper triangles ( fig .
[ fig : steady](f ) ) , black filled squares ( fig .
[ fig : steady](g ) ) , black filled lower triangles ( fig . [
fig : steady](h ) ) , and black filled triangles ( fig .
[ fig : steady](i ) ) , respectively .
the steady states represented by the cross marks show various patterns depending on the initial velocities of particles.,width=321 ] , @xmath47 , @xmath48 ) : ( a ) a dense - plate cluster parallel to @xmath49 plane , ( b ) a dense - plate cluster parallel to @xmath50 plane and ( c ) a dense - cylinder cluster parallel to @xmath51-axis.,width=302 ] , @xmath52 , @xmath53 .
( a ) @xmath54 , ( b ) @xmath55 , ( c ) @xmath56 , and ( d ) @xmath57.,width=302 ] figure [ fig : steady ] displays typical patterns formed by the particles in their steady states , which are characterized by the dimensionless parameters @xmath58 , @xmath41 , and @xmath42 as listed in table [ tab : phase ] . figure [ fig : phasediagram ] shows phase diagrams in the steady states for ( a ) @xmath59 , ( b ) @xmath43 , ( c ) @xmath44 , and ( d ) @xmath46 .
three of these phases , figs .
[ fig : steady](a ) , ( d ) and ( g ) , are similar to those observed in a quasi two - dimensional case with sllod dynamics @xcite . if the shear is dominant , the system remains in a _ uniformly sheared phase _
[ fig : steady](a ) ) .
however , if the viscous heating by the shear is comparable with the energy dissipation , we find that a _ spherical - droplet _ , a _ dense - cylinder _ , and a _ dense - plate _ coexist for extremely dilute ( @xmath59 ) , dilute ( @xmath60 ) , and moderately dense ( @xmath61 ) gases , respectively ( figs . [
fig : steady](b)(d ) ) .
these three coexistence phases are realized by the competition between the equilibrium phase transition and the dynamic instability caused by inelastic collisions .
furthermore , if the energy dissipation is dominant , there are no gas particles in steady states ( figs .
[ fig : steady](e)(g ) ) . for an extremely high density case ( @xmath62 )
, we observe an _ inverse - cylinder _ , where the vacancy forms a hole " passing through the dense region along the @xmath33-axis ( fig .
[ fig : steady](h ) ) , and an _ inverse - droplet _ , where the shape of the vacancy is spherical ( fig . [
fig : steady](i ) ) . in our simulation ,
the role of particles in a dilute system corresponds to that of vacancies in a dense system .
thus , the system has a quasi particle - hole symmetry .
+ moreover , the shape of clusters depends on the initial condition of velocities of particles , even though a set of parameters such as the density , the shear rate , the dissipation rate and the variance of the initial velocity distribution function are identical when the dissipation is strong .
we observe a dense - plate parallel to @xmath49 plane ( fig .
[ fig : initial_configuration](a ) ) , a dense - plate parallel to @xmath50 plane ( fig .
[ fig : initial_configuration](b ) ) , and a dense - cylinder parallel to @xmath33-axis ( fig .
[ fig : initial_configuration](c ) ) under the identical set of parameters .
this initial velocity dependence appears in the region far from the coexistence phases , where the system evolves from aggregates of many clusters ( see fig . [
fig : dependence ] ) .
we also measure the velocity distribution function ( vdf ) @xmath63 ( @xmath64 ) , where @xmath65 is the velocity fluctuation around the mean velocity field , @xmath66 , averaged over the time and different samples in the steady state . for simplicity
, we focus only on the following three phases ; the uniformly sheared phase ( fig . [ fig : steady](a ) ) , the dense - plate coexistence phase ( fig .
[ fig : steady](d ) ) , and the dense - plate cluster phase ( fig .
[ fig : steady](g ) ) . in this paper
, we use the width @xmath67 for bins in @xmath34-direction , while the bin sizes in both @xmath51 and @xmath33-directions are @xmath37 to evaluate vdf from our md as in fig .
[ fig : binwise ] .
it is remarkable that the vdf is almost isotropic gaussians for the phases corresponding to figs .
[ fig : steady](a ) and ( g ) as well as deep inside of both the dense and the gas regions in the coexistence phase in fig .
[ fig : steady](d ) ( see figs . [
fig : vdf_interface](a)(d ) ) .
this is because we are interested in weak shear and weak dissipation cases without the influence of gravity . on the other hand ,
vdf is nearly equal to an anisotropic exponential function @xcite in the vicinity of the interface between the dense and the gas regions in the coexistence phase corresponding to fig .
[ fig : steady](d ) as in figs .
[ fig : vdf_interface](e)(g ) . .
in addition , we introduce a new coordinate @xmath68 , and @xmath69 , which is the angle between @xmath70 and @xmath33-direction ( in the counterclockwise direction ) for later analysis.,width=226 ] ( a ) , ( b ) vdfs in the phase fig .
[ fig : steady](g ) , ( c ) vdfs in the dense region of the phase fig .
[ fig : steady](d ) , ( d ) vdfs in the dilute region of the phase fig .
[ fig : steady](d ) , ( e ) vdf of @xmath51-direction in the interface of the phase fig .
[ fig : steady](d ) , ( f ) vdf of @xmath33-direction in the interface of the phase fig .
[ fig : steady](d ) , and ( g ) vdf of @xmath34-direction in the interface of the phase fig .
[ fig : steady](d).,width=321 ] -direction ) in the plate - gases coexistence phase ( @xmath71 , @xmath72 , @xmath73 ) , where @xmath74.,width=264 ] + we now explain the non - gaussian feature near the interface by a simple stochastic model of a tracer particle subjected to coulombic friction ( the justification to use such a model is explained in appendix [ sec : coulombic ] ) .
let us consider a situation that a gas particle hits and slides on the wall formed by the particles in the dense region ( see fig .
[ fig : fcc ] ) . because the velocity gradient in the gas region is almost constant as shown in fig .
[ fig : profile ] , we may assume that a tracer particle in the gas near the interface is affected by a plane shear .
moreover , the tracer particle on a dense region may be influenced by coulombic friction ( see appendix [ sec : coulombic ] ) .
when we assume that the collisional force among gas particles can be written as the gaussian random noise @xmath75 , the equations of motion of a tracer particle at the position @xmath76 may be given by @xmath77 where @xmath78 is a peculiar momentum , which is defined by eq .
( [ eq : eom1 ] ) .
here we have introduced the friction constant @xmath79 and the effective force @xmath80 which is a function of the activation energy @xmath81 from the most stable trapped configuration of the solid crystal ( see fig .
[ fig : fcc ] ) . here , @xmath75 is assumed to satisfy @xmath82 where @xmath83 is the average over the distribution of the random variable @xmath84 .
@xmath85 is the diffusion coefficient in the momentum space , which satisfies the fluctuation - dissipation relation @xmath86 in the @xmath87-dimensional system with a temperature @xmath88 .
a set of langevin equations ( [ eq : eom1 ] ) and ( [ eq : eom2 ] ) can be converted into the kramers equation @xcite : @xmath89 where @xmath90 is the probability distribution function of the tracer particle .
+ if we multiply eq . (
[ eq : fpeq ] ) by @xmath91 and integrate over @xmath78 , we immediately obtain @xmath92 where @xmath93 . because the third term on the right hand side ( rhs ) of eq .
( [ eq : balance ] ) represents the viscous heating which is always positive as shown in eq .
( [ eq : average_pypz ] ) and the fourth term is the loss of the energy due to friction , the balance among the third , the fourth and the fifth terms on rhs of eq .
( [ eq : balance ] ) produces a steady state .
it should be noted that the first and the second terms on rhs do not contribute to the energy balance equation for the whole system . + here ,
we only consider the steady distribution , i.e. @xmath94 .
thus , eq . ( [ eq : fpeq ] ) is reduced to @xmath95 where @xmath96 .
if there is neither a shear nor a density gradient , we find that eq .
( [ eq : exponential ] ) has the steady solution obeying an exponential distribution , i.e. @xmath97 $ ] , where we have introduced @xmath98 .
we adopt the perturbative expression for @xmath99 in terms of @xmath100 , which is the ratio of the diameter @xmath9 to the interface width @xmath101 , and the dimensionless shear rate @xmath41 as ( see the derivation in appendix [ sec : perturbation ] ) @xmath102 we also adopt the expansions @xmath103 with @xmath104 and @xmath105 , where @xmath69 is the angle between @xmath78 and @xmath33-axis ( in the counterclockwise direction , see fig .
[ fig : binwise ] ) .
then , we can solve eq .
( [ eq : exponential ] ) perturbatively as @xmath106 where @xmath107 , @xmath108 and @xmath109 are , respectively , given by @xmath110 here , we have introduced @xmath111 and @xmath112 given by eq . ( [ eq : def_a ] ) .
it should be noted that the other terms except for those in eqs .
( [ eq : f_per])([eq : f_2nd ] ) automatically disappear within the linear approximation as in eq .
( [ eq : f_linear ] ) .
+ the second , the third and the fourth moments in @xmath70 and @xmath113-directions after the rotation by the angle of @xmath69 in the counterclockwise direction are , respectively , given by @xmath114 as shown in appendix [ sec : moment ] , where @xmath115 with @xmath116 or @xmath117 represents @xmath118 for a minus sign and @xmath119 for a plus sign , respectively . to reproduce the node of the third moment in md
, we phenomenologically introduce the angle @xmath120 and replace @xmath69 by @xmath121 in eqs .
( [ eq:2nd_y])([eq:4th_z ] ) . here , we choose @xmath122 to fit the node position of the third moment .
we have not identified the reason why the direction of the node is deviated from the direction at which vdf becomes isotropic .
+ now , let us compare eqs .
( [ eq:2nd_y])([eq:4th_z ] ) with md for a set of parameters @xmath123 . from the density profile ( fig .
[ fig : profile ] ) and the fitting to the second moment and the amplitude of the third moment , we obtain @xmath124 , @xmath125 , @xmath126 , and @xmath127 .
it is surprised that eqs .
( [ eq:2nd_y])([eq:4th_z ] ) can approximately reproduce the simulation results as in fig .
[ fig : moment ] except for the node positions of the second and the fourth moments . , @xmath128 , @xmath129 ( circle : @xmath70-direction , upper triangle : @xmath113-direction ) and those obtained by eqs .
( [ eq:2nd_y])([eq:4th_z ] ) ( solid line : @xmath70-direction , dashed line : @xmath113-direction).,width=325 ] + for the explicit form of vdf , at first , we convert @xmath130 to @xmath131 as in appendix [ sec : rotation_f ] : @xmath132.\label{eq : cartesian}\end{aligned}\ ] ] we obtain the peculiar velocity distribution function in each direction by integrating eq . (
[ eq : cartesian ] ) with respect to @xmath133 or @xmath134 as @xmath135 , \label{eq : vdf_y}\\ p(u_z)=&\frac{m\kappa^2}{2\pi}\int_{-\infty}^\infty du_y \exp\left(-m\kappa u\right)\nonumber\\ & \times\left[1-\frac{m\epsilon a}{3\kappa^3}\left(3+m\kappa u+m^2\kappa^2u^2\right)u_z\cos\psi \right.\nonumber\\ & \left.+\frac{m^2\dot\gamma}{4d}(u_y^2-u_z^2)\sin2\psi\right ] , \label{eq : vdf_z}\end{aligned}\ ] ] where @xmath136 .
these expressions semi - quantitatively reproduce vdf observed in our md as in fig .
[ fig : prob ] .
-direction ( left , cross ) and @xmath34-direction ( right , cross ) obtained by our md .
the dashed lines in the left and right figures are the results of eqs .
( [ eq : vdf_y ] ) and ( [ eq : vdf_z ] ) , respectively.,width=321 ]
let us discuss our results . in sec .
[ sec : phase ] , we do not discuss the time evolution of the granular temperature @xmath137 , where @xmath138 is the ensemble average velocity field @xcite .
the granular temperature abruptly decreases to zero in the cluster phases fig .
[ fig : steady](e)(i ) when a big cluster which absorbs all gas particles appears @xcite . to clarify the mechanism of abrupt change of the temperature during clusterings , we will need to study the more detailed dynamics .
+ moreover , to discuss the phase boundary between the uniformly sheared phase and the coexistence phases , we may use the stability analysis of a set of the hydrodynamic equations coupled with the phase transition dynamics @xcite .
once we establish the set of hydrodynamic equations , it is straightforward to perform weakly nonlinear analysis for this system @xcite .
it should be noted that the set of equations may be only available near the phase boundary between the uniformly sheared phase and the coexistence phases .
+ in fig .
[ fig : vdf_interface ] , the vdf in a uniformly sheared phase is almost gaussian .
this result seems to be inconsistent with the results for ordinary gases under a uniform shear flow @xcite , which showed that the vdf differs from gaussian even in a uniformly sheared phase . in this study , however , we only restrict our interest to small inelastic and weakly sheared cases .
this situation validates small deviation from gaussian .
we studied cohesive fine powders under a plane shear by controlling the density , the dimensionless shear rate and the dissipation rate .
depending on these parameters , we found the existence of various distinct steady phases as in fig .
[ fig : steady ] , and we have drawn the phase diagrams for several densities as in fig . [
fig : phasediagram ] .
in addition , the shape of clusters depends on the initial condition of velocities of particles as in fig .
[ fig : initial_configuration ] , when the dissipation is strong .
we also found that there is a quasi particle - hole symmetry for the shape of clusters in steady states with respect to the density .
+ we found that the velocity distribution functions near the interface between the dense region and the gas - like dilute region in the dense - plate coexistence phase deviate from the gaussian as in fig . [ fig : vdf_interface ] . introducing a stochastic model and its corresponding the kramers equation ( [ eq : fpeq ] )
, we obtain its perturbative vdfs as in eqs .
( [ eq : vdf_y ] ) and ( [ eq : vdf_z ] ) , which reproduce the semi - quantitative behavior of vdf observed in md as in fig .
[ fig : prob ] .
this result suggests that the motion of a gas particle near the interface is subjected to coulombic friction force whose origin is the activation energy in the dense region .
the authors thank to takahiro hatano and meheboob alam for their fruitful advices .
ks wishes to express his sincere gratitude for the yukawa institute for theoretical physics ( yitp ) to support his stay and its warm hospitality .
part of this work has been proceeded during yitp workshop yitp - ww-13 - 04 on physics of glassy and granular materials , " yitp - t-13 - 03 on physics of granular flow . "
numerical computation in this work was partially carried out at the yukawa institute computer facility .
this work is partially supported by grant - in - aid of mext , japan ( grant no .
25287098 ) .
in this appendix , we present the results of our simulations under the flat boundary condition which is one of the typical physical boundaries to clarify the influence of the boundary condition .
we prepare flat walls at @xmath139 , moving at velocities @xmath140 in @xmath33-direction , respectively .
when a particle with a velocity @xmath141 hits the walls at @xmath139 , the velocity is changed as @xmath142 after the collision , respectively .
the phase diagram of the system for the physical boundary for @xmath71 is presented in fig .
[ fig : physical ] . we have obtained three steady phases such as the uniformly sheared phase , the coexistence phase between dense - plate and gas regions , and the dense - plate cluster phase .
the phase diagram is almost same as the corresponding one under the lees - edwards boundary condition ( see figs . [
fig : phasediagram](d ) ) . this can be understood as follows : if two particles at the symmetric positions with respect to the origin of the system simultaneously collide the walls at @xmath143 and @xmath144 , the pair of velocities after collisions is same as that after passing across the boundaries at @xmath139 for the system under the lees - edwards boundary condition .
this is realized after the averaging over the collisions .
thus , the flat boundary condition is essentially equivalent to the lees - edwards boundary condition .
, uniformly sheared state ( red filled circle , fig .
[ fig : steady](a ) ) , coexistence of a dense - plate and gases ( blue open square , fig .
[ fig : steady](d ) ) , and an isolated dense - plate ( black filled square , fig .
[ fig : steady](g)).,width=188 ]
in this appendix , we try to illustrate the existence of coulombic friction force for the motion of a tracer particle near the interface .
let us consider a situation that a gas particle hits and slides on the wall formed by the particles in the dense region ( see fig . [
fig : fcc ] ) .
if the kinetic energy of the gas particle is less than the potential energy formed by the particles in the dense region , it should be trapped in the potential well .
therefore , the motion of the gas particle is restricted near the interface . in this case
, we can write the @xmath145-body distribution function near the interface @xmath146 by using the distribution function in the equilibrium system as @xcite @xmath147,\label{eq : kawasaki2}\end{aligned}\ ] ] where @xmath148 , @xmath149 is the equilibrium distribution function at time @xmath150 , and @xmath151 with @xmath152 here , we have introduced the inverse granular temperature @xmath153 and the local shear rate @xmath154 in the interface region .
if the dissipation is small and the shear rate is not large , we may assume that @xmath155 , where @xmath156 is the mean field @xmath50 component of the stress tensor .
we also assume that the stress tensor decays exponentially as @xmath157 @xcite , where @xmath158 is the relaxation time of the stress tensor . from these relationships
, we may use the approximate expression @xmath159\nonumber\\ & \times \exp\left(-\beta\tau_0\dot\gamma_l v_l\sigma_{yz}^{\rm mf}(0)\right),\label{eq : kawasaki}\end{aligned}\ ] ] where @xmath160 and @xmath161 , are respectively , the mean field hamiltonian per particle in the interface and the energy fluctuation of the particle @xmath11 which may be the activation energy from the local trap . here @xmath162 and @xmath163 are , respectively , the number of particles and the volume in the interface region and @xmath164 .
there are two characteristic time scales @xmath165 and @xmath166 corresponding to the uniform region and the interface between dense and dilute regions . because the time scale is obtained from the average over the distribution function ( [ eq : kawasaki ] ) or the local mean field distribution , the relationship between @xmath165 and @xmath166 is expected to be @xmath167,\end{aligned}\ ] ] where we have eliminated the suffix @xmath11 for the particle .
this equation can be rewritten as @xmath168 therefore , we may estimate coulombic friction constant as @xmath169 where @xmath170 , @xmath171 , @xmath172 and @xmath173 at the interface for a set of parameters @xmath174 . in this expression
, we can not determine the relaxation time @xmath158 from the simulation , which is estimated to reproduce the average value of the second moment with the aid of eq .
( [ eq:2nd_y ] ) .
in this appendix , let us calculate the average of the viscous heating term by using the distribution function near the interface . from eq .
( [ eq : kawasaki ] ) , we can rewrite the distribution function with the aid of eq .
( [ eq : def_sigma ] ) as @xmath175,\label{eq : rho_eq}\end{aligned}\ ] ] where @xmath176 $ ] .
then @xmath177 is given by @xmath178\nonumber\\ = & \int_0^\infty dp \int_0^{2\pi}d\theta p^3 \sin\theta \cos\theta \nonumber\\ & \times \exp\left[-\beta \left(\frac{p^2}{2m}+\frac{\tau_0\dot\gamma_l v_l}{m}p^2\sin\theta \cos\theta\right)\right]\nonumber\\ = & -\frac{\pi}{2}\int_0^\infty dp p^3 \exp\left(-\frac{\beta p^2}{2m}\right ) i_1\left(\frac{\beta \tau_0 \dot\gamma_l v_l}{2m}p^2\right),\label{eq : average_pypz}\end{aligned}\ ] ] where @xmath179 is the modified bessel function of the first kind @xcite . because @xmath179 is positive for @xmath180 , eq .
( [ eq : average_pypz ] ) ensures that the viscous heating term @xmath181 is always positive near the interface .
in this appendix , let us solve the kramers equation ( [ eq : exponential ] ) perturbatively to obtain the steady vdf . later
, we compare this solution with the result of md .
+ at first , we adopt the following three assumptions .
the first assumption is that the distribution function is independent of both @xmath51 and @xmath33 , the coordinates horizontal to the interface .
we also assume that the distribution function @xmath99 depends on @xmath34 , vertical to the interface , through the density and the granular temperature : @xmath182 second , we assume that the changes of the density and the granular temperature near the interface can be characterized by the interface width @xmath101 as @xmath183 where @xmath184 , @xmath185
. here , @xmath186 and @xmath187 are the density and the granular temperature in the dense region , and @xmath188 and @xmath189 are those in the dilute region , respectively .
third , we also assume that the interface width @xmath101 is much longer than the diameter of the particles @xmath9 , i.e. @xmath190 . from these assumptions , @xmath191 may be rewritten as @xmath192 + to solve eq .
( [ eq : exponential ] ) , we adopt the perturbative expression eq .
( [ eq : f_linear ] ) . equation ( [ eq : exponential ] ) ,
thus , reduces to the following three equations : for the zeroth order , @xmath193 for the first order of @xmath194 , @xmath195 and for the first order of @xmath41 , @xmath196 the solution of eq .
( [ eq:0th ] ) is given by @xmath197 where @xmath198 is the exponential integral @xmath199 @xcite , and @xmath200 and @xmath201 are the normalization constants .
here , we set @xmath202 because @xmath198 becomes infinite at @xmath203 , and @xmath204 to satisfy the normalization condition without the shear and the density gradient . using eq .
( [ eq : f_00 ] ) , equations ( [ eq:1st01 ] ) and ( [ eq:1st10 ] ) can be represented in the polar coordinates as @xmath205 and @xmath206 where we have introduced @xmath112 as @xmath207 to solve eqs .
( [ eq : f01 ] ) and ( [ eq : f10 ] ) , we adopt the expansions for @xmath208 with @xmath104 and @xmath105 @xcite .
equation ( [ eq : f01 ] ) for each @xmath36 reduces to the following equations : for @xmath209 , @xmath210 and for @xmath211 , @xmath212 the solutions of eqs .
( [ eq : f01_1 ] ) and ( [ eq : f01_n ] ) are , respectively , given by @xmath213 and @xmath214 for @xmath211 , where @xmath215 and @xmath216 are , respectively , the confluent hypergeometric function and laguerre s bi - polynomial @xcite , and the normalization constants @xmath217 and @xmath218 ( @xmath219 ) will be determined later .
similarly , equation ( [ eq : f10 ] ) for each @xmath36 reduces to the following equations : for @xmath116 , @xmath220 and for @xmath221 , @xmath222 the solutions of eqs .
( [ eq : f10_2 ] ) and ( [ eq : f10_n ] ) are , respectively , given by @xmath223 and @xmath224 for @xmath225 , where the normalization constants @xmath226 and @xmath227 ( @xmath219 ) will be determined later . + here ,
let us determine the normalization constants @xmath228 ( @xmath219 ) .
the distributions @xmath229 and @xmath230 should be finite at @xmath231 and approach zero for large @xmath232 . therefore , we obtain @xmath233 from these results , we obtain @xmath234 where @xmath107 , @xmath108 and @xmath109 are , respectively , given by @xmath235
in this appendix , we calculate the @xmath36-th moments of @xmath236 and @xmath237 using the distribution function obtained in appendix [ sec : perturbation ] . from the definition of the moment , @xmath36-th moment of an arbitrary function @xmath238
is given by @xmath239 we rotate the coordinate the coordinate @xmath240 by @xmath69 counterclockwise and introduce the new cartesian coordinate @xmath68 as in fig .
[ fig : binwise ] . from this definition ,
we obtain the @xmath36-th moments of @xmath236 , for @xmath116 , @xmath241\nonumber\\ = & \frac{3}{\kappa^2}\left(1-\frac{5\dot\gamma}{2d\kappa^2}\sin2\theta\right),\end{aligned}\ ] ] for @xmath242 , @xmath243\nonumber\\ = & -\frac{765\epsilon a}{\kappa^7}\sin\theta,\end{aligned}\ ] ] and for @xmath244 , @xmath245\nonumber\\ = & \frac{45}{\kappa^4}\left(1-\frac{7\dot\gamma}{d\kappa^2}\sin2\theta\right).\end{aligned}\ ] ] similarly , we can calculate the each moment of @xmath237 so that we obtain eqs .
( [ eq:2nd_y])([eq:4th_z ] ) .
in this appendix , we derive the velocity distribution function in the cartesian coordinate @xmath240 at first , and calculate the velocity distribution functions in @xmath33 and @xmath34-directions . the velocity distribution function in the polar
coordinates @xmath246 is given by eq .
( [ eq : f_per ] ) , where we replace @xmath69 by @xmath121 as in eqs .
( [ eq:2nd_y])([eq:4th_z ] ) , which can be converted into the form in cartesian coordinate as @xmath247\nonumber\\ = & \frac{\kappa^2}{2\pi}\exp(-\kappa p ) \left[1+\frac{\epsilon a}{3\kappa^3}(3+\kappa p+\kappa^2p^2 ) ( p_y\sin\psi - p_z\cos\psi ) + \frac{\dot\gamma}{4d}\left\{(p_y^2-p_z^2)\sin2\psi-2p_yp_z\cos2\psi\right\}\right],\label{eq : f_pypz}\end{aligned}\ ] ] where @xmath248 .
next , let us calculate the velocity distribution functions in @xmath33 and @xmath34 directions . in this paper , we focus on the vdf for the fluctuation velocity , which is defined by the deviation from the average velocity .
therefore , we can replace @xmath249 and @xmath250 by @xmath251 and @xmath252 in eq .
( [ eq : f_pypz ] ) .
the velocity distribution function in @xmath33-direction , @xmath253 , is given by integrating eq .
( [ eq : f_pypz ] ) with respect to @xmath133 as 99 h. krupp , adv .
colloid interface sci .
* 1 * , 111 ( 1967 )
. j. visser , powder technol . *
58 * , 1 ( 1989 ) . j. m. valverde , a. castellanos , a. ramos , and p. k. watson , phys . rev .
e * 62 * , 6851 ( 2000 ) .
j. tomas , particul .
technol .
* 19 * , 95 ( 2001 ) .
f. cansell , c. aymonier , and a. loppinet - serani , curr .
solid st .
m. * 7 * , 31 ( 2003 ) .
j. tomas , granul .
matter * 6 * , 75 ( 2004 ) .
a. castellanos , adv .
* 54 * , 263 ( 2005 ) .
butt , b. cappella , and m. kappl , surf .
sci . rep . * 59 * , 1 ( 2005 ) .
j. tomas , chem .
* 62 * , 1997 ( 2007 ) .
r. tykhoniuk , j. tomas , s. luding , m. kappl , l. heim , and h .- j . butt , chem .
* 62 * , 2843 ( 2007 ) .
g. calvert , m. ghadiri , and r. tweedie , adv .
powder technol .
* 20 * , 4 ( 2009 ) .
j. r. van ommen , j. m. valverde , and r. pfeffer , j. nanopart .
14 * , 1 ( 2012 ) .
j. n. israelachvili , _ intermolecular and surface forces third edition _
( academic press , new york , 2011 ) . k. l. johnson , k. kendall , and a. d. roberts , proc . r. soc .
a * 324 * , 301 ( 1971 ) .
b. v. derjaguin , i. i. abrikosova , and e. m. lifshitz , chem .
10 * , 295 ( 1956 ) .
b. v. derjaguin , v. m. muller , and y. p. toporov , j. colloid interface sci . * 53 * , 314 ( 1975 ) .
k. z. y. yen and t. k. chaki , j. appl .
* 71 * , 3164 ( 1992 ) . c. thornton , k. k. yin , and m. j. adams , j. phys .
d : appl .
* 29 * , 424 ( 1996 ) . t. mikami , h. kamiya , and m. horio , chem
. eng .
sci . * 53 * , 1927 ( 1998 ) .
g. bartels , t. unger , d. kadau , d. e. wolf , and j. kertsz , granul .
matter * 7 * , 139 ( 2005 ) .
s. luding , r. tukhoniuk , and j. tomas , chem .
technol .
* 26 * , 1229 ( 2003 ) .
s. luding , powder technol .
* 158 * , 45 ( 2005 ) .
s. luding , granul .
matter * 10 * , 235 ( 2008 ) .
j. r. royer , d. j. evans , l. oyarte , e. kapit , m. mbius , s. r. waitukatis , and h. m. jaeger , nature ( london ) , * 459 * , 1110 ( 2009 ) .
n. kumar and s. luding , arxiv:1407.6167 .
s. gonzalez , a. r. thornton , and s. luding , arxiv:1407.2370 .
d. hornbaker , r. albert , i. albert , a .-
barabsi , and p. schiffer , nature * 387 * , 765 ( 1997 ) .
l. bocquet , e. charlaix , s. ciliberto , and j. crassous , nature * 396 * , 735 ( 1998 ) . c. d. willett , m. j. adams , s. a. johnson , and j. p. k. seville , langmuir * 16 * , 9396 ( 2000 ) .
s. t. nase , w. l. vargas , a. a. abatan , and j. j. maccarthy , powder technol .
* 116 * , 214 ( 2001 ) .
s. herminghaus , adv .
* 54 * , 221 ( 2005 ) .
n. mitarai and f. nori , adv .
* 55 * , 1 ( 2006 ) .
s. ulrich and a. zippelius , phys .
* 109 * , 166001 ( 2012 ) .
f. gerl and a. zippelius , phys .
e * 59 * , 2361 ( 1999 ) .
h. hayakawa and h. kuninaka , chem .
sci . * 57 * , 239 ( 2002 ) .
a. awasthi , s. c. hendy , p. zoontjens , and s. a. brown , phys .
. lett . * 97 * , 186103 ( 2006 ) .
n. v. brilliantov , n. albers , f. spahn , and t. pschel , phys .
e * 76 * , 051302 ( 2007 ) .
m. suri and t. dumitric , phys .
b * 78 * , 081405 ( 2008 ) .
h. kuninaka and h. hayakawa , phys .
e * 79 * , 031309 ( 2009 ) .
k. saitoh , a. bodrova , h. hayakawa , and n. v. brilliantov , phys .
105 * , 238001 ( 2010 ) .
h. kuninaka and h. hayakawa , phys .
e * 86 * , 051302 ( 2012 ) .
h. tanaka , k. wada , t. suyama , and s. okuzumi , prog .
supp . * 195 * , 101 ( 2012 ) .
r. murakami and h. hayakawa , phys .
e * 89 * , 012205 ( 2014 ) . j .-
hansen and l. verlet , phys .
rev . * 184 * , 151 ( 1969 ) .
j. j. nicolas , k. e. gubbins , w. b. streett , and d. j. tildersley , mol .
37 * , 1429 ( 1979 ) .
y. adachi , i. fijihara , m. takamiya , and k. nakanishi , fluid phase equilibr . * 39 * , 1 ( 1988 ) .
a. lotfi , j. vrabec , and j. fischer , mol .
* 76 * , 1319 ( 1992 ) .
j. k. johnson , j. a. zollweg , and k. e. gubbins , mol .
phys . * 78 * , 591 ( 1993 ) .
j. kolafa , h. l. vrtler , k. aim , and i. nezbeda , mol .
simulat . *
11 * , 305 ( 1993 ) .
r. h. heist and h. he , j. phys .
chem . ref .
data * 23 * , 781 ( 1994 ) .
a. laaksonen and d. kashchiev , j. chem .
phys . * 98 * , 7748 ( 1994 ) .
k. yasuoka and m. matsumoto , j. chem .
phys . * 109 * , 8451 ( 1998 ) . i. goldhirsch and g. zanetti , phys .
lett . * 70 * , 1619 ( 1993 ) . i. goldhirsch , m .-
l . tan , and g. zanetti , j. sci .
comput . * 8 * , 1 ( 1993 ) .
s. mcnamara and w. r. young , phys .
e * 53 * , 5089 ( 1996 ) .
s. mcnamara , phys .
fluids a * 5 * , 3056 , ( 1993 ) .
n. brilliantov , c. saluea , t. schwager , and t. pschel , phys .
lett . * 93 * , 134301 ( 2004 ) .
m. e. lasinski , j. s. curtis , and j. f. pekny , phys .
fluids * 16 * , 265 ( 2004 ) .
s. l. conway and b. j. glasser , phys .
fluids * 16 * , 509 ( 2004 ) .
k. saitoh and h. hayakawa , phys .
e * 75 * , 021302 ( 2007 ) .
m. alam and p. shukla , j. fluid mech .
* 716 * , 349 ( 2013 ) .
p. shukla and m. alam , j. fluid mech .
* 718 * , 131 ( 2013 ) .
y. gu , s. chialvo , and s. sundaresan , phys .
e * 90 * , 032206 ( 2014 ) .
a. w. lees and s. f. edwards , j. phys .
c : solid state phys . * 5 * , 1921 ( 1972 ) .
s. takada and h. hayakawa , aip conf .
proc . * 1542 * , 819 ( 2013 ) .
d. j. evans and g. p. morriss , phys .
a * 30 * , 1528 ( 1984 ) .
d. j. evans and g. morriss , _ statistical mechanics of nonequilibrium liquids second edition _ ( cambridge university press , cambridge , 2008 ) .
d. chandler and j. d. weeks , phys .
lett . * 25 * , 149 ( 1970 ) .
j. d. weeks , d. chandler , and h. c. andersen , j. chem .
phys . * 54 * , 5237 ( 1971 ) .
s. takada and h. hayakawa , aip conf .
proc . * 1518 * , 741 ( 2013 ) .
k. c. vijayakumar and m. alam , phys .
e * 75 * , 051306 ( 2007 ) .
m. alam and v. k. chikkadi , j. fluid mech .
* 653 * , 175 ( 2010 ) .
r. kubo , m. toda and n. hashitsume , _ statistical physics ii : nonequilibrium statistical mechanics _
( springer , berlin , 1991 ) .
r. zwanzig , _ nonequilibrium statistical mechanics _ ( oxford university press , new york , 2001 ) .
a. kawarada and h. hayakawa , j. phys .
jpn . * 73 * , 2037 ( 2004 ) .
. de gennes , j. stat . phys .
* 119 * , 953 ( 2005 ) .
h. hayakawa , physica d * 205 * , 48 ( 2005 ) .
i. goldhirsch , annu .
fluid mech . * 35 * , 267 ( 2003 ) .
i. goldhirsch , powder technol . *
182 * , 130 ( 2008 ) .
a. onuki , _ phase transition dynamics _
( cambridge university press , cambridge , 2002 ) .
k. saitoh and h. hayakawa , granul .
matter * 13 * , 697 ( 2011 ) .
v. garz and a. santos , _ kinetic theory of gases in shear flows nonlinear transport _ ( kluwer academic publishers , dordrecht , 2003 ) . s .-
chong , m. otsuki , and h. hayakawa , prog .
suppl . * 184 * , 72 ( 2010 ) . s .-
chong , m. otsuki , and h. hayakawa , phys .
e. * 81 * , 041130 ( 2010 ) . h. hayakawa and m. otsuki , phys . rev .
e * 88 * , 032117 ( 2013 ) . m. abramowitz and i. a. stegun , _ handbook of mathematical functions : with formulas , graphs , and mathematical tables _ ( dover publications , new york , 1964 ) . | three dimensional molecular dynamics simulations of cohesive dissipative powders under a plane shear are performed .
we find the various phases depending on the dimensionless shear rate and the dissipation rate as well as the density .
we also find that the shape of clusters depends on the initial condition of velocities of particles when the dissipation is large .
our simple stochastic model reproduces the non - gaussian velocity distribution function appearing in the coexistence phase of a gas and a plate . |
the hall conductivity in the mixed state of a clean ( @xmath0 ) type - ii s - wave superconductor is determined from a microscopic calculation within a quasiclassical approximation .
we find that below the superconducting transition the contribution to the transverse conductivity due to dynamical fluctuations of the order parameter is compensated by the modification of the quasiparticle contribution . in this regime
the nonlinear behaviour of the hall angle is governed by the change in the effective quasiparticle scattering rate due to the reduction in the density of states at the fermi level .
the connection with experimental results is discussed .
the hall effect in the mixed state of type - ii superconductors has remained a theoretical puzzle for almost thirty years @xcite .
the existing phenomenological @xcite theories predict that the hall angle in the flux - flow regime is either identical to that in the normal state @xcite or constant @xcite .
theories which make use of the time - dependent ginzburg - landau equations ( tdgl ) also find no modification of the hall conductivity in the superconducting state @xcite .
these predictions are at variance with the strongly nonlinear behaviour ( as a function of magnetic field ) found in experiments performed on both low-@xmath1 materials @xcite and high-@xmath1 cuprates @xcite . for dirty superconductors ( @xmath2 , where @xmath3 is the mean free path and @xmath4 is the superconducting coherence length ) , transport coefficients can be determined from microscopic theory by a straightforward expansion in powers of the order parameter , @xmath5 .
the results of such a calculation for the transverse resistivity @xcite explain qualitatively the sharp increase in the hall angle observed in experiment (
although , to our knowledge , no systematic comparison has been made ) , and provide the physical basis for a generalized tdgl approach , in which the relaxation rate is assumed to be complex , rather than purely real , to allow for a modification of the transverse transport coefficients @xcite .
the small parameter in the expansion of the microscopic equations is proportional to both the order parameter and the mean free path , therefore , it is not small in the clean ( @xmath0 ) limit . in this regime
a straightforward expansion is not possible and the tdgl equations are not applicable @xcite , so that an alternative approach is needed for the calculation of the transverse transport coefficients . in this letter
we give the results of a calculation of the hall conductivity of a clean s - state superconductor in the mixed state near the upper critical field , @xmath6 , which uses a quasiclassical approximation to the microscopic theory .
we made this choice as both the normal state and superconducting properties of the low-@xmath1 compounds are well known , and comparison between theory and experiment is fraught with less ambiguity .
the quasiclassical approach @xcite has been applied successfully in the past to study transport phenomena in superfluids @xcite and superconductors @xcite and more recently to the unconventional superconductors @xcite . the central quantity in this method is the single - particle matrix green s function @xmath7 integrated over the quasiparticle energy @xmath8 here the @xmath9 are fermionic matsubara frequencies , @xmath10 is the normalized parameterization of the fermi surface , @xmath11 is the center of mass coordinate and @xmath12 is the relative momentum . since the green s function is strongly peaked at the fermi momentum @xmath13 , which is normally far larger than any other momenta in the problem , slower varying quantities such as the self - energy and external potential can be expanded around their values at the fermi surface .
the result of such an expansion @xcite in the small parameter @xmath14 is a set of transport - like equations for the quasiclassical propagator @xmath15 .
we have generalized these equations to include terms responsible for the hall effect in a charged superfluid .
technical details of the derivation will be reported elsewhere @xcite , here we use the equations to determine the transverse dc - conductivity we use linear response theory in the vector potential @xmath16 describing a constant electric field @xmath17 .
the magnetic field @xmath18 , chosen parallel to the z - axis , is described by the vector potential @xmath19 .
we consider a spherical fermi surface and use the born approximation for s - wave impurity scattering characterized by a collision rate @xmath20 .
the spatial dependence of the order parameter is modeled by the periodic abrikosov vortex lattice @xmath21 here @xmath22 is the lowest energy eigenfunction of the linearized ginzburg - landau equation ( i.e. the eigenfunction of a harmonic oscillator with cooper pair mass @xmath23 and frequency @xmath24 ) and @xmath25 is the magnetic length ( @xmath26 for fields @xmath27 ) .
this approach is appropriate provided that the broadening of the levels in the vortex core is large compared to their spacing @xmath28 , it breaks down in the superclean regime ( cf .
@xcite ) . in the clean limit
the finite lifetime is accounted for by replacing @xmath29 by @xmath30 ( angular brackets denote an average over the fermi surface ) ; corrections to the order parameter due to impurity renormalization are of the order @xmath31 and can be ignored . the equations for the unperturbed functions @xmath32 and @xmath33 and the linear , in @xmath34 , corrections to the propagator @xmath35 and @xmath36 are @xcite @xmath37 f = 2 i \delta g
\\ \fl \label{f1 } \bigl [ 2 \widetilde \omega_n + { \bf v}_f ( \nabla - 2 i e { \cal a})\bigr ] f_1 = i e { \bf v}_f { \bf { a } } ( f + f(- ) ) + i \delta ( g_1 - \bar g_1 ) + i \delta_1 ( g + g(- ) ) \\ \fl \label{g1 } \bigl(i \widetilde \omega_0 + i \omega_c{\partial \over \partial \phi}\bigr ) \bigl(g_1- \bar g_1\bigr)= 2 e { \bf v}_f { \bf a } ( g - g(- ) ) + \delta_1^{\star } \bigl(f - f(-)\bigr)+ \delta_1 \bigl(f^{\dagger}- f^{\dagger}(-)\bigr ) \\ \nonumber \qquad \qquad + ( 2\tau)^{-1 } \biggl ( \langle f_1^{\dagger } \rangle \bigl(f - f(-)\bigr ) + \langle f_1 \rangle \bigl(f^{\dagger}- f^{\dagger}(-)\bigr ) \biggl ) \\
\nonumber \qquad \qquad -{i \over 2}\biggl [ { \partial \delta_1^{\star } \over \partial { \bf r } } { \partial \over \partial { \bf p}_{|| } } \bigl(f+f(-)\bigr ) + { \partial \delta_1 \over \partial { \bf r } } { \partial \over \partial { \bf p}_{|| } } \bigl(f^{\dagger}+f^{\dagger}(-)\bigr ) \\
\nonumber \qquad \qquad \qquad + 2{\partial \delta^{\star } \over \partial { \bf r } } { \partial f_1^{\dagger } \over \partial { \bf p}_{|| } } + 2{\partial \delta \over
\partial { \bf r } } { \partial f_1 \over \partial { \bf p}_{|| } } \biggr],\end{aligned}\ ] ] where @xmath38 is the change in the order parameter induced by the electric field .
there are corresponding equations for @xmath39 and @xmath40 . in these equations @xmath41 is the pauli matrix , the fermi velocity @xmath42 , @xmath43 is the component of the momentum parallel to the fermi surface , @xmath44 is the cyclotron frequency , and the angular brackets denote an average over the fermi surface .
we have used the shorthand notations @xmath45 and @xmath46 , @xmath47 , and introduced @xmath48 , @xmath49 .
the superconducting order parameter satisfies the usual self - consistency condition @xmath50 , where g is the coupling constant and @xmath51 is the density of states at the fermi surface .
the unperturbed function @xmath15 obeys the normalization conditions @xmath52 and @xmath53@xcite .
we chose to write an equation for @xmath54 since the transport current is given by @xcite @xmath55 and the functions @xmath35 , @xmath56 depend on @xmath36 , @xmath57 in this combination only .
in equations ( [ f0])-([g1 ] ) we have omitted terms whose contribution to the conductivity is of order @xmath31 smaller than that of leading order terms .
equation ( [ f0 ] ) is the well - known static eilenberger equation @xcite and equations ( [ f0])-([g1 ] ) contain all the terms relevant to the hall effect for a clean type - ii superconductor in the high - field region .
it is clear from equation ( [ g1 ] ) that there are several distinct contributions to @xmath54 ( and therefore the current ) .
the first , the quasiparticle contribution , depends on the unperturbed function @xmath33 , while the second , proportional to @xmath38 , is due to the dynamical fluctuations of the order parameter induced by the perturbing electric field .
the third term in this equation describes the additional scattering of quasiparticles off these dynamic fluctuations ; it has the same origin as the thompson diagram in the analysis of transport in dirty superconductors @xcite .
the fourth term describes how as vortices move and are deformed by the transport current , the resulting gradients of the order parameter act as driving forces in the transport - like equations . finally , as the renormalization of the frequency @xmath58 in equation([g1 ] ) depends on the angular average of the quasiparticle propagator , which changes in the superconducting state , the effective transport mean free path is modified .
to proceed we approximate the diagonal part of the quasiclassical propagator by its spatial average @xcite . since the electromagnetic fields in a superconductor vary over distances of the order of the penetration depth @xmath59 ,
this is a very good approximation in the london limit @xmath60 ; even for compounds with moderate values of @xmath61 it remains valid for a wide field range below @xmath6 . in all of the following @xmath33
will stand for the averaged distribution function .
we now solve equation([f0 ] ) with the normalization condition to determine the unperturbed functions @xmath32 and @xmath33 .
we then determine @xmath35 and @xmath38 to leading order in @xmath62 by solving equation ( [ f1 ] ) together with the self - consistency condition for the order parameter . to accomplish this program the expression @xmath63 has to be evaluated .
to do this we exploit the oscillatory character of the abrikosov solution and introduce raising and lowering operators @xmath64 $ ] and @xmath65 $ ] obeying the usual bosonic commutation relations @xmath66=1 $ ] .
if the ground state equation ( [ vl ] ) is denoted by @xmath67 , the higher eigenstates ( modes of the order parameter ) are generated by the standard procedure @xmath68 .
then to make use of the properties of these operators the operator @xmath69 can be rewritten as @xmath70 $ ] and the result of its acting on any state @xmath71 evaluated explicitly @xcite .
using this approach we are able to determine the unperturbed functions @xmath72^{-1/2}\ ] ] where @xmath5 is the spatial average of the order parameter and @xmath73 , and @xmath74 the expression for @xmath33 reproduces correctly the gapped bcs - like function for quasiparticles traveling parallel to the magnetic field , while describing gapless behaviour in all other directions .
a similar expression for @xmath33 has been obtained by pesch @xcite . since @xmath32 is a fourier series in @xmath75 , the mode with @xmath76 will couple to a scalar potential , the mode with @xmath77 to a transverse potential etc .
then we find @xmath78 frequency ) .
using this value to determine the correction to anomalous propagator @xmath35 and calculate the current from equation([curr ] ) we obtain the longitudinal and transverse conductivities up to order @xmath79 .
if we use the usual notation @xmath80 for the normal state conductivity @xmath81 , the enhancement of the transverse current due to lorentz force driven fluctuations of the order parameter @xmath82 is exactly compensated by the modification of the quasiparticle hall current due to additional scattering off the vortex lattice @xmath83 similarly , the positive contribution to the transverse conductivity due to the forces generated by gradient of the excited mode of the order parameter @xmath84 is cancelled by the additional scattering introduced by the deformed and moving vortex lattice @xmath85 as a result , the behaviour of the transverse conductivity @xmath86 is determined solely by the effect of the modification of the effective elastic scattering time @xmath87 on the leading order quasiparticle contribution . for the dc conductivity
this change is due to the decrease in the number of states at the fermi surface available for scattering as the superconducting gap opens .
we find , in agreement with the result of pesch @xcite for the density of states , that the increase in the relaxation time is a non - analytic function of the small parameter @xmath88 @xmath89,\ ] ] and , up to order @xmath79 the transverse conductivity is given by @xmath90.\ ] ] the longitudinal conductivity is obtained in a similar way @xmath91.\ ] ] in the high - field region the square of the order parameter is linear in the applied magnetic field and is given by @xcite @xmath92 in figure 1 we show the qualitative behaviour of @xmath86 as a function of magnetic field , plotted using parameter values for nb .
the transverse conductivity is enhanced below the upper critical field and has negative curvature in the high field region .
while the transverse conductivity is proportional to the square of the scattering time , the hall angle @xmath93 is only linearly dependent on the scattering time and the corresponding nonlinear dependence on magnetic field is weaker , as can be seen in figure 2 . finally , as the transverse resistivity @xmath94 is independent of the effective scattering time , upon entering the superconducting state it remains linear in magnetic field with the same slope as in the normal metal .
this behaviour is to be contrasted with that of bardeen - stephen model @xcite , where the resistivity is modified but the hall angle obeys the same linear law as in the normal state . on the other hand , the nozieres - vinen theory@xcite , which predicts that the hall angle should be constant in the flux - flow regime below @xmath6 at variance with the result of this work , also finds that the transverse resistivity is identical to that of the normal state , although the individual components of the conductivity tensor are quite different from those found here .
a comparison can be made with the experimental data of fiory and serin @xcite on high purity nb .
these experiments find a transverse resistivity in the flux - flow regime which is linear in the applied magnetic field over a wide range of fields below @xmath6 .
the hall angle , however , flattens or even increases above its value at @xmath6 before decreasing at lower fields .
these results are more suggestive of the behaviour given here than the original interpretation given in terms of the nozieres - vinen theory . also , the longitudinal resistivity found in @xcite has a distinct increase in slope just below the upper critical field , which is consistent with the logarithmic behaviour given by equation ( [ sxx ] ) .
such comparisons are , of course , only qualitative , and more experimental work is needed to make a detailed comparison with the theory . to conclude , we give the results of a microscopic calculation of the hall resistivity of a clean type - ii s - state superconductor in the high - field limit .
we find that the field dependence of the hall conductivity in the high field regime , which is non - analytic , is entirely due to the change in the density of quasiparticle states at the fermi level in the superconducting state . at the same time we find that the field dependence of the transverse resistivity below the upper critical field remains unchanged .
these results are in qualitative agreement with the experimentally observed behaviour .
the approach developed here can be generalized to superconductors with unconventional order parameter symmetry , this work is now in progress .
one of us ( a h ) would like to thank a t dorsey , d rainer , and k scharnberg for important discussions .
this research was supported in part by the national science foundation through grant no dmr9008239 . |
|
x - ray photoemission spectroscopy pes is a useful tool for studying the electronic structure of solids .
the theoretical description of pes is however very complicated@xcite and almost all work has been based on the so - called sudden approximation.@xcite the photoemission spectrum is then described by the electron spectral function convoluted by a loss function , describing the transport of the emitted electron to the surface .
the sudden approximation becomes exact in the limit when the kinetic energy of the emitted electron becomes infinite.@xcite in this limit we can distinguish between intrinsic satellites , appearing in the electron spectral function , and extrinsic satellites , appearing in the loss function . for lower kinetic energy , this distinction
is blurred due to interference effects , and the satellite weights are expected to be quite different as we approach the opposite limit , the adiabatic limit , of low kinetic energy.@xcite it is then interesting to ask at what kinetic energy the sudden approximation becomes accurate .
this issue has been studied extensively for the case when the emitted electron couples to plasmons , and it has been found that the sudden approximation becomes valid only for very large ( @xmath14 kev ) kinetic energies.@xcite a semi - classical approach has been found to work exceedingly well for the study of plasmon satellites.@xcite in such a picture , one may take the emitted electron to move as a classical particle away from the region where the hole was created .
the system then sees the potential from both the created electron and hole .
initially the electron potential cancels the hole potential , but as the electron moves away , the hole potential is gradually switched on .
the switching - on of the hole potential may lead to the creation of excitations .
if the kinetic energy of the electron is sufficiently large , we can consider the hole potential as being switched on instantly , and the creation of excitations around the hole then reaches a limiting value , the sudden limit . in the semi - classical picture ,
the perturbation is turned on during a time @xmath15 , where @xmath16 is the photoelectron velocity and @xmath17 is the range of the interaction ( scattering ) potential between the emitted electron and the excitations . in this picture
we also need to determine the relevant time scale @xmath18 so that the sudden limit is reached if @xmath19 or @xmath20 .
our analysis within the semi - classical framework shows that @xmath21 is related to the energy @xmath22 of the relevant excitation of the system and to the strength @xmath23 of the scattering potential .
we find a different characteristic energy scale @xmath24 , where @xmath25 is a characteristic length scale of the scattering potential .
on dimensional grounds one may argue that the adiabatic - sudden approximation takes place when the kinetic energy of the emitted electron is comparable to @xmath12 .
this would differ dramatically from the semi - classical approach , where the transition takes place for energy of the order @xmath26 , i.e. e.g. , @xmath27 or @xmath28 .
alternatively , and again on dimensional grounds , one may argue that the sudden approximation becomes valid when the kinetic energy of the emitted electron is much larger than the energy @xmath5 of the relevant excitations of the system,@xcite in strong contrast to the two criteria above .
this latter criterion is however not true in general.@xcite for many systems with strong correlations , the core level spectrum can be understood in a charge transfer scenario.@xcite this is illustrated in fig . 1 for a cu compound , e.g. , a cu halide .
in the ground - state , cu has essentially the configuration d@xmath29 and all the ligand orbitals are filled . in the presence of a cu core hole , it becomes energetically favorable to transfer an electron from a ligand to the d - shell , obtaining a @xmath30 configuration on the cu atom with the core hole . due to the hybridization between the d@xmath29 and @xmath31 configurations ,
the states are actually mixtures of the two configurations , as indicated in fig .
[ fig : schematic ] . in the photoemission process
there is a nonzero probability that the outer electron will not stay on the ligand , but is transferred to the lower energy @xmath32-like state .
this `` shake - down '' process corresponds to the leading peak in the spectrum , while the process where the outer electron stays on the ligand corresponds to the satellite .
this kind of model has been applied to rare earth compounds,@xcite chemisorption systems,@xcite transition metal compounds,@xcite and high @xmath0 compounds.@xcite our simple model allows an accurate numerical calculation of the photocurrent either by integrating the time - dependent schrdinger equation or by directly inverting a resolvent operator ( qm ) .
we also derive analytic results with a separable potential .
these results are compared with the semi - classical theory ( sc ) , and with first order perturbation theory ( pt ) . in both these cases we have analytic results , which is very useful for understanding the physics of the problem .
the impurity model discussed here differs in certain important aspects from a real solid . to start with , for a solid
we never reach the limit of a pure intrinsic spectrum since when the cross section for extrinsic scattering goes to zero , the range from which the photoelectrons come goes to infinity . in our impurity model ,
on the other hand , the extrinsic scattering approaches zero at high kinetic energies .
secondly , for a solid , we discuss excitations in the continuum and not as here discrete energy levels . for the coupling to plasmons , the adiabatic - sudden transition takes place at large kinetic energies where the sc approximation is very accurate.@xcite the relevant length scale is given by the plasmon wave length @xmath33 and the relevant time by the inverse plasmon frequency @xmath34 .
large interference effects are then connected with a large phase velocity @xmath35 , as discussed e.g. by inglesfield.@xcite since long wave - length plasmons play an important role these large interference effects for small @xmath36 delay the approach to the sudden limit , which only is reached at very high kinetic energies ( @xmath14 kev ) . for the localized excitations studied here
the relevant length scale is much shorter .
the sc theory then predicts that the transition takes place at correspondingly smaller kinetic energies .
this is indeed what we find from the exact solution of our model .
this has two consequences .
firstly , the sc treatment itself is not valid at such small energies , and we have to rely on qm treatments .
actually , although the sc treatment correctly predicts a small transition energy , we find it predicts qualitatively wrong dependencies on the relevant parameters .
secondly , the smaller energy scale means that the energy variation of the dipole matrix elements becomes very important .
the dipole matrix element grows rapidly on an energy scale @xmath37 , which can become very important for the adiabatic - sudden transition .
we study the ratio @xmath1 between the weights of the satellite and the main peak as a function of the photon energy @xmath2 for the emission from a @xmath38 level .
first we consider the case when the scattering potential between the electron and the target is neglected .
we find that the corresponding ratio @xmath4 strongly depends on the ratio between the excitation energy @xmath5 and @xmath3 .
if @xmath39 , @xmath4 approaches its limiting value from below , while it has an overshoot if @xmath40 . in both case
the limit value is reached for photoelectron energies of the order of a few times @xmath5 .
we then study the effects of the scattering potential by focusing on @xmath41 . for small energies
there is typically an overshoot due to _ constructive _ interference in the shake - down case , contrary to the shake - up case where , like for plasmons , @xmath42 is reduced by interaction effects .
this happens on the energy scale @xmath9 .
if the scattering potential is very strong , this overshoot may extend to several times @xmath9 . depending on the parameters
there may be an undershoot for higher energies , which can extend up to quite high energies .
the undershoot is , however , rather small for the parameters we consider here , and should therefore not be very important unless we want to calculate the spectrum with a high accuracy . for cu compounds and emission from the @xmath38 core level , we find that @xmath43 and @xmath12 are comparable , and the relevant energy for @xmath11 is then given by @xmath44 .
we present our model in sec .
ii and calculate various matrix elements in sec .
the sudden approximation is described in sec .
iv and exact numerical methods are given in sec .
v. the perturbational and the semi - classical treatments are presented in sec .
vi . in sec .
vii we study the condition for the adiabatic - sudden transition qualitatively , using simple analytic matrix elements and within the framework of the semi - classical theory .
the results are discussed in sec .
we consider a hamiltonian @xmath45 describing a model with a core level @xmath46 and two valence levels @xmath47 and @xmath48 , @xmath49 the first two terms give the bare energies of the levels @xmath47 and @xmath48 , and the last term the hybridization between them .
the remaining terms involve the occupation number @xmath50 of the core level @xmath46 . in photoemission
the core level is filled in the initial state , and empty in the final , and @xmath51 only enters as a constant .
it is trivial to diagonalize @xmath52 , and one obtains two dressed energies @xmath53 and @xmath54 for the levels @xmath47 and @xmath55 in a cu compound , for instance , @xmath46 may represent the cu @xmath38 core level , @xmath47 the cu @xmath56 valence level and @xmath48 a ligand state .
this is schematically illustrated in fig .
[ fig : schematic ] . in our calculations
we almost always treat the case when @xmath57 , and @xmath58 .
the meaning of the levels @xmath47 and @xmath48 for different types of systems with localized excitations is indicated in table [ general - ct ] .
the full hamiltonian also has a one - electron part for continuum states , @xmath59 with the energies @xmath60 , and wavefunctions @xmath61 obtained from a one - electron potential corresponding to @xmath62 .
the @xmath63 are occupation numbers @xmath64 .
we use atomic units with @xmath65 and thus , e.g. energies are in hartrees ( @xmath66 ev ) .
the perturbation causing photoemission is @xmath67 where @xmath68 is an optical transition matrix element .
we take the photoelectron interaction as @xmath69c_{{\bf k}}^{\dagger } c_{{\bf k}^{^{\prime } } } .
\label{eq:5}\ ] ] here @xmath70 is a matrix element of the coulomb potential @xmath71 from the charge density @xmath72 of the orbital @xmath73 , @xmath74 it is the potential @xmath75 which determines the transition from the adiabatic to the sudden limit , with @xmath76 we are in the sudden limit .
the total hamiltonian is given by @xmath77 this hamiltonian has two conserved quantities , @xmath78 for simplicity we take the core electron and the @xmath32 electron potentials as equal , @xmath79 , and use the relation @xmath80 to obtain @xmath81 where @xmath82.\end{aligned}\ ] ] the ( scattering ) potential @xmath83 describes the change in the potential acting on the emitted electron when the electron in the target hops from level @xmath47 to @xmath48 . dropping a constant we can write @xmath45 as @xmath84 where the coulomb integral @xmath85 is given by @xmath86.\end{aligned}\ ] ] since the core level is very localized in space this leads to @xmath87 for the different types of systems in table [ general - ct ] , @xmath47 refers to a localized level and @xmath48 refers to a more extended level . for instance , for a copper dihalide compound , @xmath47 refers to a cu @xmath56 orbital and @xmath48 to a combination of orbitals on the ligand sites . for simplicity , we approximate the six ligand orbitals by a spherical shell with the radius @xmath17,@xcite where @xmath17 is the average cu - ligand separation .
the potential from the cu @xmath56 orbital @xmath88 can be considered as purely coulombic at @xmath89 .
the charge from the spherical shell gives a constant potential inside the radius @xmath17 , and we have @xmath90 here @xmath91 is a constant chosen to make @xmath92 , which may be thought of as being due to screening by the surrounding .
since @xmath93 , @xmath91 varies only weakly with @xmath17 and is approximately given by @xmath94 .
to estimate the matrix elements @xmath68 and @xmath95 we must approximate the photoelectron wavefunctions @xmath96 .
these wavefunctions are calculated from the potential of a neutral atom , which further is shifted to make the potential zero outside a muffin - tin radius @xmath97 .
the states are then described by spherical bessel functions outside @xmath97 , which are matched to a solution of the atomic potential inside @xmath97 . for the energy @xmath98
we obtain the partial wave @xmath99 & $ r > r_{mt}$,\cr } \label{eq : ma1}\ ] ] where @xmath100 is the solution of the radial schrdinger equation for the atomic potential inside the muffin - tin radius , @xmath101 is a matching coefficient and @xmath102 a phase shift .
the normalization is given by @xmath103 where @xmath104 is the radius of a large sphere to which the continuum states are normalized .
the factor @xmath105 is due to the normalization and the asymptotic behavior of @xmath106 for large @xmath107 .
slater s rules@xcite are used to generate the orbitals and charge densities , from which the potential @xmath88 is calculated .
this gives the scattering potential @xmath108 , which is shown in fig .
[ fig : potential]a .
we consider photoemission from a cu @xmath38 hole . due to the dipole selection rules ,
the core electron is then emitted into a continuum state of @xmath109-symmetry .
the matrix elements @xmath110 of the scattering potential @xmath108 are shown in fig .
[ fig : potential]b , @xmath111 where the muffin - tin radius @xmath97 is taken as the ionic radius of cu , @xmath112 a.u . , and we have dropped the @xmath113 index , since we always consider @xmath114 the dipole matrix element @xmath115 is given by @xmath116 we assume that the core level is deep , and @xmath117 much larger than the energy difference between the ligand and copper levels .
we can then take the factor @xmath118 as a constant , which drops out since we always consider relative intensities .
the result for @xmath115 is shown as the solid line in fig .
[ fig : dipole]a .
these dipole and scattering potential matrix elements are used in the following numerical calculations .
extensive calculations of dipole matrix elements for many systems were performed by yeh and lindau.@xcite to interpret the results , it is useful to also perform analytical calculations . for this purpose
we need models of the matrix elements .
below we consider the limits of low and high kinetic energies of the emitted electron . in the limit of low kinetic energies ,
we replace the spherical bessel function by its expansion for small arguments @xmath119 and the solution @xmath120 by its zero energy limit @xmath121 .
this leads to @xmath122\frac{1}{\psi_{l0}(r_{mt})}.\end{aligned}\ ] ] due to the matching , the coefficient @xmath101 contains the ratio @xmath123 .
the value of the coefficient therefore depends in an interesting way on the wave function @xmath124 and its derivative , and if @xmath125 is close to zero @xmath101 blows up .
then the matrix elements of the scattering potential also blow up and we may expect strong deviations from the sudden approximation .
in such a case the dipole matrix element @xmath126 also becomes large , i.e. , there is a resonance ( @xmath127 ) in the photoemission cross section . from eqs.([eq :
ma1 ] ) and ( [ eq : ma1b ] ) it follows that in the limit of a small @xmath128 the dipole matrix element is proportional to @xmath129 .
the main peak and the satellites in the photoemission spectrum correspond to different kinetic energies and therefore have dipole matrix elements with different @xmath128-values .
this is important at low energies , while for large energies the variation in @xmath126 is generally small over a range of the energy difference between the main peak and the satellite , and for the ratio of the peaks , the dipole matrix elements should then not play a role . for simplicity , we assume that the dipole matrix elements become independent of @xmath130 for @xmath131 , where @xmath132 is some typical length scale of the system .
for our case ( @xmath133 ) we use the model ( note that any constant factor in @xmath115 drops out in our final expressions ) @xmath134 where @xmath135 .
[ fig : dipole]a compares this model with the full calculation for a @xmath38 orbital .
we obtain @xmath136 a.u .. for a @xmath137 or @xmath138 orbital the length scale is smaller and @xmath139 a.u .. while we consider @xmath133 the behaviour for other @xmath113 is primarily modified for small energies .
we next consider the matrix elements @xmath140 . for small values of @xmath128 and @xmath141 and for @xmath133
@xmath142 ^2 \frac{(kk^{^{\prime}})^2}{(\xi+2)^2}\int_0^{r_{mt}}\psi_{l0}^2(r)v_{sc } ( r ) r^2dr \nonumber \\ & + & \frac{2}{9r}(kk^{^{\prime}})^2\int_{r_{mt}}^{r_0 } \left[r+\left(\frac { 1-\xi}{2+\xi}\right)\frac{r_{mt}^3}{r^2}\right]^2 v_{sc}(r)r^2dr .
\nonumber \\\end{aligned}\ ] ] for small values of @xmath128 and @xmath141 it then follows that @xmath143 . for large values of @xmath128 and
@xmath144 the matrix elements become very small due to destructive interference between the two wave - functions unless @xmath145 . if @xmath146 the matrix elements @xmath147 approach a constant .
these features are contained in the model @xmath148 where @xmath25 and @xmath149 are appropriate length scales and @xmath150 has the energy dimension .
we recall that @xmath151 contains information about the coefficient @xmath101 defined in eq.([eq : ma3 ] ) and therefore about the atomic potential and that @xmath151 may become particularly large close to a resonance . the dipole matrix element @xmath152 and potential matrix @xmath140 in our simplified model eqs .
( [ eq : ma4 ] ) and ( [ eq : ma6 ] ) are given in fig .
[ fig : dipole ] as compared with the exact results . for large values of @xmath128 and @xmath153 , the expression eq .
( [ eq : ma6 ] ) simplifies to @xmath154 an expression of this type can also be derived by assuming that the wave functions @xmath155 can be approximated by spherical bessel functions in all of space , and by assuming some shape of @xmath108 , e.g. , a linear dependence on @xmath156 @xmath157 for large values of @xmath128 and @xmath153 we then obtain @xmath158}{[(k - k^{^{\prime } } ) r_{0}]^{2}}}. \label{eq : ma6aa}\ ] ] for this model we relate @xmath159 to the range @xmath17 of the potential and @xmath160 . using this identification in eq .
( [ eq : ma6c ] ) , leads to the correct average value of @xmath147 and to the correct width in @xmath161 of @xmath162 . the simple form ( [ eq : ma6c ] ) , however , neglects the effects of the oscillations of the cos - function in eq .
( [ eq : ma6aa ] ) for large values of @xmath163 , and it therefore gives a worse representation of the linear potential ( [ eq : ma6a ] ) than the form ( [ eq : ma6 ] ) gives for the more realistic scattering potential ( [ eq : m4 ] ) .
later we will find that it is a reasonable approximation to put @xmath132 , @xmath164 and @xmath7 equal to the same value @xmath165 , and introduce the corresponding energy @xmath166 .
we first discuss the photoemission in the sudden limit , i.e. , we neglect the scattering potential between the emitted electron and the target ( @xmath167 ) .
the initial state @xmath168 is the ground state of @xmath169 with @xmath170 and given by @xmath171 and the corresponding ground state energy is @xmath172 the final states of the target are given by the two eigenstates of @xmath173 with @xmath174 @xmath175 with @xmath176 and the corresponding energy eigenvalues @xmath177 and @xmath178 are @xmath179 with @xmath180 being the optical excitation energy of the system .
the photocurrent @xmath181 ( @xmath182 ) is given by,@xcite @xmath183 is a final state .
according to the sudden approximation , it can be written as the final target state multiplied by the photoelectron state , @xmath184 .
this gives @xmath185 @xmath186 gives the main line ( corresponding to the quasi particle line in metal ) and @xmath187 the satellite line .
the schematic picture of the initial and final state for this system is given in fig .
[ fig : schematic ] . summing the kinetic energy distribution of the photoelectron
, we obtain the absorption spectra @xmath188 , @xmath189 where the threshold energies for @xmath190 and @xmath191 are given by @xmath192 and @xmath193 , respectively
. we can thus also write @xmath194 and @xmath195 .
the factor @xmath196 comes from the @xmath128-summation over a @xmath197-function in energy . for convenience
we introduce the quantity @xmath198 , and thus @xmath199 . in the sudden approximation the kinetic energy of
the emitted electron is large , and we can take @xmath200 .
the ratio @xmath201 of the satellite to the main peak intensity then is @xmath202 taking into account the energy dependence of the dipole matrix element according to model eq .
( [ eq : ma4 ] ) as well as the factor @xmath196 , we obtain @xmath203 ^{\frac{3}{2 } } \nonumber \\ & & \times \left [ { \frac{1+(\tilde{\omega}+\delta e)/\tilde{e}_{d}}{1+\tilde{% \omega}/\tilde{e}_{d}}}\right ] ^{2}\theta ( \tilde{\omega}).\end{aligned}\ ] ] we now require that the ratio @xmath4 should reach a fraction @xmath204 ( @xmath205 ) of its limiting value @xmath206 for @xmath207 .
this gives @xmath208 this criterion refers to the energy where @xmath4 reaches a fraction @xmath209 in its rising part , and it does not consider that there is a large overshoot for @xmath40 . in this case
we can instead require that @xmath4 is smaller than @xmath205 in its descending part .
this gives the condition @xmath210 in fig .
[ fig : sudden ] we show results for @xmath4 over a large range of values for @xmath211 .
the figure illustrates that the dipole matrix element effect alone makes the sudden approximation invalid for small kinetic energies .
it is interesting that for somewhat larger photon energies @xmath212 overshoots .
the reason is that the matrix elements @xmath115 saturate for @xmath213 , while the factor @xmath196 in eq.([eq : s10 ] ) favors the satellite . for @xmath214 ,
the result is rather close to the sudden limit for @xmath215 .
finally , for @xmath40 , there is a substantial overshoot .
as discussed in the introduction , we would like to study how the adiabatic to sudden transition depends on certain factors , like the range @xmath17 of the potential and the energy @xmath5 of the excitation causing the satellite .
we therefore keep ratio of @xmath216 , @xmath85 and @xmath217 fixed , but vary their magnitude . in this way
we can vary @xmath218 without varying the magnitude of the satellite in the sudden limit .
( [ eq : m3 ] ) requires that we vary @xmath219 as we vary @xmath5 ( via @xmath85 ) , e.g. , by varying the dielectric constant @xmath91 . in some of the calculations below , however , we do not impose eq .
( [ eq : m3 ] ) , to be able to see the effect of varying @xmath5 alone .
we furthermore vary the range @xmath17 of the potential . from the definition eq .
( [ eq : m4 ] ) it follows that this would also vary the strength of the potential .
for this reason we simultaneously vary the dielectric constant @xmath91 so that @xmath220 stays unchanged when @xmath17 is changed .
alternatively , we can use the analytical matrix elements ( [ eq : ma4 ] , [ eq : ma6 ] ) .
we can then easily vary the length scale by changing @xmath165 or the strength by changing @xmath221 . to know roughly what are interesting values for our parameters we use experimental results for some copper dihalides.@xcite we estimate the relative strength of the satellite to the main peak , and the energy difference between the peaks .
this gives two equations while in our model these quantities , @xmath201 and @xmath5 , depend on three parameters , @xmath222 and @xmath223 .
to only have two parameters we consider _ the symmetric case _ @xmath224 and @xmath225 as shown in fig.[fig : schematic ] . in the symmetric case
we are restricted to the shake - down situation since before the transition the @xmath47-level is above the @xmath48-level , @xmath226 , while after the transition the @xmath47-level is below the @xmath48-level , @xmath227 . in the symmetric case we have @xmath228 , and @xmath229 .
once we know where in the ball - park we have @xmath216 and @xmath85 , we can leave the symmetric case , and also consider , e.g. , shake - up cases when there is no level crossing , @xmath230 . in the lowest final state the electron essentially stays on level @xmath48 , while the transfer of the electron to the level @xmath47 corresponds to a shake up satellite . in this case
we have @xmath231 and @xmath232 .
our calculations usually take the cucl@xmath233 parameters as reference values . for cucl@xmath233
we have @xmath234 , which gives @xmath235 .
further @xmath236 a.u . , @xmath237 a.u .
( with @xmath238 a.u . ) , and @xmath239 a.u . ,
i.e. , @xmath240 and @xmath241 ( see also table ii ) .
to obtain exact results for model eq.([eq:6 ] ) , we use a time - dependent formulation@xcite and solve the schrdinger equation for the hamiltonian @xmath242 the interaction is switched on at @xmath243 , using @xmath244 here @xmath245 is a small quantity to assure that the external field is switched on smoothly .
the initial ( @xmath243 ) state @xmath246 is given by the ground state of @xmath247 with @xmath248 in eq.([eq:1 ] ) .
after a time @xmath249 , the state @xmath250 of the system is @xmath251 the coefficients of @xmath250 can be determined by @xmath252 which gives four differential equations for the four coefficients @xmath253 , @xmath254 , @xmath255 , and @xmath256 , @xmath257 @xmath258 @xmath259 @xmath260 where @xmath261 with @xmath262 representing the strength of the external field .
we solve the equations in the limit when @xmath263 , and thus the ratio between @xmath264 and @xmath265 is independent of @xmath262 .
the initial conditions are @xmath266 , @xmath267 , and @xmath268 .
thus the problem is reduced to solving the coupled differential equations , which is done using the runge - kutta fourth - order method .
the photoelectron currents @xmath269 and @xmath270 corresponding to main and satellite lines , respectively , are given by @xmath271 @xmath272 where @xmath249 is a sufficiently large time .
we let the system evolve for a time of the order @xmath273 to obtain converged results for a given finite @xmath274 . in principle , we should perform the calculation for a few small values of @xmath245 and then extrapolate to @xmath275 followed by an extrapolation @xmath276 . here , for simplicity we have performed the calculation for one single small value of @xmath262 .
the calculation was performed for @xmath277 ev , 0.08 ev , 0.02 ev , and the results were extrapolated to @xmath275 assuming the @xmath245 dependence @xmath278 .
the error in this approach occurs primarily for small @xmath279 ev ) , and it is then less than 5% in @xmath11 .
the approach above gives the relative intensity of the main and satellite peaks @xmath280 it can be shown that the formulas ( [ eq : t1 ] , [ eq : t2 ] ) above give identical results to the more conventional formulation ( [ eq : pt3 ] ) below , by performing derivations of the type made in , e.g. , ref . .
as an example of the results obtained in this formalism , we show in fig .
[ fig : cu - dihalides ] results for the copper dihalides cubr@xmath233 , cucl@xmath233 and cuf@xmath233 .
the corresponding parameters are shown in table [ used - parameter ] and were estimated from experiment.@xcite the figure illustrates that there is a small `` overshoot '' for small @xmath281 but that the sudden limit is reached fairly quickly as @xmath281 is further increased .
we remind that in our cucl@xmath233 reference case @xmath282=0.195 a.u.=5.3 ev .
alternatively , we can work in the energy space , and obtain the spectrum by direct inversion of a resolvent operator .
we consider the hamiltonian @xmath283 , @xmath284 where by @xmath45 we understand @xmath285 .
the exact final photoemission state @xmath286 is@xcite @xmath287 @xmath288 ( @xmath182 ) ( eq . ( [ eq : s4 ] ) ) are the exact ( target ) eigenstates of @xmath285 and @xmath289 is the energy of the final state . using eq .
( [ eq : pt2 ] ) we calculate the matrix element @xmath290 @xmath291 \delta |\psi _ { 0}\rangle . \nonumber\end{aligned}\ ] ] introducing a basis set @xmath292 the matrix elements of @xmath75 can then be written as @xmath293 here @xmath294 where we have used eqs .
( [ eq:8 ] , [ eq : s4 ] ) .
the hamiltonian matrix in this basis set is diagonalized , which gives the eigenvalues @xmath295 and the eigenvectors @xmath296 we then have @xmath297 the quantities @xmath298 were given in eq .
( [ eq : s9 ] ) , @xmath299 . by organizing the sums in eq .
( [ eq : d3 ] ) appropriately , the calculation of this expression is very fast and the main time is spent in diagonalizing the hamiltonian matrix .
we have found this method to be more efficient than the time - dependent method above . in the expression ( [ eq : d3 ] ) , we can identify the first term as the intrinsic contribution , since this is the amplitude which is obtained if there is no interaction between the photoelectron and the target .
the extrinsic effects are then determined by the square of the absolute value of the second term .
the interference between the intrinsic and extrinsic contributions is given by the cross product of these terms .
it is interesting to consider a separable potential @xmath300 since it is then possible to obtain an analytical expression for @xmath1 .
the operator in the denominator of eq .
( [ eq : pt2 ] ) is written as @xmath301 where again @xmath302 is a combined index for the target state @xmath303 and the continuum state @xmath128 and @xmath304 is a ( complex ) number . then @xmath305 and @xmath306 using the fact that @xmath75 is separable , it is then straightforward to invert the expression in eq .
( [ eq : sep2 ] ) and obtain @xmath307_{ij } \label{eq : sep5 } \\
\nonumber & = & { \frac{\delta _ { ij}}{d_{i}(z)}}+\tilde{v}{\frac{c_{i}c_{j}}{% d_{i}(z)d_{j}(z)(1-\tilde{v}\sum_{l}c_{l}^{2}/d_{l}(z))}}.\end{aligned}\ ] ] this leads to @xmath308 where @xmath309 is defined in eq .
( [ eq : s10 ] ) , @xmath310 } } , \label{eq : sep7}\ ] ] with @xmath6 and @xmath311 and @xmath312 to obtain a model for @xmath313 we can , for instance , put @xmath314 compared with the expression in eq .
( [ eq : ma6 ] ) , there is no term @xmath315 in the corresponding expression for @xmath162 .
the neglect of this term means that @xmath313 goes to zero more slowly as one of the arguments @xmath128 or @xmath153 goes to infinity . to compensate for this we use the power three for @xmath316 in the denominator of ( [ eq : sep10 ] ) , while in eq .
( [ eq : ma6 ] ) the corresponding power is two .
this is a reasonable approximation for small @xmath128 , but it breaks down for large @xmath128 . for the satellite to main line intensity ratio we have , @xmath317 this ratio does not depend on any constant factor in @xmath115 , since @xmath318 is proportional to @xmath115 .
if we take the parameters @xmath132 , @xmath319 , and @xmath320 equal to a common typical radius @xmath321 ( as will be motivated later ) , and use the analytic expressions in eqs .
( [ eq : ma4 ] ) and ( [ eq : ma6 ] ) then @xmath322 or @xmath323 , apart from @xmath324 and @xmath325 , becomes a function of @xmath211 , @xmath326 , and @xmath327 , with @xmath328 .
we can see this since @xmath329 is a function of @xmath330 , and @xmath331 a function of @xmath332 and @xmath333 apart from their prefactors .
the prefactor of @xmath334 is @xmath335 , while that for @xmath329 has no influence . for each @xmath75 in a perturbation expansion of eq .
( [ eq : pt3 ] ) we have an energy denominator and a @xmath128-summation .
the @xmath128-summation gives an integral and a factor @xmath336 using variables @xmath337 , the @xmath338 in the prefactor vanishes .
factoring out @xmath282 in the energy denominator , we have a factor @xmath326 for each @xmath339 , and instead of @xmath5 and @xmath281 we have @xmath340 and @xmath341 . with @xmath342 and @xmath343 given ,
@xmath5 is proportional to @xmath85 .
@xmath85 in turn is equal to @xmath344 , and thus somehow related to the strength of the scattering potential @xmath345 if we fix the value of the sudden limit @xmath346 by choosing one of the angles , we still have an independent parameter left .
this parameter can be used to decouple the relation between @xmath5 and @xmath23 ( whatever it is ) . summarizing , we have found that the parameters of our model system appear as the angles @xmath342 and @xmath324 , and the excitation energy @xmath347 ( or @xmath85 ) , while the coupling between the photoelectron and the model system only appears in one parameter , @xmath348 , provided we use @xmath349 as variable .
we have further motivated that we can vary the parameters @xmath5 and @xmath221 independently .
the same problem can be also studied using the standard perturbation approach .
we consider the expression for the matrix elements @xmath318 in eq .
( [ eq : pt3 ] ) .
to lowest order in @xmath75 , we can neglect @xmath75 in the denominator of eq .
( [ eq : pt3 ] ) . inserting the completeness relation @xmath351 in terms of eigenstates @xmath352_{k^{^{\prime } } s^{^{\prime } } ,
k^{^{\prime } } s^{^{\prime } } } \langle s^{^{\prime } } |\langle k^{^{\prime } } |\delta _ { 0}\rangle .
\nonumber\end{aligned}\ ] ] using eqs .
( [ eq : d1a ] , [ eq : d1b ] ) we obtain @xmath353 \nonumber \\ & -&\frac{\sin 2\varphi \cos ( \varphi + \theta ) } { 2}\sum_{k^{\prime } } \left [ \frac{v_{kk^{\prime } } m_{k^{\prime } } } { e - e_{2}-\epsilon _ { k^{\prime } } + i\eta } \right ] , \label{eq : pt4}\end{aligned}\ ] ] @xmath354 \nonumber \\ & + & \frac{\sin 2\varphi \sin ( \varphi + \theta ) } { 2}\sum_{k^{\prime } } \left [ \frac{v_{kk^{\prime } } m_{k^{\prime } } } { e - e_{1}-\epsilon _ { k^{\prime } } + i\eta } \right ] , \label{eq : pt5}\end{aligned}\ ] ] where @xmath355 and @xmath356 .
we can then immediately calculate the photoemission spectra using eq .
( [ eq : s8 ] ) . in figs .
[ fig : ex_energy ] and [ fig : range ] we compare the perturbation expansion with the exact time dependent calculation for a realistic scattering potential in the symmetric case . in the symmetric case we have @xmath357 , and @xmath92 . since the ratio @xmath358 was discussed extensively in sec .
[ sec : sudden ] , we here focus on @xmath11 , which describes the effect of the scattering potential .
we vary the excitation energy @xmath5 by varying @xmath216 , while keeping @xmath201 constant .
we also vary the potential range by replacing @xmath359 in eq .
( [ eq : m4 ] ) by @xmath360 and then varying @xmath361 . with
@xmath201 fixed , @xmath5 is proportional to @xmath219 .
thus a small @xmath5 and a small @xmath361 make the perturbation weak .
the calculations are made for a range of parameter values around those given for cucl@xmath362 in table [ used - parameter ] .
we can also perform the photoemission calculation by assuming a classical trajectory of the emitted photoelectron@xcite , producing a time - dependent potential which drives the dynamics of the the model .
it has been reported that the semi - classical approach can give the unexpectedly good results for the systems with coupling to plasmons.@xcite the essence of the semi - classical approach is to replace the scattering potential @xmath108 by a time - dependent potential using the charge density @xmath363 of the emitted electron , _
i.e. _ @xmath364 where we have used @xmath365 .
we can then write the hamiltonian as @xmath366 where @xmath285 can be expressed in terms of the exact final states ( @xmath367 and @xmath368 ) in the presence of a core hole @xmath369 the time - dependent potential takes the form @xmath370 where ( cf . eq .
( [ eq : d1b ] ) ) @xmath371 the remaining system ( target ) is still purely quantum mechanical , and we write its time - dependent wave function @xmath372 as @xmath373 the classical electron velocity @xmath16 is determined by energy conservation , that is , @xmath374 .
we have here chosen the velocity corresponding to the satellite .
we could also have performed two calculations , with the velocities corresponding to the leading peak and to the satellite , respectively .
in contrast to the approach used here , this would , however , lead to the problem that the spectral weight would not be normalized . applying the time - dependent schrdinger equation to @xmath372
, we obtain @xmath375 and @xmath376 , @xmath377 @xmath378 where @xmath5 ( eq . ( [ eq : s7 ] ) )
is the optical excitation energy .
( [ eq : c7 ] ) and ( [ eq : c8 ] ) are subject to the initial conditions ( cf eq .
[ eq : s9 ] ) @xmath379 the final photoemission currents @xmath380 is @xmath381 it is sufficient to perform the calculation up to @xmath382 , since the potential vanishes for larger values of @xmath383 . the relative intensity between main and satellite contributions
is given by @xmath384 as before . in figs .
[ fig : ex_energy ] and [ fig : range ] we compare the semi - classical and exact results for a realistic potential in the symmetric case .
the semi - classical theory is inaccurate over most of the energy range considered here . for large energies however , the semi - classical theory comes much closer to the exact result than does the pt .
it is also clear that an increasing @xmath5 ( @xmath385 ) does not noticeably affect the energy for the adiabatic - sudden transition , where it strongly effects the maximum deviation .
an increasing @xmath361 on the other hand not only strongly increases the maximum deviation , but also makes the adiabatic - sudden transition energy smaller .
the dependence on the parameters will be investigated more extensively in the next section .
we are now in a position to address the adiabatic - sudden transition and its dependence on the parameters .
the calculations are performed with the analytical matrix elements in eqs .
( [ eq : ma4 ] , [ eq : ma6 ] ) .
first we study the different characteristic lengths , @xmath132 for the dipole matrix elements , and @xmath7 and @xmath320 for the scattering potential matrix elements .
we find that it makes sense to use only one effective length @xmath165 , and the corresponding energy @xmath166 . as we remarked in sec .
vd , @xmath386 as a function of @xmath387 depends on the parameters @xmath211 and @xmath326 , and also on the system parameters @xmath342 and @xmath388 .
we vary @xmath5 independently of @xmath23 , although for a given model there is a direct relation between these two quantities .
part of this relation can be offset by using different @xmath342 and @xmath343 ( with @xmath201 constant ) but we do not explore this possibility . the exact solution with a separable potential is used to discuss the validity and breakdown of perturbation theory .
we find that @xmath326 has a large effect on the deviation from the sudden limit , but little effect on the value of @xmath341 where the deviation becomes small , while @xmath211 has a comparatively small effect on both magnitude and range of the deviation .
for simplicity we use the cucl@xmath233 parameters @xmath389 , which gives @xmath235 .
for cucl@xmath362 we further have @xmath23=-0.36 a.u . , @xmath282=0.195 a.u .
( @xmath165=1.6 a.u . ) , and @xmath239 a.u .
, i.e. , @xmath390 and @xmath391 . in our calculations , we vary @xmath392 and @xmath393 by typically a factor of two around these reference values .
we first illustrate the dependence on the ratio between the length scales @xmath394 , @xmath395 and @xmath320 . in fig .
[ fig : rd ] we show the results for different ratios @xmath396 keeping @xmath397 .
these results are obtained for @xmath398 and @xmath399 , where @xmath400 is the energy scale set by the scattering potential length scale .
[ fig : rd ] shows that as @xmath401 is reduced the magnitude of the `` overshoot '' is increased .
there are , however , no qualitative changes .
[ fig : rsd ] shows results for different values of @xmath402 for a fixed @xmath403 . from eq .
( [ eq : ma6 ] ) it can been seen that this corresponds to varying the range of values @xmath161 where @xmath313 is large , without changing the range over which @xmath147 varies .
the figure illustrates that the overshoot becomes larger as @xmath404 is reduced .
this is natural , since decreasing @xmath320 effectively makes the scattering potential stronger by expanding the range of values @xmath405 with large scattering matrix elements .
the qualitative behaviour , however , is not changed . in view of figs .
[ fig : rd ] and [ fig : rsd ] , we study below the case when @xmath406 , as mentioned in sec . vd .
[ fig : strength ] shows such results for different values of the strength of the scattering potential @xmath392 and for different values of the excitation energy @xmath393 . in each panel @xmath407
is kept fixed , but the ratio is varied by a factor of four from fig .
[ fig : strength]a to fig .
[ fig : strength]c
. typically @xmath11 has an overshoot for small values of @xmath281 . for somewhat larger
@xmath408 the ratio approaches unity and possibly becomes smaller than unity .
the overshoot can be fairly large and happens on a small energy scale ( @xmath409 ) .
in a few cases of a large overshoot , @xmath11 does not become approximately unity until @xmath281 is several times @xmath410 , although the relevant energy scale is still @xmath410 . in the case of an undershoot , @xmath11 approaches unity from below very slowly ( energy scale much larger than @xmath410 ) .
the undershoot is , however , relatively small , and if we do not require a high accuracy , we consider the sudden approximation is valid when the overshoot becomes small .
this means that as the range of the scattering potential is made larger , the sudden limit is reached at a smaller energy .
this is the opposite to what one would expect from the semi - classical theory .
the figure illustrates that @xmath5 is not the relevant energy scale . since in each panel
we keep @xmath411 fixed , there is a variation of @xmath211 by a factor of four .
furthermore there is a variation of @xmath412 by a factor of four in going from the top to the bottom panel in fig .
[ fig : strength ] .
there is no corresponding change in the energy for the adiabatic to sudden transition .
it is interesting to study a separable potential , since it is then possible to obtain an analytical solution .
this makes it easier to interpret the results .
it also allows the study the effects of multiple scattering , i.e. the deviations from first order perturbation theory .
[ fig : separable ] shows results of the exact and first order theory using the same values of @xmath413 and @xmath326 as in fig . [
fig : strength]b .
the separable potential overestimates the magnitude of the overshoot in @xmath414 quite substantially .
otherwise the results are rather similar . for a qualitative discussion
, we can therefore use the separable potential . 0.8 cm for simplicity , we consider @xmath415 .
we further put @xmath416 = b_{k}$ ] .
this is a poor approximation for large @xmath128 , but then anyhow also @xmath331 is poorly represented by the separable potential .
our approximations lead to simple results for the functions @xmath417 , @xmath361 and @xmath418 entering in eqs .
( [ eq : sep6]-[eq : sep10 ] ) . @xmath419 and @xmath420 .
then @xmath421 since @xmath422 is independent of @xmath128 in this approximation , we have dropped the index @xmath128 .
the function @xmath423 is shown in fig .
[ fig : universal ] .
for @xmath424 we can then write @xmath425}{1-\frac{1}{2}\frac{% \tilde{v}}{\tilde{e}}c(\tilde{\omega})/[1+\frac{\tilde{v}}{\tilde{e}}c(% \tilde{\omega})]}}\right| ^{2}. \label{eq : as3}\ ] ] for a level ordering as indicated in fig .
[ fig : schematic ] , @xmath426 and cos@xmath427 . in our standard case with @xmath428 , @xmath429 .
thus the term in the numerator of eq .
( [ eq : as3 ] ) dominates .
the factor @xmath430 gives the multiple scattering , which is not included in the first order perturbation theory , i.e. the first order result is @xmath431 _ { { \rm pt}}=\left| \frac{1-% \frac{\cos ^{2}\varphi } { \cos 2\varphi } \frac{\tilde{v}}{\tilde{e}}c(\tilde{% \omega})}{1-\frac{1}{2}\frac{\tilde{v}}{\tilde{e}}c(\tilde{\omega})}\right| ^{2}. \label{eq : as3a}\ ] ] we now compare the behaviors of eqs .
( [ eq : as3 ] ) and ( [ eq : as3a ] ) , to see the effects of using perturbation theory . for small values of @xmath281
, re@xmath432 is positive and then changes sign at about @xmath433 .
i m @xmath432 is always positive . due to our crude approximations for @xmath434 and @xmath115
, @xmath432 rapidly becomes unreliable beyond @xmath435 . both for
the exact and perturbative expressions @xmath386 goes from over- to undershoot approximately when re@xmath436 .
this is somewhat earlier than in fig .
11 , where however @xmath437 .
comparing @xmath438 for the exact ( eq .
( [ eq : as3 ] ) ) and the first order result ( eq .
( [ eq : as3a ] ) ) , we find that the exact solution is larger when @xmath439 , cf fig .
[ fig : separable ] .
this is consistent with second order perturbation theory , which is found to enhance @xmath8 for small @xmath440 and reduce it for large @xmath281 . for @xmath441 ,
@xmath432 is purely real and slightly larger than 0.1 , thus multiple scattering gives a divergence in both @xmath442 and @xmath443 when @xmath444 .
this is due to the attractive potential @xmath75 forming a bound state from the continuum states .
the important conclusion from analysing the separable potential is that if @xmath151 is not too large , first order perturbation theory gives roughly the correct range over which there are essential deviations from the sudden limit , while multiple scattering increases the magnitude of these deviations for small @xmath281 and slightly decreases them for larger @xmath281 .
we next discuss the physical interpretation of the expression ( [ eq : as3 ] ) ( or the perturbational expressions in ( [ eq : pt0 ] , [ eq : pt4 ] , [ eq : pt5 ] ) ) . here
unity ( the first term in ( [ eq : pt0 ] , [ eq : pt4 ] , [ eq : pt5 ] ) ) corresponds to a direct transition into the final continuum state corresponding to energy conservation .
the second term ( the last two terms in ( [ eq : pt0 ] , [ eq : pt4 ] , [ eq : pt5 ] ) ) corresponds to a virtual transition into some other continuum state followed by one or several scattering events with the electron ending up in the continuum state corresponding to energy conservation .
let us consider the virtual emission into a continuum state with a larger energy than the final state and let this be followed by one scattering event into the final state .
for a negative @xmath151 the interference with the direct event is then constructive . for small photon energies such events dominate for two reasons .
firstly , there are many more states available above the energy corresponding to energy conservation than below , and secondly the dipole matrix elements suppress the transitions to the energies below . as a result , both the main peak and the satellite are enhanced by the scattering effects . for the values of @xmath324 and @xmath325 considered here ( @xmath445 ) ,
the relative effect is stronger for the satellite . as a result @xmath11
is enhanced . for larger photon energies
re@xmath432 becomes negative .
the density of states of partial waves with given @xmath113 and @xmath446 quantum numbers decreases with energy ( @xmath447 ) .
this favours virtual emissions to states below the final continuum state . depending on the model for @xmath448 and @xmath152
, these matrix elements may have the same effect . as a result , re @xmath432 becomes slightly negative for large energies and the ratio @xmath11 is slightly smaller than one . the relevant energy scale for @xmath449 is @xmath410 .
this is a combination of the two effects discussed above .
the turn on of the dipole matrix elements on an energy scale of the order @xmath410 favors an increasing value of @xmath417 over this energy scale , while the density of states effects becomes more important for larger energies . as a result , both re @xmath449 and i m @xmath449 have a maximum at an energy of the order @xmath410 . in this section
we study the perturbation theory expression in more detail and without relying on a separable potential .
instead we consider the more realistic matrix elements in eqs .
( [ eq : ma4 ] ) and ( [ eq : ma6 ] ) , assuming that @xmath450 .
we define a function @xmath451 by @xmath452 which is possible due to the simple form of @xmath313 and @xmath453 .
explicitly we have @xmath454 ^{2}\left [ 1+\left ( \tilde{r}k - x\right ) ^{2}\right ] \left [ x^{2}-\epsilon -i\eta \right ] } .
\label{eq : fk2}\ ] ] from eqs .
[ eq : pt4 ] and [ eq : pt5 ] we have @xmath455 since @xmath456 and @xmath457 we see , as stated earlier , that @xmath386 depends only on @xmath458 , @xmath211 and @xmath459 first we consider the limit of small values of @xmath128 and @xmath5 . for @xmath460 we have @xmath461 , and @xmath462 ^{2}. \label{eq : apt7}\ ] ] if the more localized level @xmath47 is above @xmath48 ( see fig . [
fig : schematic ] ) in the initial state and below @xmath48 in the final state ( `` shake - down '' ) , we have @xmath463 and @xmath464 , and the factor in the brackets is larger than unity .
thus interaction effects enhance the ratio @xmath8 .
this corresponds to a constructive interference between intrinsic and extrinsic effects .
this is in contrast to the destructive interference found for the plasmon case.@xcite the present treatment , however , refers to the `` shake - down '' scenario , and it is more appropriate to compare the plasmon case with the shake up case ( @xmath231 and @xmath465 ) .
then the expression ( [ eq : apt7 ] ) for @xmath8 indeed becomes smaller than one , and the relative weight of the satellite is reduced for small energies .
we notice , however , that both the satellite and the main peak are enhanced by the interference , but that the main peak is enhanced more in the shake - up situation .
we next consider the case when @xmath128 is large .
@xmath466 for large @xmath128 and @xmath467 is ( with @xmath468 ) @xmath469 .
\label{eq : apt8}\ ] ] for the case when @xmath470 we obtain @xmath471 thus the approach to the sudden limit goes as @xmath472 with a coefficient which depends on the parameters . with our standard cucl@xmath233 parameters , we have @xmath473 we note that for large @xmath474 , and when @xmath475 is large enough compared to @xmath476 the last ( positive ) term dominates .
the approach to the sudden limit is then set by @xmath477 . to evaluate eq.([eq : as4 ] ) when @xmath415 we only need the function @xmath478 , @xmath479 ^{2}\left [ 1+\left ( \sqrt{\epsilon}-x\right ) ^{2}\right ] \left [ x^{2}-\epsilon -i\eta \right ] } .\ ] ] we show @xmath480 in the lower panel of fig.[fig : universal ] . the results eq . ( [ eq : apt8 ] ) for large values of @xmath467 are shown by the dotted lines in the figure .
clearly the approach of re @xmath481 to its asymptote is very slow .
if we take @xmath415 we have the same form as in first order perturbation theory with a separable potential eq .
( [ eq : as3a ] ) , @xmath482 where we have put @xmath483 . as shown in fig .
[ fig : universal ] , @xmath484 has a qualitatively similar behavior as @xmath423 for @xmath485 . as in the case of the function @xmath417 ,
the relevant energy scale is @xmath410 . 0.8 cm in fig .
[ fig : pert ] we show @xmath8 as a function of @xmath387 for a few values of @xmath412 and @xmath486 .
we see that @xmath8 starts at a positive value ( cf eq .
( [ eq : apt7 ] ) ) , and reaches a broad maximum at about @xmath487 .
compared to the separable potential solution in fig .
[ fig : separable ] , the overshoot behavior is robust up to fairly large energies , which is due to im@xmath488 decaying more slowly than im@xmath432 . for much larger values of @xmath489 @xmath490 decays as @xmath491 , as shown in eq .
( [ eq : apt8a ] ) . here , based on the discussion in sec .
viib , we can expect that as multiple scattering becomes important , the region where there is an overshoot is substantially reduced and the region with an undershoot becomes larger . at the same time
, the overshoot intensity will be enhanced .
these behaviors are actually confirmed by comparing with the exact calculations given in fig .
[ fig : strength ] .
[ fig : pert ] illustrates that for intermediate values of @xmath492 , when re@xmath493 dominates , @xmath490 goes as roughly latexmath:[$% too large . for larger ( but not too large ) values of @xmath281 , re@xmath493 becomes small and im@xmath493 dominates .
since re@xmath493 is positive for small energies , this leads to a constructive interference between intrinsic and extrinsic effects . for energies of the order @xmath282 , re@xmath493 changes sign , and
the interference becomes weakly destructive .
for somewhat larger energies the extrinsic effects are mainly determined by the imaginary part of @xmath493 . from eq .
( [ eq : apt8a ] ) it follows that in perturbation theory the extrinsic effects become small on the energy scale @xmath495 . in this section
we analyze the adiabatic - sudden transition within the semi - classical framework . from the coupled differential equations eqs .
( [ eq : c7 ] ) and ( [ eq : c8 ] ) we can obtain differential equations for @xmath496 , @xmath497 .
integration of these equations , leads to @xmath498 where @xmath499 is the time at which the emitted electron with the velocity @xmath16 leaves the range @xmath17 of the scattering potential .
latexmath:[$% from the sudden limit .
for small values of @xmath383 both the coefficients @xmath501 and the exponent @xmath502 are approximately real and there is a small contribution to the imaginary part of the integral in eq .
( [ eq : a1 ] ) . as @xmath383 grows
there is , however , a contribution from both these sources . to obtain a qualitative understanding of the semi - classical approximation , we solve the schrdinger equations eqs .
( [ eq : c7 ] ) and ( [ eq : c8 ] ) to lowest order in @xmath503 .
this leads to @xmath504 \nonumber\end{aligned}\ ] ] and a similar result for @xmath376 .
this gives @xmath505 ^{2}+\sin ( 2\varphi + 2\theta ) \delta e\int_{0}^{\tau _
{ 0}}d\tau \tau v_{sc}(\tau ) \right\ } .
\nonumber\end{aligned}\ ] ] to discuss the result , we for a moment assume a simple @xmath506-dependence of @xmath507 @xmath508 which corresponds to the @xmath156-dependence used in eq .
( [ eq : ma6a ] ) .
we note , however , that this form is too simple to describe the behavior of the more realistic potential in eq .
( [ eq : m4 ] ) . inserting eqs .
( [ eq : a1a ] ) and ( [ eq : a3 ] ) in eq .
( [ eq : a1b ] ) gives @xmath509 \tau _ { 0}^{2}.\end{aligned}\ ] ] we now extend this treatment to intermediate values of @xmath16 where the adiabatic to sudden transition takes place . using the expressions eqs .
( [ eq : s2 ] ) , ( [ eq : s5 ] ) and ( [ eq : s7 ] ) to relate @xmath5 and @xmath510 we obtain @xmath511 _ within the semi - classical theory _ the condition for the sudden approximation is then @xmath512 we are now in a position to discuss the approach to the sudden limit . within a semi - classical framework it seems clear that we have to require that the hole potential is fully switched on after a `` short '' time @xmath513 , i.e. , that the emitted electron leaves the range of the scattering potential after a short time .
the question is , however , what we mean by `` short '' . from eq .
( [ eq : a4 ] ) it follows that the time - scale is set by both the inverse of @xmath5 and the inverse of @xmath219 . in eq .
( [ eq : a6 ] ) we have used the relation between @xmath5 and @xmath220 to remove @xmath219 from eq .
( [ eq : a6 ] ) . from eq .
( [ eq : a6 ] ) we obtain the condition for the sudden approximation within the sc theory @xmath514 thus , according to the semi - classical theory , the sudden approximation requires that @xmath515 .
comparison with the full quantum mechanical calculations in fig .
[ fig : strength ] shows that this criterion is not appropriate for the range of parameters considered here .
the reason is that we have considered a parameter range where the semi - classical theory is not very accurate .
it is interesting that the sc theory correctly predicts that the weight of the satellite goes to zero at threshold
. nevertheless , the sc theory does not give the correct physics at the threshold .
in the full quantum mechanical calculation the weight of the satellite goes to zero due to the effects of the dipole matrix element , which becomes very small at small photoelectron energies .
this effect is not included in the sc theory . in the semi - classical treatment ,
the small weight of the satellite is due to the fact that the scattering potential between the outgoing slow electron and the excitation means that the hole potential is only switched on slowly . in the quantum mechanical treatment , on the other hand
, the scattering potential leads to an enhancement of the relative weight of the satellite close to the threshold for the shake - down case .
we have studied the photoemission spectrum of a simple model with a localized charge transfer excitation .
we have obtained exact numerical results for the spectrum as a function of the photon energy @xmath408 and in particular focussed on the ratio @xmath322 between the weights of the satellite and the main peak .
these calculations are compared with perturbational and semi - classical treatments .
the results have been analyzed using the latter two approaches .
an important effect in the ratio @xmath322 is due to the energy dependence of the dipole matrix elements and a factor @xmath516 in the expression for the spectrum .
this leads to a suppression of the satellite close to the threshold , but can lead to an overshoot further away from the threshold .
this effect was discussed in sec .
[ sec : sudden ] and is described by @xmath358 .
if the interaction between the emitted electron and the target is weak , this effect dominates .
it is determined by the excitation energy @xmath5 and the relevant energy scale @xmath43 of the dipole matrix element .
if @xmath517 is small , @xmath518 reaches its limiting value from below , while there is an overshoot if @xmath519 . in both cases @xmath520 reaches its limiting value for a kinetic energy of the order a few times @xmath5 . to study the effects of the scattering potential between the emitted electron and the target we have focussed on the ratio @xmath11 .
this quantity shows an overshoot for small values of @xmath408 in the `` shake - down '' situation studied here .
depending on the parameters there may be an undershoot for larger energies , which extends over a large energy range .
this undershoot is , however , fairly small for the cases considered here .
the sudden approximation is then valid to a reasonable accuracy when the overshoot has become small .
we show that this happens on the energy scale @xmath521 , where @xmath522 is a typical length scale of the scattering potential .
one of the main results of this paper is that for a coupling to localized excitations , the adiabatic to sudden transition takes place at quite small kinetic energies of the photoelectron .
this is in contrast to the large kinetic energies needed for the case of coupling to plasmons . in the plasmon case , the kinetic energy is typically so large that the semi - classical treatment is a very good approximation .
the adiabatic to sudden transition is then expected to happen on the energy scale @xmath523,@xcite where @xmath524 and @xmath525 are the plasmon frequency and wavelength , respectively .
since the long wavelength plasmons dominate the transition , this happens at very large energies . for a localized excitation ,
the relevant length scale of the scattering potential is smaller , and the transition is expected to take place at a smaller energy scale .
actually , the transition takes place at such a small energy that the semi - classical theory is usually not valid any more .
it is interesting that the semi - classical theory therefore predicts the opposite dependence on the range @xmath522 of the scattering potential , namely as @xmath526 instead of @xmath527 . for the shake - down scenario considered here ( the two outer levels cross as the hole is created ) ,
we find constructive interference ( increase of @xmath8 ) between the intrinsic and extrinsic processes at low photoelectron energies .
this is in contrast to the destructive interference found in the plasmon case and to the reduction of @xmath42 found here for the shake - up case ( no level crossing ) .
this work has been supported by the max - planck forschungspreis .
one of us ( lh ) carried out part of his contribution to this work at the max - planck institute for festkrperforschung .
berglund and w.e .
spicer , phys .
rev . * 136 * , a1030 ( 1964 ) ; j.j .
chang and d.c .
langreth , phys .
b * 5 * , 3512 ( 1972 ) ; _ ibid _ * 8 * , 4638 ( 1973 ) ; c. caroli , d. lederer - rosenblatt , b. roulet , and d. saint - james , phys .
b * 8 * , 4552 ( 1993 ) ; a.c .
simonsen , f. yubero , and s. tougaard , phys .
b * 56 * , 1612 ( 1997 ) .
m. kaupp , private communications ; for example , in the atomic configuration of the cu - cl octahedral cluster in cucl@xmath233 , @xmath528 a.u . , @xmath529 a.u . , and the average distance of them will @xmath530 a.u .
( see the inset of fig.[fig : potential]a ) .
.the model hamiltonian eq.([eq:1 ] ) can describe various charge transfer systems .
the table indicates the meaning of the states @xmath47 and @xmath48 for different cases . [ cols="^,^,^",options="header " , ] | we consider core electron photoemission in a localized system , where there is a charge transfer excitation .
examples are transition metal and rare earth compounds , chemisorption systems , and high @xmath0 compounds .
the system is modelled by three electron levels , one core level and two outer levels . in the initital state the core level and one outer level is filled ( a spinless two - electron problem ) .
this model system is embedded in a solid state environment , and the implications of our model system results for solid state photoemission are disussed .
when the core hole is created , the more localized outer level ( d ) is pulled below the less localized level ( l ) .
the spectrum has a leading peak corresponding to a charge transfer between l and d ( shake - down ) , and a satellite corresponding to no charge transfer .
the model has a coulomb interaction between these levels and the continuum states into which the core electron is emitted .
the model is simple enough to allow an exact numerical solution , and with a separable potential an analytic solution .
analytic results are also obtained in lowest order perturbation theory , and in the high energy limit of the semiclassical approximation .
we calculate the ratio @xmath1 between the weights of the satellite and the main peak as a function of the photon energy @xmath2 .
the transition from the adiabatic to the sudden limit is found to take place for quite small kinetic energies of the photoelectron .
for such small energies , the variation of the dipole matrix elements is substantial and described by the energy scale @xmath3 . without the coupling to the photoelectron , the corresponding ratio @xmath4 shows a smooth turn - on of the satellite intensity , due to the turn on of the dipole matrix element .
the characteristic energy scales are @xmath3 and the satellite excitation energy @xmath5 . when the interaction potential with the continuum states is introduced a new energy scale @xmath6 enters , where @xmath7 is a length scale of the interaction ( scattering ) potential . at threshold
there is typically a ( weak ) _ constructive _ interference between intrinsic and extrinsic contributions , and the ratio @xmath8 is larger than its limiting value for large @xmath2 .
the interference becomes small or weakly destructive for photoelectron energies of the order @xmath9 . for larger photoelectron energies
@xmath10 therefore typically has a weak undershoot .
if this undershoot is neglected , @xmath11 reaches its limiting value on the energy scale @xmath12 for the parameter range considered here . in a
shake - up
scenario , where the two outer levels do not cross as the core hole is created , we instead find that @xmath13 is typically reduced for small @xmath2 by interference effects , as in the case of plasmon excitation .
for the shake - down case , however , the results are very different from those for a simple metal , where plasmons dominate the picture .
in particular , the adiabatic to sudden transition takes place at much lower energies in the case of a localized excitation .
the reasons for the differences are briefly discussed . 2 |
the mechanical response of most cells arises from the mechanics of its cytoskeleton , a polymeric scaffold that spans the interior of these cells , and its interaction with the extra - cellular environment .
the cytoskeleton is made up of complex assemblies of protein filaments crosslinked and bundled together by a variety of accessory proteins .
for example , there are approximately 23 distinct classes of accessory proteins such as fascin , @xmath0-actinin , and filamin a @xcite that crosslink filamentous - actin ( f - actin ) , a major component of the cytoskeleton that is resposible for the mechanical integrity and motility of cells .
given the multitude of crosslinkers , several natural questions arise : are the different types of crosslinkers redundant , or do they each serve specific functions ?
do they act independently or cooperatively ?
what are the consequences of their mechanics for the mechanical integrity and response of the cell ?
a mutation study of _ dictyostelium discoideum _ cells lacking a particular actin crosslinking can still grow , locomote , and develop , though with some defects , thereby suggesting at least partial redundancy in the crosslinker s mechanical function @xcite . on the other hand , two types of crosslinkers working cooperatively
may produce enhanced mechanical response .
this cooperativity has been demonstrated in stress fibers crosslinked with the actin binding proteins ( abp ) @xmath0-actinin and fascin , where stress fibers containing both @xmath0-actinin and fascin were more mechanically stable than stress fibers containing only @xmath0 -actinin or fascin @xcite .
in addition , it has been found that two different crosslinkers are required for actin bundle formation _ in vivo _
@xcite . it could also be the case that different crosslinkers work independently of one another such that the dominant crosslinker dictates the mechanical response of the network @xcite . given these various possibilities , how the cell uses different crosslinking proteins to optimize for certain mechanical characteristics is an important open issue in cytoskeletal mechanics . here
, we address this redundancy versus cooperativity issue by studying a model network of semiflexible filaments crosslinked with two types of crosslinkers .
we first study the mechanical properties of the model network with one type of crosslinker and then add the second type of crosslinker and look for mechanical similarities and differences with the original model network .
in addition , we also address the redundancy versus cooperativity issue of two types of crosslinkers for networks made of flexible filaments . as for the two types of crosslinkers ,
we consider crosslinkers that allow the crossing filaments to rotate freely ( freely - rotating crosslinks ) and crosslinkers that constrain the angle between two filaments .
the abp @xmath0-actinin is a candidate for the former type of crosslinking mechanics : optical trapping studies demonstrate that two filaments bound by @xmath0-actinin can rotate easily @xcite . as an example of the latter , we consider filamin a ( flna ) , which binds two actin filaments at a reasonably regular angle of ninety degrees , suggesting that flna constrains the angular degrees of freedom between two filaments @xcite . here
, we do not take into account the possible unfolding of flna since the energy to unfold filamin a is large @xcite , nor do we take into account the kinetics of flna since we seek to understand fully the mechanics in the static regime first .
there exist other possible examples of angle - constraining crosslinkers such as arp2/3 that serves a dual role as an f - actin nucleator and a crosslinker @xcite .
while its role as a nucleator has been emphasized in lamellipodia formation @xcite , its role constraining the angle between the mother and daughter filaments is presumably also important for lamellipodia mechanics .
better understanding of the mechanical role of arp2/3 in lamellipodia may also help to distinguish between the dendritic nucleation model for lamellipodia formation and a new model where
arp2/3 only nucleates new filaments but does not produce branches @xcite .
in studying the mechanical properties of compositely crosslinked filamentous networks , we focus on the onset of mechanical rigidity as the filament concentration is increased above some critical threshold .
this onset is otherwise known as rigidity percolation @xcite . above this critical threshold ,
both experiments and theoretical studies of f - actin networks have observed distinct mechanical regimes . for dense , stiff networks
the mechanical response is uniform or affine and the strain energy is stored predominantly in filament stretching modes . while for sparse , floppy networks one finds a non - affine response dominated by filament bending
where the observed mechanical response of the network is inhomogeneous and highly sensitive to the lengthscale being probed @xcite .
it has been recently reported that there exists a _
bend - stretch _ coupled regime for intermediate crosslinking densities and filament stiffnesses @xcite .
while considerable progress has been made in understanding the mechanics of cytoskeletal networks that are crosslinked by one type of crosslinkers , compositely crosslinked networks are only beginning to be explored experimentally @xcite as are composite filament networks with one type of crosslinker theoretically @xcite . here
we investigate the mechanics of such networks as a function of the concentration and elasticity of the crosslinkers and the filaments .
we arrange infinitely long filaments in the plane of a two - dimensional triangular lattice .
the filaments are given an extensional spring constant @xmath0 , and a filament bending modulus @xmath1 .
we introduce finite filament length @xmath2 into the system by cutting bonds with probability @xmath3 , where @xmath4 , with no spatial correlations between these cutting points .
the cutting generates a disordered network with a broad distribution of filament lengths .
when two filaments intersect , there exists a freely - rotating crosslink preventing the two filaments from sliding with respect to one another .
next , we introduce angular springs with strength @xmath5 between filaments crossing at @xmath6 angles with a probability @xmath7 , where @xmath8 denotes non - collinear .
these angular springs model the second type of crosslinker .
see fig.[fig0 ] for a schematic .
we study the mechanical response of this disordered network under an externally applied strain in the linear response regime . for simplicity we set the rest length of the bonds to unity .
let @xmath9 be the unit vector along bonds and @xmath10 the strain on the bond @xmath11 . for small deformation @xmath12 ,
the deformation energy is [ energies ] e & = & _ ij p_ij ( _ ij . _
ij ) ^2 + _
= p_ij p_jk ( ( _ ji + _ jk ) _ ji ) ^2 + & + & _
p_ij p_jk p_nc ^2 where @xmath13 is the probability that a bond is occupied , @xmath14 represents sum over all bonds and @xmath15 represents sum over pairs of bonds sharing a node .
the first term in the deformation energy corresponds to the cost of extension or compression of the bonds , the second term to the penalty for the bending of filament segments made of pairs of adjacent collinear bonds , and the last term to the energy cost of change in the angles between crossing filaments that meet at @xmath6 angle .
furthermore , for small deformations @xmath16 .
it is straightforward to see that the angular spring @xmath17 between @xmath11 and @xmath18 will contribute to an effective spring in parallel with @xmath19 , giving rise to an enhanced effective spring constant @xmath20 .
we study the effective medium mechanical response for such disordered networks following the mean field theory developed in @xcite for central force networks and @xcite for filament bending networks .
the aim of the theory is to construct an effective medium , or ordered network , that has the same mechanical response to a given deformation field as the depleted network under consideration .
the effective elastic constants are determined by requiring that strain fluctuations produced in the original , ordered network by randomly cutting filaments and removing angular springs vanish when averaged over the entire network .
let us consider an ordered network with each bond having a spring constant @xmath21 , a filament bending constant for adjacent collinear bond pairs @xmath22 , and an angular bending constant @xmath23 between bonds making @xmath6 angles . under small applied strain ,
the filament stretching and filament bending modes are orthogonal , with stretching forces contributing only to deformations along filaments ( @xmath24 ) and bending forces contributing only to deformations perpendicular to filaments ( @xmath25 ) , and hence we can treat them separately . the angular forces due to the angular ( non - collinear ) springs , when present , contribute to stretching of filaments as discussed earlier , where we only consider three body interactions . for these springs to contribute to bending one needs to consider four - body interactions which is outside the scope of this paper and will be addressed in future work .
we start with the deformed network and replace a pair of adjacent collinear bonds with bending rigidity @xmath22 by one with a rigidity @xmath1 , and a bond spring with extensional elastic constant @xmath21 by a spring with an elastic constant @xmath26 and the facing @xmath6 angular spring by @xmath5 .
this will lead to additional deformation of the above filament segments and the angle which we calculate as follows .
the virtual force that needs to be applied to restore the nodes to their original positions before the replacement of the bonds will have a stretching , a bending and an angular contribution : @xmath27 , @xmath28 , and @xmath29 .
the virtual stretching force is given by @xmath30 , the virtual filament bending force is @xmath31 , while the virtual force to restore the angle is @xmath32 , where @xmath33 , @xmath34 and @xmath35 are the corresponding deformations in the ordered network under the applied deformation field . by the superposition principle , the strain fluctuations introduced by replacing the above bending hinges and bonds in the strained network are the same as the extra deformations that result when we apply the above virtual forces on respective hinges and segments in the unstrained network .
the components of this `` fluctuation '' are , therefore , given by : d _ & = & + d _ & = & + d&= & the effective medium spring and bending constants , @xmath21 , @xmath22 and @xmath23 , respectively , can be calculated by demanding that the disordered - averaged deformations @xmath36 , @xmath36 , and @xmath37 vanish , i.e. @xmath38 , @xmath39 , and @xmath40 . to perform the disorder averaging , since the stretching of filaments is defined in terms of spring elasticity of single bonds @xmath0 , the disorder in filament stretching
is given by @xmath41 .
filament bending , however , is defined on pairs of adjacent collinear bonds with the normalized probability distribution @xmath42 .
similarly , for the angular springs , the normalized probability distribution is given by @xmath43 .
this disorder averaging gives the effective medium elastic constants as a function of @xmath44 and @xmath7 as @xmath45 the constants @xmath46 , @xmath47 and @xmath48 for the network contribution to the effective spring constant @xmath49 of bonds , to the filament bending rigidity @xmath50 , and the bending rigidity @xmath51 of angular springs making @xmath6 angles respectively , are given by @xmath52 $ ] .
the sum is over the first brillouin zone and @xmath53 is the coordination number .
the stretching , filament bending and non - collinear bending contributions , @xmath54 respectively , to the full dynamical matrix @xmath55 , are given by : @xmath56 \rbold_{ij } \rbold_{ij } \nonumber \\ { \dbold_b}(q ) & = & \kappa_m \sum_{\langle ij \rangle } \left [ 4(1 - \cos(\qbold.\rbold_{ij } ) ) \right .
\nonumber \\ & & \left . - ( 1 - \cos(2 \qbold.\rbold_{ij } ) ) \right ] \left(\itens-\rbold_{ij } \rbold_{ij } \right ) \nonumber \\ \dbold_{nc } ( q ) & = & \frac{3}{2 } \kappa_{nc , m } \sum \left [ 2 ( 1 - \cos(\qbold.\rbold_{ij } ) ) + 2 ( 1 - \cos(\qbold.\rbold_{ik } ) ) \right .
\nonumber \\ & & \left .
- 2 ( 1 - \cos(\qbold.\rbold_{jk } ) ) \right ] \rbold_{ij } \rbold_{ik}\label{dnc}\end{aligned}\ ] ] with @xmath57 the unit tensor and the sums are over nearest neighbors @xcite .
note that for small @xmath58 , @xmath59 and @xmath60 have the expected wavenumber dependencies for bending and stretching . by definition , @xmath61 , where @xmath62 is the dimensionality of the system . at the rigidity percolation
threshold @xmath63 , @xmath21 , @xmath22 and @xmath23 vanish , giving @xmath64 , @xmath65 and @xmath66 . for semiflexible filament networks with only freely - rotating crosslinks i.e. filament stretching and bending interactions
only , the rigidity percolation threshold is given by @xmath67 . for networks with angle - constraining crosslinks , at @xmath68 , we obtain rigidity percolation thresholds @xmath69 for the case of flexible filament networks , and @xmath70 for semiflexible filament networks .
we also calculate how @xmath71 changes on continuously increasing @xmath7 from @xmath72 to @xmath73 .
simulations were carried out on a triangular lattice with half periodic boundary conditions along the shear direction for the energetic terms whose small deformation limit is given in eq . .
networks were constructed by adding bonds between lattice sites with probability @xmath44 .
next , a shear deformation was applied to the two fixed boundaries of magnitude @xmath74 .
the lattice was then relaxed by minimizing its energy using the conjugate gradient method @xcite allowing the deformation to propagate into the bulk of the lattice .
once the minimized energetic state was found within the tolerance specified , in this case the square root of the machine precision @xmath75 , the shear modulus was then measured using the relation , @xmath76 , using small strains @xmath77 , with @xmath2 denoting the system length and @xmath78 denoting the area of the unit cell for a triangular lattice which is equal to @xmath79 in our units .
system size @xmath80 was studied , unless otherwise specified , and sufficient averaging was performed .
* mechanical integrity as measured by the shear modulus : * on a triangular lattice , networks made solely of hookean springs lose rigidity at a bond occupation probability around @xmath81 @xcite .
this result corresponds to the central force isostatic point at which the number of constraints is equal to the number of degrees of freedom on average .
in contrast , networks made of semiflexible filaments become rigid at a smaller @xmath44 due to extra constraints placed on the system via filament bending . for semiflexible networks with freely - rotating crosslinks , our effective medium theory shows that the shear modulus , @xmath82 , approaches zero at @xmath67 as shown in fig.[fig1 ] @xmath83 .
this result is in good agreement with our simulation results yielding @xmath84 and previous numerical results @xcite .
see fig.[fig1 ] @xmath85 . a different formulation of the emt yields @xmath86 @xcite .
by introducing additional crosslinks that constrain angles between filaments at @xmath6 , the rigidity percolation threshold is lowered .
our emt yields @xmath70 and our simulations yield @xmath87 for @xmath68 ( fig.[fig1 ] @xmath88 and @xmath89 ) .
the cooperative mechanical interplay between these crosslinks and their interaction with filaments allows the network to form a rigid stress - bearing structure at remarkably low crosslinking densities , almost immediately after it attains geometric percolation , @xmath90 , which agrees with a calculation by kantor and webman @xcite . for flexible filament networks , introducing angle - constraining crosslinkers also lowers the rigidity percolation threshold as compared to the isostatic point with the network attaining rigidity at @xmath69 for our emt and @xmath91 in the simulations ( ( fig.[fig1 ] @xmath92 and @xmath93 ) .
incidentally , our result agrees very well with a previous simulation @xcite .
we also compute analytically and numerically how @xmath71 changes with @xmath7 .
see fig.[fig2]@xmath83 , @xmath92 and @xmath88 .
note that @xmath71 is lowered continuously as the concentration of angle - constraining crosslinks is increased . just above the rigidity percolation threshold ,
for a semiflexible network with freely - rotating crosslinks , we find a bending - dominated regime for sparse networks with the shear modulus eventually crossing over to a stretch dominated affine regime at higher filament densities .
the purely stretch dominated regime is represented by the macroscopic shear modulus @xmath82 staying almost constant with increasing @xmath44 , while in the purely bend dominated regime the network is highly floppy and @xmath82 is a sensitive function of @xmath44 , decreasing rapidly as @xmath44 is lowered . this behavior has been observed previously in @xcite . for @xmath94 , both the effective medium theory and
the simulations yield a bend - stretch coupled regime , which is characterized by an inflection in @xmath82 as a function of @xmath44 as observed most clearly for @xmath95 ( with @xmath96 ) .
we find a similar non - affine to affine crossover for the compositely crosslinked flexible filament networks and semflexible filament networks as @xmath44 is increased .
for the flexible filament networks , however , the bend - stretch coupling regime occurs for @xmath97 , i.e. @xmath5 replaces @xmath1 . for semiflexible filament networks , as long as @xmath98 , the bend - stretch coupled regime is robust ( for fixed @xmath7 ) .
in contrast , for @xmath99 , the angle - constraining crosslinker suppresses the bend - stretch coupled regime and enhances the shear modulus to that of an affinely deforming network ( for fixed @xmath7 ) .
the mechanics of the network has been altered with the introduction of the second type of crosslinker .
* non - affinity parameter : * to further investigate how the interaction of the crosslinkers affects the affine and non - affine mechanical regimes , we numerically study a measure for the degree of non - affinity in the mechanical response , @xmath100 , defined in ref.@xcite as : = _ i^n(_i-_aff)^2 .
[ na ] the non - affinity parameter can be interpreted as a measure of the proximity to criticality , diverging at a critical point as we approach infinite system size .
we find that @xmath100 develops a peak at the rigidity percolation threshold , which progressively moves to smaller values of @xmath44 as the concentration of angular crosslinkers @xmath7 is increased ( fig.[fig3 ] @xmath83 ) .
a second peak develops near the isostatic point for @xmath98 as seen in fig.[fig3 ] @xmath92 . as both the collinear and non - collinear bending stiffnesses tend to zero , the network mechanics approaches that of a central force network , and the second peak in @xmath100 at the isostatic point becomes increasingly more pronounced . on the other hand , this second peak can be suppressed by increasing @xmath101 ( fig.[fig3 ] @xmath92 ) , or by increasing the concentration @xmath7 ( fig.[fig3 ] @xmath83 ) even for very small values of @xmath102 .
this further corroborates that adding angle - constraining crosslinkers to non - affine networks can suppress non - affine fluctuations , provided they energetically dominate over filament bending .
the reason for this suppression can be understood by considering the effect of adding a constraint which prohibits the free rotation of crossing filaments .
as the concentration of these non - collinear crosslinks @xmath7 is increased ( at fixed avg .
filament length ) microscopic deformations will become correlated .
the lengthscale associated with this correlation will increase on increasing either @xmath44 or @xmath7 , and will eventually reach a lengthscale comparable to system size even at @xmath103 at large enough concentration and/or stiffness of the angular springs . as a result the mechanical response of the network will approach that of an affinely deforming network . upon decreasing the value of @xmath104 relative
to @xmath102 we again recover the second peak because energetically the system can afford to bend collectively near the isostatic point .
* scaling near the isostatic point : * finally , using scaling analysis we quantify the similarity in mechanics between freely - rotating crosslinked semiflexible networks and compositely crosslinked flexible networks . to do this , we examine the scaling of the shear modulus @xmath82 near the isostatic point with @xmath105 . for @xmath106 ( or @xmath107 ) , the shear modulus scales as @xmath108 ( or @xmath109 ) @xcite . for both @xmath110,@xmath111 and @xmath112 , @xmath113 ,
the emt predicts @xmath114 and @xmath115 as shown in fig.[fig4](a ) and ( b ) , indicating that both types of networks demonstrate redundant , or generic , mechanics . to compare the emt results with the simulations , we use the position in the second peak in @xmath100 to determine the central force percolation threshold , @xmath116 , and then vary @xmath117 and @xmath118 to obtain the best scaling collapse .
for case ( a ) , @xmath119 , @xmath120 and @xmath121 . for case
( b ) , @xmath122 , @xmath120 and @xmath123 .
both sets of exponents are reasonably consistent with those found in ref .
@xcite for a semiflexible network with freely - rotating crosslinks only .
preliminary simulations for compositely crosslinked semiflexible networks indicate that the shear modulus scales as @xmath124 also with a similar @xmath117 and a similar @xmath118 with @xmath125 .
in the limit of small strain , we conclude that the presence of multiple crosslinkers in living cells can be simultaneously cooperative and redundant in response to mechanical cues , with important implications for cell mechanics .
redundant functionality helps the cytokeleton be robust to a wide range of mechanical cues . on the other hand , different crosslinkers
can also act cooperatively allowing the system to vary the critical filament concentration above which the cytoskeleton can transmit mechanical forces .
this may enable the cytoskeleton to easily remodel in response to mechanical cues via the binding / unbinding of crosslinkers ( tuning concentration ) or their folding / unfolding ( tuning stiffness and type of crosslinker ) . since the cytoskeleton consists of a finite amount of material , the ability to alter mechanics without introducing major morphological changes or motifs may play important role in processes such as cell motility and shape change . * cooperativity : * in our study of two types of crosslinkers , crosslinkers that allow free rotations of filaments and crosslinkers that do not , we find two types of cooperative effects in the mechanics of such compositely crosslinked networks .
the first cooperative effect depends on the relative concentration of the two types of crosslinkers and second depends on the relative stiffness of the angle - constraining crosslinkers to the bending stiffness of the individual filaments .
the first cooperative effect can be most strikingly observed beginning with an actin/@xmath0-actinin network and increasing the concentration of flna , with @xmath0-actinin representing the freely - rotating crosslinker @xcite and flna representing the angle - constraining crosslinker @xcite . by tuning the concentration of flna
, the cell can modulate the minimum concentration of actin filaments necessary to attain mechanical rigidity , which can be essentially as low as the filament concentration required to form a geometrically percolating structure .
this is in good agreement with the experimental observation that flna creates an f - actin network at filament concentrations lower than any other known crosslinker @xcite . increasing the flna concentration
also suppresses the non - affine fluctuations near the rigidity percolation threshold by increasing the shear modulus of the network and giving rise to a more affine mechanical response while keeping the filament concentration fixed .
moreover , the cooperativity of @xmath0-actinin and flna working to ehance the mechanical stiffness of actin networks has recently been observed in experiments @xcite .
the addition of angle - constraining crosslinkers to flexible filament networks also decreases the concentration threshold required for mechanical rigidity , though the lower bound on the threshold is not as close as to geometric percolation as it is for semiflexible filaments .
the lowering of the rigidity percolation threshold is independent of the energy scale of the crosslinker .
it depends purely on the number of degrees of freedom the crosslinker can freeze out between two filaments , i.e. the structure of the crosslinker .
the second cooperative interplay between the two crosslinkers depends on the energy scale of the angle - constraining crosslinker to the filament bending energy . for @xmath126 ,
the freely - rotating semiflexible filament system exhibits large non - affine fluctuations near the isostatic point .
upon addition of the angle - constraining crosslinkers , for @xmath127 , the non - affine fluctuations near this point become suppressed and the mechanics of the angle - constraining crosslinker dominates the system . once again , with a small change in concentration of the second crosslinker , the mechanical response of the network is changed dramatically .
* redundancy : * we observe two redundant effects in these compositely crosslinked networks , the first of which depends on energy scales . for @xmath128 with @xmath126 ,
the non - affine fluctuations near the isostatic point in the freely - rotating crosslinker semiflexible filament network remain large even with the addition of the angle - constraining crosslinker . in other words ,
the angle - constraining crosslinkers are redundant near the isostatic point .
their purpose is to decrease the amount of material needed for mechanical rigidity as opposed to alter mechanical properties at higher filament concentrations .
redundancy is also evident in the mechanics of these networks sharing some important , generic properties .
all three networks studied here ( free - rotating crosslinked semibflexible networks and compositely crosslinked semiflexible and flexible networks ) have three distinct mechanical regimes : a regime dominated by the stretching elasticity of filaments , a regime dominated by the bending elasticity of filaments and/or stiffness of angle - constraining crosslinkers , and an intermediate regime which depends on the interplay between these interactions .
the extent of these regimes can be controlled by tuning the relative strength of the above mechanical interactions .
in particular , the ratio of bending rigidity to extensional modulus of an individual actin filament is @xmath129 @xcite .
since the bend - stretch coupled regime has not been observed in prior experiments on _ in - vitro _ actin networks crosslinked with flna only , we conjecture that the energy cost of deformation of angles between filaments crosslinked with flna is larger than the bending energy of filaments .
the qualitative redundancy becomes quantitative , for example , near the isostatic point where we obtain the same scaling exponents for @xmath82 as a function of @xmath130 and @xmath1(or @xmath5 ) for the free - rotating crosslinked semiflexible network and the compositely crosslinked flexible network .
preliminary data suggests the same scaling extends to compositely crosslinked semiflexible networks .
this result is an indication of the robustness of these networks and should not be considered as a weakness . whether or not this robustness extends to systems experiencing higher strains such that nonlinearities emerge is not yet known . * lamellipodia mechanics : * the interplay between cooperative and redundant mechanical properties may be particularly important for the mechanics of branched f - actin networks in lamellipodia . within lamellipodia
, there exist some filament branches occuring at an angle of around @xmath131 with respect to the plus end of the mother filament ( referred to as @xmath132 junctions ) .
these branches are due to the abp arp2/3 @xcite . during lamellipodia formation
, these branches are presumed to be the dominant channel for filament nucleation .
the mechanics of arp2/3 can be modeled as an angular spring between the mother and daughter filament with an angular spring constant of approximately @xmath133 @xcite . in other words
, arp2/3 is an angle - constraining crosslinker for @xmath132junctions ( as opposed to @xmath134junctions ) , and thereby plays an important role in lamellipodia mechanics as demonstrated in this work .
the mechanical role of arp2/3 in lamellipodia has not been investigated previously and may help to discriminate between the dendritic nucleation model @xcite and a new model @xcite by predicting the force transmitted in lamellipodia as a function of the arp2/3 concentration .
in addition to arp2/3 , flna localizes at @xmath134junctions in the lamellipodia and is thought to stabilize the dendritic network @xcite .
both angle - constraining crosslinkers lower the filament concentration threshold required for mechanical rigidity in the system .
depending on the energy scale of flna as compared to the energy scale of arp2/3 , addition of the flna may or may not modulate , for example , the bend - stretch coupling regime at intermediate filament concentrations .
again , at times mechanical redundancy is needed and at times not . with three crosslinkers ,
the system can maximize the redundancy and the cooperativity .
of course , lamellipodia are dynamic in nature and are anisotropic since the arp2/3 is activated from the leading edge of a cell .
both attributes will modulate the mechanical response .
* outlook : * we have demonstrated both cooperativity and redundancy in the mechanics of compositely crosslinked filamentous networks .
we have done so while maintaining the structure of an isotropic , unbundled filament network .
of course , crosslinkers can alter the morphology of the network via bundling , for example .
in other words , different crosslinkers serve specific functions .
this specificity results in a change in microstructure .
this will presumably affect the mechanics such that the cooperative and redundant interactions between multiple crosslinkers may differ from the above analysis .
for example , the crosslinker that dominates in terms of creating the morphology will presumably dominate the mechanics .
schmoller and collaborators @xcite suggest that crosslinker with the higher concentration determines the structure and , therefore , the mechanics . instead of redundancy or cooperativity
, the specificity leads to the simple additivity of two types of crosslinkers in that different crosslinkers act independently of one another . in this study
, however , we find both cooperativity and redundancy in the network mechanics even in the absence of such structural changes @xcite , which , is arguably less intuitive and , therefore , more remarkable . finally ,
while our focus here has been on the actin cytoskeleton as an example of a filamentous network , our results can be extended to collagen networks as well @xcite .
daq would like to thank silke henkes and xavier illa for useful discussions regarding lattice simulations .
md would like to thank alex j. levine , f. c. mackintosh , c. broedersz , t.c .
lubensky , c. heussinger and a zippelius for discussions on the mechanics of semiflexible networks .
md and jms also acknowledge the hospitality of the aspen center for physics where some of the early discussions took place .
jms is supported by nsf - dmr-0654373 .
md is supported by a veni fellowship from nwo , the netherlands .
rivero f , furukawa r , fechheimer m , noegel aa ( 1999 ) three actin cross - linking proteins , the 34 kda actin - bundling protein , alpha - actinin and gelation factor ( abp-120 ) have both unique and redundant roles in the growth and development of _ dictyostelium_. _
j. cell sci .
_ 112:2737 - 2751 .
tseng y , kole tp , lee jsh , fedorov e , almo sc , schafer bw , wirtz d ( 2005 ) how actin crosslinking and bundling proteins cooperate to generate an enhanced mechanical response . _
_ 334 : 183 - 192 .
tilney lg , connelly ka , vranich mk , shaw mk , guild gm ( 1998 ) why are two different cross - linkers necessary for actin bundle formation in vivo and what does each cross - link contribute ?
_ j. cell biol .
_ 143 : 121 - 133 .
nakamura f , osborn tm , hartemink ca , hartwig jh , stossel tp ( 2007 ) structural basis of filamin a functions _ j. cell biol _ 179 : 1011 - 1025 ; stossel tp , condeelis j , cooley l , hartwig jh , noegel a , schleicher m , shapiro ss ( 2001 ) filamins as integrators of cell mechanics and signalling _ nat rev mol cell biol _ 2:138 - 45 .
gardel ml , nakamura f , hartwig jh , stossel tp , weitz da ( 2006 ) prestressed f - actin networks cross - linked by hinged filamins replicate mechanical properties of cells _
_ 103:1762 - 1767 .
blanchoin l , amann kj , higgs hn , marchand j - b , kaiser da , pollard td ( 2000 ) direct observation of dendritic actin filament networks nucleated by arp2/3 complex and wasp / scar proteins _ nature _ 404 : 1007 - 1011 .
latva - kokko m , timonen j ( 2001 ) rigidity of random networks of stiff fibers in the low - density limit _ phys .
64:066117 , 5 pgs ; latva - kokko m , mkinen j , timonen j , rigidity transition in two - dimensional random fiber networks _ phys .
63:046113 , 10 pages . head da , levine aj , mackintosh fc ( 2003 ) deformation of cross - linked semiflexible polymer networks _ phys .
lett . _ 91:108102 , 4 pgs . ; head da , mackintosh fc , levine aj ( 2003 ) distinct regimes of elastic response and deformation modes of cross - linked cytoskeletal and semiflexible polymer networks _ phys .
68 : 061907 , 15 pgs . ;
head , f.c .
mackintosh and a.j .
levine ( 2003 ) nonuniversality of elastic exponents in random bond - bending networks _ phys .
68:025101(r ) , 4 pgs .
gardel ml , shin jh , mackintosh fc , mahadevan l , matsudaira pa , weitz da scaling of f - actin network rheology to probe single filament elasticity and dynamics ( 2004 ) _ phys .
_ 93 : 188102 , 4 pgs . ; gardel ml , shin jh , mackintosh fc , mahadevan l , matsudaira pa , weitz da , elastic behavior of cross - linked and bundled actin networks _ science _ 304:1301 - 1305 .
wagner b , tharmann r , haase i , fischer m , bausch a r ( 2006 ) cytoskeletal polymer networks : the molecular structure of cross - linkers determines macroscopic properties _ proc .
_ 103:13974 - 13978 . , and angle - constraining crosslinker occupation probability @xmath135 the purple lines denote semiflexible filaments , the red arcs denote angle - constraining crosslinks , the black circles represent nodes where all crossing filaments are free to rotate , while the grey circles denote nodes where some of the crossing filaments are free to rotate . the filament bending stiffness relative to stretching stiffness @xmath136 and the stiffness of angular crosslinks relative to stretching stiffness @xmath137 . ] for semiflexible networks with freely - rotating crosslinks ( ( a ) and ( d ) )
, flexible networks with freely - rotating and angle - constraning crosslinks ( ( b ) and ( e ) ) , and semiflexible networks with both crosslinkers ( ( c ) and ( f)).the top panels show results from the effective medium theory and bottom panels show results from the simulations.,scaledwidth=100.0% ] for the semiflexible filaments and @xmath80 for the flexible filaments . figures ( b ) and ( c ) show the shear modulus ( in logarithmic scale described by the colorbar ) as a function of @xmath44 and @xmath7 for flexible networks ( b ) and semiflexible networks ( c ) .
the parameter values studied are ( b ) @xmath138 and ( c ) @xmath139 , @xmath140 .
the black dashed lines in ( b ) and ( c ) correspond to the effective medium theory prediction of the rigidity percolation threshold .
, scaledwidth=100.0% ] as a function of @xmath44 for semiflexible networks with both types of crosslinkers . in ( a )
we show the effect of changing the concentration @xmath7 of the angle - constraining crosslinkers for @xmath139 , @xmath140 and @xmath141 , while in ( b ) we show the effect of changing their stiffness @xmath5 for @xmath80 . ]
scales with @xmath142 and @xmath1 ( @xmath5 ) as @xmath143 .
the effective medium theory predicts mean field exponents @xmath114 and @xmath115 for both semiflexible networks with freely - rotating crosslinkers ( a ) and compositely crosslinked flexible networks ( b ) , while simulations predict @xmath120 and @xmath123 for semiflexible networks with freely - rotating crosslinkers ( c ) and @xmath120 and @xmath121 for compositely crosslinked flexible networks ( d ) . ] | the actin cytoskeleton in living cells has many types of crosslinkers .
the mechanical interplay between these different crosslinker types is an open issue in cytoskeletal mechanics .
we develop a framework to study the cooperativity and redundancy in the mechanics of filamentous networks with two types of crosslinkers : crosslinkers that allow free rotations of filaments and crosslinkers that do not .
the framework consists of numerical simulations and an effective medium theory on a percolating triangular lattice .
we find that the introduction of angle - constraining crosslinkers significantly lowers the filament concentrations required for these networks to attain mechanical integrity .
this cooperative effect also enhances the stiffness of the network and suppresses non - affine deformations at a fixed filament concentration .
we further find that semiflexible networks with only freely - rotating crosslinks are mechanically very similar to compositely crosslinked flexible networks with both networks exhibiting the same scaling behavior .
we show that the network mechanics can either be redundant or cooperative depending on the relative energy scale of filament bending to the energy stored in the angle - constraining crosslinkers , and the relative concentration of crosslinkers .
our results may have implications for understanding the role of multiple crosslinkers even in a system without bundle formation or other structural motifs . |
many physical systems yield wave - like solutions which satisfy the wave equation with a speed of propagation @xmath0 , @xmath1\phi ( \mathbf{x},t ) = 0\ ] ] which is also the massless klein - gordon equation . to leading order
, the lagrangian of any such field obeys a lorentz - like symmetry of lorentz transformations scaled with the limiting speed @xmath0 , even if the underlying system does not have that invariance .
however , if there are multiple fields , they will in general have different limiting velocities , and there will not be a global lorentz symmetry . if the fundamental interactions are emergent phenomenon from an underlying theory without lorentz invariance@xcite , we might expect that particles would display different limiting speeds . in this paper
we show how interactions between the fields can lead to a universal limiting velocity , i.e. the speed of light , at low energies .
we calculate how the different fields influence each other s propagation velocity through the self - energy diagrams , and then use the renormalization group to scale the results to low energy . using several examples
we show that the condition of equal velocities is the low energy endpoint of renormalization group evolution .
a heuristic explanation for this is that because fields can split into other types of fields , the propagation velocity of one field approaches that of the related other fields
. this result could be useful if the fields of the standard model are emergent from an underlying theory that is not lorentz invariant .
of course , lorentz invariance is conventionally taken as one of the foundational principles underlying all our fundamental interactions . however , the weinberg - witten @xcite theorem is usually interpreted as telling us that non - abelian gauge bosons and gravitons can not be emergent fields arising from any underlying lorentz invariant non - gauge theory .
all known examples@xcite satisfy this property . therefore if the idea of emergent fields has any application in the fundamental interactions it appears required that lorentz invariance is also emergent .
our results show that a universal limiting velocity can be an emergent property in the low energy limit .
however , in general the difference in velocity runs towards zero only logarithmically .
this means that the underlying scale of emergence needs to be exponentially far away , making it difficult to test any feature of that theory which is power - suppressed .
for example , we estimate that the scale where differences in the velocity are of order 10% would be beyond @xmath2 gev .
because of this feature we propose a model that produces much faster power - law running .
the model involves a large number of fields which accelerate the running .
another consequence of the running speeds is that observable differences in the velocities would be greatest if the interactions are the weakest .
this suggests that the measurement of the velocity of gravitational waves would be the most sensitive test of this aspect of emergence .
this paper has the following structure . in the next section
we give some general comments on our procedure .
then in sec . 3 , 4 , 5 and 6
we calculate the beta functions for yukawa theories , electrodynamics , yang - mills and mixed theories respectively .
all cases yield beta functions such that the limiting velocities run towards each other at low energy . in sec 7
, we analyze the general effects of logarithmic running and address the phenomenological constraints . because of the difficulties posed by logarithmic running
, we display a model with power - law running in sec . 8 .
we close with a summary and discussion .
some of the more technical details are described in a pair of appendices .
we assume that different species of fermions , scalars and gauge fields emerge at some uv scale with different speeds of light . in condensed matter systems , phonons and magnons
do not propagate at the same speed .
similarly , the same behavior is expected to carry on in an emergent theory of nature . in the absence of any form of interactions between particles ,
their speeds are expected to be frozen as we run down to the ir .
however , these particles are interacting due to yukawa and gauge forces .
hence , the total lagrangian of such a system will be given by the sum of kinetic and interaction terms with certain _ bare _ coefficients specified initially at the uv .
the parameter space of the system is spanned by the different speeds and interaction strengths . according to the principle of self - similarity and wilsonian renormalization
, the same lagrangian will continue to describe the system at different energy scales provided that we replace the bare parameters with the _ renormalized _ ones .
this can be achieved by integrating out the high momentum modes as we run down from uv to ir .
quantum loops are sensitive to high momenta , and hence can be used to track the evolution of trajectories of the different speeds and interaction strengths in the parameter space .
the evolution of these trajectories are encoded in the @xmath3 functions that are given by the gell - mann low equations @xmath4 where @xmath5 is the mass scale we introduce in dimensional regularization . in theories with a universal limiting velocity , the lorentz symmetry prevents the renormalization of the speed of light , and one can set @xmath6 as a definition of natural units . however ,
if different species carry different limiting velocities , @xmath0 , then these parameters also get renormalized and must be treated in the same manner as coupling constants .
they carry a scale dependence through the renormalization procedure , and also generate their own beta function .
we exploit this property to study the running of the limiting velocities . throughout the paper we use dimensional regularization ( dim - reg ) .
the high energy part of the quantum loops can be isolated by retaining only the @xmath7 pieces that arise upon using dim - reg .
finally , we notice that our treatment is limited to one - loop corrections , and that the @xmath3 functions of the speeds require only self - energy corrections , while those of the couplings require vertex corrections as well ( see fig .
[ quantum loops ] ) .
we consider a two - species system , namely scalars and fermions , having different speeds of light at the uv and coupled through yukawa interaction .
the lagrangian density reads @xmath8 where the subscript @xmath9 denotes the renormalized values of the fields .
the momentum - space propagators for scalars and fermions are given by @xmath10 the self - energies of fermions and scalars are respectively @xmath11 and @xmath12\,.\end{aligned}\ ] ] in the following , we will be interested only in the divergent pieces of ( [ fermionse ] ) and ( [ bosonse ] ) . the integral ( [ bosonse ] ) is trivial to perform upon using the substitutions @xmath13 , @xmath14 , @xmath15 , and @xmath16 . then
, one readily finds @xmath17\left(\frac{2}{\epsilon}+\mbox{finite}\right)\,.\ ] ] on the other hand , the integral ( [ fermionse ] ) is more involved and it needs a bit more attention . using the substitution @xmath13 , and @xmath14 we find @xmath18\,,\end{aligned}\ ] ] where the integrals @xmath19 and @xmath20 are given by ( the details are in appendix a ) @xmath21 and @xmath22 where @xmath23 .
now , we move to the vertex correction which , to one - loop order , reads @xmath24 using the change of variables @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , and @xmath30 , and retaining only the divergent part of the integral we obtain ( see apendix a ) @xmath31 at this point , the total lagrangian including the one - loop effect is @xmath32\left(\frac{2}{\epsilon}\right)\\ \nonumber & & + \frac{g^2}{\left(4\pi\right)^2c_f}\left[\frac{1}{c_f^2}\partial_0\phi_r\partial_0\phi_r-\vec\partial\phi_r\cdot \vec\partial\phi_r \right]\left(\frac{2}{\epsilon } \right)\\ \label{full yukawa } & & + \frac{g^3}{\left(4\pi\right)^2c_bc_f^2}\frac{2}{1+a}\left(\frac{2}{\epsilon } \right)\phi_r\bar \psi_r\psi_r\,,\end{aligned}\ ] ] where @xmath33 is the counter lagrangian @xmath34 at this point , we can read off the different @xmath35s that are required to absorb the infinities .
furthermore , we define the bare fields @xmath36 , and @xmath37 , bare speeds @xmath38 , and @xmath39 and bare coupling @xmath40 such that the lagrangian density reads @xmath41 comparing ( [ full yukawa ] ) and ( [ bare yukawa ] ) we find @xmath42 where @xmath43 @xmath44 to proceed , we regard all the renormalized quantities above as functions of the scale @xmath5 that occurs in dim - reg .
then , we differentiate the system in eq .
[ relation between the bare and the dressed in yukawa ] w.r.t @xmath5 and solve simultaneously for @xmath45 , @xmath46 and @xmath47 to find @xmath48 notice that the @xmath3 functions of @xmath49 and @xmath50 do not depend on the vertex correction @xmath51 .
finally , we calculate the @xmath3 function of the ratio @xmath23 to find @xmath52}{c_bc_f^2(1+a)^2}\,,\end{aligned}\ ] ] from which we see that @xmath53 is an ir attractive line .
we can also see that by studying the jacobian @xmath54 , for @xmath55 at the fixed line @xmath56 .
the eigenvalues of @xmath57 are @xmath58 .
the positivity of the second value ensures that @xmath53 is an ir attractive fixed line .
we have seen the existence of an attractive ir fixed line corresponding to a common limiting speed .
we will address more details about the running in sec .
in this section , we study the rg flow of the limiting speeds of fermions and photons .
the non - covariant and gauge invariant lagrangian density reads @xmath59 where @xmath60 , and @xmath61 , and @xmath62 is the photon speed .
the photon propagator in the feynman gauge is given by @xmath63 to find the photon and fermion self - energies , it proves easier to write the interaction lagrangian in the form @xmath64 , where @xmath65 .
hence , the fermion self - energy is @xmath66\gamma^\nu\,,\\\label{fermion self energy in general electrodynamics}\end{aligned}\ ] ] where @xmath67 and @xmath68 are given in ( [ i0 ] ) and ( [ i1 ] ) after replacing @xmath49 with @xmath62 . while the photon self - energy is given by @xmath69\,.\end{aligned}\ ] ] using the substitution @xmath70 , @xmath14 , @xmath71 , and @xmath16 we can put @xmath72 in a standard integral form .
hence , @xmath73 where @xmath74 .
explicit calculations shows that @xmath75 , where @xmath76 , and hence @xmath72 is gauge invariant as expected .
the counter lagrangian reads @xmath77 and @xmath78 is the counter term for the gauge sector .
then , from ( [ counter lagrangian general electrodynamics ] ) and ( [ fermion self energy in general electrodynamics ] ) , and after using the properties of @xmath79 matrices , we can immediately read @xmath80 and @xmath81 @xmath82 now we come to the counter terms in the gauge sector . a general counter term written in the momentum space
takes the form @xmath83a_{0,r}\\ \nonumber & & + a_{i\,r}\left[c_g^2\delta_{g_b}p^ip^j\right.\\ \nonumber & & \quad\quad\quad\left.+\delta^{ij}\left(\delta_a \left(p^0\right)^2-c_g^2\delta_{g_b}\vec p^{\,2}\right ) \right]a_{j\,r}\\ & & -2a_{i\,r}\delta_a c_g p^0p^ia_{0\,r}\,.\end{aligned}\ ] ] one can show that all the infinities in ( [ final form of photon self energy ] ) can be absorbed using @xmath84 and @xmath85 @xmath86 where as usual @xmath43 .
we write @xmath87 in the compact form @xmath88 with @xmath89 it is trivial to see that @xmath90 , and hence @xmath91 is gauge invariant . now , defining the bare fields @xmath37 , @xmath92 , and @xmath93 , and bare speeds @xmath94 , and @xmath95 the lagrangian density reads @xmath96 the bare gauge field lagrangian in the momentum space is given by @xmath97 , and @xmath98 the relations between the bare and renormalized speeds are @xmath99 from which we obtain @xmath100 these @xmath3 functions have the same structure as in the case of yukawa interactions , and we immediately conclude that @xmath101 is an ir attractive line .
in this section we generalize the results of qed to the case of non - abelian gauge theories .
we take the gauge group to be @xmath102 , and the fermions in the fundamental representation @xmath103 where @xmath104 is the coupling constant , @xmath105 are the group generators and @xmath106 are the group structure constants .
@xmath107 is the free part of the lagrangian @xmath108 where as in the case of qed @xmath109 , @xmath61 , and @xmath62 is the gauge boson speed .
the fermion self - energy is identical to the case of qed , one just includes the quadratic casimir operator in the fundamental representation @xmath110 into eq .
[ fermion self energy in general electrodynamics ] to find @xmath111 in calculating the gauge boson self - energy @xmath112 one encounters , in addition to the fermion loop , gauge boson and ghost loops @xmath113\delta^{ab}\\ & & \quad \times \left(\frac{2}{\epsilon}\right)\,,\end{aligned}\ ] ] where @xmath114 , and @xmath115 are group factors , @xmath116 and @xmath117 .
as we did in qed , the infinites can be absorbed into the counter term @xmath118 where @xmath119 are given in eq .
[ counter terms in the gauge sector of qed ] .
hence , we find @xmath120 gluon loops will not modify their own propagation speed , due to the lorentz - like symmetry of that sector when considered in isolation .
this is visible in the formulas above . since @xmath121 ,
the gauge bosons and ghosts contributions cancel in obtaining @xmath122 .
overall , the @xmath3 functions read @xmath123 which , apart from group factors , are identical to the qed case .
the emergence of a universal lorentz symmetry in the above examples is intriguing to explore a more general setup consisting of multi - species and/or mixing between fermions , bosons and gauge fields . before delving into the most general case
we derive a general formula that enables us to calculate the @xmath3 functions of such complex systems .
this is done in appendix b. now , let us consider the more general case of yukawa - electrodynamics .
in this theory a fermion couples to a scalar through yukawa interaction , and minimally to a @xmath124 gauge field .
the scalar is neutral under the @xmath124 field .
we assume that the fermion , scalar and gauge field all have different speeds of light , @xmath50 , @xmath49 and @xmath62 respectively .
this is the simplest generalization of the above cases .
the scalar and gauge field self - energies are identical to their expressions in yukawa and qed sections , while the fermion self - energy is the sum of the contributions from the scalar and gauge field .
the calculations of the corresponding @xmath125 renormalizations are very straight forward , and can be obtained directly from the previous two sections .
thus , @xmath126 , @xmath127 , @xmath128 , and @xmath129 are given by their expressions in eq .
[ z for yukawa system ] and eq . [ z for qed ] respectively , while @xmath130 the relations between the bare and renormalized quantities are given as before @xmath131 in order to find the @xmath3 functions of @xmath49 , @xmath50 , and @xmath62 we use eq .
( [ the grand formula ] ) to find @xmath132(\rho_f-\rho_\psi)\,,\\ \nonumber \beta_{c_b}&=&-\frac{c_bg}{2}\frac{\partial}{\partial g}(\rho_\phi-\rho_b)\,,\\ \label{structure of beta for yukawa electrodynamics } \beta_{c_g}&=&-\frac{c_ge}{2}\frac{\partial}{\partial e}(\rho_a-\rho_{g_b})\,,\end{aligned}\ ] ] where @xmath133 . finally , the @xmath3 functions read @xmath134 this is exactly expected : since only fermions can couple to both scalars and gauge fields , we find that the photon and scalars speeds of light are identical to those found before , while the fermion speed gets contributions from both yukawa and gauge sectors
. we can see that @xmath135 is an ir attractive fixed line by computing the eigenvalues of the jacobian @xmath136 , where @xmath137 , @xmath138 we consider @xmath139 fermions interacting with @xmath140 scalars or gauge bosons .
although we shall carry out the calculation in the case of yukawa interactions , the abelian and non - abelian @xmath3-functions have the same structure as we pointed out above .
the general lagrangian density reads @xmath141 where summation over repeated indices is implied . denoting @xmath142 and noticing that @xmath143
as we move @xmath144 across the vertex @xmath145 @xcite we find @xmath146 @xmath147 , @xmath148 can be read directly from the yukawa expressions in eq .
[ z for yukawa system ] , while @xmath149 and @xmath150 are obtained using a series of integrals similar to those given in appendix a. notice that quantum loops can also generate off - diagonal corrections @xmath151 if the couplings @xmath152 contain off - diagonal components .
these corrections will induce kinetic mixing terms of the form @xmath153 .
in lorentz invariant theories , where @xmath154 , we can always find basis where @xmath154 are diagonal by performing @xmath155 rotations .
however , in the present case , and since in general @xmath156 , we have the freedom to diagonalize either the time - time or the space - space components . diagonalizing the time - time component , and hence working in a basis where we have canonical kinetic terms , will always leave space - space mixing terms .
we assume that these terms are always small compared to the diagonal speeds , i.e. @xmath157 for all @xmath158 , and we ignore their evolution in the following analysis .
the same thing can also be said about kinetic mixing terms for bosons . to be able to use the grand formula ( [ the grand formula ] )
we define @xmath159 and @xmath160 now , we write ( [ the grand formula ] ) as @xmath161 where the greek incices run from @xmath162 to @xmath163 and @xmath164 , and the upper case lattin letters run over @xmath158 and @xmath165 .
the non zero values in ( [ modified grand formula ] ) are @xmath166 , @xmath167 , @xmath168 , @xmath169 , @xmath170 , and @xmath171 .
hence , we obtain @xmath172}{(c_{f_a}+c_{f_b})^3}\,,\\ \beta(c_{f_a})&=&\frac{4}{3\left(4\pi\right)^2}\sum_{ib}\frac{z_{ab}^iz_{ba}^{*i}\left[3c_{f_a}-c_{f_b}-2c_{b_i}\right]}{c_{b_i}\left(c_{f_b}+c_{b_i}\right)^2}\,.\end{aligned}\ ] ] again , we find in this general setup that @xmath173 for all @xmath174 is a fixed line . to study the nature of this line we perform a perturbation to the system about this line , i.e. we construct the jacobian matrix at @xmath173 . defining @xmath175 , and @xmath176
we find @xmath177\,,\end{aligned}\ ] ] where @xmath178}{3\left(4\pi\right)^2c^3}\,,\quad j^1_{a , c}=-\frac{\sum_i z_{ac}^iz_{ca}^{*i}}{3\left(4\pi\right)^2c^3}\,,\\ \nonumber j^2_{a , i}&=&\frac{-2\left[\lambda_a^i+\gamma_a^i\right ] } { 3\left(4\pi\right)^2c^3}\,,\quad j^3_{i , a}=\frac{-4\left[\lambda_a^i+\gamma_a^i\right ] } { \left(4\pi\right)^2c^3}\,,\\ j^4_{i , j}&=&\delta_{i , j}\frac{4\sum_{a}\left[\lambda_a^i+\gamma_a^i\right ] } { \left(4\pi\right)^2c^3}\,.\end{aligned}\ ] ] although we were not able to diagonalize @xmath57 analytically , numerical calculations show that we always have a spectrum of positive eigenvalues on the top of a zero mode . in figure ( [ eigenvalue vs nf ] ) we plot the smallest eigenvalue @xmath179 of @xmath57 , which governs the behavior of @xmath180 and @xmath181 , against the number of fermions @xmath139 for fixed number of bosons @xmath140 .
we find that @xmath182 for large @xmath139 .
hence , the effective running increases with the number of species as expected .
let us consider how the differing speeds approach each other by studying the situation where the speeds are relatively close , but not identical , at some scale @xmath183 .
we treat this problem to first order in the speed difference .
if we define the relative speed difference as @xmath184 the generic beta function has the form @xmath185 where @xmath186 is a constant of order unity and where we denote the common low energylimit as @xmath187 .
if the coupling @xmath104 were to be treated as a constant , this running could be integrated to yield @xmath188 the rescaling is power - law in form . if the coupling constant is small , the power - law exponent is also small and the running is slow .
however , if the coupling is large ( and constant ) the running would be rapid with a power - law form , leading quickly to a universal speed of light
. however , the coupling itself also runs . for example , the yukawa coupling beta function can be integrated to yield @xmath189 with @xmath190 . while this coupling could be large at high energy it runs to smaller values at low energy
this produces a quite different form for the running of the relative speeds .
the correct form for the running of @xmath191 is @xmath192^{\frac{2b}{5 } } = \eta _ * \left[\frac{g^2({\mu})}{g^2(\mu_*)}\right]^{\frac{2b}{5}}\ ] ] .
this implies that the difference in the speeds runs only logarithmically towards each other .
there are tight constraints on the equality of the limiting velocities for the different particles . for direct measurement of the velocities
, we can look at timing accomplished at high energy accelerators .
for example at lep , the electron beam travels at essentially the limiting velocity , since @xmath193 .
the timing of the accelerator relies on this limiting velocity being the speed of light .
because the timing of each bunch is recorded within @xmath194 ns over about 1000 revolutions in the 27 km accelerator@xcite , we estimate that this constrains @xmath195 for electrons . however , indirect constraints are more powerful , and these have been described by altschul@xcite . for @xmath196
, energetic electrons traveling faster than the speed of light will radiate cherenkov light , losing energy until they move at only the speed of light .
this effect produces a maximum energy for subluminal motion , which is constrained by the observation of energetic electrons in astrophysics . for @xmath197
, there is a constraint from the cutoff frequency in synchrotron emission .
these constraints are more powerful than direct measurements because they bound factors of @xmath198 rather than the linear bounds on @xmath191 from the velocity measurements .
altschul s bounds are @xmath199 . in order to achieve this close equality of the different speeds with logarithmic running , the running needs to occur over an exponentially large energy range .
for example , even if we take @xmath200 and @xmath201 ( which barely allows perturbation theory to be used near the energy @xmath183 ) , we would need @xmath202 , where we have generously used @xmath203 as the low energy scale .
this clearly poses a problem for model building .
in any realistic model of emergence without an intrinsic lorentz invariance , we do not expect the different species to emerge with the same limiting speed . in the above sections we showed that there is a potential mechanism to overcome this problem in a class of models whenever we run the renormalization group down to lower energies .
however , we found that the speeds of light are forced to run logarithmically along with the running coupling constants .
this is a relatively slow running if we want to meet the stringent constraints on lorentz violations without having to fine tune the speeds at the uv . in this section
, we propose a way out of this situation . in order to increase the effect of rg
running there are two options .
one is to keep the coupling constant large and unchanged with energy scale .
such a nearly conformal theory would convert logarithmic running into power - law running , as we saw in the last section .
we also need the large coupling such that the exponent is large .
such theories are under active investigation @xcite in the context of `` walking technicolor '' where slowly running but strongly interacting theories are used to provide dynamical breaking of the electroweak theory while not producing excessive flavor changing processes .
should walking technicolor theories prove successful , it would be quite interesting to tie those results with the idea of an emergent limiting velocity .
the other option is if there are a very large number of fields of different scales , such that the running is increased by a large ( and energy dependent ) factor .
we explore this option below .
we introduce a large number @xmath204 of hidden fermions in addition to the standard model ( sm ) ones @xcite .
moreover , we assume that all these fermions ( hidden and sm ) have the same origin , and hence all have the same initial speed of light @xmath205 , with @xmath206 , at some uv emergence scale @xmath183 . as a warm up calculation
, we assume that the fermions have a common initial charge @xmath207 under a single @xmath124 gauge sector .
the gauge photon emerges with some initial speed @xmath208 , with @xmath209 , that is different from the speed of fermions . at the uv scale
the fermions are taken to be massless and hence will participate in the running of the gauge coupling as well as photon and fermions speeds .
as we run down our rg equations , some of the hidden fermions become massive and decouple from the rg equations .
we model the dependence on the mass scale using a power law @xmath210 where @xmath211 and @xmath212 are positive constants , and @xmath213 is an ir mass scale .
@xcite . in this context
, @xmath214 is the lowest kk mode @xmath215 , where @xmath216 is the size of the extra dimension , and @xmath217 is the number of extra dimensions . ]
since the fermions have a common initial speed and a common initial coupling strength , the evolution of the system can be modeled with a single @xmath50 and @xmath218 common to all fermions .
therefore , to one - loop order we have @xmath219 integrating this system yields the running charge @xmath220}\,,\end{aligned}\ ] ] and speeds @xmath221^{\frac{e^2(0)}{3\pi^2\alpha_f}}\,\\ & \approx & \frac{e^2(\mu)}{e_*^2}\,.\end{aligned}\ ] ] hence , we see that both @xmath222 and @xmath223 experience power - law running with ir values given by @xmath224 therefore , we can choose @xmath225 in order to meet the stringent requirement @xmath226 .
however , taking @xmath227 , so that we can trust our perturbation theory , weakens the coupling strength to values @xmath228 .
this is way below any interesting phenomenology . in order to cure this problem
, we introduce a large number of hidden @xmath124 sectors in addition to the sm @xmath229 .
we also assume that all these gauge sectors emerge with the same initial speed of light .
in addition , we take all fermions to be charged under the different @xmath124s with the same initial charge . as in the case of fermions ,
we assume that the hidden @xmath124 s are massless at the uv scale , then they become massive and decouple as we run down the rg equations .
hence , the gauge sector obeys the scale - dependent relation @xmath230 under these assumptions the second eq . in ( [ rg phenomenology ] )
is replaced by @xmath231 the solution of @xmath222 is still given by eq .
[ running of u(1 ) charge ] , while that of @xmath223 can be expressed in terms of the hypergeometric function .
instead , we numerically integrate our system setting appropriate values of the parameters @xmath232 , @xmath233 , @xmath234 , and @xmath235 . from fig .
[ rg flow of speeds and coupling ] we see that introducing many copies of @xmath124s achieves power - law suppression of @xmath191 in a very short interval .
since the many @xmath124s do not intrude the evolution of @xmath218 , we can still get reasonable coupling strength in the ir . to understand the choice of parameters used in the simulation in fig .
[ rg flow of speeds and coupling ] , it is instructive to calculate the total number of fermions and gauge fields as seen in the uv . using eqs .
[ power law running formulas ] and [ power law running for gauge bosons ] , and the numerical coefficients given in fig .
[ rg flow of speeds and coupling ] ( take @xmath236 ) we find @xmath237 , and @xmath238 , which explains the above findings . because the rg flow of @xmath218 is enhanced only by a few number of fermionic species , the coupling constant runs only logarithmically . while the huge number of gauge fields that participate in the rg equation of speeds force the running of @xmath239 to be extremely fast .
because the freezing of the speeds sets in almost immediately , we can try to replace the power law numbers in eqs .
[ power law running formulas ] and [ power law running for gauge bosons ] by constant numbers .
the idea is that we just need to have @xmath240 for a short interval until the speeds freeze to their desired values .
then , we can immediately integrate the rg equations to find @xmath241 now if we take @xmath242 , @xmath243 , @xmath244 , and @xmath245 we find that @xmath246 , and @xmath247 .
moreover , assuming that the masses of the hidden sectors are larger than @xmath248 , these masses decouple below @xmath248 and drop from the rg equations .
therefore , one needs only a constant large number of copies , @xmath249 to accomplish the emergence of an ir lorentz - invariant fixed point in a short interval of running . moreover choosing the ratio @xmath250
, we can meet the stringent constraints on the parameter @xmath191 .
this opens up the possibility that many copies of hidden sectors may suppress lorentz - violating effects already present at the tev scale .
achieving a universal speed of light is a challenge for theories which do not postulate an fundamental lorentz symmetry .
this problem is visible in known emergence models@xcite and also in hoava - lifshitz gravity@xcite . for emergent gauge fields
, the weinberg - witten theorem@xcite suggests that this will be a continual challenge as a lorentz - noninvariant initial theory may be required .
we have shown through several examples that a common limiting velocity can be emergent at low energies even if the original high - energy theory involves fields satisfying the wave equation with different velocities .
there is a heuristic rationale for this in that since fields can transform into each other through interactions , the endpoint where all the fields travel in unison is preferred .
the renormalization group treatment indeed produces this outcome .
because the running is only logarithmic for simple systems , it would take an exponentially large amount of running in order that the limiting velocities be close enough to agree with experiment .
we addressed the phenomenological constraints in sec .
7 . however , power law running is also possible if the coupling is large and constant , or if there are a very large number of interacting degrees of freedom .
we have reported on a model with this latter property .
it is important to note that not all forms of lorentz - violation disappear at low energies .
the renormalization of a general parameterization of lorentz - violation of qed has been studied in ref .
@xcite , and some operators that grow at low energy are found . a well behaved emergent theory must avoid those operators .
the running of the limiting velocities only happens due to the interactions that couple one type of particle to another .
this implies that the running will be weakest if the coupling is weak . at low energies the gravitational coupling is by far the weakest of all the fundamental forces
this implies that the most plausible velocity difference would be that of gravity .
while there have been some claims that the speed of gravity has been indirectly measured@xcite , the consensus appears to be that there is no experimental constraint on the speed of gravity@xcite .
however indirectly there is a stringent limit at the @xmath251 level on the difference of the speeds of gravity and that of light from gravitational cherenkov radiation @xcite which is valid if the speed of gravity is less than that of light .
future experiment with gravitational wave detectors provide the best opportunity to measure or constrain the difference if the speed of gravity is greater than that of light .
m.a . would like to thank erich poppitz for stimulating discussions and suggestions , and bob holdom and amanda peet for useful conversations .
we thank lorenzo sorbo for bringing ref .
@xcite to our attention .
the work of m.a . is supported by nserc discovery grant of canada .
the work of j.d . has been supported in part by the nsf grant phy - 0855119 , and in part by the foundational questions institute .
in this appendix we work out the details of the integrals @xmath19 and @xmath20 appearing in eq .
[ primary integrals ] .
these integrals are given by @xmath252 and @xmath253 to perform the integral @xmath19 , we first use the feynman trick to find @xmath254 where @xmath255 .
next , we interchange the integrals @xmath256 and @xmath257 and perform the integral over @xmath257 to find @xmath258 further , we exchange the integrals @xmath256 and @xmath259 , and rearrange the integrands to find @xmath260 where @xmath261 then , we perform the integral over @xmath259 , after analytically continuing from @xmath162 to @xmath262 dimensions , to obtain @xmath263^{3/2-\epsilon}}\\ \nonumber & & \times\frac{\gamma(\epsilon/2)}{\pi^{\epsilon/2}\left[x(p^0)^2/c_b^2-x\left(x+(1-x)c_f^2/c_b^2\right)\right]^{\epsilon/2}}\,.\\\end{aligned}\ ] ] finally , we find @xmath264 where @xmath23 .
similarly , we can show @xmath265 the vertex correction in eq .
[ vertex correction in yukawa ] results in the integral @xmath266}\\ & & \times \frac{1}{\left[\left(p_2 ^ 0-k^0\right)^2-\left(\vec p_2-\vec k\right)^2\right]}\,.\end{aligned}\ ] ] next , we use the feynman trick to get @xmath267 ^ 3}\,,\end{aligned}\ ] ] where @xmath268(k^0)^2\\ \nonumber & & -2\left[(1-x - y)p_1 ^ 0+p_2 ^ 0x \right]k^0\\ \nonumber & & + \left[(1-x - y)\vec p_1+x\vec p_2 \right]^2\\ & & + ( 1-x - y ) p_1 ^ 2+x p_2 ^ 2\,.\end{aligned}\ ] ] then , proceeding as we did before , we finally obtain the result in eq.[result for yukawa vertex ] .
we assume that the parameter space is spanned by @xmath269 , @xmath270 couplings ( these could be coupling strengths as well as speeds ) .
quantum loops will generate @xmath271 , @xmath272 corrections to the original lagrangian , and we restrict our treatment to one loop order . in general , we may write @xmath273 taking the derivative of of ( [ general expression for bare coupling ] ) w.r.t .
@xmath5 we obtain @xmath274 writing @xmath275 we find @xmath276 . also , using the definition @xmath277 , eq . (
[ next step in general expression ] ) becomes @xmath278 since we are only interested in one - loop corrections , we can ignore all terms @xmath279 .
hence , eq . ( [ third step general expression ] ) reads @xmath280\\ \nonumber & & + \frac{2}{\epsilon}g_i(\mu)\sum_{l=1}^{d}\sum_{j=1}^{c}n_{l , i}\beta_j(\mu)\frac{\partial \rho_l}{\partial g_j}\\ \label{fourth step general expression } & & + g_i(\mu)p_i\epsilon\left[1+\frac{2}{\epsilon}\sum_{m=1}^d\rho_m(\mu)n_{m , i}\right]=0\,.\end{aligned}\ ] ] now , we can rewrite eq .
( [ fourth step general expression ] ) in the following simple expression @xmath281+\frac{2}{\epsilon}\sum_{j\neq
i}^c{\cal c}_{ij}(\mu)\beta_j(\mu)\\ \label{simple expression general beta } & & = -g_i(\mu)p_i\left[\epsilon+2\sum_{m=1}^d\rho_m(\mu)n_{m , i}\right]\,,\end{aligned}\ ] ] where @xmath282n_{m , i}\,,\\ { \cal c}_{ij}(\mu)&=&g_i(\mu)\sum_{m=1}^d\frac{\partial \rho_m}{\partial g_j}n_{m , i}\,.\end{aligned}\ ] ] eq .
( [ simple expression general beta ] ) can also be written in the matrix form @xmath283\,,\ ] ] where @xmath284 the inverse of @xmath285 is given by @xmath286 hence , solving for @xmath287 from ( [ matrix form for beta ] ) we obtain @xmath288 finally , we rearrange the terms to find @xmath289 z. c. gu and x. g. wen , `` emergence of helicity + /-
2 modes ( gravitons ) from qbit models , '' arxiv:0907.1203 [ gr - qc ] .
+ m. levin and x. g. wen , `` colloquium : photons and electrons as emergent phenomena , '' rev .
phys . * 77 * , 871 ( 2005 ) .
+ m. levin and x. g. wen , `` quantum ether : photons and electrons from a rotor model , '' phys .
b * 73 * ( 2006 ) 035122 [ arxiv : hep - th/0507118 ] .
+ x. g. wen , `` artificial light and quantum order in systems of screened dipoles , '' phys .
b * 68 * , 115413 ( 2003 ) [ arxiv : cond - mat/0210040 ] .
s. s. lee , `` emergence of gravity from interacting simplices , '' int .
j. mod .
a * 24 * , 4271 ( 2009 ) [ arxiv : gr - qc/0609107 ] . c. xu , `` algebraic liquid phase with soft graviton excitations , '' arxiv : cond - mat/0602443 .
+ n. seiberg , `` emergent spacetime , '' arxiv : hep - th/0601234 .
+ l. sindoni , f. girelli and s. liberati , `` emergent gravitational dynamics in bose - einstein condensates , '' arxiv:0909.5391 [ gr - qc ]
. + s. liberati , f. girelli and l. sindoni , `` analogue models for emergent gravity , '' arxiv:0909.3834 [ gr - qc ] .
+ s. weinfurtner , m. visser , p. jain and c. w. gardiner , `` on the phenomenon of emergent spacetimes : an instruction guide for experimental cosmology , '' pos * qg - ph * , 044 ( 2007 ) [ arxiv:0804.1346 [ gr - qc ] ] .
+ g. e. volovik , `` emergent physics : fermi point scenario , '' phil .
a * 366 * , 2935 ( 2008 ) [ arxiv:0801.0724 [ gr - qc ] ] .
+ f. r. klinkhamer and g. e. volovik , `` coexisting vacua and effective gravity , '' phys .
a * 347 * , 8 ( 2005 ) [ arxiv : gr - qc/0503090 ] .
+ g. e. volovik , `` superfluid analogies of cosmological phenomena , '' phys .
rept . * 351 * , 195 ( 2001 ) [ arxiv : gr - qc/0005091 ] .
+ t. konopka , f. markopoulou and s. severini , `` quantum graphity : a model of emergent locality , '' phys . rev .
d * 77 * , 104029 ( 2008 ) [ arxiv:0801.0861 [ hep - th ] ] .
+ o. dreyer , `` emergent general relativity , '' arxiv : gr - qc/0604075 .
+ s. -s .
lee , `` tasi lectures on emergence of supersymmetry , gauge theory and string in condensed matter systems , '' [ arxiv:1009.5127 [ hep - th ] ] p. horava , `` quantum gravity at a lifshitz point , '' phys . rev .
d * 79 * , 084008 ( 2009 ) [ arxiv:0901.3775 [ hep - th ] ] .
+ p. horava , c. m. melby - thompson , `` general covariance in quantum gravity at a lifshitz point , '' phys .
* d82 * , 064027 ( 2010 ) .
[ arxiv:1007.2410 [ hep - th ] ] .
+ p. horava , `` general covariance in gravity at a lifshitz point , '' [ arxiv:1101.1081 [ hep - th ] ] .
j. w. moffat , `` superluminary universe : a possible solution to the initial value problem in cosmology , '' int .
j. mod .
phys . *
d2 * , 351 - 366 ( 1993 ) .
[ gr - qc/9211020 ] .
+ j. w. moffat , `` lorentz violation of quantum gravity , '' class .
* 27 * , 135016 ( 2010 ) .
[ arxiv:0905.1668 [ hep - th ] ] .
r. iengo , j. g. russo , m. serone , `` renormalization group in lifshitz - type theories , '' jhep * 0911 * , 020 ( 2009 ) .
[ arxiv:0906.3477 [ hep - th ] ] .
+ r. iengo , m. serone , `` a simple uv - completion of qed in 5d , '' phys .
rev . * d81 * , 125005 ( 2010 ) .
[ arxiv:1003.4430 [ hep - th ] ] .
v. a. kostelecky , c. d. lane and a. g. m. pickering , `` one - loop renormalization of lorentz - violating electrodynamics , '' phys .
d * 65 * , 056006 ( 2002 ) [ arxiv : hep - th/0111123 ] .
s. weinberg and e. witten , `` limits on massless particles , '' phys . lett .
b * 96 * , 59 ( 1980 ) .
g. baribaud , j. borer , c. bovet _ et al .
_ , `` the lep beam orbit measurement system : status and running - in results , '' in _ epacs90 : second european particle acclerator conference , vol . 1 _ , ( editions frontires , gif - sur - yvette , 1990 ) ed by p. marin , p. 137
. b. altschul , `` synchrotron and inverse compton constraints on lorentz violations for electrons , '' phys .
rev . * d74 * , 083003 ( 2006 ) . [ hep - ph/0608332 ] .
+ b. altschul , `` laboratory bounds on electron lorentz violation , '' phys
. rev . * d82 * , 016002 ( 2010 ) .
[ arxiv:1005.2994 [ hep - ph ] ] .
t. appelquist , `` tev physics and conformality , '' int .
j. mod .
* a25 * , 5114 - 5127 ( 2010 ) .
+ e. t. neil , t. appelquist , g. t. fleming , `` the conformal window in su(3 ) yang - mills , '' pos * lattice2008 * , 057 ( 2008 ) .
+ a. deuzeman , m. p. lombardo , e. pallante , `` evidence for a conformal phase in su(n ) gauge theories , '' phys .
rev . * d82 * , 074503 ( 2010 ) .
[ arxiv:0904.4662 [ hep - ph ] ] .
+ b. holdom , `` raising the sideways scale , '' phys .
* d24 * , 1441 ( 1981 ) .
g. dvali , m. redi , `` phenomenology of @xmath290 dark sectors , '' phys .
* d80 * , 055001 ( 2009 ) .
[ arxiv:0905.1709 [ hep - ph ] ] .
+ g. dvali , i. sawicki , a. vikman , `` dark matter via many copies of the standard model , '' jcap * 0908 * , 009 ( 2009 ) .
[ arxiv:0903.0660 [ hep - th ] ] . k. r. dienes , e. dudas and t. gherghetta ,
`` grand unification at intermediate mass scales through extra dimensions , '' nucl . phys .
b * 537 * , 47 ( 1999 ) [ arxiv : hep - ph/9806292 ] .
+ m. masip , `` ultraviolet dependence of kaluza - klein effects on electroweak observables , '' phys .
d * 62 * , 105012 ( 2000 ) [ arxiv : hep - ph/0007048 ] . c. m. will , `` propagation speed of gravity and the relativistic time delay , '' astrophys .
j. * 590 * , 683 - 690 ( 2003 ) .
[ astro - ph/0301145 ] .
+ s. carlip , `` model dependence of shapiro time delay and the speed of gravity / speed of light controversy , '' class .
* 21 * , 3803 - 3812 ( 2004 ) .
[ gr - qc/0403060 ] .
+ s. samuel , `` on the speed of gravity and the jupiter / quasar measurement , '' int .
j. mod .
phys . * d13 * , 1753 - 1770 ( 2004 ) .
[ astro - ph/0412401 ] . | we display several examples of how fields with different limiting velocities ( the `` speed of light '' ) at a high energy scale can nevertheless have a common limiting velocity at low energies due to the effects of interactions .
we evaluate the interplay of the velocities through the self - energy diagrams and use the renormalization group to evolve the system to low energy .
the differences normally vanish only logarithmically , so that an exponentially large energy trajectory is required in order to satisfy experimental constraints .
however , we also display a model in which the running is power - law , which could be more phenomenologically useful .
the largest velocity difference should be in system with the weakest interaction , which suggests that the study of the speed of gravitational waves would be the most stringent test of this phenomenon . |
magnetic structures are promising platforms for many modern devices , e.g. memory,@xcite sensors,@xcite and quantum computation hardware .
@xcitethe opportunities to get an ordered magnetic phase in the bulk and low - dimensional systems due to ruderman - kittel - kasuya - yosida ( rkky ) interaction @xcite were studied in a number of theoretical and experimental works.@xcite the prominent feature of rkky interaction in 1d systems is the ordering of localized spins into a helix.@xcite when the current is driven through the system of electrons and nuclei , the spin polarization can be swapped between the two subsystems through the hyperfine interaction , leading to dynamic nuclear polarization effects.@xcite if the polarization of current carrying electrons and localized spins differ , the spin - transfer torque arises,@xcite important for dynamics of domain walls@xcite and enhancing the tilting of the spiral structure in helimagnets.@xcite closely related is the dynamic nuclear polarization , arising e.g. in helical edge states of topological insulator .
the backscattering of helical electrons can be of different origins , such as assisted by phonons,@xcite magnetic impurities,@xcite or absence of axial spin symmetry.@xcite it was shown that nuclear - assisted backscattering of electrons due to hyperfine interaction induces nuclear polarization when the current is driven through the edge states of topological insulator.@xcite the main motivation for our work comes from the recent experiment by scheller _
et al._,@xcite where the conductance of a cleaved edge overgrowth gaas quantum wire was measured .
the measurements showed that the conductance of the first mode becomes @xmath0 at low temperatures instead of the naively expected @xmath1 .
this suggests the lifting of electron spin degeneracy .
the possible explanation is the presence of a helical nuclear spin polarization that gaps out one subband and thus provides an electron spin selection .
further ways to confirm the presence of the nuclear spin helix were suggested theoretically , by means of nuclear magnetic resonance,@xcite nuclear spin relaxation,@xcite and quantum hall effect anisotropies.@xcite in this work we propose and study a complementary method to detect nuclear spin helical polarization in the wire .
it is based on the effect of bias voltage applied to the wire and therefore straightforward to perform experimentally .
we investigate how the bias voltage applied to the wire affects its nuclear spin polarization .
we assume that at zero bias and finite temperature , nuclear spins are partially polarized into a helix due to the rkky interaction .
we find that an applied voltage induces a uniform nuclear polarization from both helical and non - polarized nuclear spins available for nuclear spin flips via electrons .
therefore , upon increasing the voltage the helical nuclear polarization drops , while the uniform polarization grows , and the total polarization grows too . for small voltages and increasing temperature , the uniform polarization grows because of thermal activation of electrons , while the helical polarization dramatically drops in magnitude .
once a macroscopic uniform polarization has developed , the remaining nuclear spin helix rotates as a whole around the axis along the uniform polarization . since the helical polarization affects the conductance of such systems , @xcite these predicted features
are expected to show up in the voltage and temperature dependence of the transport currrent and thus they can be tested experimentally .
recently , cantilever - based magnetic sensing techniques have been reported which enable nuclear spin magnetometry of nanoscale objects such as the nanowires considered here.@xcite such powerful techniques offer promising perspectives for direct experimental tests of the results obtained in this work . the paper is organized as follows . in sec .
[ sec : themodel ] we present the hamiltonian of our model . in sec .
[ sec : bathasonedimensionalchiralelectrons ] we describe the properties of the electron bath .
the derivation of the bloch equation for the total nuclear spin in the wire is discussed in sec .
[ sec : blochequation ] .
the resulting nuclear spin polarization and its dependence on the parameters of the system are presented and discussed in sec .
[ sec : polarization ] .
our conclusions follow in sec .
[ sec : conclusions ] .
additional information about our calculation is given in appendix .
we consider a one - dimensional electron gas and localized spins in a semiconductor nanowire .
we will refer to these localized spins as nuclear spin in the following , however , they can be also of other origins , such as e.g. magnetic impurities , etc .
the electrons and nuclei interact via the hyperfine interaction described by the hamiltonian @xmath2 where @xmath3 is a hyperfine constant of the material , @xmath4 is the nuclear spin density , @xmath5 is the transverse part of electron wavefunction , @xmath6 denotes the electron position along the wire , @xmath7 is the position of the nucleus along the wire and in the transverse direction respectively , @xmath8 is an electron spin operator , and @xmath9 is a nuclear spin operator ( in units of @xmath10 ) with the magnitude @xmath11 .
we assume that the transverse part of the electron wavefunction @xmath12 is constant in the wire cross - section , @xmath13 , where @xmath14 is the wire crossection area .
we parametrize it alternatively by the number of nuclear spins in the cross - section , @xmath15 , with @xmath16 being the lattice constant . in gaas @xmath17 ,
@xmath18 nm , @xmath19 , @xmath20 , and @xmath21 is typically of the order of @xmath22 .
finally , we introduce @xmath23 with @xmath24 the wire length ( typically of order microns ) , which gives @xmath25 as the total number of nuclear spins in the wire .
the total hamiltonian reads @xmath26 where @xmath27 is electron effective mass and @xmath10 is the planck constant . if the hyperfine interaction , eq .
( [ eq : hyperfinehamiltonian ] ) , is weak on the energy scale of the electrons , its effects can be treated perturbatively .
the condition is quantified by @xmath28 , where @xmath29 is the fermi energy of the electron system .
this condition is well satisfied in the cases we consider here . a schrieffer - wolff transformation on @xmath30 perturbatively in @xmath31 , _
i.e. _ , in order @xmath32 , results to leading order in an effective interaction between the localized spins , the rkky interaction,@xcite @xmath33 here , the indexes @xmath34 label the nuclear spins and the rkky coupling @xmath35 is related to the static spin susceptibility of electrons ( see eq .
( c1 ) and below in ref . ) , giving rise to the spatially dependent rkky interaction .
let us rewrite eq
. in the momentum representation , defined through the fourier transforms @xmath36 j_{ij}$ ] , with @xmath37 , and @xmath38 , with @xmath39 , and in both cases @xmath40 .
we get @xmath41 in one dimension , the rkky coupling @xmath42 has a sharp minimum at momentum @xmath43 , with @xmath44 the electron fermi wavevector.@xcite consider an approximation in which we neglect all values of @xmath45 with respect to the large ( negative ) value at this minimum , @xmath46 to understand the spectrum of this hamiltonian , we introduce linearly transformed spin operators , @xmath47 with @xmath48 the matrix corresponding to a rotation by angle @xmath49 around a unit vector @xmath50 . inserting eq . into eq .
we get @xmath51 where we define the vector components along @xmath52 as @xmath53 , and perpendicular to it as @xmath54 , and we separated the terms bilinear in the spin operators at finite momenta , @xmath55 .
\label{eq : ugly b}\ ] ] the first term in the bracket of eq
. describes the energy of ferromagnetically coupled spins @xmath56 : a configuration in which all these spins are collinear , along a vector perpendicular to @xmath52 , gives a minimal possible energy , of value @xmath57 .
this configuration corresponds to a classical ground state of eq . as well , as it saturates the energy lower bound obtained using the parseval s identity @xmath58 . going back to the laboratory frame according to eq .
, the ground state corresponds to a helical configuration where the nuclear spins are oriented parallel to each other in the wire cross - section , along a direction which rotates in a fixed plane as one moves along the wire ( for illustration , see fig .
[ fig : nanowire ] ) .
we shall refer to this plane as the helical plane , with @xmath52 being its normal unit vector .
a unit vector @xmath59 gives the direction of the polarization within this plane at position @xmath60 .
the finite momenta components , eq . , arise from the choices of a definite helicity and the vector @xmath52 in eq . , which break the full spin rotational symmetry of eq .
namely , choosing a frame with helicity opposite to the ground state helicity would lead to a swap of the roles of @xmath61 and @xmath62 .
further , configurations where both helicities are populated lead to a lower energy gain .
for example , choosing both with the same weight , gives in the laboratory frame a spin - density wave , i.e. , a cos - like oscillation along a fixed vector , @xmath63 , which gives only half of the energy gain of a helical order .
such oscillating , rather than rotating , configuration corresponds to the first term in eq . . to conclude , up to the spin rotational symmetry , which allows for arbitrary directions of @xmath52 , and @xmath64 , the ground state with
ferromagnetically aligned @xmath56 ( helically ordered @xmath65 ) is unique . if the order is established , the expectation value of @xmath66 is macroscopic , and we parametrize it by a polarization @xmath67 , @xmath68 so that @xmath69 corresponds to a completely ordered state . with this
we reduce eq . by the mean field approximation to a hamiltonian describing a set of non - interacting spins @xmath70 in the presence of the position - dependent internal field @xmath71 this concludes a simplified derivation of the reduction of the rkky hamiltonian , eq . , into a set of non - interacting spins ,
( [ eq : rkkyaveraged ] ) , in an effective ( mean ) field , eq . ( [ eq : omega0 ] ) .
a detailed analysis of the applicability of such an approximation was given in ref .
, based on the derivation of the spectrum of the full hamiltonian eq . , without employing a mean field ansatz .
there it was found that this approximation , in essence neglecting the long wavelength magnons , is well justified for sub - kelvin temperatures and wire lengths relevant for mesoscopic experiments .
as we consider the limit @xmath28 , we adopt the bohr - oppenheimer approximation , assuming that electrons react instantaneously to the changes in nuclear spin subsystem .
consequently , we can consider the effect of the nuclear polarization on electrons as an overhauser field@xcite @xmath72 where @xmath73 is an electron magnetic moment .
thus , the electron hamiltonian is @xmath74 in eq .
( [ eq : hb ] ) we do not include electron - electron interactions explicitly . in the following to evaluate the internal field @xmath75 we use eqs .
( c4 ) , and ( c5 ) from ref . .
in these equations electron - electron interaction is significant ( for example , for the critical temperature of the helical polarization @xcite ) and therefore is included .
to describe the nuclear polarization in the wire when a bias voltage is applied , we will first investigate the behaviour of one nuclear spin placed in an effective field of all others , eq .
( [ eq : rkkyaveraged ] ) , and interacting with the bath of electrons described by eq .
( [ eq : hb ] ) .
to find how a nuclear spin is affected by the electrons when the bias voltage is applied , we first consider the properties of the electron bath in the wire .
as already mentioned in sec .
[ sec : themodel ] , the electrons are moving in the overhauser field produced by the nuclear spins [ see eqs .
( [ eq : bov ] ) , ( [ eq : hb ] ) ] . as the nuclear spins form a helix in equilibrium , this particular overhauser field , denoted by @xmath76 , is also helical .
consequently , the electron spectrum is @xmath77 where @xmath78 is the electron wavevector , and @xmath79 and @xmath80 denote the lower and upper subbands respectively .
they are split by the gap @xmath81 at @xmath82 .
the corresponding wavefunctions are @xmath83 , \mbox{\ \ \ \ } \\ \label{eq : wavefunctionfull2 } \psi_{k,+}(r)&=&\frac{e^{ikr}}{\sqrt{l}}[e^{ik_fr}\cos{\frac{\theta_k}{2}}\ket{\downarrow}-e^{-ik_fr}\sin{\frac{\theta_k}{2}}\ket{\uparrow } ] , \mbox{\ \ \ \ } \end{aligned}\ ] ] where @xmath84 and @xmath85 , and @xmath86 , @xmath87 denote the spin states with spin up and spin down respectively , where @xmath52 sets the quantization axis .
these expressions of the wavefunctions can be simplified since typically the ratio @xmath88 , where @xmath89 is the maximum overhauser field when all nuclei are fully polarized along a given direction .
for example , for a gaas quantum wire @xmath90 , while @xmath91 , which gives @xmath92 .
consequently , we can use @xmath93 as a small parameter .
we expand eq .
( [ eq : wavefunctionfull ] ) in leading order of @xmath93 and for the states within the partial gap we get @xmath94 where for right - moving electrons @xmath95 the spinor is @xmath96 , and for left - moving @xmath97 it is @xmath98 . therefore , within our approximation the electronic states in the partial gap are helical : the spin is determined by the propagation direction , and is opposite for left - moving and right - moving electrons .
next , we consider the voltage applied to the wire and define it as the difference between the chemical potentials for the left- and right - moving electrons ( see fig .
[ fig : spectrum ] ) . assuming a ballistic wire ,
the chemical potential of a given branch is constant in space . concerning the nuclear spins , only the electron states from the partial gap are relevant , because we consider the case of small voltage , i.e. @xmath99 , where @xmath100 is the electron charge , and small temperatures @xmath101 , _
i.e. _ , @xmath102 .
this means that we adopt two approximations .
first , we neglect the influence from electron states which are not in the partial gap ( the upper ( @xmath103 ) subband is neglected completely ) , because their contribution to transport is exponentially small , proportional to @xmath104 $ ] .
second , we use eq .
( [ eq : simplifiedwavefunction ] ) for the electron wavefunctions , which means that we consider it in leading order of @xmath105 , @xmath106 , and @xmath93 .
therefore , for a description of the electron system in terms of a heat bath that causes the relaxation of the nuclear spins , we take into account two branches : left- and right - moving electrons with spins @xmath87 and @xmath86 , respectively . with the polarity as assumed in fig .
[ fig : spectrum ] , the applied voltage depletes the left ( l ) branch and increases the population of the right ( r ) branch .
this imbalance in population opens up an additional phase space for the electrons to backscatter - predominantly from r to l. because of the helical character of the states , such backscattering is accompanied by an electron spin flip ( from @xmath86 to @xmath87 ) .
this , in turn , is enabled by the total spin - conserving hyperfine interaction eq .
( [ eq : hyperfinehamiltonian ] ) , so that each electron spin flip is compensated by a nuclear spin flip in the opposite direction . in this way a
_ uniform _ nuclear polarization along the @xmath52 direction is built up .
we denote @xmath107 as the overhauser field corresponding to this uniform polarization .
the spectrum of electrons moving in the total ovehauser field @xmath108 reads @xmath109 ^ 2 } , \label{eq : spectrum}\ ] ] see fig .
[ fig : spectrum ] .
the asymmetry of the spectrum is due to the uniform overhauser field @xmath107 .
the corrections to the wave functions , eq .
( [ eq : simplifiedwavefunction ] ) , due to @xmath107 are negligible in leading order of @xmath93 . ) and the direction of the electron spins in the presence of the helical overhauser field @xmath110 and the uniform overhauser field @xmath111 perpendicular to the plane of the helix .
red arrows denote the spin directions of the electrons in the lower subband , and the blue arrows label the spin directions for the upper subband .
the coordinate system for the spins is formed by @xmath112 and @xmath52 shown in the right lower corner .
the chemical potentials for left- and right - movers are denoted as @xmath113 and @xmath114 , respectively .
the voltage applied to the wire is @xmath115 . ]
we note that from eqs .
( [ eq : wavefunctionfull ] ) and ( [ eq : wavefunctionfull2 ] ) it follows that the electron spins become also polarized , thereby producing a knight shift acting as an effective magnetic field @xmath116 back on the nuclear spins .
this knight shift is defined as @xmath117 , where @xmath118 denotes averaging over the eigenstates of the hamiltonian in eq .
( [ eq : hb ] ) with populations defined by the voltage . in this work , however , we can neglect @xmath116 with respect to @xmath119 produced by the rkky interaction.@xcite
to investigate the time - dynamics of the nuclear spins , we apply the standard bloch - redfield theory to our problem , which is valid for weak coupling between spin system and bath degrees of freedom,@xcite as is the case here .
first , we write down the bloch equation for the average @xmath120 of the @xmath121th nuclear spin . by applying eqs .
( 7)-(11 ) from ref . to our eqs .
( [ eq : hyperfinehamiltonian ] ) , ( [ eq : rkkyaveraged ] ) , and ( [ eq : hb ] ) , we get ( for more details see appendix [ app : blochequation ] ) @xmath122 where @xmath123 determines the precession , the relaxation tensor @xmath124 the decay and the inhomogeneous vector term @xmath125 the stationary value of @xmath126 .
both , @xmath124 and @xmath125 are expressed in terms of time correlators ( see app . a ) @xmath127 where @xmath128 is time , the indexes @xmath129 label the components of the effective fluctuating internal field @xmath130 defined via @xmath131 .
the time - dependence follows from the interaction representation @xmath132 .
we note that above equations are valid for a spin 1/2 .
however , it is well - known @xcite that the relaxation time of a spin into its stationary value does not depend on the spin length ( in born approximation ) .
thus , we will assume that our results apply for arbitrary spins . as follows from secs .
[ sec : themodel ] and [ sec : bathasonedimensionalchiralelectrons ] , we can define the expectation value of a nuclear spin at position @xmath60 as @xmath133 where @xmath134 denote the polarizations along the two orthogonal directions @xmath112 and @xmath52 , respectively .
we also introduce position - independent tensors @xmath135 and @xmath136 in the rotated frame defined by the rotation matrix @xmath137 , via @xmath138 having eqs .
( [ eq : omega0 ] ) , ( [ eq : blochequationi ] ) , ( [ eq : gamma0 ] ) , and ( [ eq : upsilon0 ] ) , we can describe the time - evolution of the nuclear spin @xmath139 in the rotated frame .
eventually we are interested in the dynamics of the total ( macroscopic ) polarizations , rather than the one of an individual nuclear spin .
we therefore introduce the total nuclear spin in the rotated frame @xmath140 , and write the equation of motion for it using eqs .
( [ eq : omega0 ] ) , ( [ eq : blochequationi ] ) , ( [ eq : gamma0 ] ) , and ( [ eq : upsilon0 ] ) .
we get @xmath141 where we denoted @xmath142 .
the first term implies a rotation of the helical direction @xmath112 , around the axis @xmath52 with frequency @xmath143 .
this can be seen by introducing a time - dependent vector @xmath144 , where @xmath145 . in the born - oppenheimer approximation , the tensors @xmath124 and @xmath146 are functions of the instantaneous values of @xmath112 and @xmath52 , so we write @xmath147 with this the time evolution of @xmath148 in the rotating frame , @xmath149 , is described by @xmath150 using this equation and properties of the electron bath discussed in sec .
[ sec : bathasonedimensionalchiralelectrons ] , we can describe the polarization of the nuclear spins in the wire as function of temperature and voltage .
to find the polarizations @xmath67 and @xmath151 from eq .
( [ eq : blochequationfinal ] ) we now evaluate the tensors @xmath135 and @xmath136 explicitly . for that we first evaluate the correlator @xmath152 .
using eqs .
( [ eq : simplifiedwavefunction ] ) and ( [ eq : spectrum ] ) we get @xmath153 , \mbox { \ \ \ \ } \label{eq : iab}\end{aligned}\ ] ] where @xmath154 is the difference between chemical potentials of branch @xmath155 and @xmath16 , with @xmath16 and @xmath155 denoting @xmath24 ( left - movers ) or @xmath156 ( right - movers ) .
here we also use the fermi distribution function @xmath157 + 1\right]^{-1}$ ] . as was mentioned in sec .
[ sec : bathasonedimensionalchiralelectrons ] , we consider voltages and temperatures smaller than the partial gap @xmath81 given by the helical polarization .
therefore the term @xmath158 $ ] allows us to consider only the energy window of @xmath159 around @xmath29 , because @xmath160 decays exponentially for @xmath161 .
consequently , we approximate the electron density of states ( per spin ) by @xmath162 . up to first order in @xmath93 , we have @xmath163 , where @xmath164 is the fermi velocity of the electrons . having obtained @xmath152 , it is straightforward to calculate @xmath135 and @xmath136 , using eqs .
( [ eq : gammad])-([eq : upsilon ] ) and ( [ eq : correlatorelectrons])-([eq : iab ] )
. we can then solve eq .
( [ eq : blochequationfinal ] ) for the steady state polarizations ( keeping @xmath165 as a constant ) and obtain @xmath166 however , from eq .
( [ eq : omega0 ] ) it follows that @xmath167 , _
i.e. _ , @xmath165 depends on @xmath67 .
this leads to non - linear algebraic equations for two unknowns , @xmath151 and @xmath67 , which we solve numerically using material parameters for gaas ( analytical expressions for small deviations of the polarizations are given below ) .
we plot the values obtained in this way and discuss their behaviour as a function of voltage and temperature , the experimental parameters that are most directly accessible . along helical direction @xmath112 ( blue ) , polarization @xmath151 in the direction of @xmath52 perpendicular to the helix plane ( red ) , and the overall polarization of nuclei @xmath168 ( black ) .
( b ) enlarged from ( a ) the voltage dependence of @xmath67 and @xmath168 .
we use @xmath169 and other parameters as given in the text . ]
the voltage dependence of the polarizations is shown in fig .
[ fig : polarizationvoltage ] .
we can see that the polarization @xmath151 grows faster with voltage than @xmath67 decays , therefore the overall polarization of the nuclei @xmath168 grows with voltage , too .
this means that the nuclear spins are more polarized when a voltage is applied than when they are in equilibrium at the same temperature .
we also note that having a non - zero component @xmath151 means that nuclear spins have a conical polarization , rather than a helical one .
to plot fig .
[ fig : polarizationvoltage ] we used eqs .
( c4 ) , and ( c5 ) from ref . as was mentioned above , where the dependence of @xmath170 on temperature is described in detail . to evaluate @xmath170 we used the characteristic values for gaas : the fermi velocity @xmath171 , and the number of nuclei in the wire cross - section @xmath172 . for the expression for @xmath173 taken from ref .
we use the electron - electron interaction luttinger liquid parameter @xmath174 and the absolute value of spin @xmath20 . for the constants described above and at @xmath169 and @xmath175 the rotation frequency of the nuclear spin helix is @xmath176 .
( upper panel , red ) and the polarization @xmath67 ( lower panel , blue ) . for these plots
the same parameters were used as in fig .
[ fig : polarizationvoltage ] and the applied voltage is @xmath177 .
we note that our calculation is valid for @xmath178 , therefore the smallest value of @xmath67 we consider here is @xmath179 . ]
it is natural to expect that high temperature destroys the nuclear helical order.@xcite indeed , fig .
[ fig : polarizationtemperature ] shows that the helical polarization @xmath67 decays with temperature and then drops in magnitude around @xmath180 .
as our calculation is valid for @xmath178 , the smallest value of @xmath67 allowed by self - consistency for our parameters is @xmath179 . from fig .
[ fig : polarizationtemperature ] it also follows that the polarization @xmath151 grows with temperature .
this growth is explained by the fact that due to higher temperature the electron states with higher energy become occupied .
this makes the nuclear spin flip more probable .
it is obvious that there is a temperature where the polarization @xmath151 gets destroyed , however , for the range of temperatures given in fig .
[ fig : polarizationtemperature ] @xmath151 does grow , whereas the helical polarization @xmath67 decays significantly .
the decay of @xmath67 with temperature is rapid , while the growth of @xmath151 is less pronounced .
therefore , the overall nuclear polarization in the wire strongly decays with increasing temperature . for the parameters we used for fig .
[ fig : polarizationtemperature ] the effect of temperature on @xmath67 is stronger than the one of a finite voltage .
the initial temperature scaling of @xmath67 away from unity ( see fig . [
fig : polarizationtemperature ] ) can be obtained readily from eq .
( [ polarization_ph ] ) by treating @xmath181 as a small perturbation .
this yields @xmath182 where the first equality holds well for the temperature interval @xmath183 , while the second one is a good approximation for @xmath184 .
here we denoted @xmath185 , and the temperature - independent parameter @xmath186 depends on the material and geometrical properties of the sample ( see eq . ( 11 ) and eqs .
( c4 ) and ( c5 ) of ref . ) . for @xmath174 ( chosen for the plots )
we get @xmath187 ( we recall that @xmath188 corresponds to vanishing electron - electron interactions ) . the initial decrease of @xmath67 due to voltage in fig .
[ fig : polarizationvoltage ] for @xmath189ev scales as @xmath190 where @xmath191 and @xmath192 depend on material and geometrical parameters of the nanowire and on temperature .
finally , we mention that recent progress in nuclear spin magnetometry on nanowires@xcite has opened the perspective to measure the nuclear spin polarizations directly and thus to test the predictions made here . moreover , due to the helical nuclear polarization which acts on electrons as an overhauser field @xmath76 there is a partial gap in the electron spectrum [ see eq .
( [ eq : spectrum ] ) ] . as a result ,
the conductance of a ballistic nanowire is less than @xmath1 for sufficiently low temperatures and @xmath193 .
@xcite as was shown above , the polarization @xmath67 , and consequently @xmath76 , decrease with increasing voltage and temperature .
we thus expect qualitatively that the conductance of the wire will increase with the decrease of the partial gap @xmath194 .
we have shown that due to the hyperfine interaction between electrons and nuclei in the wire the applied voltage changes the form of the nuclear polarization and its amplitude . assuming that in equilibrium there is a helical nuclear polarization @xmath67 present in the wire due to rkky interaction , a bias voltage induces a uniform polarization @xmath151 perpendicular to the helix plane . due to this polarization
the nuclear spin helix starts to rotate around the axis perpendicular to the helical plane .
when a non - zero polarization @xmath151 buidls up , the nuclear polarization changes from helical to conical .
we have also presented the voltage dependence of @xmath151 and @xmath67 and seen that @xmath151 increases with voltage , whereas @xmath67 decreases . following from these two effects the overall nuclear polarization in the wire grows with voltage .
remarkably , @xmath151 grows with temperature in the considered range of temperatures .
this is because the nuclear spin flip becomes more probable as electrons occupy higher energy states .
this thermal activation effect is noticeable for the considered regime @xmath195 .
the growth of the overall polarization @xmath168 with voltage and the growth of @xmath151 with temperature are intriguing and _ a priori _ non - obvious effects .
the polarization effects predicted here might be observed in transport experiments @xcite or more directly via cantilever based nanoscale magnetometry .
@xcite we thank v. n. golovach , d. becker , b. braunecker , l. glazman , c. kloeffel , t. meng , and c. orth for helpful discussions and acknowledge support from the swiss nf , nccr qsit , and s@xmath196nano .
to write down the bloch equation for the total nuclear spin in the wire , we use eqs .
( 7)-(11 ) from ref . .
here we present them adopted to our case of a nuclear spin interacting with the bath of electrons and placed into the effective field produced by all other nuclear spins in the wire .
the bloch equation for the @xmath197th nuclear spin reads @xmath198 to express tensors @xmath199 and @xmath200 we introduce a unit vector @xmath201 along @xmath202 , i.e. @xmath203 . the tensor @xmath199 consists of a dephasing part @xmath204 which comes from energy conserving processes and a pure relaxation part @xmath205 , which comes from the energy exchange with the bath @xcite ( played here by the electron system ) , @xmath206 , \\ & \bm{\gamma}^r_{n , ij}&=[\delta_{ij}(\delta_{pq}-l_pl_q)\mathcal{j}_{pq}^+(\omega_n)-\\ \nonumber & -&(\delta_{ip}-l_il_p)\mathcal{j}_{pj}^+(\omega_n)-\delta_{ij}\epsilon_{kpq}l_k\mathcal{i}_{pq}^-(\omega_n)+\\ \nonumber & + & \epsilon_{ipq}l_p\mathcal{i}^-_{qj}(\omega_n)].\end{aligned}\ ] ] here , the indexes @xmath207 denote components of tensors , and we use the einstein convention of summation over repeated indexes .
further , @xmath208 is the levi - civita symbol and @xmath209 the kronecker delta , while @xmath210 denotes the @xmath78th component of vector @xmath201 .
the inhomogeneous part of the bloch equation @xmath200 reads @xcite @xmath211),\mbox{\ \ \ } \end{aligned}\ ] ] where @xmath212 denotes the component of @xmath200 .
the terms @xmath213 @xmath214 are defined as @xmath215 , \\ \mathcal{i}_{ij}^\pm(\omega)&=&\mbox{im}[\mathcal{j}_{ij}(\omega)\pm \mathcal{j}_{ij}(-\omega)].\end{aligned}\ ] ] the term @xmath216 is the laplace transformation of the correlator of the fluctuating fields @xmath217 at different times , @xmath218 where @xmath219 . using eq .
( [ eq : appblochequation ] ) we expressed the bloch equation for the total nuclear spin in the wire resulting in eq .
( [ eq : blochequationtotal ] ) . however , we caution that the behaviour of the polarizations @xmath220 can only indicate the trend for the conductance but does not give its precise dependence on temperature and voltage . indeed , in transport new temperature and voltage effects emerge , especially in the transition region where the partial gap is comparable to temperature and voltage . | we study the effect of bias voltage on the nuclear spin polarization of a ballistic wire , which contains electrons and nuclei interacting via hyperfine interaction . in equilibrium ,
the localized nuclear spins are helically polarized due to the electron - mediated ruderman - kittel - kasuya - yosida ( rkky ) interaction . focusing here on non - equilibrium , we find that an applied bias voltage induces a uniform polarization , from both helically polarized and unpolarized spins available for spin flips .
once a macroscopic uniform polarization in the nuclei is established , the nuclear spin helix rotates with frequency proportional to the uniform polarization .
the uniform nuclear spin polarization monotonically increases as a function of both voltage and temperature , reflecting a thermal activation behavior .
our predictions offer specific ways to test experimentally the presence of a nuclear spin helix polarization in semiconducting quantum wires . |
galaxy redshift surveys such as the 2dfgrs ( colless et al . 2001 ) or the sloan digital sky survey ( york et al .
2000 ) and n body simulations of cosmic structure formation ( for example springel et al . 2005 and references therein ) demonstrate the existence of a complicated network of matter . at the most prominent positions in this network , massive clusters of galaxies can be found , which are interconnected by filaments and , to a lesser degree , sheets . in both observations and simulations clusters
are aspherical systems , a fact which has lead to investigations of the degree and origin of asphericity and of possible alignments between neighbouring objects . in this paper
we use an n body simulation of a @xmath0cdm universe to study the link between this network of structure and the shapes and alignments of galaxies and clusters .
binggeli ( 1982 ) first investigated the alignment of galaxy clusters , finding that for the 44 abell clusters in a sample there was a strong longitudinal alignment signal ( clusters tend to point towards each other , the `` binggeli effect '' ) .
this was seen for cluster - cluster separations of up to @xmath1 15 @xmath2 mpckm / sec / mpc . ] .
most follow up studies , optical and otherwise , have confirmed binggeli s results ( flin 1987 , rhee & katgert 1987 , west 1989 , rhee et al .
1992 , plionis 1994 , west et al .
1995 , chambers et al . 2000 and 2002 ) , although there were also some negative reports ( struble & peebles 1985 , ulmer et al .
1989 ) . on the theoretical side
, much attention has been devoted to the shapes of dark matter halos ( for example , the radial density profile was examined by navarro et al .
1997 , moore et al .
1999 , and the asphericity of galaxies and/or clusters by jing & suto 2002 , bailin & steinmetz 2004 , hopkins et al .
2004 , kasun & evrard 2004 , allgood et al .
2005 , lee et al . 2005b , and paz et al .
the alignments of halos have been getting somewhat less attention .
splinter et al .
( 1997 ) , onuora & thomas ( 2000 ) , and faltenbacher et al .
( 2002 ) have all found significant alignments of galaxy cluster halos , as did kasun & evrard ( 2004 ) , hopkins et al .
( 2005 ) , and basilakos et al .
sized halos also have a strong tendency to be aligned in the same direction as other nearby halos ( see e.g. heavens , refregier & heymans 2000 , croft & metzler 2000 ) , as well as pointing along the direction vector to nearby halos ( e.g. , li & croft 2005 ) .
this latter signal , the intrinsic density shear correlation has been recently seen in observational galaxy data by mandelbaum ( 2005 ) , and agustsson & brainerd ( 2005 ) .
it is not immediately obvious what causes the alignment of halos .
as shown in van haarlem & van de weygaert ( 1993 ) , clusters tend to orient themselves toward the direction of the last matter infall ( as shown in colberg et al .
1999 , matter falls into cluster predominantly from filaments ) .
but there is also a positive correlation between the inertia tensor of a cluster and its surrounding tidal field ( bond et al .
1996 ; as shown in lee et al .
2005b , the axis
ratio distribution of halos can be modelled analytically on the basis of this , also see lee et al .
in reality , both effects will probably be intertwined .
hopkins et al .
( 2005 ) attempt to connect the alignment between cluster pairs and large scale structure by looking at the number density of clusters contained in a cylinder that connects the two clusters .
they find that as the number density rises , so does the average cluster alignment .
while this indicates that filaments between clusters might cause increased alignment , for a more general analysis it is necessary to investigate cluster samples for which the inter - cluster filaments are found with reference to the density field itself .
such a set of filaments would also make possible an investigation of shapes and alignments of halos much smaller than galaxy clusters .
colberg et al .
( 2005 ) investigated inter
cluster configurations of matter in a high resolution simulation of cosmic structure and found a complete set of filaments . here
we will use their data as the basis for a detailed investigation of the connection between large scale structure and halo alignments .
in particular , we will investigate whether there is a connection between the alignment of pairs of clusters and the existence ( or non existence ) of a filament between them .
we will also study whether halos of mass smaller than that of a massive cluster are aligned with filaments and/or the tidal fields of the clusters .
the latter is interesting in the light of an algorithm proposed by pimbblet ( 2005 ) to locate filaments in galaxy redshift catalogues .
this work is organized as follows . in section [ simulation ]
we describe the simulation , in section [ alignments ] we study the alignments of halos between pairs of clusters connected by filaments or with voids in between them , and in section [ ellip ] , we briefly re visit ellipticities of halos .
section [ summary ] contains a summary and discussion .
we make use of the high resolution @xmath0cdm simulation introduced in jenkins et al .
the simulation parameters ( @xmath3 = 0.3 , @xmath0 = 0.7 , @xmath4 = 0.7 , and @xmath5 = 0.9 ) are in good agreement with the currently accepted standard cosmology ( colless et al . 2001 , spergel et al . 2003 , seljak et al . 2005 ) .
the simulation follows the evolution of 256@xmath6 dark matter particles in a cubical volume of size ( 141.3@xmath2mpc)@xmath6 on a side , resulting in a particle mass of @xmath7 .
the group catalogue is obtained by running a standard friends of friends group finder on the full particle set , using a linking length of @xmath8 times the mean interparticle separation .
we use all groups with 50 particles or more for the analysis , which results in a total sample of 17461 halos .
our choice of the minimum halo mass is motivated by the fact that for less than 50 particles , the structure of a halo can not be reliably determined ( see e.g. the tests in kasun & evrard 2004 ) . in the following , we will refer to the 170 most massive groups as clusters and to all others as halos .
the cluster sample was designed to match the space density of abell clusters as outlined in colberg et al .
( 2005 ) . for each halo
, we compute its principal axes by diagonalizing the moment of inertia tensor @xmath9 where the sum is over all particles in the halo , and the coordinates are defined with respect to the center of mass of the group .
the resulting eigenvalues @xmath10 , @xmath11 , and @xmath12 are sorted by size , in descending order .
the ellipticity of a group is then defined by @xmath13 we make use of the filament catalogue described by colberg et al .
( 2005 ) . in that paper , intercluster filaments were found by investigating the configuration of matter between neighbouring clusters .
the mass distributions between all pairs of clusters ( up to the 12th nearest neighbour of each cluster ) were projected onto orthogonal planes .
the results were then visually classified according to the appearance of the projections ( essentially into either a filament , absence of a structure , or rarely , a sheet ) .
we will use the colberg ( 2005 ) classifications to examine alignments of halos and clusters with the cluster
cluster axes , and how it depends on the presence of a filaments .. figure [ fig : image ] shows a small piece of the simulation volume , with halos being plotted as ellipsoids . the ellipsoids are scaled according to their mass , and the orientations of the ellipsoids correspond to those of the actual mass distributions .
the region shown in the plot encompasses a pair of clusters connected by a filament .
the same filament is shown bottom
center in figure 1 of colberg et al .
the filament is clearly visible , although any coherent alignments of the halos is difficult to pick out by eye .
we will examine this statistically in the next section .
cc cc cc
for each cluster pair , we examine the halos that lie in the cylinder whose central axis is defined by the cluster cluster axis and whose radius extends 70@xmath2mpc from the cluster cluster axis . by going out to these large radii
we cover as much of the simulation volume for each cluster pair as possible .
we define two vectors @xmath14 and @xmath15 such that @xmath14 lies on the cluster cluster axis , and
@xmath15 points along one of the halo principal axes .
our measure of the alignment between them is defined by @xmath16 where @xmath17 is the angle between the two vectors . for each of the three eigenaxes
we compute this alignment . in figure [
fig : halo - fil - perpd - align ] , we plot the alignments of the principal axes of the halos with the cluster
cluster axis as a function of the perpendicular distance from that axis .
the rows show the major , intermediate , and minor axes . in the leftmost , center , and rightmost columns we plot alignments of halos , of halos plus clusters , and of clusters , respectively .
cluster pairs which are connected by a filament are shown using squares , whereas asterisks are used for cluster pairs for which no coherent structure was found along the cluster cluster axis .
errors bars are computed assuming poissonian statistics .
the dotted line shows the expectation for a random sample with no alignments . for small separations from the cluster
cluster axis , the major and minor axes of halos are aligned and anti aligned with that axis , respectively ( leftmost column of figure [ fig : halo - fil - perpd - align ] ) .
interestingly , there is no clear difference between those cluster pairs that are connected by a filament and those that are not connected at all . for galaxy
size halos therefore the fact that they lie in a filament or not does not affect their alignments ( we will examine the effect on their ellipticities in section 4.2 ) .
the alignment signals become small at larger separations from the cluster cluster axis . at separations of around 4@xmath2mpc
the alignment signal is almost absent . as can be seen from figure 8 in colberg et al .
( 2005 ) , at this scale , the averaged density profile of filaments has dropped strongly from its central value .
however , given that we do not find a difference in the alignments between connected and unconnected cluster pairs , this finding has to be treated as a mere coincidence .
we note that the ( anti ) alignment signal is somewhat stronger for the minor axes than for major axes .
also , for the minor axes , there is a small difference between halos in filaments and halos elsewhere , although the error bars are relatively large .
the intermediate halo axes are neither aligned nor anti aligned with the cluster cluster axes .
the rightmost column of figure [ fig : halo - fil - perpd - align ] shows the alignment results for clusters .
the major and minor axes of the clusters connected by filaments are aligned and anti aligned with those filaments , respectively .
although the statistical uncertainty is quite large , the signal is significantly stronger than for galaxy size halos .
however , no such signal exists for cluster pairs that are not connected by filaments , which again is very different from the galaxy halo result .
this finding provides an explanation for the bingelli effect , which is directly connected to the presence of filaments .
cc cc having examined alignments as a function of the radial distance from cluster
cluster axes , we now turn our attention to the dependence of any alignment signal on the distance of a halo from its nearest cluster , i.e. along the cluster cluster axis .
we restrict this analysis to all halos that are within 1.5@xmath2mpc from the axis .
figure [ fig : halo - fil - projd - align ] shows the alignments for the three halo axes , both for halos ( top row ) and for halos plus clusters ( bottom row ) .
the symbols are the same as in figure [ fig : halo - fil - perpd - align ] .
we scale the halo cluster separations on the @xmath18axis such that a halo position just outside the virial radius of the nearest cluster is at position 0 , whereas a halo right in the center between both clusters sits is at position 1 .
@xmath19 is the total length of a cluster
cluster axis , and @xmath20 corresponds to the distance of a halo from the nearer one of the clusters that define the cluster cluster axis .
as before , the dotted line shows the expectation for a random sample with no alignments .
figure [ fig : halo - fil - projd - align ] shows that the magnitude of the alignments does not depend on the distance of a halo from the nearest cluster .
there also is no difference between cluster pairs that are connected by a filament and those that intersect a void .
cc cc the presence of tidal fields can provide a very natural explanation for the alignments and shapes of objects . in the following sections
, we will study the influence of the tidal forces exterted by the clusters that are used to define the cluster cluster configurations . in order to make a simple model of the tidal field , we treat the two clusters as spherical objects and we ignore the mass contained in the filaments .
given the large masses of the clusters and the relatively low average overdensity of filaments ( see the discussion in colberg et al .
2005 ) this simple model is reasonable . for this dumbell
shaped configuration , use of an analytical expression for the gravitational potential is possible .
we use cylindrical coordinates ( @xmath21 ) and take one of the clusters to lie at the origin with the other one on the @xmath22axis .
the gravitational potential @xmath23 then becomes @xmath24 where @xmath19 is the the length of the cluster
cluster axis , and the masses of the two clusters are given by @xmath25 and @xmath26 .
the components of the tidal field tensor @xmath27 can be computed through @xmath28 since we are only interested in the absolute magnitude of the tidal field components we neglect the sign , thus taking the absolute value to be a measure of the stretching a body would feel . in this axially symmetric model ,
there are two components of @xmath29 that stretch a spherical mass into an ellipse .
they are @xmath30 } { [ \rho^2+(l - z)^2]^{5/2 } } - \frac { m_1 ( \rho^2 - 2z^2 ) } { ( \rho^2+z^2)^{5/2 } } \\ \mbox{and}\nonumber\\ t_{\rho\rho } & = & \frac { m_1 ( 2\rho^2-z^2 ) } { ( \rho^2+z^2)^{5/2 } } - \frac{m_2 [ ( l - z)^2 - 2\rho^2]}{[\rho^2+(l - z)^2]^{5/2 } } \end{aligned}\ ] ] @xmath31 stretches along the cluster cluster axis , whereas @xmath32 stretches perpendicular to the cluster
cluster axis .
we normalize the tidal field components by dividing them by the strength of the tidal field of a milky way size halo with @xmath33 at a radius of @xmath34mpc .
we use a procedure similar to the one previously outlined to compute the alignments , as follows . to study @xmath32 , we take @xmath14 to be perpendicular to the cluster cluster axis and pointing towards the halo and @xmath15 along one of the principal axes of the halo as before . figure [ fig : tidal - align ] shows the alignments for the two tidal field components @xmath31 and @xmath32 . as before , the leftmost , center , and rightmost columns show the major , intermediate , and minor halo axes , respectively .
the two rows give the alignments along @xmath31 and @xmath32 . in each plot
, the dotted line shows the expectation for a random sample with no alignments .
the stretching perpendicular to the cluster
cluster axis is about twice that parallel to the cluster
cluster axis .
we note that for this analysis we do not differentiate between cluster pairs with filaments or voids in between .
this also results in smaller error bars , since the sample sizes are bigger .
the magnitude of this effect is smaller than alignments discussed in the previous sections .
however , unlike the alignment as a function of distance from the cluster - cluster axis , the alignment with the tidal field is a much more abrupt function of tidal field strength . from figure
[ fig : tidal - align ] we can see that when the tidal field is less than @xmath35 of the mw field the results are statistically consistent with no alignment .
the sudden change for larger values of the tidal field indicates the direct role that tidal distortions must play in aligning halos . cc
in the light of the results obtained so far , in particular the alignment signals and their dependence on the distance from cluster cluster axes , it is worthile to examine ellipticities of the halos .
figure [ fig : ellip - mass ] shows the ellipticities of the halos and clusters in our sample as a function of their mass .
the error bars assume poissonian distributions .
there is a trend for more massive clusters to be more elliptical than smaller halos , although the sample size is small .
this finding agrees with the results obtained by e.g. warren et al .
1992 .
figure [ fig : halo - fil - perpd - align ] shows that the alignment of halos with a cluster
cluster axis depends on the distance from that axis .
it is thus interesting to see whether the ellipticities are correlated with that distance as well .
figure [ fig : ellip - perpd ] shows halo and cluster ellipticities as a function of the distance from cluster cluster axes .
shown are all halos ( left column ) , all halos plus the clusters ( center column ) , and only the 170 clusters ( right column ) .
square and asterisk symbols denote groups associated with cluster pairs conncected by a filament or with a void in between them , respectively . from the figure
we see that within the errors there is no correlation between the ellipticities and the distance of the galaxy halos from from the cluster
cluster axes .
if tidal forces are responsible for the alignments , they do not cause a difference in the overall shape of halos .
furthermore , we also find that there is no difference between halos lying along the axes of cluster pairs connected by a filament and those cluster pairs which have a void in between them .
many studies of the alignments of galaxy clusters have been carried out both with observational data ( binggeli 1982 , struble & peebles 1985 , flin 1987 , rhee & katgert 1987 , ulmer et al . 1989 ,
west 1989 , rhee et al .
1992 , plionis 1994 , west et al .
1995 , chambers et al . 2000 and 2002 ) and theoretically ( splinter et al .
1997 , onuora & thomas 2000 , faltenbacher et al .
2002 , kasun & evrard 2004 , hopkins et al .
2005 , basilakos et al .
most observations and all theoretical studies indicate that neighbouring galaxy clusters indeed appear to be aligned , with some uncertainties due to the facts that shapes of clusters are notoriously hard to measure observationally , and that the simulations used for the theoretical studies follow the evolution of the dark matter , which may or may not trace the distribution of the galaxies used to define observational cluster shapes .
the causes of such an alignment are not clear . both infall of material ( van haarlem & van de weygaert 1993 ) and tidal fields ( bond et al .
1996 ; also see lee et al .
2005a , 2005b ) have been suggested as explanations . as shown in colberg et al .
( 1999 ) , the formation of clusters happens along filaments ( also compare the recent direct observations of this process for the @xmath36 cluster clj0152.71357 in maughan et al .
2005 and tanaka et al . 2005 ) .
if the infall of material causes alignments of clusters , then the presence of filaments is a prerequisite for alignments to exist .
we have investigated this scenario by studying the alignments of neighbouring clusters , separating cluster pairs into those that are connected by a filament and those that are not . for this study , we have made use of the filaments identified in colberg et al .
as the rightmost column of figure [ fig : halo - fil - perpd - align ] shows , there is a clear difference between the two cases .
clusters connected by a filament are clearly aligned , whereas unconnected ones are not .
on the basis of this result it appears the binggeli effect can be explained by the presence of filaments , along which material falls into the clusters .
note that this result provides another indicator of the likely presence of filaments .
as colberg et al .
( 2005 ) indicated , close pairs of clusters are candidates for the presence of a filament in between them .
the presence of alignment of the clusters adds another strong indicator
. however , given the difficulties involved when measuring cluster shapes , this theoretical prediction might not be much help in practice .
it must be stressed that the presence of filaments does not mean that tidal fields play no role whatsoever in the process of aligning the halos of galaxy clusters or of galaxies themselves . as can be seen from the leftmost column of figure [ fig : halo - fil - perpd - align ] , while filaments cause alignments of cluster size halos , they have no discernible influence on smaller halos . an alternative way to phrase
this result is to say that halos of galaxies being aligned has nothing to do with whether those halos are embedded in a filament or not .
there also is no dependence of the alignments of galaxy size halos on their distance from the nearest cluster , as shown in figure [ fig : halo - fil - projd - align ] .
a somewhat esoteric question which one can ask is whether the strong alignments of halos seen in simulations by e.g. heavens et al ( 2000 ) , croft & metzler ( 2000 ) are due to halos being formed largely in filaments and being aligned with the filament direction . in this case
one would expect halos to be aligned with each other as a result . in this study , we have found , however that halos are aligned with each other whether they fall in a filament or not , so that this explanation is not valid .
the tidal field around each galaxy sized halo has a much more direct impact on the alignments of the halos .
we have investigated the contribution of tidal fields by modeling the tidal field of a pair of clusters analytically as a simple model of two spherical masses .
figure [ fig : tidal - align ] shows the alignments for the two tidal field components @xmath31 and @xmath32 .
the stretching perpendicular to the cluster
cluster axis is about twice that parallel to the cluster
cluster axis .
the magnitude of this effect is smaller than that of the alignment through filaments discussed above , but the abruptness of the increase in alignment signal once tidal fields become substantial leads one to the conclusion that tidal fields have a very direct influence on halo alignments . in the light of the theoretical model of lee
et al . ( 2005a , 2005b ) , it is interesting to study whether there exists a correlation of the magnitudes of the ellipticities of halos with large - scale structure .
as we have already seen , the orientiations of the ellipticities are clearly influenced by tidal fields .
figure [ fig : ellip - perpd ] shows halo and cluster ellipticities as a function of the distance from cluster cluster axes .
there is no correlation between the ellipticities and the distance of the halos from the cluster cluster axes , whether filament or void .
furthermore , there is no difference between cluster pairs connected by a filament and those which have a void in between them .
we are thus left to conclude that while the alignment of clusters is dominated by the infall of matter along filaments , for the vast majority of halos only tidal fields determine alignments of halos . as an aside
, we note that this implies that pimbblet ( 2005 ) s filament finding mechanism which is based on using alignments to find filaments is unlikely to be successful .
this judgment is however based on the assumption that the stellar parts of galaxies will align in a similar way to the dark matter , something which must be tested .
this project is supported by the national science foundation , nsf ast-0205978 .
the simulation discussed here was carried out as part of the virgo consortium programme , on the cray t3d / es at the rechenzentrum of the max planck - gesellschaft in garching , germany , and at the edinburgh parallel computing center .
we are indebted to the virgo consortium for allowing us to use it for this work .
agustsson , i , & brainerd , t. , 2005 , apj submitted , astro - ph/0509405 allgood b. , flores r.a . , primack j.r .
, kravtsov a.v . ,
wechsler r.h .
, faltenbacher a. , bullock j.s . , astro - ph/0508497 bailin j. , steinmetz m. , 2004 , , 616 , 27 basilakos s. , plionis m. , yepes g. , gottlber s. , turchanino v. , astro - ph/0505620 binggeli b. , 1982 , a&a , 107 , 338 bond j.r . , kofman l. , pogosyan d. , 1996 , nature , 380 , 603 chambers s.w .
, melott a.l . , miller c.j . , 2000 , apj , 544 , 104 chambers s.w . ,
melott a.l . ,
miller c.j . , 2002 ,
apj , 565 , 849 colberg j.m . ,
white s.d.m . , jenkins a. , pearce f.r . , 1999 , mnras , 308 , 593 colberg j.m . ,
krughoff k.s .
, connolly a.j . , 2005 , mnras , 359 , 272 croft , r.a.c . , & metzler c. , 2000 , apj , 545 , 561 colless m. et al .
( 2df collaboration ) , 2001 , mnras , 328 , 1039 faltenbacher a. , gottlber s. , kerscher m. , mller v. , 2002 , a&a , 395 , 1 flin p. , 1987 ,
mnras , 228 , 941 heavens , a. , refregier , a. , heymans , c. , 2000 , mnras , 319 , 649 hopkins p.f . , bahcall n.a .
, bode p. , 2005
, , 618 , 1 jenkins a. et al .
( virgo consortium ) , 1998 , apj , 499 , 20 jing y.p .
, suto y. , 2002 , apj , 574 , 538 kasun s.f .
, evrard a.e . , 2005 ,
apj , 629 , 781 lee j. , kang x. , jing y.p . , 2005a , apj , 629 , 5 lee j. , jing y.p . , suto y. , 2005b , apj , 632 , 706 li , s.f . , & croft , r.a.c , 2005 , mnras , _
submitted_. mandelbaum , r. , hirata , c.m .
, ishak , m. , seljak , u. , & brinkmann , j. , 2005 , mnras submitted , astro - ph/0509026 maughan b.j . , ellis s.c .
, jones l.r . , mason k.o . , cordova f.a .
, priedhorsky w. , accepted for publication in apj , astro
ph/0511466 moore b. , quinn t. , governato f. , stadel j. , lake g. , 1999 , mnras , 310 , 1147 navarro j.f . , frenk c.s .
, white s.d.m . , 1997 , apj , 490 , 493 onuora l.i . , thomas p.a . , 2000 ,
mnras , 319 , 614 paz d.j .
, lambas d.g . , padilla n. , merchan m. , astro - ph/0509062 pimbblet k.a . , 2005 , mnras , 358 , 256 plionis m. , 1994 , apj , 95 , 401 plionis m. , basilikos s. , 2002 , mnras , 329 , l47 rhee g.f.r . , katgert p. , 1987 ,
a&a , 183 , 217 rhee g.f.r . , van harleem m. , katgert p. , 1992
, aj , 103 , 1721 seljak u. et al . , 2005 ,
phrvd , 71 , 10 , 103515 spergel d.n .
et al . , 2003 ,
apjs , 148 , 175 splinter r.j .
, melott a.l . , linn a.m. , buck c. , tinker j. , 1997 , apj , 479 , 632 springel v. et al .
( virgo consortium ) , 2005 , nature , 435 , 629 struble m.f . , peebles p.j.e .
, 1985 , aj , 90 , 582 tanaka m. , kodama t. , arimoto n. , tanaka i. , astro ph/0511296 ulmer m.p . ,
mcmillan s.l.w .
, kowalski m.p . , 1989 ,
apj , 338 , 711 van haarlem m. , van de weygaert r. , 1993 , 1993 , apj , 418 , 544 warren , m.s . , quinn , p.j . ,
salmon , j.k .
, & zurek , w. , 1992 , apj , 399 , 405 west m.j . , 1989 , apj , 347 , 610 west m.j . ,
jones c. , forman w. , 1995 , apj , 451 , l5 york d.g .
( sdss collaboration ) , 2000 , aj , 120 , 1579 | alignments of galaxy clusters ( the binggeli effect ) , as well as of galaxies themselves have long been studied both observationally and theoretically . here
we test the influence of large - scales structures and tidal fields on the shapes and alignments of cluster size and galaxy size dark matter halos .
we use a high resolution n body simulation of a @xmath0cdm universe , together with the results of colberg et al .
( 2005 ) , who identified filaments connecting pairs of clusters .
we find that cluster pairs connected by a filament are strongly aligned with the cluster - cluster axis , whereas unconnected ones are not . for smaller , galaxy size halos , there also is an alignment signal , but its strength is independent of whether the halo is part of an obvious large scale structure . additionally , we find no measureable dependence of galaxy halo shape on membership of a filament .
we also quantify the influence of tidal fields and find that these do correlate strongly with alignments of halos .
the alignments of most halos are thus caused by tidal fields , with cluster
size halos being strongly aligned through the added mechanism of infall of matter from filaments .
[ firstpage ] cosmology : observations large - scale structure of universe |
primarily through measurements in the optical , active galactic nuclei ( agns ) have been categorized as type 1 or 2 , respectively with and without very broad emission lines .
the unification model @xcite successfully explains this behavior in terms of a dusty circumnuclear torus that hides the broad - line region ( blr ) from our line of sight for type 2 agn .
it also appears that extinction in the host galaxy can hide the blr ( e.g. , @xcite ) in some active galaxies .
extinction therefore affects the optical properties of agn in fundamental ways .
these effects can be minimized by studying these objects using infrared ( ir ) emission lines , both to test the predictions of the unification model and to characterize agns in a uniform way . toward this end , we have used mid - ir observations with the infrared spectrograph ( irs , * ? ? ?
* ) on to examine the behavior of a large sample of agn .
important high - ionization lines accessible in the mid - ir include [ ] ( 97.1ev ) at 14.32 and 24.32 , [ ] ( 54.9ev ) at 25.89 , and [ ] ( 41ev ) at 15.56 .
because of their very high ionization potential , the [ ] lines are considered to be reliable signposts for an agn @xcite .
these lines have been used to estimate the accretion power in the local universe @xcite and to identify low - luminosity agns in local galaxies @xcite . however , they are also produced in supernova remnants ( snr , * ? ? ?
* ; * ? ? ?
* ) , planetary nebulae ( pn , * ? ? ?
* ) , and wolf - rayet stars ( wr , * ? ? ?
although the [ ] luminosities of these objects are low , @xmath5erg s@xmath6 @xcite , several thousands of them might also produce detectable [ ] emission . despite the lower ionization potential of [
] , the 25.89 emission line appears to be an accurate indicator of agn power , since it correlates well with the hard ( @xmath7 14kev ) x - ray luminosity @xcite and the mid - ir [ ] emission lines @xcite .
however , this line also appears in the spectra of starburst galaxies with no other evidence for agn @xcite , where it is attributed to wolf - rayet stars @xcite and/or to shocks @xcite .
the [ ] 15.56 line is excited by young , massive stars @xcite .
nonetheless , it seems to be a reasonably good proxy for agn luminosity , at least for reasonably high luminosities @xcite .
quantifying when its excitation is dominated by an active nucleus would allow it to be used in concert with the other high - ionization lines to probe conditions in the neighborhoods of agns . in this paper , we present a statistical study of the behavior of these lines in a sample of 426 galaxies .
the data are compiled from the literature on high spectral resolution ( @xmath8600 ) irs measurements , as well as our own new measurements of low - luminosity seyfert galaxies and luminous infrared galaxies ( lirgs ) .
we describe the sample and the data reduction respectively in sections 2 and 3 .
section 4 explores the correlations among the high - ionization emission lines . in section 5
, we discuss the star formation contribution to these lines .
section 6 compares the star formation in seyfert 1 and 2 galaxies .
finally , in section 7 we use photoionization models to study the physical conditions in the narrow line regions ( nlrs ) . throughout this paper
we assume a flat cosmology with @xmath9 km s@xmath6mpc@xmath6 , @xmath10 , and @xmath11 .
the sample contains 426 galaxies ( table [ tbl_fluxes_sample ] ) for which high spectral resolution ( @xmath8600 ) /irs spectra were available , either in the literature ( @xcite ; pereira - santaella et al .
2011 , in prep . ) , in the archive or observed through the programs 40936 and 50597 ( pi : g. h. rieke ) .
agn type classifications were taken from nasa extragalactic database ( ned ) .
we designated only pure cases as type 2 and included within type 1 all intermediate cases ( i.e. : 1.2 , 1.5 , 1.8 , 1.9 ) based on observations and modeling of the nuclear spectra energy distributions @xcite
. we could not find the nuclear activity classification for 32 of the galaxies .
the sample includes 28 qsos , 76 seyfert 1 , 125 seyfert 2 , 55 liners and 110 hii or starburst galaxies ( see table [ tbl_sample ] ) .
the luminosity of the active galaxies ranges from qso ( quest sample , @xcite ) to typical seyfert galaxies ( 12 sample , @xcite ) , relatively low - luminosity seyferts ( rsa sample , @xcite ) and liners @xcite . according to their infrared luminosities
71 sample members are classified as lirgs ( @xmath12 ) and 54 as ultraluminous infrared galaxies ( ulirgs , @xmath13 ) .
we found in the revised bright galaxy sample catalog @xcite the _ iras _ fluxes for 196 of our galaxies which we used to calculate their @xmath14 as defined in @xcite . note that the liner group includes very different galaxies .
ir - bright liners ( @xmath14/@xmath15 @xmath16 1 ) have infrared seds similar to starbursts , although high - ionization lines and hard x - ray cores are detected in some , suggesting the presence of an agn . in comparison , ir - faint liners ( @xmath14/@xmath15 @xmath17 1 )
seem to be powered by an agn @xcite .
figure [ fig_distance ] shows the distance distribution of the sample .
the median distance is 60 mpc and most ( 68% ) of the galaxies are between 15 and 400 mpc .
lccccl qso & 28 & 22 & 16 & 25 + seyfert 1 & 76 & 59 & 42 & 63 + seyfert 2 & 125 & 89 & 69 & 101 + liner & 55 & 10 & 6 & 27 + hii / starburst & 110 & 4 & 2 & 59 + unknown & 32 & 2 & 1 & 7 + total & 426 & 186 & 136 & 282 ccccccccc / & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + / & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + / & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + / & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + / & [email protected] & [email protected] & [email protected] & 28@xmath1814 & 57@xmath1842 + / & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + / & [email protected] & [email protected] & [email protected] & 6@xmath182 & 22@xmath1816 + / & [email protected] & [email protected] & [email protected] & [email protected] & [email protected]
in addition to results from the literature , we utilized the measurements for the 88 galaxies of the rsa sample observed with the high spectral resolution modules as well as for the 34 lirgs of the @xcite sample ( 3 of these lirgs are also members of the rsa sample ) .
we retrieved the basic calibrated data ( bcd ) from the archive processed by the pipeline version s18.7 .
we subtracted the background contribution when a dedicated sky observation was available .
note however that the sky subtraction is not important to measure these fine structure emission lines and it just improves the final quality of the spectra by removing bad pixels .
then we extracted the spectra using the standard programs included in the irs custom extraction ( spice ) package provided by the science center ( ssc ) .
we assumed the point source calibration for all the galaxies , which is a good approximation for most of the galaxies at least for the high ionization lines .
for the galaxies observed in the mapping mode we extracted the nuclear spectra from the data cubes using a square aperture of 134@xmath19134 and then we applied an aperture correction ( pereira - santaella et al .
2011 , in prep ) .
we used a gaussian profile to fit the emission lines . for 9 galaxies
( ngc 777 , ngc 3254 , ngc 3486 , ngc 3941 , ngc 4138 , ngc 4378 , ngc 4472 , ngc 4698 , ngc 5631 ) we were not able to measure any spectral feature due to the low signal - to - noise ratio of the spectra .
they are not listed in table [ tbl_fluxes_sample ] nor included in the analyzed sample .
table [ tbl_sample ] includes the number of detections of the , and emission lines and in table [ tbl_ratios ] we show the observed median line ratios for each galaxy type both for galaxies from literature and those analyzed by us .
note that there are 4 detections of the line and 2 of the line in galaxies classified as hii .
for these galaxies ( ngc 613 , ngc 1792 , ngc 3621 and ngc 5734 ) the detection of the [ ] lines is the only evidence of agn activity .
we decided to keep their hii classification because : ( 1 ) the agn may be extremely obscured or very low luminosity and , thus , the nuclear spectra might be dominated by star formation features ; and ( 2 ) these lines are also detected in snr , pn and wr stars which could be the origin of the [ ] emission in these galaxies .
table [ tbl_ratios ] gives the median and deviation of all the line ratios we discuss in this paper for each type of galaxy .
the line fluxes for all the galaxies are listed in table [ tbl_fluxes_sample ] .
the and the emission lines are commonly detected in active galaxies @xcite .
the detection of the former is considered a good indicator of agn activity because of its high ionization potential . however using the line as an agn tracer is not as straightforward . for optically classified seyfert galaxies
the luminosity is a good proxy for the agn intrinsic luminosity @xcite and there is a good correlation between the and the emission in agns @xcite .
however , the line is also produced by wr stars @xcite and is observed in starburst galaxies with no other evidence for the presence of an agn @xcite .
the left panel of figure [ fig_ne5o4 ] shows the tight correlation between the and lines spanning at least 5 orders of magnitude in luminosity ( 38 @xmath0 log @xmath20}$](erg s@xmath6 ) @xmath0 43 ) .
we do not find any dependence between the optical classification of the nuclear activity and this correlation .
this implies that the ionization parameter is similar in type 1 and type 2 seyfert galaxies ( see section [ s : models ] ) .
the slope of the best fit to the luminosity data is 0.96 @xmath18 0.02 with a 0.14dex dispersion .
it is reasonable to assume a linear correlation between these two emission lines , and this yields a ratio / @xmath213.5 , with rms scatter of 0.8 .
the line is detected in @xmath2250% of the optically classified seyfert galaxies ( see table [ tbl_sample ] ) .
however , for a considerable number of galaxies we have the measurement and an upper limit for the flux .
we plot these fluxes and upper limits ( middle panel of figure [ fig_ne5o4 ] ) for all the galaxies .
the upper limits to the flux are compatible with the correlation for most of the seyfert galaxies .
however , the line is detected in more than 50% of the hii galaxies in our sample . in 90% of these galaxies ,
the sensitivity of the measurement ( or upper limits ) is inadequate to probe whether the [ ] emission is associated with a hidden agn .
in short , we find that the luminosity is well correlated with the intrinsic agn luminosity ( i.e. , [ ] luminosity ) for seyfert galaxies and quasars with @xmath20 } > 10^{39}$]erg s@xmath6 to , at least , @xmath20 } \sim 10^{43}$]erg s@xmath6 . for lower luminosities
the fraction of the emission produced by star formation may be considerable .
thus the may not be an accurate tracer of the agn luminosity for galaxies with @xmath20 } < 10^{39}$]erg s@xmath6@xcite .
the linear correlation also holds for the and the luminosities ( figure [ fig_ne5o4_2 ] ) , although the dispersion is larger ( / @xmath213.4 , , with rms scatter of 1.4 assuming a linear fit ) .
.,scaledwidth=45.0% ] due to the intermediate ionization potential of ne@xmath23 ( 41.0ev ) , the emission line may be produced by young stars or by agns .
this line is observed in star forming galaxies @xcite , and it is also correlated with the agn intrinsic luminosity @xcite . the middle panel of figure [ fig_ne5ne3 ] shows the relation between the observed flux of the and emission lines for our sample of agns and for the 4 hii galaxies with [ ] detections .
we compare these lines because they are close in wavelength and the effect of differential extinction is minimized .
it is clear that some galaxies have an excess of emission relative to that of the , which we attribute to star formation ( see section [ s : ne3sf ] ) .
therefore to obtain the fit to the data we use the outliers resistant linear fit algorithm provided by the idl function robust_linefit .
the slope of this fit is [email protected] with a scatter of about 0.2 dex around the fit .
the result is consistent with a simple correlation ; by fixing the slope to unity , we obtain the ratio / @xmath211.9 , , with rms scatter of 0.8 . this correlation was previously reported by @xcite , although they found a slight dependence of the ratio with the luminosity ( slope 0.89 ) .
their sample included 53 x - ray selected agns with luminosities between 10@xmath24 and 10@xmath25erg s@xmath6 .
the larger number of galaxies in our sample allows us to minimize the contribution of the galaxies contaminated by star formation to the fit and the luminosity dependence of the correlation is substantially reduced .
where the agn clearly dominates , the points fall around the line determined for the whole sample by robust_linefit , thus supporting our interpretation that this fit determines the relationship for agns in general .
the emission traces young ( @xmath010myr ) star formation @xcite .
the ionization potential of this emission line is 21ev and thus it is mainly produced in hii regions .
figure [ fig_ne5ne2 ] shows that the scatter in the / ratio ( 0.7dex ) is larger than that found in the other line ratios , for the lower agn luminosities there is a large scatter .
there is a weak trend of / ratios decreasing with increasing agn luminosity . assuming that there is an intrinsic ratio for agn
, this behavior implies that the agn does not dominate the output and that there is a wide range of star formation rates contributing to its luminosity .
however , for the most luminous objects ( @xmath2614\mu m}~({\rm erg~s^{-1 } } ) > $ ] 41.5 ) , the smaller scatter indicates that the relative contribution of the agn to the total excitation is higher .
@xcite determined that the pure agn / ratio is 1.1 , but we find many seyfert galaxies with higher ratios .
the largest / ratio in our sample is @xmath225.5 , which is compatible with the largest ratio predicted by nlr photoionization models ( @xcite , their figure 11 ) . with our data ( figure [ fig_ne5ne2 ] )
we are not able to determine the pure agn ratio , which also depends on the ionization parameter @xcite , but it is likely that it is between that found by @xcite and the largest ratio in our sample . the / ratio can be used to separate agns from star - forming galaxies @xcite .
we plot this ratio versus the luminosity in figure [ fig_o4ne2 ] .
most of the active galaxies have ratios larger than 0.35 whereas the ratios for hii galaxies are lower than 0.05 ( figure 5 of @xcite ) .
liners appear in the region between these ratios ( figure [ fig_o4ne2 ] ) . a considerable number of hii galaxies also have / ratios above 0.05 .
the [ ] lines are not detected for them although their upper limits are compatible with the versus [ ] correlations and thus a low - luminosity agn could be the origin of the emission . the large scatter in the / ratio , due to the different star formation rates in the agns , does not allow us to determine the value of the pure agn ratio .
the largest / ratios that we find in agns are @xmath2210 . .
the dashed orange ( [ ] / [ ] @xmath27 ) and green ( [ ] / [ ] @xmath28 ) lines are the boundaries of the hii and seyfert galaxies respectively.,scaledwidth=45.0% ]
in figures [ fig_ne5o4 ] and [ fig_ne5o4_2 ] there are few outliers above the versus [ ] correlations .
most of them are represented by open symbols , which indicate that their / ratios ( see section [ ss : ne3sf_estimation ] ) are larger than those found in agn dominated galaxies ( qso , sy1 and sy2 ) .
it is likely that the emission of these galaxies is contaminated by star formation .
we now compare the emission of star forming regions and agn .
the / ratio in star forming regions is @xmath0 0.05 ( section [ s : ne3_agn ] ) . taking this upper limit and using the relation between the luminosity and the total infrared ( 8 - 1000 ) luminosity ( @xmath14 ) from @xcite we obtain the following relation for star - forming galaxies .
@xmath29}~({\rm erg~s^{-1 } } ) < \log l_{\rm ir}~({\rm erg~s^{-1 } } ) - 4.7 \pm 0.6\ ] ] @xcite calculated the ratio between the total agn luminosity and the emission .
we use the ratio for type 1 agns ( @xmath222500 ) because the hard x - ray ( e @xmath7 10kev ) emission used for the bolometric corrections may be affected by extinction in seyfert 2s @xcite .
@xmath30}~({\rm erg~s^{-1 } } ) = \log l_{\rm agn}~({\rm erg~s^{-1 } } ) - 3.4 \pm 0.4\ ] ] then , combining both relations we find that the luminosity due to star formation , @xmath14 , has to be at least 20 times brighter than the agn luminosity ( that is , agn contribution to the total luminosity below 5% ) for the star formation to dominate the emission .
we note that the number of hii galaxies falls rapidly above log @xmath20}$](erg s@xmath6 ) @xmath21 40.2 ( see figure [ fig_o4ne2 ] ) . assuming that star formation dominates the emission at this luminosity in these galaxies , the equivalent infrared luminosity ( @xmath14 ) would be 2@xmath1910@xmath31@xmath32 .
this value of the @xmath14 seems reasonable since most of the hii galaxies with log @xmath20}$](erg s@xmath6 ) between 39.6 and 41.0 are classified as lirgs .
the @xmath14 limit calculated above is one order of magnitude larger than @xmath33 @xcite so the reason for the lower number of hii galaxies above this luminosity can be that the space density of luminous star forming galaxies decreases rapidly with increasing luminosity . ratio vs. @xmath14 luminosity .
galaxy symbols are as in figure [ fig_ne5o4 ] .
the shaded region is the upper limit to the /@xmath14 ratio of star - forming galaxies .
galaxies above this region need an agn contribution to the emission in order to explain the observed ratio .
, scaledwidth=45.0% ] out of 196 galaxies with estimates for @xmath14 , is detected in 122 .
only 2@xmath182% of the hii galaxies have luminosities above those expected from their @xmath14 .
likewise star - formation can explain the observed luminosities for 9@xmath184% of the seyfert galaxies ( figure [ fig_lo4_lir ] ) .
thus the emission of seyferts is generally dominated by the agn , whereas for optically classified hii galaxies star formation is likely to be the origin of the emission .
the emission may be produced by star formation and/or agn . in this section
we try to quantify the contribution of each one to the total output . for this purpose
we used the correlations between , and ( see section [ s : o4ne5 ] ) . combining the typical / and / ratios found in agn we estimate a ratio / @xmath21 0.6 , with rms scatter of 0.3 for agns , or log [ ] / [ ] @xmath21 .
this estimated ratio is very similar to that obtained using a linear fit between the and luminosities ( table [ tbl_ratios ] ) .
the median of this ratio for hii galaxies is 4.7 , or log [ ] / [ ] @xmath21 0.67 , which is @xmath228 times larger than that estimated for active galaxies .
figure [ fig_lo4_o4ne3 ] shows that the observed ratio for seyfert galaxies and qsos falls around the predicted value for agn . only a small fraction ( 4% ) of the hii galaxies have / ratios in the agn range .
it is also apparent that hii galaxies , a large fraction of liners and some seyfert galaxies have an excess of relative to their emission which can also be attributed to star formation .
we find that the four hii galaxies with detections have the / and / ratios much larger than those of seyfert galaxies ( table [ tbl_ratios ] ) .
this indicates that the star - formation contribution is larger than that in seyfert galaxies .
thus as suggested in section [ s : sample ] the agn does not dominate the nuclear spectra of these galaxies . as can be seen in the left panel of figure [ fig_ne5ne3 ] , most of the outliers ( galaxies more than 3@xmath34 above the versus flux correlation ) have @xmath35 } < 10^{42}$]erg s@xmath6 .
that is , for the most luminous objects the agn dominates the emission . using the relation by @xcite that relates star formation rate ( sfr ) and the luminosity of the and emission lines and assuming a / ratio @xmath36 ( see section [ ss : ne3sf_ratio ] )
we find that the sfr of the outliers is between 0.1 and 100@xmath37 yr@xmath6 . .
the solid red line is the / ratio calculated from the linear fit to the vs. luminosities .
the dotted black lines are the 1@xmath34 deviation .
, scaledwidth=45.0% ] the / ratio traces the hardness of the radiation field and the age of the stellar population @xcite .
figure [ fig_ne3ne2_histo ] compares the / ratio observed in hii galaxies ( median 0.2 ) with that observed in seyfert galaxies with star formation ( median 0.6 ) .
the larger ratio found in seyfert galaxies can be explained if the agn contributes noticeably to the total emission but not to the emission ( see section [ s : ne3_agn ] ) . by assuming this we calculate the star formation / ratio in these seyfert galaxies using the non - agn emission that we estimated using the method described in section [ s : ne3sf_sy1sy2 ] .
the median of this ratio becomes @xmath220.3 , which is somewhat larger than that found in pure hii galaxies but it is in the range of the observed ratio in this class of galaxies @xcite .
the blue histogram shows the ratios calculated for seyfert galaxies after subtracting the agn contribution to the emission.,scaledwidth=45.0% ]
the agn unification scenario predicts that there should be no differences between the star formation activity in seyfert 1 and 2 galaxies .
a number of studies compared the star formation rates in seyfert 1 and 2 galaxies .
some of them found enhanced star formation activity in seyfert 2 with respect seyfert 1 galaxies , for a given agn luminosity @xcite , whereas others did not find any differences between the two types @xcite .
however , as discussed by @xcite , the different methods used to study the star formation activity are sensitive to different stellar age ranges .
therefore these apparently contradictory results may be consistent with each other .
the line is sensitive to young ( @xmath05myr ) massive star formation @xcite .
studies using this line ( and the line ) found that star formation is enhanced in seyfert 2 galaxies @xcite .
@xcite did not find differences between seyfert 1 and seyfert 2 star formation , but their method ( the 4000 break and the h@xmath38 stellar absorption ) was sensitive to stellar populations older than 100 myr @xcite .
the same applies to the pah features used by @xcite , which can be affected by the presence of the agn ( * ? ? ?
* and references therein ) .
it is also important to use the appropriate indicator for the agn luminosity .
we will use the emission , as it is a good isotropic indicator @xcite , as long as the agn dominates the emission as we showed in section [ ss : o4contamination ] .
we estimate the star formation contribution to the emission by using the observed / ratio and subtracting the estimated agn contribution from the total line strength .
further details are provided in appendix [ apxcontribution ] .
we used the / instead of the / ratio because we could not calculate the typical agn / ratio due to the larger relative contribution of star formation activity to the emission ( section [ s : ne3_agn ] ) .
we only subtract this contribution from those active galaxies more than 1@xmath34 above the typical agn / ratio ( see section [ ss : ne3sf_estimation ] and figure [ fig_lo4_o4ne3 ] ) , since the galaxies below this ratio are presumably dominated by the agn and their star formation activity , if any , is masked .
this criterion basically selects only those active galaxies with at least 25% of the emission arising from star formation . in our sample 45% of the agns with
the line detected have excess emission due to star formation .
the right panel of figure [ fig_comparison_sy1sy2_all ] shows that the fraction of seyfert 2s with star formation is slightly higher than that of seyfert 1s . in all the luminosity bins except for one ,
the difference in the fraction with a excess is less than 2@xmath34 significant .
however , for the overall proportion of excess detections , the fisher s exact test indicates that there is a probability of less than 0.03 that the incidence of this behavior is the same in the sy1 and sy2 samples .
therefore , the full sample in our study supports , with a moderate statistical significance , previous indications that sy2 host galaxies tend to have higher rates of star formation than sy1 hosts .
however the full sample is heterogeneous and may be subject to a number of biases .
for example , it includes a number of radio galaxies , which are known to have low star formation rates in fact none of the radio galaxies have excess .
it also contains ulirgs with seyfert spectral characteristics , most of which do have excess .
the observed differences depend to some extent on the relative number of galaxies in these two classes . in the following section
, we discuss this trend in two relatively complete but smaller subsamples where the biases should be reduced . at this point
we study separately the 12 sample and the rsa sample .
both are complete samples of seyfert galaxies selected with homogeneous criteria .
the former is selected based on their _ iras _ 12 fluxes whereas the latter is selected based on the optical magnitude of the host galaxy .
given the different selection criteria they they might be affected by star formation in different ways . as can be seen from table [ tbl_comparison_samples ] and figure [ fig_comparison_rsa12 ] ,
the fraction of galaxies with a excess in the rsa sample is higher ( by a factor of 1.4 ) than in the 12 sample .
the lower number of galaxies with a excess in the 12 sample is because the galaxies of this sample are brighter and the sensitivity to star formation of our method is reduced at high luminosities .
both effects are clearly seen in figure [ fig_comparison_rsa12 ] .
first , the rsa sample contains a larger fraction of low luminosity seyferts ( @xmath39 } \le 10^{40}$]erg s@xmath6 ) than the 12 sample .
second , for each agn luminosity , we are sensitive to a different sfr limit . in other words , for the brightest seyferts we are only sensitive to sfr above 100@xmath40 yr@xmath6 , while for low luminosity agn we are sensitive to sfr @xmath41 yr@xmath6 .
we note that the fraction of galaxies with star formation in each luminosity bin is similar for both samples ( right panel of figure [ fig_comparison_rsa12 ] ) . in both samples
the fraction of seyfert 2s with excess is slightly larger than that of seyfert 1s .
however due to the small size of the samples the statistical significance of this excess is low .
based on the two sample k - s test , the probability for these differences being due to chance is 0.25 and 0.4 for the 12 and rsa samples respectively .
lcccc seyfert 1 & 15/45 & 12/25 + seyfert 2 & 23/52 & 25/45 + total & 38/97 & 37/70
we studied the physical conditions in the narrow line region of the agns with the photoionization code mappingsiii @xcite .
we used a radiation pressure dominated model that includes the effects of dust . for the input parameters we followed the prescription given by @xcite .
briefly , we assumed a plane parallel geometry and solar abundances .
we modeled the input ionizing spectrum with two power - laws with exponential cut - offs @xcite .
we explored the effect of the variations in the total hydrogen column density ranging from @xmath1 @xmath3 ( @xmath4 ) = 19.0 to 22.0 .
this range was chosen since it reproduces the observed line ratios and is in good agreement with the hydrogen column density determined using uv and x - ray observations @xcite .
for each hydrogen column density we varied the total pressure ( @xmath42_k _ ) and the incident ionizing flux ( @xmath43 ) .
the values for the total pressure are @xmath1 @xmath44/_k_(k @xmath45 ) = 6 , 7 , 8 , 9 .
these correspond to electron densities , as traced by [ ] @xmath466716/@xmath466731 , of @xmath22 @xmath010@xmath47 , 10@xmath48 , 10@xmath49 and @xmath7 10@xmath50@xmath45 respectively .
the explored range for the ionizing flux is @xmath43 = 0.1 to 0.55 .
the incident ionizing flux is scaled by the factor 2.416@xmath19((@xmath44/_k_)/10@xmath51)erg @xmath4 s@xmath6 which gives a range in the ionization parameter , @xmath1 @xmath2 , from @xmath22 -3 to -2 .
figure [ fig_hardness ] shows the ratio between the two [ ] emission lines available in the mid - infrared .
the critical densities of the and lines are 3.5@xmath1910@xmath50@xmath45 and 6.2@xmath1910@xmath49@xmath45 , respectively @xcite , thus they can be used to trace the density of the nlr .
we find however that a large fraction of the galaxies ( @xmath22 30% ) lies above the low density limit ratio .
this issue and its possible cause are discussed in detail by @xcite and also addressed by @xcite for a sample of seyfert galaxies .
@xcite concluded that the ratios above the low density limit are due to differential extinction from the obscuring torus , and thus can indicate the inclination angle to our line of sight .
however , similar to @xcite , we do not find any significant difference between the proportion of seyfert 1 and seyfert 2 galaxies above the low density limit , as would be predicted by this hypothesis .
we calculated the @xmath52/@xmath53 ratio to quantify the extinction needed to move the points above the low density limit ratio to this value ( which would be a lower limit to the extinction since the real ratio may be lower than the low density limit ) .
the ratios are summarized in table [ tbl_extinction ] for some infrared extinction laws .
as can be seen depending on the extinction law chosen the extinction can increase or decrease the / ratio . since the construction of an extinction law implies some interpolation we also used direct measurements of @xmath54/@xmath55 and @xmath56/@xmath55 .
values and those at 14.3 and 24.3 would be small ( @xmath225% ) . ] @xcite measured the extinction at 15 and found @xmath56/@xmath55 @xmath21 0.40 , although it ranges from 0.25 to 0.55 . for the @xmath54/@xmath55 ratio the range is 0.28 to 0.65 for different @xmath55 bins @xcite and the average value is @xmath220.5 @xcite . from these measurements we estimate @xmath52/@xmath53 @xmath21 [email protected] .
again it is close to unity and it is not clear if the extinction would increase or decrease the [ ] ratio .
moreover the extinction is nearly neutral between the lines with all the extinction laws and therefore there is very little potential effect on the ratio unless the extinction is extremely large .
lc rosenthal & 1.25 + mcclure 0.3 @xmath0 @xmath55 @xmath0 1 & 0.83 + mcclure 1 @xmath0 @xmath55 @xmath0 7 & 0.83 + chiar galactic center & 1.38 + chiar local ism & 0.97 we also checked if aperture effects can explain the large values of this ratio .
the line is observed with the sh module ( 57@xmath19113 ) , whereas the is observed with the lh module ( 111@xmath19223 ) .
we would expect a correlation between the / ratio and the distance if the [ ] emission region was larger than the sh slit .
however we do not find this correlation thus we rule out any aperture bias . alternatively the atomic parameters for the ne@xmath57 ion may not be sufficiently accurate for the comparison with these observations and the calculated low density limit may be incorrect . .
the dashed black line indicates the low density limit calculated at 10000 k for the / ratio ( galaxies above this line have densities below the low density limit ) .
the blue lines trace constant pressure and the orange lines constant ionization parameter ( @xmath43 ) .
the input parameters for the model are @xmath1 @xmath44/_k _ = 6 , 7 , 8 , 9 and @xmath43 = 0.1 , 0.15 , 0.20 , 0.25 , 0.32 , 0.4 , 0.55 for constant hydrogen column density @xmath1 @xmath3 ( @xmath4 ) = 20.3 .
the galaxy symbols are as in figure [ fig_ne5o4].,scaledwidth=45.0% ] in figure [ fig_hardness ] we plot the model grid together with the data . in this plot
we vary the ionizing flux , @xmath43 , and the total pressure , @xmath42_k _ , for a constant total hydrogen column density ( @xmath1 @xmath3(@xmath4 ) = 20.3 ) .
we see that the [ ] ratio can be used to measure the pressure .
however the pressure ( or density ) values derived from this ratio have some issues , since a considerable fraction of galaxies are above the low density limit ratio ( section [ ss : ne5ratio ] ) .
also , figure [ fig_hardness ] shows that the / ratio traces the ionization parameter .
the / ratio spans a narrow range ( less than 1 dex ) which is related to the tight correlation found between the and luminosities ( section [ s : o4ne5 ] ) .
this implies a very small range for the ionization parameter ( -2.8 @xmath58 -2.5 ) .
this range is similar to that found when modeling the optical [ ] and h@xmath59 emission @xcite .
the critical density of the [ ] lines is larger than that of these mid - ir lines and for instance the @xcite models predict electron densities @xmath60(@xmath45 ) @xmath21 [email protected] well above the critical density of the mid - ir lines studied here .
thus it is possible that these lines are produced in different gas clouds .
we use the and the lines to estimate the total hydrogen ( @xmath61 ) column density , @xmath3 .
the line is produced in the inner part of the nlr while the is produced in a more external region .
thus the / ratio is subject to change depending on the column density . as expected , the models predict larger / ratios for larger @xmath3 .
in figure [ fig_columndesnity ] we plot the model grid together with the observed ratios .
the / ratio is not sensitive to column densities larger than @xmath1 @xmath3(@xmath4 ) @xmath7 21 .
we find that the column density for most of the galaxies is in the range 20.3 @xmath0 @xmath1 @xmath3(@xmath4 ) @xmath0 21 .
there is no difference between the predicted hydrogen column density for type 1 and 2 galaxies .
this column density range is compatible with the values derived from uv observations although lower than that determined using x - ray data @xcite .
a few galaxies are above the model grid .
the / ratio of these galaxies is larger than that predicted by the models .
as we proposed in sections [ s : ne3_agn ] and [ ss : ne3sf_estimation ] , star formation acting in these galaxies is likely producing the extra emission .
@xmath3(@xmath4 ) = 19.0 , 20.0 , 20.3 , 20.6 , 21.0 , 22.0 and @xmath43 = 0.1 , 0.15 , 0.20 , 0.25 , 0.32 , 0.4 , 0.55 for constant pressure ( @xmath1 @xmath44/_k _ = 8) .
the galaxy symbols are as in figure [ fig_ne5o4].,scaledwidth=45.0% ]
we studied a sample of 426 galaxies observed with the /irs spectrograph in the high resolution mode .
our analysis includes published data for qso , seyfert , liner and hii galaxies as well as unpublished measurements for the rsa seyfert sample and lirgs .
we explored the relationship between the high - ionization ( , and ) and intermediate ionization ( and ) emission lines present in the mid - ir spectra .
the median ratios for each class of galaxies are listed in table [ tbl_ratios ] .
the main results are as follows : 1 .
there is a tight linear correlation between the fluxes and luminosities of the high - ionization emission lines and .
this correlation spans 5 orders of magnitude ( 38 @xmath0 @xmath1 @xmath6224}$](erg s@xmath6 ) @xmath0 43 ) , ranging from liners , low - luminosity seyfert galaxies , types 1 and 2 , to qsos .
the typical / ratio for agns is 3.5 , with rms scatter of 0.8 .
the correlation also holds between the and the lines , although the scatter is larger .
the typical / ratio for agns is 3.4 , with rms scatter of 1.4 . 2 .
there is also a good linear correlation between the and the fluxes and luminosities for seyfert galaxies and
the / ratio is 1.9 , with rms scatter of 0.8 . using this correlation and the previous one we calculated the typical / ratio for agns of 0.6 , with rms scatter of 0.3 .
we calculate the emission due to star formation ( @xmath63}~({\rm erg~s^{-1 } } ) < \log l_{\rm ir}~({\rm erg~s^{-1 } } ) - 4.7 \pm 0.6 $ ] ) and we estimate that it may dominate the total emission when the intrinsic agn luminosity is a factor of 20 smaller than the star forming luminosity . in general , we find that the emission is dominated by the agn in seyfert galaxies , whereas star - formation can explain the emission of optically classified hii galaxies
we do not find any significant difference between the mid - ir high - ionization emission lines in type 1 and in type 2 seyfert galaxies .
either there are no differences in the conditions in the nlr of the two agn types , or the effects associated with the different line of sight ( i.e. , dust extinction ) are minimized in the mid - infrared spectra .
we find that a significant number of seyfert galaxies from the 12 sample ( 30% ) and the rsa sample ( 40% ) show evidence for excess emission relative to their emission associated with star formation .
the larger fraction in the rsa sample is explained because the rsa sample contains a larger fraction of low - luminosity agns ( @xmath20}$ ] @xmath0
10@xmath64erg s@xmath6 ) in which these star formation excesses are easier to detect . 6 .
the fraction of seyfert 2 galaxies with excess is larger than that of seyfert 1 with a moderate statistical signicance in the full sample although this sample may be affected by several bias .
however when we consider the 12 or rsa seyfert samples separately these differences are not statistically significant due to the smaller size of these samples .
7 . a considerable fraction ( 30% ) of the galaxies have / ratios above the low density limit .
we did not find a connection between the seyfert type and the ratio of galaxies above the limit .
we are not able to explain this in terms of differential extinction .
our modeling shows that the nebular conditions in the nlrs are remarkably similar among all the agns in our sample .
this similarity allows us to compare conditions critically in the nlrs of the type 1 and type 2 galaxies .
there appear to be no significant overall differences , consistent with the unified model .
we constrained the ionization parameter in the range -2.8 @xmath0 @xmath1 @xmath2 @xmath0 -2.5 and the hydrogen column density 20 @xmath0 @xmath1 @xmath3(@xmath4 ) @xmath0 21 .
the relationships presented in this paper provide an important benchmark for the interpretation of the future mid - ir observations of agns and star forming galaxies with jwst / miri .
the authors thank b. groves and t. daz - santos for their help and enlightening discussion .
we thank the anonymous referee for useful comments and suggestions .
mp - s acknowledges support from the csic under grant jae - predoc-2007 .
mp - s also thanks the steward observatory for their hospitality during his stay while part of this work was done .
aa - h and mp - s acknowledge support from the spanish plan nacional del espacio under grant esp2007 - 65475-c02 - 01 .
this work was partially supported by caltech / jpl through contract 1255094 to university of arizona .
aa - h also acknowledges support for this work from the spanish ministry of science and innovation through proyecto intramural especial under grant number 200850i003 and from plan
nacional de astronomia y astrofisica under grant number aya2009 - 05705-e .
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 .
this research also used the vizier catalogue service @xcite .
in this appendix we explain briefly the method used in sections [ s : ne3sf ] and [ s : ne3sf_sy1sy2 ] to calculate the star formation fraction of the emission . if we assume that all the is produced by the agn we can use the typical / ratio observed in seyfert galaxies with low star formation ( figure [ fig_lo4_o4ne3 ] ) to estimate the amount of emission coming from the agn .
@xmath65 = \rm [ ne~{\scriptstyle iii}]_{sf } + \rm [ o~{\scriptstyle iv}]\times \left ( \frac{\rm [ ne~{\scriptstyle iii}]}{[\rm o~{\scriptstyle iv } ] } \right)_{agn}\ ] ] finally we obtain the fraction of from star formation . @xmath66_{sf}}{\rm
[ ne~{\scriptstyle iii } ] } = 1 - \frac{\rm [ o~{\scriptstyle iv}]}{\rm [ ne~{\scriptstyle iii } ] } \times \left ( \frac{\rm [ ne~{\scriptstyle iii}]}{\rm [ o~{\scriptstyle iv } ] } \right)_{agn}\ ] ] this method can be applied as well to other line ratios .
the only assumption is that one of the lines in the ratio is uniquely produced by the agn while star formation and the agn contributes to the other line ( e.g. the and [ ] lines ) .
also note that this method makes use of a typical agn ratio thus the estimated star formation contribution to an emission line for a single object will be uncertain and dependent on the ionization paremeter .
mrk 335 @xmath67 & 00 06 19.5 & + 20 12 10 & 113 & sy1 & 0.25 & 0.38 & 0.61 & 2.0 & 7.2 & 4 + ngc 23 & 00 09 53.4 & + 25 55 25 & 64.5 & & 96 & @xmath00.50 & 13 & @xmath00.44 & 1.4 & 2 + ngc 24 & 00 09 56.5 & 24 57 47 & 7.30 & & 3.1 & @xmath00.88 & 1.3 & & @xmath00.42 & 5 + mrk 938 @xmath67 & 00 11 06.5 & 12 06 26 & 85.3 & sy2 & 52 & @xmath02.2 & 6.4 & @xmath00.37 & @xmath00.66 & 4 + iras f001880856 & 00 21 26.5 & 08 39 26 & 600 & liner & 4.7 & @xmath00.18 & 0.69 & @xmath01.6 & @xmath00.90 & 3 + iras 001987926 @xmath67 & 00 21 53.6 & 79 10 07 & 329 & sy2 & 6.2 & 12 & 14 & 11 & 33 & 4
+ pg 0026 + 129 & 00 29 13.6 & + 13 16 03 & 672 & qso & 0.23 & 0.47 & 0.79 & @xmath00.33 & 2.1 & 1 + eso 012-g021 @xmath67 & 00 40 46.2 & 79 14 24 & 144 & sy1 & 12 & 3.2 & 6.4 & 4.6 & 16 & 4 + iras f003971312 & 00 42 15.5 & 12 56 02 & 1329 & & 4.4 & @xmath00.20 & 2.7 & @xmath01.5 & @xmath01.2 & 3 + ngc 253 & 00 47 33.1 & 25 17 17 & 2.50 & & 2832 & @xmath021 & 205 & @xmath073 & 155 & 8 + mrk 348 @xmath67 & 00 48 47.1 & + 31 57 25 & 65.1 & sy2 & 16 & 5.8 & 20 & 4.9 & 18 & 4 + ngc 278 & 00 52 04.3 & + 47 33 01 & 11.4 & & 18 & @xmath00.070 & 3.0 & @xmath00.59 & @xmath00.62 & 7 + pg 0050 + 124 & 00 53 34.9 & + 12 41 36 & 273 & qso & 1.9 & 5.5 & 4.5 & @xmath02.1 & 2.7 & 1 + iras 005217054 @xmath67 & 00 53 56.1 & 70 38 04 & 310 & sy2 & 5.8 & 5.8 & 8.1 & 2.4 & 8.6 & 4 + mcg + 12 - 02 - 001 & 00 54 03.6 & + 73 05 11 & 68.1 & & 242 & @xmath01.00 & 43 & @xmath01.4 & 3.8 & 2 + ugc 556 & 00 54 50.3 & + 29 14 47 & 66.1 & & 40 & @xmath00.21 & 7.1 & @xmath00.73 & @xmath01.6 & 2 + ngc 337 & 00 59 50.1 & 07 34 40 & 22.4 & & 19 & @xmath00.74 & 8.0 & & @xmath00.49 & 5 + eso 541-ig012 @xmath67 & 01 02 17.5 & 19 40 08 & 253 & sy2 & 1.9 & 2.2 & 2.0 & 1.2 & 5.0 & 4 + iras 010032238 & 01 02 50.0 & 22 21 57 & 550 & & 3.1 & @xmath00.30 & 1.3 & @xmath00.30 & @xmath00.30 & 3 + 3c 33 & 01 08 52.9 & + 13 20 13 & 269 & sy2 & 3.9 & 2.0 & 5.3 & & 8.1 & 6 + ngc 404 & 01 09 27.0 & + 35 43 04 & 3.24 & liner & 3.1 & @xmath00.16 & 1.7 & @xmath00.31 & 0.90 & 2 + ngc 424 @xmath67 & 01 11 27.6 & 38 05 00 & 50.8 & sy2 & 8.7 & 16 & 18 & 6.4 & 26 & 4 + ngc 526a @xmath67 & 01 23 54.4 & 35 03 55 & 83.0 & sy1 & 5.8 & 6.3 & 10 & 5.9 & 19 & 4 + ngc 513 @xmath67 & 01 24 26.9 & + 33 47 58 & 85.0 & sy2 & 13 & 1.9 & 4.4 & 1.1 & 6.5 & 4 + ngc 520 & 01 24 35.1 & + 03 47 32 & 30.2 & & 45 & @xmath00.64 & 7.5 & @xmath01.4 & 8.1 & 8 + ngc 584 & 01 31 20.8 & 06 52 05 & 20.0 & & & @xmath00.92 & 1.7 & & @xmath00.30 & 5 + ngc 613 & 01 34 18.2 & 29 25 06 & 15.0 & & 131 & 0.67 & 16 & 3.2 & 9.1 & 7 + ngc 633 & 01 36 23.4 & 37 19 17 & 74.2 & & 50 & @xmath00.50 & 6.1 & @xmath00.59 & 1.9 & 2 + ngc 628 & 01 36 41.8 & + 15 47 00 & 7.30 & & 6.2 & @xmath01.00 & & & @xmath00.38 & 5 + iras 013641042 & 01 38 52.9 & 10 27 11 & 215 & & 7.9 & @xmath00.32 & 1.5 & @xmath00.60 & @xmath00.91 & 2 + ngc 660 & 01 43 02.4 & + 13 38 42 & 12.1 & & 346 & @xmath00.57 & 37 & @xmath04.5 & 28 & 2 + mrk 573 & 01 43 57.8 & + 02 20 59 & 73.8 & sy2 & & 18 & 24 & & @xmath079 & 6 + iii zw 035 & 01 44 30.5 & + 17 06 05 & 120 & & 3.8 & @xmath00.17 & 0.19 & @xmath01.9 & @xmath01.3 & 2 + iras 014750740 @xmath67 & 01 50 02.7 & 07 25 48 & 76.7 & sy2 & 14 & 6.4 & 9.9 & 1.9 & 6.5 & 4 + 3c 55 & 01 57 10.5 & + 28 51 37 & 4522 & sy2 & & 1.1 & 2.0 & & & 6 + mrk 1014 & 01 59 50.2 & + 00 23 40 & 781 & sy1 & 6.6 & 7.4 & 9.7 & 5.0 & 13 & 3 + ngc 788 @xmath68 & 02 01 06.4 & 06 48 55 & 58.9 & sy2 & 6.1 & 5.3 & 14 & 7.9 & 23 & 2 + ngc 855 & 02 14 03.6 & + 27 52 37 & 9.60 & & 5.5 & @xmath00.72 & & & @xmath00.40 & 5 + ngc 891 & 02 22 33.4 & + 42 20 56 & 8.60 & & 8.6 & @xmath00.040 & 0.84 & @xmath00.62 & @xmath01.0 & 7 + mrk 1034 ned01 @xmath67 & 02 23 18.9 & + 32 11 18 & 148 & sy1 & 19 & @xmath01.1 & 1.5 & @xmath00.59 & @xmath00.77 & 4 + mrk 1034 ned02 @xmath67 & 02 23 22.0 & + 32 11 49 & 148 & sy1 & 35 & 1.1 & 3.6 & @xmath00.60 & 2.7 & 4 + ugc 1845 & 02 24 08.0 & + 47 58 11 & 66.7 & & 106 & @xmath00.59 & 11 & @xmath00.82 & 3.1 & 2 + eso 545-g013 @xmath67 & 02 24 40.6 & 19 08 31 & 148 & sy1 & 10 & 4.8 & 11 & 3.2 & 12 & 4 + ngc 925 & 02 27 16.9 & + 33 34 45 & 9.10 & & 10 & @xmath01.0 & 5.1 & & @xmath00.36 & 5 + ngc 931 @xmath67 & 02 28 14.5 & + 31 18 41 & 72.2 & sy1 & 5.5 & 14 & 15 & 14 & 43 & 4 + ngc 1055 & 02 41 45.2 & + 00 26 35 & 11.3 & sy2 & 26 & @xmath00.19 & 2.6 & @xmath00.58 & 1.2 & 7 + ngc
1068 @xmath68 @xmath67 & 02 42 40.7 & 00 00 47 & 16.3 & sy2 & 461 & 895 & 1355 & 808 & 2061 & 2 + ngc 1056 @xmath67 & 02 42 48.3 & + 28 34 27 & 22.1 & sy2 & 34 & @xmath01.8 & 10 & @xmath01.2 & 1.4 & 4 + ngc 1058 @xmath68 & 02 43 30.0 & + 37 20 28 & 7.42 & sy2 & 1.0 & @xmath00.46 & & @xmath00.33 & @xmath00.50 & 2 + ugc 2238 & 02 46 17.5 & + 13 05 44 & 94.5 & & 65 &
@xmath00.47 & 10 & @xmath02.6 & 5.5 & 2 + ngc 1097 @xmath68 @xmath67 & 02 46 19.1 & 30 16 29 & 17.7 & sy1 & 165 & @xmath00.74 & 20 & @xmath02.4 & 5.0 & 2 + iras 02438 + 2122 & 02 46 39.1 & + 21 35 10 & 102 & liner & 18 & @xmath00.33 & 1.6 & @xmath02.2 & @xmath02.6 & 2 + ngc 1125 @xmath67 & 02 51 40.3 & 16 39 03 & 47.2 & sy2 & 16 & 5.1 & 16 & 9.7 & 40 & 4 + ngc 1142 @xmath67 & 02 55 12.2 & 00 11 00 & 126 & sy2 & 17 & 0.92 & 5.4 & 1.8 & 5.3 & 4 + mrk 1066 & 02 59 58.6 & + 36 49 14 & 51.9 & sy2 & & 18 & 52 & & & 6 + mcg 02 - 08 - 039 @xmath67 & 03 00 30.6 & 11 24 56 & 131 & sy2 & 3.9 & 6.6 & 9.8 & 5.2 & 14 & 4 + ngc 1194 @xmath67 & 03 03 49.1 & 01 06 13 & 58.9 & sy2 & 3.8 & 4.3 & 7.4 & 3.8 & 15 & 4 + ngc 1204 & 03 04 39.9 & 12 20 28 & 64.8 & & 54 & @xmath00.34 & 5.0 & @xmath02.0 & @xmath01.6 & 2 + ngc 1222 & 03 08 56.7 & 02 57 18 & 32.3 & & 81 & @xmath00.57 & 89 & @xmath00.78 & 9.9 & 8 + 3c 79 & 03 10 00.1 & + 17 05 58 & 1295 & sy2 & & 0.11 & 0.20 & & 0.70 & 6 + ngc 1241 @xmath68 @xmath67 & 03 11 14.6 & 08 55 19 & 58.5 & sy2 & 14 & 2.5 & 8.4 & 1.5 & 4.8 & 2 + ngc 1266 & 03 16 00.8 & 02 25 38 & 30.0 & liner & 29 & @xmath00.92 & 9.7 & & @xmath01.4 & 5 + ngc 1291 & 03 17 18.6 & 41 06 29 & 10.8 & & 5.2 & @xmath00.86 & 7.2 & & 1.4 & 5 + iras f03158 + 4227 & 03 19 12.4 & + 42 38 28 & 631 & & 5.8 & @xmath01.1 & 0.94 & @xmath01.4 & @xmath01.8 & 3 + ngc 1275 @xmath68 & 03 19 48.2 & + 41 30 42 & 76.2 & sy2 & 43 & @xmath01.2 & 21 & @xmath02.5 & 9.6 & 2 + ngc 1316 & 03 22 41.7 & 37 12 29 & 24.3 & liner & 13 & @xmath00.61 & 11 & & 2.6 & 5 + ngc 1320 @xmath67 & 03 24 48.7 & 03 02 32 & 38.3 & sy2 & 9.6 & 11 & 14 & 7.5 & 27 & 4 + ngc
1365 @xmath68 @xmath67 & 03 33 36.4 & 36 08 25 & 23.9 & sy1 & 156 & 19 & 61 & 40 & 151 & 2 + ngc 1358 @xmath68 & 03 33 39.7 & 05 05 21 & 58.2 & sy2 & 5.1 & 3.5 & 8.2 & @xmath01.6 & 9.1 & 2 + ngc 1377 & 03 36 39.1 & 20 54 08 & 25.2 & & 4.1 & @xmath01.9 & 2.6 & & @xmath01.7 & 5 + ngc 1386 @xmath68 @xmath67 & 03 36 46.2 & 35 59 57 & 12.5 & sy2 & 14 & 37 & 39 & 34 & 97 & 2 + ngc 1404 & 03 38 51.9 & 35 35 39 & 18.5 & & 1.4 & @xmath00.97 & 1.00 & & @xmath00.35 & 5 + ngc 1433 @xmath68 @xmath67 & 03 42 01.6 & 47 13 19 & 15.4 & sy2 & & @xmath01.6 & & @xmath01.9 & @xmath06.9 & 2 + ngc 1448 & 03 44 31.9 & 44 38 41 & 11.5 & & 8.2 & 3.7 & 5.0 & 3.6 & 16 & 7 + ic 342 & 03 46 48.5 & + 68 05 46 & 4.60 & & 615 & @xmath02.4 & 37 & @xmath04.9 & @xmath07.7 & 8 + iras
03450 + 0055 @xmath67 & 03 47 40.2 & + 01 05 14 & 136 & sy1 & 1.1 & @xmath01.5 & 1.8 & @xmath01.9 & 2.5 & 4 + ngc 1482 & 03 54 38.9 & 20 30 08 & 23.2 & & 457 & @xmath02.6 & 56 & & @xmath06.4 & 5 + iras 03521 + 0028 & 03 54 42.2 & + 00 37 03 & 724 & liner & 2.8 & @xmath00.45 & 1.3 & @xmath00.48 & @xmath00.90 & 3 + ngc 1512 & 04 03 54.3 & 43 20 55 & 11.8 & & 31 & @xmath01.1 & 4.7 & & @xmath00.83 & 5 + 3c 109 & 04 13 40.4 & + 11 12 13 & 1589 & sy1 & 0.030 & 0.070 & 0.11 & & 0.27 & 6 + ic 2056 & 04 16 24.5 & 60 12 24 & 13.8 & & 13 & @xmath00.16 & 2.2 & @xmath00.10 & 0.75 & 7 + ngc 1559 & 04 17 35.8 & 62 47 01 & 12.7 & & 15 & @xmath00.44 & 2.3 & @xmath00.36 & @xmath00.69 & 7 + ngc 1566 @xmath68 @xmath67 & 04 20 00.4 & 54 56 16 & 21.3 & sy1 & 15 & 1.1 & 9.6 & @xmath01.9 & 8.4 & 2 + ngc 1569 & 04 30 49.1 & + 64 50 52 & 4.60 & sy1 & 19 & @xmath00.38 & 188 & @xmath01.4 & 29 & 7 + 3c 120 @xmath67 & 04 33 11.1 & + 05 21 15 & 145 & sy1 & 7.8 & 17 & 28 & 29 & 123 & 4 + mrk 618 @xmath67 & 04 36 22.2 & 10 22 33 & 156 & sy1 & 16 & 3.9 & 5.4 & @xmath01.3 & 10 & 4 + iras f043850828 @xmath67 & 04 40 55.0 & 08 22 22 & 65.4 & sy2 & 14 & 2.3 & 7.1 & @xmath01.4 & 8.6 & 4 + ngc
1667 @xmath68 @xmath67 & 04 48 37.1 & 06 19 11 & 65.9 & sy2 & 10.0 & @xmath00.32 & 6.9 & 1.7 & 6.3 & 2 + ngc 1705 & 04 54 13.5 & 53 21 39 & 5.10 & & 3.2 & @xmath00.91 & 12 & & 2.5 & 5 + eso 033-g002 @xmath67 & 04 55 59.0 & 75 32 28 & 78.6 & sy2 & 2.1 & 6.3 & 9.2 & 5.3 & 14 & 4 + ngc 1792 & 05 05 14.4 & 37 58 50 & 12.5 & & 23 & 0.41 & 2.1 & @xmath00.25 & 0.96 & 7 + ngc 1808 & 05 07 42.3 & 37 30 47 & 12.6 & sy2 & 177 & @xmath00.91 & 17 & @xmath08.4 & @xmath09.5 & 7 + eso 362-g018 @xmath67 & 05 19 35.8 & 32 39 27 & 53.8 & sy1 & 12 & 3.3 & 7.5 & 2.6 & 8.9 & 4 + iras f051892524 & 05 21 01.5 & 25 21 45 & 190 & sy2 & 21 & 18 & 18 & 12 & 24 & 3 + iras 051871017 & 05 21 06.5 & 10 14 46 & 126 & liner & 10 & @xmath00.15 & 1.9 & @xmath00.66 & @xmath00.92 & 2 + eso 253-g003 @xmath67 & 05 25 18.1 & 46 00 21 & 188 & sy2 & 16 & 6.8 & 18 & 7.2 & 24 & 4 + ugc 3351 & 05 45 47.9 & + 58 42 03 & 64.1 & & 67 & @xmath00.50 & 9.6 & @xmath00.31 & 0.83 & 2 + ugca 116 & 05 55 42.6 & + 03 23 31 & 11.6 & & 6.0 & @xmath00.28 & 117 & @xmath01.1 & 8.5 & 7 + iras f055633820 @xmath67 & 05 58 02.0 & 38 20 04 & 149 & sy1 & 3.9 & 2.5 & 4.8 & @xmath00.79 & 5.3 & 4 + iras f060357102 & 06 02 54.0 & 71 03 10 & 358 & & 7.0 & @xmath00.48 & 1.7 & @xmath00.81 & @xmath03.0 & 3 + eso 121-g006 & 06 07 29.9 & 61 48 27 & 14.5 & & 16 & 0.65 & 3.2 & @xmath00.59 & 4.6 & 7 + mrk 3 & 06 15 36.4 & + 71 02 15 & 60.6 & sy2 & 100 & 64 & 179 & & 214 & 6 + ngc 2146 & 06 18 37.7 & + 78 21 25 & 16.5 & & 625 & @xmath02.8 & 91 & @xmath03.7 & 19 & 8 + ugca 127 & 06 20 55.7 & 08 29 44 & 10.2 & & 9.3 & @xmath00.15 & 1.5 & @xmath00.97 & 3.4 & 7 + iras f062066315 & 06 21 01.2 & 63 17 23 & 421 & sy2 & 6.6 & 2.3 & 2.9 & 2.0 & 3.0 & 3 + ngc
2273 @xmath68 & 06 50 08.7 & + 60 50 44 & 26.5 & sy2 & 40 & 4.5 & 19 & 6.1 & 14 & 2 + mrk 6 @xmath67 & 06 52 12.3 & + 74 25 37 & 81.7 & sy1 & 28 & 9.4 & 49 & 10 & 48 & 4 + iras 071452914 & 07 16 31.2 & 29 19 28 & 25.8 & sy2 & & 83 & 168 & & & 6 + ngc 2369 & 07 16 37.7 & 62 20 37 & 46.2 & & 124 & @xmath00.51 & 10 & @xmath00.91 & 1.8 & 2 + ngc 2388 & 07 28 53.4 & + 33 49 08 & 58.9 & & 121 & @xmath01.2 & 9.8 & @xmath00.79 & 1.4 & 2 + mcg + 02 - 20 - 003 & 07 35 43.4 & + 11 42 33 & 71.1 & & 60 & @xmath00.47 & 6.3 & @xmath00.53 & @xmath01.2 & 2 + ngc 2403 & 07 36 51.4 & + 65 36 09 & 3.20 & & 5.1 & @xmath00.71 & 2.1 & & @xmath00.36 & 5 + mrk 9 @xmath67 & 07 36 57.0 & + 58 46 13 & 176 & sy1 & 3.2 & 2.2 & 1.9 & 2.2 & 5.6 & 4 + 3c 184 & 07 39 24.5 & + 70 23 10 & 6559 & sy2 & 0.19 & 0.17 & 0.65 & & & 6 + mrk 79 @xmath67 & 07 42 32.8 & + 49 48 34 & 96.7 & sy1 & 10 & 6.6 & 20 & 13 & 42 & 4 + eso 209-g009 & 07 58 15.4 & 49 51 15 & 11.8 & & 9.8 & @xmath00.26 & 1.4 & @xmath00.12 & @xmath00.29 & 7 + iras f07598 + 6508 & 08 04 33.1 & + 64 59 48 & 703 & & 3.9 & @xmath00.75 & 2.5 & @xmath03.0 & @xmath01.8 & 3 + 3c 196 & 08 13 36.0 & + 48 13 02 & 5572 & sy1 & & 0.30 & 0.46 & & & 6 + iras 083112459 & 08 33 20.6 & 25 09 33 & 460 & sy1 & 24 & 12 & 23 & 9.8 & 26 & 3 + ugc 4459 & 08 34 07.2 & + 66 10 54 & 2.78 & & & @xmath00.89 & 0.98 & & @xmath00.37 & 5 + ngc 2623 & 08 38 24.1 & + 25 45 16 & 77.4 & sy2 & 56 & 2.7 & 15 & 2.2 & 9.6 & 8 + ngc 2639 @xmath68 @xmath67 & 08 43 38.1 & + 50 12 20 & 44.6 & sy1 & 8.9 & @xmath00.45 & 4.9 & @xmath00.82 & 1.8 & 2 + pg 0838 + 770 & 08 44 45.3 & + 76 53 09 & 615 & qso & 0.41 & 0.32 & 0.56 & @xmath00.19 & 1.3 & 1 + pg 0844 + 349 & 08 47 42.5 & + 34 45 04 & 287 & qso & 0.42 & 0.30 & 1.0 & 0.42 & 1.5 & 1 + ngc 2681 & 08 53 32.7 & + 51 18 49 & 12.5 & & 6.9 & @xmath00.15 & 3.4 & @xmath00.36 & 2.2 & 7 + ngc 2685 @xmath68 & 08 55 34.7 & + 58 44 03 & 12.7 & sy2 & & @xmath00.28 & & @xmath00.22 & 0.43 & 2 + ngc 2655 @xmath68 @xmath67 & 08 55 37.7 & + 78 13 23 & 20.2 & sy2 & 13 & @xmath01.3 & 7.2 & @xmath01.3 & @xmath08.0 & 2 + iras f08572 + 3915 & 09 00 25.4 & + 39 03 54 & 259 & & 7.2 &
@xmath00.75 & 2.0 & @xmath05.4 & @xmath06.0 & 3 + iras 090223615 & 09 04 12.7 & 36 27 01 & 269 & & 57 & @xmath02.4 & 40 & @xmath03.6 & 6.7 & 3 + ngc 2798 & 09 17 22.9 & + 41 59 59 & 26.2 & & 215 & @xmath01.9 & 33 & & @xmath05.1 & 5 + mrk 704 @xmath67 & 09 18 26.0 & + 16 18 19 & 128 & sy1 & & 3.9 & 5.6 & @xmath03.7 & 12 & 4 + ngc 2841 & 09 22 02.6 & + 50 58 35 & 14.1 & liner & 5.7 & @xmath04.5 & 6.7 & & 0.78 & 5 + ngc 2915 & 09 26 11.5 & 76 37 34 & 3.80 & & 3.2 & @xmath00.75 & 11 & & 0.45 & 5 + ngc 2903 & 09 32 10.1 & + 21 30 03 & 8.30 & & 182 & @xmath00.82 & 14 & @xmath00.94 & @xmath01.7 & 7 + ugc 5101 & 09 35 51.7 & + 61 21 11 & 172 & liner & 34 & 2.6 & 14 & 2.8 & 7.3 & 3 + ngc 2992 @xmath68 @xmath67 & 09 45 42.1 & 14 19 34 & 33.2 & sy2 & 46 & 22 & 61 & 28 & 131 & 2 + ngc 2976 & 09 47 15.5 & + 67 54 59 & 3.60 & & 7.4 & @xmath00.99 & 2.8 & & @xmath00.31 & 5 + ngc 3059 & 09 50 08.2 & 73 55 19 & 14.2 & & 27 &
@xmath00.18 & 3.1 & @xmath00.69 & @xmath01.5 & 7 + mrk 1239 @xmath67 & 09 52 19.1 & 01 36 43 & 86.7 & sy1 & 9.4 & 3.4 & 9.4 & 3.2 & 16 & 4 + ngc 3049 & 09 54 49.6 & + 09 16 15 & 23.9 & & 39 & @xmath00.89 & 6.0 & & @xmath01.00 & 5 + ngc
3031 @xmath68 @xmath67 & 09 55 33.2 & + 69 03 55 & 3.64 & sy1 & 27 & @xmath00.50 & 21 & @xmath01.4 & 4.5 & 2 + m 82 & 09 55 52.7 & + 69 40 45 & 3.60 & & 506 & @xmath00.89 & 81 & @xmath015 & 46 & 7 + holmberg ix & 09 57 32.0 & + 69 02 45 & 2.64 & & & @xmath00.64 & & & @xmath00.38 & 5 + ngc
3081 @xmath68 & 09 59 29.5 & 22 49 34 & 34.3 & sy2 & 12 & 30 & 34 & 39 & 118 & 2 + 3c 234 @xmath67 & 10 01 49.5 & + 28 47 08 & 897 & sy2 & & 2.3 & 3.4 & 2.9 & 9.0 & 4 + ngc 3079 @xmath68 @xmath67 & 10 01 57.8 & + 55 40 47 & 16.2 & sy2 & 108 & 0.91 & 24 & @xmath00.89 & 11 & 2 + ngc 3110 & 10 04 02.1 & 06 28 29 & 72.4 & & 75 & @xmath00.68 & 9.9 & @xmath00.62 & 1.2 & 2 + pg 1001 + 054 & 10 04 20.1 & + 05 13 00 & 771 & qso & 0.40 & @xmath00.21 & 0.21 & @xmath00.30 & 0.52 & 1 + ngc 3175 & 10 14 42.1 & 28 52 19 & 13.4 & & 34 & @xmath00.35 & 2.5 & @xmath00.86 & 1.1 & 7 + ic 2560 @xmath68 & 10 16 18.7 & 33 33 49 & 41.7 & sy2 & 16 & 19 & 36 & 18 & 56 & 2 + ngc 3147 @xmath68 @xmath67 & 10 16 53.7 & + 73 24 02 & 40.3 & sy2 & 4.5 & @xmath00.59 & 3.4 & @xmath00.91 & @xmath02.5 & 2 + ngc 3185
@xmath68 & 10 17 38.6 & + 21 41 17 & 17.7 & sy2 & 9.6 & 1.0 & 3.6 & 1.8 & 8.2 & 2 + ngc 3190 & 10 18 05.6 & + 21 49 55 & 20.9 & liner & 8.3 & @xmath00.67 & 4.8 & & 0.93 & 5 + ngc 3184 & 10 18 17.0 & + 41 25 27 & 11.1 & & 18 & @xmath00.83 & 2.2 & & @xmath00.29 & 5 + ngc 3198 & 10 19 54.9 & + 45 32 59 & 13.7 & & 15 & @xmath01.00 & 0.81 & & @xmath00.32 & 5 + ngc 3227 @xmath68 @xmath67 & 10 23 30.6 & + 19 51 54 & 15.7 & sy1 & 66 & 23 & 74 & 18 & 70 & 2 + ngc 3245 & 10 27 18.4 & + 28 30 26 & 18.2 & & 9.9 & @xmath00.21 & 3.0 & @xmath00.76 & 0.71 & 2 + ngc 3256 & 10 27 51.3 & 43 54 13 & 40.1 & & 780 & @xmath01.9 & 119 & @xmath07.9 & @xmath020 & 2 + ngc 3265 & 10 31 06.8 & + 28 47 48 & 23.2 & & 30 & @xmath00.78 & 5.4 & & @xmath00.71 & 5 + ngc 3281 @xmath68 & 10 31 52.1 & 34 51 13 & 49.6 & sy2 & 17 & 47 & 58 & 46 & 173 & 2 + mrk 33 & 10 32 31.9 & + 54 24 03 & 22.9 & & 59 & @xmath01.0 & 45 & & 0.89 & 5 + 3c 244.1 & 10 33 34.0 & + 58 14 35 & 2353 & sy2 & 1.4 & 0.60 & 0.30 & & & 6 + ngc 3310 & 10 38 45.9 & + 53 30 12 & 19.8 & & 28 & @xmath00.25 & 28 & @xmath01.0 & 4.0 & 8 + iras 10378 + 1109 & 10 40 29.2 & + 10 53 18 & 641 & liner & 3.5 & @xmath01.1 & 0.60 & @xmath02.4 & 2.0 & 3 + ngc 3351 & 10 43 57.7 & + 11 42 13 & 9.30 & & 219 & @xmath02.3 & 18 & & 2.5 & 5 + ngc 3368 & 10 46 45.7 & + 11 49 11 & 10.5 & liner & 4.3 & @xmath00.29 & 2.8 & @xmath00.74 & 0.99 & 7 + ngc 3393 & 10 48 23.5 & 25 09 43 & 56.2 & sy2 & & 42 & 95 & & & 6 + iras f10565 + 2448 & 10 59 18.1 & + 24 32 34 & 190 & & 64 & @xmath01.00 & 7.6 & @xmath00.90 & @xmath01.2 & 3 + ngc 3511 @xmath67 & 11 03 23.8 & 23 05 12 & 15.9 & sy1 & 8.7 & @xmath00.49 & 1.00 & @xmath01.7 & @xmath01.8 & 4 + ngc 3507 & 11 03 25.4 & + 18 08 07 & 14.2 & liner & 3.7 & @xmath00.12 & 1.4 & @xmath00.55 & @xmath00.21 & 2 + 3c 249.1 & 11 04 13.7 & + 76 58 58 & 1619 & sy1 & & 0.090 & 0.14 & & @xmath00.23 & 6 + ngc 3521 & 11 05 48.6 & 00 02 09 & 10.1 & & 14 & @xmath00.95 & 7.4 & & 2.2 & 5 + ngc
3516 @xmath68 @xmath67 & 11 06 47.5 & + 72 34 06 & 38.0 & sy1 & 7.2 & 7.3 & 16 & 9.5 & 44 & 2 + eso 265-g007 & 11 07 49.6 & 46 31 27 & 11.7 & & 2.6 & @xmath00.41 & 1.1 & @xmath01.2 & @xmath01.0 & 7 + ngc 3556 & 11 11 31.0 & + 55 40 26 & 13.9 & & 21 & @xmath00.37 & 3.2 & @xmath01.2 & @xmath01.5 & 8 + iras f110950238 & 11 12 03.4 & 02 54 22 & 495 & liner & 6.1 & @xmath00.48 & 1.9 & @xmath01.8 & @xmath00.90 & 3 + iras f11119 + 3257 & 11 14 38.9 & + 32 41 33 & 920 & sy1 & 3.0 & 0.75 & 2.0 & @xmath00.90 & 1.9 & 3 + ngc 3621 & 11 18 16.5 & 32 48 50 & 6.60 & & 16 & 1.0 & 5.7 & & 4.6 & 5 + ngc 3627 & 11 20 15.0 & + 12 59 29 & 9.40 & sy2 & 23 & @xmath00.92 & 8.5 & & 1.7 & 5 + ngc 3628 & 11 20 17.0 & + 13 35 22 & 10.00 & liner & 126 & 0.90 & 10 & @xmath01.0 & @xmath02.4 & 8 + mcg + 00 - 29 - 023 @xmath67 & 11 21 12.3 & 02 59 03 & 109 & sy2 & 47 & @xmath01.0 & 4.4 & @xmath05.4 & @xmath06.0 & 4 + pg 1119 + 120 & 11 21 47.1 & + 11 44 18 & 222 & qso & 0.46 & 1.7 & 2.7 & 1.3 & 5.9 & 1 + ngc 3642 & 11 22 17.9 & + 59 04 28 & 22.3 & liner & 2.8 & @xmath00.11 & 0.97 & @xmath00.54 & @xmath00.18 & 2 + ngc 3660 @xmath67 & 11 23 32.3 & 08 39 30 & 53.1 & sy2 & 6.5 & 0.98 & 1.5 & 1.7 & 3.6 & 4 + iras 112152806 @xmath68 & 11 24 02.8 & 28 23 15 & 59.5 & sy2 & 1.5 & 2.6 & 5.2 & 2.8 & 12 & 2 + ngc 3675 & 11 26 08.6 & + 43 35 09 & 12.7 & & 8.1 & @xmath00.23 & 2.2 & @xmath00.48 & @xmath00.58 & 7 + ic 694 & 11 28 27.3 & + 58 34 42 & 44.5 & & 294 & @xmath01.6 & 69 & @xmath04.7 & @xmath019 & 2 + ngc 3690 & 11 28 32.2 & + 58 33 44 & 44.0 & sy2 & 151 & @xmath01.5 & 77 & @xmath04.6 & 21 & 2 + pg 1126 - 041 & 11 29 16.7 & 04 24 07 & 269 & qso & 1.4 & 4.3 & 5.2 & 4.7 & 16 & 1 + ngc 3726 & 11 33 21.1 & + 47 01 45 & 14.5 & & 5.5 & @xmath00.41 & 0.43 & @xmath00.52 & @xmath00.44 & 7 + ngc 3735 @xmath68 @xmath67 & 11 35 57.3 & + 70 32 08 & 38.9 & sy2 & 9.4 & 7.1 & 13 & 9.8 & 37 & 2 + ngc 3773 & 11 38 12.9 & + 12 06 43 & 11.9 & & 18 & @xmath00.81 & 16 & & 0.36 & 5 + ngc 3783 @xmath68 & 11 39 01.7 & 37 44 18 & 41.6 & sy1 & 18 & 16 & 27 & 12 & 40 & 2 + ngc 3884 & 11 46 12.2 & + 20 23 29 & 101 & liner & 1.5 & @xmath00.12 & 0.86 & @xmath00.72 & @xmath00.72 & 2 + ngc 3898 & 11 49 15.4 & + 56 05 03 & 16.1 & & 1.0 & @xmath00.13 & 1.3 & @xmath00.51 & 0.40 & 2 + ngc 3938 & 11 52 49.5 & + 44 07 14 & 13.3 & & 5.4 & @xmath00.66 & 1.0 & & 0.25 & 5 + eso 320-g030 & 11 53 11.7 & 39 07 48 & 43.6 & & 110 & @xmath00.53 & 11 & @xmath01.6 & @xmath00.88 & 2 + ngc 3949 & 11 53 41.7 & + 47 51 31 & 13.6 & & 5.8 & @xmath00.11 & 1.1 & @xmath00.11 & 1.5 & 7 + ngc 3976 @xmath68 & 11 55 57.6 & + 06 45 02 & 36.0 & sy2 & 2.3 & @xmath00.26 & 1.1 & @xmath00.35 & 0.73 & 2 + ngc 3982 @xmath68 @xmath67 & 11 56 28.1 & + 55 07 30 & 15.2 & sy2 & 9.2 & 3.0 & 6.7 & 1.7 & 4.6 & 2 + ngc 3998 & 11 57 56.1 & + 55 27 12 & 14.6 & liner & 11 & @xmath00.18 & 6.9 & @xmath00.65 & 0.74 & 2 + ngc 4013 & 11 58 31.4 & + 43 56 47 & 13.8 & & 13 & @xmath00.090 & 2.8 & @xmath00.89 & @xmath01.0 & 7 + ngc 4036 & 12 01 26.8 & + 61 53 44 & 19.1 & liner & 4.4 & @xmath00.060 & 2.8 & @xmath00.37 & 1.4 & 2 + ngc 4051 @xmath68 @xmath67 & 12 03 09.6 & + 44 31 52 & 9.27 & sy1 & 17 & 11 & 16 & 10 & 32 & 2 + iras f12018 + 1941 & 12 04 24.5 & + 19 25 09 & 813 & liner & 3.0 & @xmath00.80 & 0.35 & @xmath00.10 & @xmath00.63 & 3 + ngc 4085 & 12 05 22.7 & + 50 21 10 & 14.6 & & 23 & @xmath00.15 & 2.9 & @xmath00.92 & @xmath00.44 & 7 + ngc 4088 & 12 05 34.2 & + 50 32 20 & 13.4 & & 37 & @xmath00.39 & 2.5 & @xmath00.50 & 0.73 & 8 + ngc 4125 & 12 08 06.0 & + 65 10 26 & 22.9 & liner & 2.3 & @xmath00.62 & 3.3 & & 0.75 & 5 + iras 120710444 & 12 09 45.1 & 05 01 13 & 600 & sy2 & 5.2 & 2.9 & 5.1 & 3.7 & 6.6 & 3 + ngc 4151 @xmath68 @xmath67 & 12 10 32.6 & + 39 24 20 & 14.0 & sy1 & 131 & 77 & 205 & 68 & 244 & 2 + ngc 4157 & 12 11 04.4 & + 50 29 04 & 13.3 & & 12 & @xmath00.070 & 1.4 & @xmath00.70 & 1.1 & 7 + ngc 4192 & 12 13 48.3 & + 14 54 01 & 19.3 & & 19 & @xmath00.18 & 4.7 & @xmath00.48 & 1.6 & 2 + ngc 4194 & 12 14 09.5 & + 54 31 36 & 40.3 & sy2 & 165 & 3.0 & 54 & 4.0 & 27 & 8 + pg 1211 + 143 & 12 14 17.7 & + 14 03 12 & 368 & qso & 0.32 & @xmath00.43 & 0.73 & @xmath00.56 & 2.4 & 1 + ngc 4236 & 12 16 42.1 & + 69 27 45 & 4.36 & & & @xmath01.5 & & & @xmath00.23 & 5 + ngc 4235 @xmath68 & 12 17 09.9 & + 07 11 29 & 33.5 & sy1 & 3.8 & @xmath00.43 & 3.3 & @xmath00.54 & 3.3 & 2 + mrk 766 @xmath67 & 12 18 26.5 & + 29 48 46 & 55.9 & sy1 & 23 & 21 & 24 & 18 & 46 & 4 + ngc 4254 & 12 18 49.6 & + 14 24 59 & 16.6 & & 53 & @xmath00.92 & 6.1 & & 2.5 & 5 + ngc 4258 @xmath68 & 12 18 57.5 & + 47 18 14 & 6.60 & sy1 & 12 & @xmath01.3 & 8.1 & @xmath00.82 & 7.5 & 2 + ngc 4278 & 12 20 06.8 & + 29 16 50 & 9.17 & liner & 5.4 & @xmath00.23 & 3.9 & @xmath00.29 & 0.88 & 2 + ngc 4321 & 12 22 54.9 & + 15 49 20 & 14.3 & liner & 152 & @xmath01.2 & 18 & & @xmath01.9 & 5 + mrk 52 & 12 25 42.8 & + 00 34 21 & 30.1 & & 29 & @xmath00.57 & 3.8 & @xmath00.83 & 1.3 & 8 + ngc 4388 @xmath68 @xmath67 & 12 25 46.7 & + 12 39 43 & 36.3 & sy2 & 75 & 45 & 106 & 68 & 308 & 2 + ngc
4395 @xmath68 & 12 25 48.9 & + 33 32 48 & 4.33 & sy1 & 4.9 & 0.93 & 6.8 & 1.4 & 6.9 & 2 + ngc 4419 & 12 26 56.4 & + 15 02 50 & 17.2 & & 36 & @xmath00.46 & 8.5 & @xmath01.0 & @xmath01.6 & 2 + ngc 4435 & 12 27 40.5 & + 13 04 44 & 11.9 & & 6.8 & @xmath00.24 & 2.1 & @xmath00.67 & 1.1 & 2 + ngc 4450 & 12 28 29.6 & + 17 05 05 & 16.6 & liner & 3.5 & @xmath00.53 & 2.1 & & 0.63 & 5 + ngc 4457 & 12 28 59.0 & + 03 34 14 & 12.4 & liner & 7.8 & @xmath00.18 & 3.2 & @xmath00.47 & 2.2 & 2 + 3c 273 & 12 29 06.7 & + 02 03 08 & 755 & sy1 & 1.5 & 3.4 & 6.0 & 2.9 & 8.5 & 3 + ngc 4477 @xmath68 & 12 30 02.2 & + 13 38 11 & 19.3 & sy2 & 3.2 & @xmath00.33 & 2.8 & @xmath03.5 & 1.3 & 2 + ngc 4490 & 12 30 36.4 & + 41 38 37 & 10.5 & & 3.5 & @xmath00.10 & 3.5 & @xmath00.10 & 1.1 & 7 + ngc 4486 & 12 30 49.4 & + 12 23 28 & 17.0 & sy2 & 6.9 & @xmath00.19 & 5.4 & @xmath00.57 & 1.2 & 2 + ngc 4501 @xmath68 @xmath67 & 12 31 59.2 & + 14 25 13 & 32.7 & sy2 & 5.0 & @xmath00.47 & 4.7 & @xmath00.43 & 2.5 & 2 + pg 1229 + 204 & 12 32 03.6 & + 20 09 29 & 283 & qso & 0.61 & 0.91 & 1.3 & 0.99 & 2.8 & 1 + ngc 4536 & 12 34 27.1 & + 02 11 17 & 14.4 & & 386 & @xmath01.2 & 60 & & @xmath07.9 & 5 + ngc 4507 @xmath68 & 12 35 36.6 & 39 54 33 & 50.9 & sy2 & 30 & 12 & 29 & 8.6 & 34 & 2 + ngc 4552 & 12 35 39.8 & + 12 33 22 & 15.9 & liner & 1.7 & @xmath00.89 & 2.9 & & @xmath00.53 & 5 + ngc 4559 & 12 35 57.7 & + 27 57 35 & 10.3 & & 8.3 & @xmath01.0 & 2.3 & & 0.12 & 5 + ngc 4565 @xmath68 @xmath67 & 12 36 20.8 & + 25 59 15 & 18.1 & sy1 & 2.6 & 0.32 & 3.3 & @xmath00.82 & 4.1 & 2 + ngc 4569 & 12 36 49.8 & + 13 09 46 & 16.6 & liner & 36 & @xmath00.93 & 16 & & 2.9 & 5 + ngc 4579 @xmath68 @xmath67 & 12 37 43.5 & + 11 49 05 & 21.5 & sy1 & 23 & @xmath00.40 & 12 & @xmath00.68 & 3.2 & 2 + ngc 4593 @xmath68 @xmath67 & 12 39 39.4 & 05 20 39 & 38.8 & sy1 & 6.8 & 3.5 & 7.4 & 5.4 & 13 & 2 + ngc 4594 @xmath68 @xmath67 & 12 39 59.4 & 11 37 23 & 15.7 & sy1 & 14 & @xmath00.55 & 16 & @xmath00.71 & 2.4 & 2 + ngc 4602 @xmath67 & 12 40 36.9 & 05 07 58 & 36.5 & sy1 & 7.6 & 0.82 & 0.63 & @xmath01.2 & @xmath02.3 & 4 + tol 1238 - 364 @xmath68 @xmath67 & 12 40 52.9 & 36 45 21 & 47.7 & sy2 & 42 & 9.0 & 27 & 3.5 & 17 & 2 + ngc 4625 & 12 41 52.7 & + 41 16 26 & 9.20 & & 5.9 & @xmath00.84 & 1.2 & & @xmath00.26 & 5 + ngc 4631 & 12 42 08.0 & + 32 32 29 & 8.10 & & 133 & @xmath00.52 & 30 & & 3.0 & 5 + ngc 4639 @xmath68 & 12 42 52.4 & + 13 15 26 & 14.0 & sy1 & 0.84 & @xmath00.27 & 1.1 & @xmath00.28 & @xmath01.2 & 2 + ngc 4666 & 12 45 08.6 & 00 27 42 & 21.5 & liner & 35 & @xmath00.30 & 8.3 & 1.4 & 6.3 & 2 + ngc 4676 & 12 46 10.1 & + 30 43 55 & 94.0 & & 28 & @xmath00.25 & 4.7 & @xmath00.33 & 1.4 & 8 + pg 1244 + 026 & 12 46 35.2 & + 02 22 08 & 213 & qso & 0.94 & 0.53 & 1.2 & 1.1 & 1.5 & 1 + ngc 4725 @xmath68 @xmath67 & 12 50 26.6 & + 25 30 03 & 17.4 & sy2 & 1.2 & @xmath00.38 & 2.4 & @xmath00.50 & 1.8 & 2 + ngc 4736 & 12 50 53.1 & + 41 07 13 & 5.00 & liner & 13 & @xmath00.81 & 14 & & 3.3 & 5 + ngc 4748 @xmath67 & 12 52 12.5 & 13 24 53 & 63.4 & sy1 & 7.4 & 6.7 & 16 & 20 & 82 & 4 + iras f12514 + 1027 & 12 54 00.8 & + 10 11 12 & 1667 & sy2 & 2.3 & 1.9 & 2.7 & 1.7 & 2.7 & 3 + mrk 231 & 12 56 14.2 & + 56 52 25 & 186 & sy1 & 20 & @xmath03.0 & 3.0 & @xmath018 & @xmath09.5 & 3 + ngc 4826 & 12 56 43.7 & + 21 40 57 & 5.00 & sy2 & 105 & @xmath01.3 & 23 & & 4.3 & 5 + ngc 4818 & 12 56 48.9 & 08 31 31 & 9.40 & & 185 & @xmath01.2 & 14 & @xmath02.0 & @xmath01.6 & 8 + ngc 4922 @xmath67 & 13 01 24.9 & + 29 18 40 & 103 & liner & 36 & 2.4 & 9.2 & @xmath01.9 & 4.3 & 2 + mcg 02 - 33 - 098-e & 13 02 19.7 & 15 46 03 & 68.9 & & 32 & @xmath00.58 & 12 & @xmath00.61 & 1.2 & 2 + mcg 02 - 33 - 098-w & 13 02 20.4 & 15 45 59 & 68.1 & & 67 & @xmath00.51 & 12 & @xmath00.79 & @xmath01.2 & 2 + ngc 4941 @xmath68 @xmath67 & 13 04 13.1 & 05 33 05 & 15.9 & sy2 & 13 & 7.2 & 24 & 6.9 & 26 & 2 + ngc 4939 @xmath68 & 13 04 14.4 & 10 20 22 & 44.8 & sy2 & 8.1 & 12 & 24 & 17 & 65 & 2 + ngc 4945 @xmath68 & 13 05 27.5 & 49 28 05 & 8.19 & sy2 & 586 & 3.4 & 69 & @xmath05.1 & 35 & 2 + pg 1302 - 102 & 13 05 33.0 & 10 33 19 & 1422 & qso & 0.36 & 0.49 & 0.66 & 0.39 & 2.6 & 1 + ugc 8201 & 13 06 24.9 & + 67 42 25 & 3.68 & & & @xmath00.75 & & & @xmath00.24 & 5 + ngc
4968 @xmath67 & 13 07 06.0 & 23 40 37 & 42.6 & sy2 & 25 & 18 & 34 & 11 & 34 & 4 + pg 1307 + 085 & 13 09 47.0 & + 08 19 48 & 739 & qso & 0.40 & 0.56 & 0.98 & 0.62 & 0.74 & 1 + ngc 5005 @xmath68 @xmath67 & 13 10 56.2 & + 37 03 33 & 13.8 & sy2 & 37 & @xmath01.1 & 13 & @xmath01.2 & 4.6 & 2 + pg 1309 + 355 & 13 12 17.8 & + 35 15 21 & 893 & qso & 0.51 & 0.27 & 1.3 & @xmath00.30 & @xmath00.49 & 1 + ngc 5033 @xmath68 @xmath67 & 13 13 27.5 & + 36 35 38 & 12.4 & sy1 & 31 & 1.2 & 14 & 2.0 & 9.2 & 2 + ic 860 & 13 15 03.5 & + 24 37 07 & 55.8 & & 3.6 & @xmath00.14 & 0.68 & @xmath00.48 & @xmath00.62 & 2 + iras 131205453 & 13 15 06.3 & 55 09 22 & 136 & sy2 & 150 & 1.7 & 18 & @xmath020 & 6.4 & 3 + ngc 5055 & 13 15 49.3 & + 42 01 45 & 7.80 & liner & 22 & @xmath00.95 & 9.6 & & 2.3 & 5 + ugc 8387 & 13 20 35.3 & + 34 08 22 & 101 & & 113 & @xmath00.54 & 19 & @xmath02.3 & 6.2 & 2 + ngc 5104 & 13 21 23.1 & + 00 20 32 & 80.0 & liner & 43 & @xmath00.39 & 5.1 & @xmath01.2 & 2.4 & 2 + mcg 03 - 34 - 064 @xmath67 & 13 22 24.5 & 16 43 42 & 71.7 & sy1 & 56 & 63 & 119 & 38 & 115 & 4 + ngc 5128 @xmath68 & 13 25 27.6 & 43 01 08 & 7.85 & sy2 & 197 & 23 & 150 & 28 & 135 & 2 + ngc 5135 @xmath68 @xmath67 & 13 25 44.1 & 29 50 01 & 58.9 & sy2 & 112 & 14 & 58 & 18 & 73 & 2 + ngc 5194 @xmath68 @xmath67 & 13 29 52.7 & + 47 11 42 & 6.77 & sy2 & 44 & 3.0 & 34 & 3.9 & 19 & 2 + ngc 5195 & 13 29 59.6 & + 47 15 58 & 8.00 & liner & 18 & @xmath01.0 & 6.9 & & @xmath01.9 & 5 + 3c 286 & 13 31 08.3 & + 30 30 32 & 5399 & sy1 & & 0.75 & 0.69 & & & 6 + ngc 5218 & 13 32 10.4 & + 62 46 03 & 41.1 & & 45 & @xmath00.16 & 4.9 & @xmath00.57 & @xmath02.3 & 2 + mcg 06 - 30 - 015 @xmath67 & 13 35 53.8 & 34 17 44 & 33.4 & sy1 & 5.0 & 5.0 & 5.9 & 7.4 & 26 & 4 + iras f13342 + 3932 & 13 36 24.1 & + 39 17 31 & 866 & sy1 & 5.7 & 3.5 & 5.0 & 4.2 & 10 & 3 + ngc 5236 & 13 37 00.9 & 29 51 55 & 3.60 & & 503 & @xmath00.61 & 29 & @xmath01.2 & 5.7 & 7 + mrk 266 @xmath67 & 13 38 17.5 & + 48 16 37 & 116 & sy2 & 57 & 8.0 & 28 & 11 & 53 & 8 + ngc 5273 @xmath68 & 13 42 08.3 & + 35 39 15 & 15.2 & sy1 & 2.0 & @xmath01.5 & 3.4 & @xmath00.88 & 4.9 & 2 + mrk 273 & 13 44 42.1 & + 55 53 12 & 167 & liner & 42 & 12 & 34 & 15 & 56 & 3 + iras f13451 + 1232 & 13 47 33.4 & + 12 17 24 & 565 & sy2 & 5.0 & @xmath01.0 & 5.1 & @xmath02.1 & 2.1 & 3 + ic 4329a @xmath67 & 13 49 19.3 & 30 18 34 & 69.6 & sy1 & 28 & 29 & 57 & 35 & 117 & 4 + pg 1351 + 640 & 13 53 15.8 & + 63 45 45 & 402 & qso & 1.8 & @xmath00.92 & 2.7 & @xmath01.00 & @xmath00.90 & 1 + ngc 5347 @xmath67 & 13 53 17.8 & + 33 29 26 & 33.6 & sy2 & 4.2 & 2.1 & 4.1 & @xmath01.7 & 7.6 & 4 + ngc 5371 & 13 55 39.9 & + 40 27 42 & 37.1 & liner & 1.7 & @xmath00.090 & 1.1 & @xmath00.32 & @xmath00.50 & 2 + mrk 463 & 13 56 02.9 & + 18 22 19 & 222 & sy1 & 9.3 & 18 & 41 & 20 & 69 & 3 + ngc 5395 @xmath68 & 13 58 38.0 & + 37 25 28 & 50.3 & sy2 & & @xmath00.54 & & @xmath00.31 & @xmath00.18 & 2 + ngc 5398 & 14 01 21.6 & 33 03 49 & 10.8 & & & @xmath00.50 & & & @xmath00.31 & 5 + ngc 5408 & 14 03 20.9 & 41 22 39 & 4.85 & & & @xmath00.93 & 1.2 & & @xmath00.51 & 5 + ngc 5427 @xmath68 & 14 03 26.1 & 06 01 50 & 37.5 & sy2 & 10 & 1.7 & 5.1 & 1.2 & 4.2 & 2 + ngc 5474 & 14 05 01.6 & + 53 39 44 & 6.40 & & & @xmath00.83 & & & @xmath00.31 & 5 + iras f14070 + 0525 & 14 09 31.3 & + 05 11 31 & 1341 & sy2 & 1.8 & @xmath00.15 & 0.58 & @xmath00.60 & @xmath01.5 & 3 + 3c 295 & 14 11 20.6 & + 52 12 09 & 2571 & sy2 & 0.060 & 0.070 & 0.050 & & & 6 + circinus @xmath68 & 14 13 09.9 & 65 20 20 & 6.09 & sy2 & 427 & 219 & 379 & 261 & 871 & 2 + ngc
5506 @xmath68 @xmath67 & 14 13 14.9 & 03 12 27 & 26.1 & sy1 & 81 & 58 & 151 & 63 & 239 & 2 + pg 1411 + 442 & 14 13 48.3 & + 44 00 13 & 411 & qso & 0.36 & 0.96 & 0.92 & 0.55 & 1.5 & 1 + ngc 5548 @xmath67 & 14 17 59.5 & + 25 08 12 & 74.5 & sy1 & 8.5 & 5.4 & 7.3 & 3.9 & 17 & 4 + pg 1426 + 015 & 14 29 06.6 & + 01 17 06 & 392 & qso & 1.3 & 1.2 & 2.3 & 0.75 & 3.4 & 1 + ngc 5653 & 14 30 10.4 & + 31 12 55 & 50.6 & & 70 & @xmath00.55 & 5.9 & @xmath00.25 & @xmath01.2 & 2 + ngc 5643 @xmath68 & 14 32 40.8 & 44 10 28 & 16.9 & sy2 & 38 & 25 & 56 & 30 & 121 & 2 + mrk 817 @xmath67 & 14 36 22.1 & + 58 47 39 & 138 & sy1 & 3.8 & 1.9 & 4.6 & 3.6 & 6.5 & 4 + iras f143481447 & 14 37 38.4 & 15 00 22 & 378 & liner & 11 & @xmath00.21 & 2.6 & @xmath01.5 & @xmath03.3 & 3 + ngc 5713 & 14 40 11.5 & 00 17 20 & 29.4 & & 127 & @xmath00.80 & 17 & & 2.8 & 5 + iras f143783651 & 14 40 59.0 & 37 04 32 & 306 & liner & 11 & @xmath00.90 & 1.2 & @xmath02.3 & @xmath03.8 & 3 + pg 1440 + 356 & 14 42 07.5 & + 35 26 22 & 358 & qso & 4.1 & 1.3 & 3.9 & 1.9 & 6.3 & 1 + ngc 5728 @xmath68 & 14 42 23.9 & 17 15 11 & 40.9 & sy2 & 28 & 22 & 53 & 27 & 116 & 2 + 3c 303.1 & 14 43 14.9 & + 77 07 28 & 1358 & sy2 & 0.040 & @xmath00.020 & 0.070 & & 0.090 & 6 + ngc 5734 & 14 45 09.1 & 20 52 13 & 58.0 & & 77 & @xmath00.96 & 8.2 & @xmath00.23 & 0.93 & 2 + ngc 5743 & 14 45 11.0 & 20 54 48 & 59.7 & & 50 & 1.3 & 8.9 & 0.78 & 3.2 & 2 + pg 1448 + 273 & 14 51 08.8 & + 27 09 26 & 292 & qso & 0.51 & 2.7 & 3.1 & 4.1 & 10 & 1 + ic 4518w & 14 57 41.2 & 43 07 55 & 68.5 & sy1 & 40 & 25 & 49 & 24 & 85 & 2 + ic 4518e & 14 57 44.6 & 43 07 53 & 65.4 & & 18 & @xmath00.63 & 1.7 & @xmath00.26 & 0.33 & 2 + 3c 309.1 & 14 59 07.6 & + 71 40 19 & 5833 & sy2 & & 0.30 & 0.40 & & & 6 + iras f15001 + 1433 & 15 02 31.9 & + 14 21 35 & 781 & sy2 & 6.8 & 1.1 & 2.6 & 0.66 & 1.7 & 3 + pg 1501 + 106 & 15 04 01.2 & + 10 26 16 & 158 & qso & 3.6 & 8.0 & 11 & 8.0 & 24 & 1 + ngc 5866 & 15 06 29.5 & + 55 45 47 & 15.1 & & 7.8 & @xmath00.97 & 4.9 & & 0.93 & 5 + iras f150912107 @xmath67 & 15 11 59.8 & 21 19 01 & 198 & sy1 & 12 & 8.5 & 16 & 8.1 & 31 & 4 + cgcg 049 - 057 & 15 13 13.1 & + 07 13 31 & 55.8 & & 20 & @xmath00.58 & 1.5 & @xmath00.82 & @xmath00.59 & 2 + ngc 5899 @xmath68 @xmath67 & 15 15 03.2 & + 42 02 59 & 37.0 & sy2 & 11 & 6.8 & 16 & 6.7 & 22 & 2 + ngc 5907 & 15 15 53.8 & + 56 19 43 & 12.1 & & 6.1 & @xmath00.0100 & 1.3 & @xmath00.070 & 1.6 & 7 + vv 705 & 15 18 06.3 & + 42 44 36 & 178 & & 62 & @xmath00.33 & 14 & @xmath00.79 & @xmath02.8 & 2 + iras f15206 + 3342 & 15 22 38.0 & + 33 31 35 & 580 & & 13 & @xmath00.40 & 21 & @xmath01.5 & @xmath02.4 & 3 + ngc 5929 @xmath67 & 15 26 06.2 & + 41 40 14 & 35.8 & sy2 & 13 & 1.1 & 9.8 & 2.2 & 5.3 & 4 + iras 15250 + 3609 & 15 26 59.4 & + 35 58 37 & 245 & liner & 10 & @xmath01.2 & 2.7 & @xmath02.4 & @xmath01.5 & 3 + ngc 5936 & 15 30 00.8 & + 12 59 21 & 57.1 & & 82 & @xmath00.81 & 6.1 & @xmath00.63 & 1.1 & 2 + 3c 321 & 15 31 43.5 & + 24 04 19 & 441 & sy2 & & 0.70 & 0.57 & & 1.9 & 6 + ngc 5953 & 15 34 32.4 & + 15 11 37 & 28.9 & liner & 51 & 1.4 & 17 & 3.0 & 17 & 2 + arp 220 & 15 34 57.1 & + 23 30 11 & 78.2 & & 65 & @xmath02.9 & 7.8 & @xmath014 & @xmath021 & 3 + iras 153350513 & 15 36 11.7 & 05 23 52 & 119 & & 32 & @xmath00.19 & 6.1 & @xmath00.95 & @xmath01.4 & 2 + 3c 323.1 & 15 47 43.5 & + 20 52 16 & 1341 & sy1 & 0.12 & @xmath00.060 & 0.11 & & 0.17 & 6 + mcg 02 - 40 - 004 @xmath67 & 15 48 25.0 & 13 45 27 & 110 & sy2 & 16 & 6.1 & 8.5 & 3.1 & 13 & 4 + iras f154620450 & 15 48 56.8 & 04 59 33 & 460 & sy1 & 7.4 & @xmath00.30 & 1.4 & @xmath01.0 & @xmath03.6 & 3 + iras fsc154800344 @xmath67 & 15 50 41.5 & 03 53 18 & 133 & sy2 & 5.6 & 6.1 & 9.4 & 8.9 & 35 & 4 + 3c 330 & 16 09 36.6 & + 65 56 43 & 3170 & sy2 & & 0.40 & 0.40 & & & 6 + iras f160900139 & 16 11 40.5 & 01 47 05 & 631 & liner & 7.8 & @xmath00.12 & 2.0 & @xmath02.0 & @xmath01.4 & 3 + pg 1613 + 658 & 16 13 57.2 & + 65 43 09 & 605 & qso & 3.9 & 1.1 & 3.3 & 0.65 & 4.9 & 1 + iras 161640746 & 16 19 11.8 & 07 54 02 & 102 & liner & 47 & 1.2 & 14 & @xmath01.2 & 7.2 & 2 + pg 1617 + 175 & 16 20 11.3 & + 17 24 27 & 520 & qso & 0.29 & @xmath00.17 & 0.36 & 0.28 & 0.39 & 1 + 3c 334 & 16 20 21.8 & + 17 36 23 & 3212 & sy1 & 0.66 & 1.2 & 1.6 & & & 6 + pg 1626 + 554 & 16 27 56.1 & + 55 22 31 & 626 & qso & 0.069 & @xmath00.069 & 0.11 & @xmath00.16 & @xmath00.20 & 1 + 3c 343 & 16 34 33.8 & + 62 45 35 & 6510 & sy1 & 0.70 & 0.83 & 1.4 & & & 6 + ngc 6221 @xmath68 & 16 52 46.1 & 59 13 07 & 20.8 & sy2 & 196 & @xmath01.6 & 24 & @xmath03.1 & 20 & 2 + ngc 6240 & 16 52 58.9 & + 02 24 03 & 105 & liner & 171 & 4.4 & 61 & @xmath05.7 & 27 & 3 + ngc 6285 & 16 58 24.0 & + 58 57 21 & 82.3 & & 18 & @xmath00.080 & 3.0 & @xmath00.48 & 1.2 & 2 + ngc 6286 & 16 58 31.4 & + 58 56 10 & 79.2 & liner & 29 & @xmath00.43 & 3.3 & 1.6 & @xmath00.75 & 2 + pg 1700 + 518 & 17 01 24.8 & + 51 49 20 & 1505 & qso & 1.2 & @xmath00.23 & 1.6 & @xmath00.43 & 1.7 & 1 + 3c 351 & 17 04 41.4 & + 60 44 30 & 1988 & sy1 & 0.100 & 0.22 & 0.21 & & & 6 + iras 171381017 & 17 16 35.8 & 10 20 39 & 74.6 & & 124 & @xmath00.81 & 20 & @xmath01.3 & @xmath03.4 & 2 + ngc 6300 @xmath68 & 17 16 59.5 & 62 49 14 & 15.9 & sy2 & 11 & 13 & 15 & 9.1 & 31 & 2 + iras f17179 + 5444 & 17 18 54.4 & + 54 41 48 & 698 & sy2 & 4.5 & 2.2 & 2.9 & 0.82 & 2.1 & 3 + iras 172080014 & 17 23 22.0 & 00 17 00 & 190 & & 41 & @xmath01.00 & 8.1 & @xmath03.2 & @xmath02.4 & 3 + ngc 6500 & 17 55 59.8 & + 18 20 17 & 43.1 & liner & 4.8 & @xmath00.11 & 2.6 & @xmath00.49 & @xmath00.35 & 2 + ic 4687 & 18 13 39.6 & 57 43 31 & 73.8 & & 146 & @xmath00.80 & 45 & @xmath01.2 & 2.5 & 2 + ic 4710 & 18 28 38.0 & 66 58 56 & 9.00 & & 0.88 & @xmath01.0 & 4.3 & & @xmath00.30 & 5 + 3c 381 & 18 33 46.3 & + 47 27 02 & 771 & sy1 & 0.060 & 0.070 & 0.13 & & 0.22 & 6 + eso 103-g035 & 18 38 20.3 & 65 25 39 & 56.2 & sy2 & & 18 & 41 & & & 6 + ic 4734 & 18 38 25.7 & 57 29 25 & 66.7 & & 65 & @xmath00.48 & 6.1 & @xmath00.66 & @xmath01.8 & 2 + ngc 6701 & 18 43 12.5 & + 60 39 12 & 56.2 & & 73 &
@xmath00.59 & 7.0 & @xmath00.63 & 2.1 & 2 + ngc 6744 & 19 09 46.1 & 63 51 27 & 9.90 & liner & 1.1 & @xmath00.030 & 1.5 & @xmath00.13 & @xmath00.73 & 7 + eso 593-ig008 & 19 14 30.9 & 21 19 07 & 218 & & 32 & @xmath00.72 & 8.9 & @xmath00.29 & @xmath00.78 & 2 + eso 141-g055 @xmath67 & 19 21 14.2 & 58 40 13 & 158 & sy1 & 2.2 & 2.2 & 5.6 & 1.6 & 7.3 & 4 + iras f192547245 & 19 31 21.4 & 72 39 18 & 283 & sy2 & 31 & @xmath02.8 & 13 & @xmath01.6 & 6.3 & 3 + iras f192970406 & 19 32 21.3 & 03 59 56 & 392 & & 18 & @xmath00.92 & 2.5 & @xmath02.2 & @xmath00.90 & 3 + ngc 6814 @xmath68 & 19 42 40.6 & 10 19 24 & 22.4 & sy1 & 7.2 & 3.2 & 15 & 6.1 & 27 & 2 + ngc 6810 @xmath67 & 19 43 34.4 & 58 39 20 & 29.2 & sy2 & 103 & @xmath01.1 & 13 & @xmath02.3 & 2.5 & 4 + ngc 6860 @xmath67 & 20 08 46.9 & 61 06 00 & 64.5 & sy1 & 5.6 & 2.8 & 6.7 & 2.4 & 12 & 4 + iras f200870308 & 20 11 23.9 & 02 59 50 & 490 & liner & 14 & @xmath00.75 & 1.6 & @xmath01.9 & @xmath01.6 & 3 + iras f201004156 & 20 13 29.5 & 41 47 34 & 610 & & 7.3 & @xmath00.48 & 2.8 & @xmath01.3 & @xmath04.8 & 3 + ngc 6890 @xmath67 & 20 18 18.1 & 44 48 25 & 34.8 & sy2 & 11 & 5.8 & 6.6 & 3.8 & 10 & 4 + ngc 6946 & 20 34 52.3 & + 60 09 14 & 6.80 & & 430 & @xmath01.5 & 39 & & @xmath07.2 & 5 + ngc 6951 @xmath68 & 20 37 14.1 & + 66 06 20 & 19.8 & sy2 & 40 & @xmath01.8 & 11 & @xmath02.0 & 8.2 & 2 + mrk 509 @xmath68 @xmath67 & 20 44 09.7 & 10 43 24 & 151 & sy1 & 11 & 5.1 & 16 & 7.4 & 28 & 2 + iras f204141651 & 20 44 18.2 & 16 40 16 & 397 & liner & 6.8 & 1.00 & 1.6 & @xmath01.5 & @xmath01.8 & 3 + ic 5063 @xmath67 & 20 52 02.3 & 57 04 07 & 49.0 & sy2 & 27 & 30 & 66 & 24 & 114 & 4 + iras f205514250 & 20 58 26.8 & 42 39 00 & 190 & & 13 & @xmath00.75 & 2.8 & @xmath01.5 & @xmath02.0 & 3 + mrk 897 @xmath67 & 21 07 45.8 & + 03 52 40 & 115 & sy2 & 24 & 1.1 & 4.4 & @xmath00.80 & 0.62 & 4 + 3c 433 & 21 23 44.5 & + 25 04 11 & 470 & sy2 & 1.9 & 2.7 & 5.2 & & @xmath07.9 & 6 + pg 2130 + 099 & 21 32 27.8 & + 10 08 19 & 283 & qso & 1.4 & 3.7 & 5.7 & 4.0 & 11 & 1 + ngc 7130 @xmath68 @xmath67 & 21 48 19.5 & 34 57 04 & 69.8 & sy2 & 82 & 7.1 & 27 & 4.2 & 15 & 2 + ngc 7177 & 22 00 41.2 & + 17 44 17 & 16.1 & & 4.2 & @xmath00.11 & 1.5 & @xmath00.48 & 0.54 & 2 + ngc 7172 @xmath68 @xmath67 & 22 02 01.9 & 31 52 11 & 37.2 & sy2 & 31 & 8.9 & 17 & 11 & 40 & 2 + b2 2201 + 31a & 22 03 15.0 & + 31 45 38 & 1523 & qso & 0.096 & 0.53 & 0.32 & @xmath00.22 & 0.56 & 1 + iras f22017 + 0319 @xmath67 & 22 04 19.2 & + 03 33 50 & 274 & sy2 & 5.9 & 8.3 & 14 & 9.4 & 29 & 4 + ngc 7213 @xmath68 @xmath67 & 22 09 16.2 & 47 10 00 & 25.8 & sy1 & 24 & @xmath00.94 & 13 & @xmath00.73 & 2.4 & 2 + ic 5179 & 22 16 09.1 & 36 50 37 & 48.4 & & 113 & @xmath00.44 & 11 & @xmath00.81 & 2.0 & 2 + pg 2214 + 139 & 22 17 12.3 & + 14 14 20 & 297 & qso & 0.23 & 0.27 & 0.63 & @xmath00.31 & 1.3 & 1 + ngc 7252 & 22 20 44.8 & 24 40 41 & 66.4 & & 42 & @xmath00.33 & 3.7 & @xmath00.51 & 1.3 & 8 + 3c 445 @xmath67 & 22 23 49.6 & 02 06 12 & 251 & sy1 & 2.3 & 2.0 & 6.2 & 5.8 & 23 & 4 + eso 602-g025 & 22 31 25.5 & 19 02 04 & 110 & liner & 44 & 1.6 & 7.6 & @xmath02.3 & 6.8 & 2 + ngc 7314 @xmath68 @xmath67 & 22 35 46.2 & 26 03 00 & 20.6 & sy1 & 8.3 & 17 & 24 & 22 & 67 & 2 + ngc 7331 & 22 37 04.1 & + 34 24 56 & 14.5 & liner & 19 & @xmath00.52 & 10 & & 3.0 & 5 + mcg 03 - 58 - 007 @xmath67 & 22 49 37.1 & 19 16 26 & 138 & sy2 & 8.5 & 6.6 & 9.3 & 3.9 & 8.8 & 4 + iras f224911808 & 22 51 49.3 & 17 52 23 & 354 & & 5.4 & @xmath00.45 & 1.9 & @xmath00.90 & @xmath02.4 & 3 + pg 2251 + 113 & 22 54 10.4 & + 11 36 38 & 1709 & qso & 0.17 & 0.49 & 0.80 & 0.63 & 3.1 & 1 + ngc 7410 @xmath68 & 22 55 00.9 & 39 39 40 & 24.0 & sy2 & 3.7 & @xmath01.1 & & @xmath01.9 & @xmath012 & 2 + ngc 7469 @xmath68 @xmath67 & 23 03 15.6 & + 08 52 26 & 69.8 & sy1 & 250 & 15 & 45 & 13 & 31 & 2 + mcg 02 - 58 - 022 & 23 04 43.5 & 08 41 08 & 213 & sy1 & & 2.4 & 8.6 & & & 6 + cgcg 453 - 062 & 23 04 56.5 & + 19 33 08 & 109 & liner & 22 & 2.0 & 6.6 & 2.1 & 5.5 & 2 + ngc 7479 @xmath68 @xmath67 & 23 04 56.6 & + 12 19 22 & 34.2 & sy2 & 9.4 & @xmath02.4 & 5.8 & @xmath02.8 & 4.9 & 2 + ngc 7496 @xmath68 @xmath67 & 23 09 47.3 & 43 25 40 & 22.8 & sy2 & 46 & @xmath00.41 & 6.3 & @xmath00.98 & @xmath00.48 & 2 + iras f231285919 & 23 15 46.8 & 59 03 15 & 199 & sy2 & 27 & 2.6 & 20 & 3.0 & 18 & 3 + ngc 7552 & 23 16 10.8 & 42 35 05 & 21.0 & liner & 834 & @xmath03.6 & 70 & & @xmath010 & 5 + ngc 7591 & 23 18 16.3 & + 06 35 08 & 70.7 & liner & 56 & @xmath00.62 & 5.5 & @xmath00.76 & @xmath00.54 & 2 + ngc 7582 @xmath68 @xmath67 & 23 18 23.5 & 42 22 14 & 22.7 & sy2 & 248 & 36 & 102 & 60 & 220 & 2 + ngc 7590 @xmath68 @xmath67 & 23 18 54.8 & 42 14 20 & 23.0 & sy2 & 6.8 & @xmath00.36 & 3.7 & @xmath00.52 & 3.0 & 2 + iras f232306926 & 23 26 03.6 & 69 10 18 & 490 & liner & 7.4 & @xmath00.75 & 2.0 & @xmath01.2 & @xmath01.5 & 3 + ngc 7674 @xmath67 & 23 27 56.7 & + 08 46 44 & 127 & sy2 & 20 & 21 & 35 & 17 & 49 & 4 + iras f232535415 & 23 28 06.1 & 53 58 31 & 610 & sy2 & 5.5 & @xmath00.33 & 1.9 & 1.2 & 1.2 & 3 + ngc 7714 & 23 36 14.1 & + 02 09 18 & 38.2 & liner & 103 & @xmath01.0 & 77 & @xmath01.8 & 5.5 & 8 + iras f23365 + 3604 & 23 39 01.3 & + 36 21 08 & 287 & liner & 8.6 & @xmath00.80 & 0.73 & @xmath00.54 & @xmath02.0 & 3 + ngc 7743 @xmath68 & 23 44 21.1 & + 09 56 02 & 24.3 & sy2 & 4.1 & @xmath00.28 & 4.2 & 0.95 & 2.9 & 2 + cgcg381 - 051 @xmath67 & 23 48 41.7 & + 02 14 23 & 134 & sy2 & 19 & @xmath00.70 & 1.4 & @xmath01.2 & @xmath01.3 & 4 + ngc 7769 & 23 51 04.0 & + 20 09 01 & 60.2 & & 13 & @xmath00.66 & 3.4 & @xmath00.34 & 0.41 & 2 + ngc 7771 & 23 51 24.9 & + 20 06 42 & 61.9 & & 108 & @xmath00.71 & 8.8 & @xmath00.51 & 1.2 & 2 + pg 2349 - 014 & 23 51 56.1 & 01 09 13 & 840 & qso & 1.4 & 0.71 & 2.0 & @xmath00.96 & 3.9 & 1 + iras f23498 + 2423 & 23 52 26.0 & + 24 40 16 & 1046 & sy2 & 3.2 & 1.1 & 2.7 & 2.0 & 5.0 & 3 + ngc 7793 & 23 57 49.8 & 32 35 27 & 3.80 & & 10 & @xmath01.0 & 2.6 & & @xmath00.32 & 5 | we used /irs spectroscopic data on 426 galaxies including quasars , seyferts , liner and hii galaxies to investigate the relationship among the mid - ir emission lines .
there is a tight linear correlation between the [ ] 14.3 and 24.3 ( 97.1ev ) and the [ ] 25.9 ( 54.9ev ) high - ionization emission lines .
the correlation also holds for these high - ionization emission lines and the ( 41ev ) emission line , although only for active galaxies .
we used these correlations to calculate the [ ] excess due to star formation in seyfert galaxies .
we also estimated the [ ] luminosity due to star formation in active galaxies and determined that it dominates the [ ] emission only if the contribution of the active nucleus to the total luminosity is below 5% .
we find that the agn dominates the [ ] emission in most seyfert galaxies , whereas star - formation adequately explains the observed [ ] emission in optically classified hii galaxies .
finally we computed photoionization models to determine the physical conditions of the narrow line region where these high - ionization lines originate .
the estimated ionization parameter range is -2.8 @xmath0 @xmath1 @xmath2 @xmath0 -2.5 and the total hydrogen column density range is 20 @xmath0 @xmath1 @xmath3 ( @xmath4 ) @xmath0 21 . |
ever since the first polarization measurements have been made for astronomical objects , scientists are using polarimetry as a tool to probe astrophysical processes in celestial environments .
it has been the case for the comets since the first observation of the great comet c/1819 n1 tail polarization by arago ( von humboldt 1858 ) and his conclusions that comet was not a self - luminous object and that its polarization was the result of reflection of the sunlight on material lost by the comet .
it has also been the case for the interstellar dust , evidenced independently by hall ( 1949 ) and hiltner ( 1949 ) from their measurements of stars polarization that was interpreted as starlight scattered by elongated small particles aligned in the magnetic field ( greenberg 1951 ) .
nowadays , among the various processes that may cause polarization from remote astronomical sources , light scattering by small solid particles is expected to be the most common case .
such cosmic dust may appear naturally under the form either of dispersed particles ( comet tails , reflection nebula , etc . ) or of deposited particles ( regoliths , pond - like deposits , etc . ) .
information about shape and composition of the cosmic dust particles should come from theory , which tells us how the light polarization is related to the various particle properties and to the scattering geometries .
+ historically , the theory of light polarization from dusts should start from the successful explanation of the sky polarization by rayleigh ( 1871 ) .
indeed , at a time where the concept of molecules was still unclear ( maxwell 1873 ) , rayleigh studied explicitly light scattered by minute solid particles in order to skirt questions about the nature of the atmospheric particles .
thus , the rayleigh dipolar scattering theory applies fully to any particle much smaller than the wavelength and of refractive index close to 1 possibly the case for dusty media in the infrared domain .
later , mie proposed in 1905 the analytical solution of the scattering of a plane electromagnetic wave by a perfect homogeneous dielectric sphere of any radius .
this solution ( van de hulst 1957 ) provided a real breakthrough in the field , since it allowed for the first time to tackle the relation between polarization from dusts and astrophysical particle sizes ( in all the following , `` particle size '' means the typical radius of the particle , thus we consider particles which are not essentially anisotropic ) . generally , dust particles are not expected to be perfect spheres , then the mie theory is only approximation of the real process though a very important one indeed .
the advent of the computer in the 1970s has allowed to deal with light scattered by other shapes and structures . over the years
, numerical tools were made available to compute light scattered by irregularly shaped particles which were expected to be most naturally occurring in astrophysical conditions .
such numerical simulation codes include t - matrix ( mishchenko et al .
1994 ) , ddscat ( draine 2000 ) , fdtd ( taflove & hagness 2005 ) , to mention just a few . over the years , as we have understood that astrophysical dusts should be made of disordered porous / fluffy particles , scientists came up with various numerical codes to simulate such dust particles . that way , algorithms were developed to simulate formation processes of irregularly shaped dust particles under realistic physical conditions
the most popular code , the cluster - cluster aggregation mode ( jullien & botet , 1987 ) and its variants , provide reliable numerical analogues for astrophysical dusts made of aggregates of small grains .
such aggregates are often fractal hence of fluffy structures , though they may exhibit irregular compact shapes as well , depending on the conditions .
there is a set of commonly accepted terminologies that is worth recalling here , since it may depend on the domain of physics where dust particles are considered . in the context of astrophysics ,
a `` grain '' is tiny solid particle much smaller than a characteristic length scale ( in the present context : much smaller than the wavelength ) .
a grain is generally assumed to be spherical in shape because the surface irregularities of such a minute solid are irrelevant .
an absence of a grain in a compact particle will introduce porosity .
we use the words _ fluffy _ or _ porous _ with the same intention to mean voids in the aggregates .
finally , a grain can be represented by a collection of dipoles to comply with the discrete - dipole approximation ( dda ) , which is the base of the ddscat calculations ( draine 2000 ) .
the dipoles are the electromagnetic counterparts of nanoscopic compact domains of standard shape , occupying a definite position in a three dimensional regular lattice structure and having complex refractive index assigned to each individual dipole ( shen et al .
2008 , 2009 ) .
+ porous / fluffy aggregates are expected to exist everywhere in outer space , from our solar system to the interstellar medium and also in other galaxies . however , in this paper the results that we obtain from our simulations will be compared with the data from comets . indeed , there is huge amount of data for comets observed at various scattering geometries , and the ground - based data are complemented by in situ space measurements .
this provides a very stable and large data base to compare with the simulated results .
existence of such porous / fluffy particles in comets , in interplanetary medium and in our upper atmosphere was confirmed by various in situ ( brownlee 1985 ) and space measurements .
their existence was also predicted by astronomers from the analysis of ground - based data ( greenberg & hage 1990 ; xing & hanner 1997 ; kimura 2001 ; kimura et al .
2006 etc . ) .
both laboratory and computer simulations of solar radiation scattered from cometary dusts have been carried out to understand the structures of cometary dust particles ( wurm & blum 1998 ; gustafson & kolokolova 1999 ; hadamcik et al .
2010,2013,2015 ; xing & hanner 1997 ; kimura 2001 ; petrova , tishkovets & jockers 2004 ; tishkovets , petrova & jockers 2004 ; gupta et al 2006 ; kimura , kolokolova & mann 2006 ; bertini , thomas & barberi 2007 ; kolokolova et al .
2007 ; moreno et al . 2007
; das & sen 2011 ; das et al .
2011 and others ) .
+ in most of the astrophysical conditions , the gases are so diluted that grain aggregation is expected to follow the ballistic aggregation procedure ( meakin 1983 , 1984 ) , that is all the particles and grains move along random straight trajectories . according to the physcial conditions , the process can be either successive sticking of grains on aggregates ( this is the ballistic particle - cluster aggregation model , bpca ) , or diffusion is similar for all the particles and the aggregates grow essentially sticking together ( this is the ballistic cluster - cluster aggregation model , bcca ) .
the bpca aggregates are compact in the fractal sense ( that is fractal dimension of the bpca aggregates is equal to 3 ) , while the bcca aggregates are fractal with the fractal dimension @xmath10 .
laboratory diagnostics of particle coagulation in the solar nebula suggests that the particles grow preferentially under the bcca process ( wurm & blum 1998 ) .
a variant of the bcca is the reaction - limited cluster - cluster aggregation model ( rcca ) . in the rcca , sticking is hard to occur , for example because the aggregates are electrically charged .
then , two fractal particles may randomly collide and interpenetrate because of their fluffy structures , but without sticking because of the electric charges distributed on their grains .
when by chance entanglement is strong , the aggregates may stay long together , and after some time , the relative kinetic energy of the entangled aggregates allow to overcome the repulsive forces and it results in effective contact , that is sticking by van der waals forces .
this process leads to aggregates of fractal dimension @xmath11 . in astrophysical clouds ,
bcca is expected to hold when the grain electric charges are irrelevant , and rcca when the grain electric charges are relevant .
+ as we are here more interested in compact shapes , we may recall that shen et al.(2008,2009 ) considered interesting modifications to standard bpcas in which local grains migrate along the surface of the aggregates by rolling or sliding over the first - contact grain .
these reorganizations produce even more compact aggregates than bpca .
the models are called ballistic aggregation with migration ( bam ) . allowing for one ( bam1 ) or two ( bam2 )
migrations produce aggregates of increasing density .
the bam models are not based on realistic rearrangement of the grains inside the particles , and they are used as `` cheap approach '' ( seizinger & kley 2013 ) to generate aggregates denser than those obtained using the realistic bpca process .
other aggregation models were proposed for specific purposes .
a famous example is the off - lattice eden model , used to simulate tumor growth . in this model
, grains are added randomly on any location tangent to a grain of the aggregate and realizing no overlaps with other grains .
it results in very compact structure with irregular surface . unlike the fractal particles ,
the compact - shape particles are not characterized by their fractal dimension ( since it realizes the maximum value , @xmath12 ) , but by their porosity , that is the fraction of voids inside the particle .
for instance , large bpca aggregates porosity ( fraction of void inside the particle ) is about @xmath13 , whereas the bam2 aggregates are denser with porosity about @xmath14 .
the porosity of eden - model particles is about @xmath15 ( wang et al 1995 ) .
as a starting point of the problem , we consider the list , by order of importance , of the parameters relevant for characterizing the particle scattering profile .
the most important parameter is the total volume of the particle , that is the amount of matter interacting with the electromagnetic wave .
this parameter was already highlighted in the rayleigh theory , and has be written in various forms , worth mentioning under the form of the effective radius , @xmath16 , defined by the equality between the total volume of the matter inside the particle , @xmath17 , and the volume of the perfect sphere of radius @xmath16 , that is : @xmath18 porosity is another parameter ubiquitous in the physics of the disordered small particles , but it is not yet clear whether the parameter is relevant or not for the light scattering process . the present study tries to put ideas in order about that quantity .
the first question is the proper porosity definition .
actually , one should consider two cases : * the closed porosity is the case where the particle can be defined as a compact volume with voids inside it . in this context
, voids are called : ` pores ' .
the porosity , @xmath19 , of a particle of volume @xmath17 is naturally defined as : @xmath20 with @xmath21 the total volume of the pores .
note that @xmath19 depends only on the non - dimensional ratio @xmath22 between the total volume of the pores and the volume of the matter inside a given geometric volume .
+ condensation of droplets in which gas bubbles are dispersed gives an example of such porous particles .
this is the case for ash produced by coal pyrolysis ( jiang 2006 ) .
* the open porosity is the case where the outer surface irregularities are of same order as the overall size of the particle .
such open structures going from surface to inner places are sometimes called ` fjords ' by reference to the long , narrow , inlets going into the coasts of norway ( barabsi & vicsek 1990 ) .
a generic example is the case of fractal particles .
if @xmath23 denotes the fractal dimension of the matter distribution inside the particle , then the total volume of the matter inside a sphere of radius @xmath24 smaller than the overall radius of the particle , scales as @xmath25 , with @xmath26 a numerical constant and @xmath27 the typical size of the grains the particle is made of .
then , generalizing the concept of pores to the general voids inside a meaningful geometrical sphere , the porosity , as defined by eqn .
( [ f ] ) , writes : @xmath28 taking @xmath24 as the typical radius of the particle , one notes that the porosity eqn .
( [ ff ] ) is size - dependent , and goes to @xmath29 for the large particles since @xmath30 .
this is in opposition to the fractal dimension , @xmath31 , which is characteristic of the process and not of the size of the particles . despite the embarrassing possible size - dependence of the porosity
, one can argue that @xmath19 is a relevant parameter for light scattering .
indeed , @xmath17 is the amount of homogeneous matter surrounded by empty space , interacting with the light , and @xmath21 is similarly the amount of voids surrounded by homogeneous matter , interacting with the light ( from babinet s principle ) .
then , using @xmath17 and @xmath19 appears to be the most important shape parameters to categorize light scattering of finite particles . the next step ( not considered in the present work ) would be to consider the various correlations either in the matter or in the void distribution .
in the following , we will discuss the case of the closed porosity , leaving for a companion article the case of the open porosity .
as our knowledge about cometary dust becomes more complete thanks to extensive observations and in - situ cometary missions , the very high - porosity of the particles ejected by comet nucleus , appears to be essential result .
a simple way to model very porous particles is to use the ballistic cluster - cluster aggregation model ( wurm & blum 1998 ) .
this approach may be relevant as the bcca model corresponds to general physical conditions possibly met in the cometary environment ( meakin 1983 , 1984 ) .
however , we have to be aware of a possible flaw : forcing the fractal particles to realize a given value of the particle porosity requires to fix the average radius of the particles , because of eqn .
( [ ff ] ) . here , we take the opposite point of view : we do not mind the fractal structure of the particles , but we fix the porosity of the particles to a given value . taking a model of closed porosity ,
the value of @xmath19 is a constant whatever the size of the particle is . in other words ,
@xmath17 and @xmath21 are the two independent parameters of the model .
this class of model corresponds to the following process : outer layer of the comet nucleus becomes more and more porous through mechanisms which may include release of trapped gases ( including gas sublimation ) , irradiation by solar radiation , interaction with solar wind , etc .
therefore , microscopic voids form naturally down to the penetration depth of heat and solar radiations .
the dust particles are formed in a second process from the breakup of the porous layer by the gas escaping through the inner parts of the nucleus . considering the dda method to obtain light scattering properties ,
one starts from a large number of electric dipoles regularly placed on the nodes of a cubic lattice .
two dipoles are said to be connected if they are nearest neighbors .
the first ensemble of @xmath32 dipoles is arranged initially in the shape of a sphere of radius @xmath33 .
voids are generated according to the following procedure : one dipole is chosen randomly and removed .
if this change leaves the ensemble of the remaining dipoles all connected ( that is every couple of dipoles can be linked by a sequence of connected dipoles ) , one keeps the change .
if it is not the case ( removing that dipole leads to isolated dipole or isolated group of dipoles ) , then removal is canceled , the dipole is put back in its original position . then , one removes another dipole .
the process stops when the desired porosity is obtained .
previously , similar studies were carried out with the dipoles unconnected and hence , such aggregates were unrealistic ( vilaplana et al .
2006 and 2011 ) .
the test for connection of the ensemble of dipoles is realized through epidemic algorithm ( demers et al 1987 ) . more precisely : a dipole is chosen randomly .
then , the list of all the dipoles connected to it is calculated ( infection ) .
one considers all the dipoles in the list and do the same calculation , that is determining all the dipoles , not considered before , and connected to them ( spreading ) . at the end of the process , one knows the list of all the dipoles connected to the very first dipole .
if all the dipoles of the actual ensemble are in the final list , the ensemble is connected .
otherwise the ensemble is disconnected .
+ this model allows to generate solid spherical particles of definite porosity , taking into account cohesion of the structure .
these particles are homogeneous in average because of the random removal process . at the end , most of the pores are totaly included inside the structure .
other pores impinge on the particle surface , but they are a few in number and not relevant ( unlike the ` fjords ' of a fractal structure ) . according to the definition by eqn .
( [ f ] ) , the final ensemble of dipoles contains @xmath34 dipoles .
a slice of a particle of porosity @xmath35 is shown in the fig.1 . in order to calculate the polarization and intensities of scattered radiation
, we use ddscat 7.3 light scattering code .
we find that this code has the desired level of flexibility in order to be applied on the porous aggregate structures that we have reported in the section above .
even if our approach is quite general and can be used similarly in many different situations , we will concentrate on applications to the cometary dust particles since large and reliable data base is currently available .
the dust particles ejected from comets have a wide range of sizes .
we take here the lower limit for the particles diameter : @xmath36 m that is the lower limit used for halley and other types of comets in the past ( sen et al .
ideally the upper limit for size distribution can go up to several micrometers , according to conclusions from ground - based observations of recently studied comets , such as 67p ( hadamcik et al 2011 , 2013 , 2015 ) . also , for the comet 67p , the on - board giada instrument during rosetta mission had detected the presence of large particles ( rotundi et al .
2015 ) , as well as many porous dust particles . for computational reasons ,
we limit here the upper bound of particle size distribution to @xmath37 m , in accordance with many other authors ( das et al .
2011 and references therein ) .
indeed , the size distribution is often represented as an inverse power law , @xmath38 with large index @xmath39 values ranging from @xmath40 to @xmath41 ( das et al 2011 ; kolokolova et al 1997,2001 ; mukai et al 1987 ; sen et al 1991 ) . then , the very large particles are so infrequent that they are expected not to influence the averaged light scattering properties . in our analysis
below , we shall use the following values of the size - distribution exponent , @xmath39 , namely : @xmath42 , @xmath43 and @xmath41 , in order to check how the simulated values can match with the observed polarization data for comets . for the composition of cometary dusts , we choose pyroxene with complex refractive index values as @xmath44 at the wavelength @xmath45 m , and @xmath46 at the wavelength @xmath47 m ( das et al .
2011 ) , with the justification for using this material given in section 4 .
a large data base on cometary polarization for p / halley is available at these two wavelengths . running the ddscat 7.3 code allows to generate the mueller matrix elements @xmath48 @xmath49 etc . needed to compute all the optical quantities , including light polarization ( bohren and hoffman 1983 ; draine and flatau 1994 ) .
the calculations are done performing averages over @xmath50 random orientations , and by integration over the dust size distribution .
the linear polarization , @xmath51 , is calculated using the relation : @xmath52 in principle , one should take into account the @xmath49 coefficient , using instead ( bohren and hofman 1983 ) : @xmath53 however , the random aggregates generated by our code have spherical symmetry in average , then we expect to have : @xmath54 when averaged over angular randomization and the definition by eqn .
( [ p0 ] ) can be used as a good approximation . more precisely , difference between eqns .
( [ p1 ] ) and ( [ p0 ] ) is a factor @xmath55 . in our present simulations , the highest @xmath56 value , that we encountered is @xmath57
this has occurred for the @xmath58 porosity particle at @xmath59 m and with power law index @xmath60 .
this @xmath56 value should influence the estimated polarization value by a multiplicative factor of @xmath61 .
we are very much aware from observational astronomers that the uncertainty in observed polarization is never below @xmath62 , and thus a factor of @xmath63 will have no effect on our simulation and modeling work .
we also checked on some examples that both ( [ p0 ] ) and ( [ p1 ] ) gave very similar values of the polarization .
however , under a different situation , the linear polarization should be expressed by ( [ p1 ] ) .
in recent years vilaplana et al ( 2006 , 2011 ) had addressed this problem of effect of porosity on the polarization ( @xmath51 ) and light scattering efficiency ( @xmath64 ) .
the authors generated particles with fluffiness / porosity by uniformly randomly removing dipoles , without constraining particles to remain connected .
they defined the fluffiness degree as the percentage of dipoles removed and generated aggregates with fluffiness degrees as @xmath65 and @xmath66 .
as can be seen in the figs .
9 , 10 and 11 of vilaplana et al.(2011 ) , a displacement to the right and rising of the first maximum of @xmath64 is observed as the fluffiness degree increases .
the result is that both the real and imaginary parts of the _ effective _ refractive index decrease simultaneously as the fluffiness degree is increased .
this behaviour can be expected as the authors calculated the effective refractive index of the porous dust by maxwell - garnet ( 1907 ) theory . in the present work ,
we consider connected aggregate structures with various porosities .
we start with compact structures where we have 12,0000 dipoles and then we generate structures with various porosity values @xmath67 and @xmath66 . while running ddscat code , we do averaging over @xmath50 random orientations .
we use the power law size distribution with lower bound and upper bound as @xmath68 and @xmath37 m as discussed earlier .
we perform integration over this size range numerically , by doing the calculation in 71 steps and then using simpson s 1/3 rule .
the index for power law dust size distribution is considered as 1.8 , 2.8 and 3.8 as these are the most commonly used values in literature ( das et al 2011 and references therein ) . as far the compositions
are concerned , there can be many possible options , and scientists often consider mixtures of various types .
but since we want here to focus our studies on @xmath69 , we restrict ourselves to only one type of composition which is pyroxene marked by the complex refractive index values of ( 1.692 , 0.0492 ) and ( 1.672 , 0.0185 ) at the wavelengths @xmath70 m and 0.684 @xmath71 m respectively ( das et al .
this material is most commonly used as a candidate for cometary dust . in this section
we discuss the effect of porosity , @xmath19 , on cometary polarizations , with the porosity defined generally through eqn .
( [ f ] ) .
it was noticed ( das et al 2011 ) that the linear - polarization curve of the halley s comet is bell - shaped which is a general feature among all the comets , with max - value , @xmath72 , and min - value , @xmath73 , about @xmath74 and @xmath75 respectively . however ,
these values are not very accurate , since it was not possible to make polarization measurements at all the scattering angles .
we should also note that @xmath72 can be even higher than @xmath76 and @xmath73 even more negative than @xmath77 .
crossover from positive to negative polarization values takes place at a scattering angle @xmath78 degrees . in the figs .
2- 7 , we draw simulated polarization curves at various porosities and various power law indices of the particle size distribution .
we find that with the increase in porosity ( from @xmath79 to @xmath35 ) both the maximum and minimum polarization values change . the negative polarization branch becomes deeper with the increase in porosity and
then again it becomes shallower .
this non - monotonic phenomenon appears for both the wavelengths @xmath45 m and @xmath47 m , and with the three values of power law index @xmath80 and 3.8 . among all the different sets of parameters considered here ,
the matching with the observed polarization data of comet is better obtained with 40% porosity and @xmath81 power law index at both the wavelengths . in table
i , we list only those simulated polarization values which are close to the above observed polarization values of @xmath72 , @xmath73 and consistent with crossover angles .
we could even fine tune the porosity and power law index values , and try to match the observed polarization data more closely . by a @xmath82 minimization technique
, one can find the optimum values of porosity ( @xmath19 ) and power law index ( @xmath39 ) , which realize the best fit to the observed polarization data .
this we postpone for future , which will be more detailed work .
+ way back in 1987 , immediately after the apparition comet halley , mukai et al ( 1987 ) , carried out simulation work by varying the refractive index ( that is composition ) of cometary dust .
as one can see from the fig . 1 of their article , various shapes for the polarization versus scattering - angle curve are possible , as we change the refractive index of cometary dust .
in the present work , we can see in a similar way that various shapes for the polarization versus scattering - angle curve are possible by changing the porosity @xmath19 . these two results may be related since an effective refractive index can be found dependent on the actual porosity using the maxwell - garnett effective medium theory . however , this procedure may be somewhat empirical and it will be discussed in more details below in the section 5 . on the other hand ,
the procedure developed in the present work , provides a definite and analytical way of dealing with porous dust .
in addition to polarization features , comets also show interesting properties of photometric _
color_. since , there are many cometary molecules having their own emission bands within the visible region , we have to observe the comets only through some specific wavelength bands , where we can measure the light scattered by cometary dust alone and uncontaminated by cometary molecular emissions .
the ratio of the scattered intensities as observed through such two continuum bands , when expressed in suitable log scale is called _
kolokolova et al ( 1997 ) had worked on cometary color and elaborated on the definition of the concept of cometary color . as stated by the authors ,
the _ color _ , @xmath83 , can be suitably defined as : @xmath84 the quantity @xmath85 and @xmath86 are the two recorded intensities of light scattered by cometary dusts through the two wavebands centered around @xmath45 m and @xmath47 m respectively . in other words ,
cometary _ color _ can be measured simply by the relation : @xmath87 where the b(0.485 ) and r(0.684 ) are the cometary magnitudes observed through the above cometary filters , which were standardized by international halley watch ( ihw ) and widely used for cometary photometry . the _ color _ , @xmath88 ,
can be estimated theoretically using eqn .
( [ d ] ) by considering certain specific cometary dust population .
the detailed procedure has been described by kolokolova et al ( 1987 ) . from their equations ( 9 ) and ( 10 )
, we arrive at the following two equations below , which will be applicable directly in our case : @xmath89 @xmath90 where @xmath91 is the total intensity scattering matrix element obtained after running ddscat code at the given wavelengths @xmath92 m and @xmath47 m and for the given dust radius @xmath33 expressed in @xmath71 m .
the dust size follows the power law distribution @xmath38 .
the lower and upper limit of dust size distributions are @xmath93 and @xmath94 are respectively @xmath36 m and @xmath37 m in the present case . + here ,
the _ color _ values are calculated from the simulation data using the eqs .
( [ d ] ) , ( [ i1 ] ) and ( [ i2 ] ) for different porosity values of cometary dust .
the _ _ color__s are plotted against the scattering angles on the figs .
8 , 9 , 10 for different power law indices .
looking at these figures , we find that the _ color _ values , @xmath88 , remain almost constant for different scattering angles at a given porosity , except sudden increase in the domain of the negative branch of the polarization curve ( that is for scattering angles > 160 degrees ) . still more remarkable , the constant value of the _ color _ with the scattering angle does not change appreciably for the different porosities at a given power index value .
one can conclude from these remarks that the color of the cometary dust depends essentially on the size - distribution of the dust particles and not on their porosity , with the noticeable exceptions for the backward scattering directions .
this is confirmation of previous conclusions by kolokolova et al ( 1997 ) . indeed , kolokolova et al .
( 1997 , 2001 ) using power law index @xmath95 or still higher , concluded that the cometary @xmath96 is mainly determined by the size distribution of the cometary dust , whereas the cometary polarization is mainly determined by the complex refractive index .
+ however , we can not explain quantitatively at the same time both the polarization curve of the comet halley and its _
color_. indeed , the average @xmath96 of the comet is known to be @xmath97 ( kolokolova et al 1997 ) , which is a value much smaller than the @xmath96 found in the simulations for the best fit of the polarization curve ( namely : @xmath98 for porosity @xmath99 and power law index @xmath81 ) .
actually , as the estimated @xmath96 values are the @xmath100 , @xmath101 and @xmath102 respectively for the three power law index values @xmath80 and 3.8 , then the @xmath103 is always found to be much more negative than the observed halley s
comet @xmath96 .
a possible interpretation of the numerical data is that big dust particles ( more numerous for the small values of @xmath39 ) tend to be white , whereas small - size particles ( dominant for the large values of @xmath39 ) are essentially reddish .
therefore , the number of large particles may well have been underestimated compared to the real halley comet dust size - distribution . larger value of the size cutoff @xmath94 should be tempted , but that means adding a free parameter to the model , then making the simulations and their analysis much more lengthy . by the way , let us note that , till today , there have been no successful attempts where a single dust model could explain both polarization and @xmath96 .
as mentioned in the previous sections and also discussed by maxwell - garnett ( 1907 ) , the effect of porosity , @xmath19 , of dust particles can be approximately replaced by replacing the actual matter refractive index @xmath104 for an equivalent homogeneous matter of refractive index @xmath105 .
such approach is known as _ effective medium theory_. various schemes are available for determining effective refractive index . following voshchinnikov
et al ( 2007 ) , we adopt here the mixing rule of bruggeman , since it has been found to work for wide porosity range and deviations to the exact results do not exceed @xmath106 .
this symmetric mixing rule is stated below with our set of symbols : @xmath107 solving this complex - value equation allows to compute the effective refractive index for any set of the parameters @xmath108 .
then , using the mie result for the homogeneous spheres with radii distributed as @xmath109 , approximate averaged polarization and _ color _ of the ensemble of porous particle characterized by @xmath110 and @xmath109 can be calculated .
this scheme was used previously for polarization ( _ viz_. sen et al .
+ the values of the effective refractive index are shown in the table 2 . the corresponding polarization and _ color _ values
are plotted along with those previously obtained using porous aggregates in the figs .
3 - 11 . from these figures
it is clear that the effective medium theory can not exactly reproduce the numerical polarization and color values obtained using dda method on the simulated porous aggregates , though most of the data are comparable .
the differences become more noticeable , as the porosity increases .
vaidya et al ( 2007 ) had also observed similar limitations of effective medium theory .
m - @xmath37 m , with a power - law size index 1.8 and wavelength @xmath111 = 0.485 @xmath71 m .
the curve marked by o is the one drawn with the dda method , and the curve joined by + is the one drawn with the mie theory . ]
m - @xmath37 m , with a power - law size index @xmath112 and wavelength @xmath113 = 0.485 @xmath71 m .
the curve ( marked by o symbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
m - @xmath37 m , with a power - law size index @xmath81 and wavelength @xmath113 = 0.485 @xmath71m.the curve ( marked by osymbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
m - @xmath37 m , with a power - law size index @xmath42 and wavelength @xmath113 = 0.684 @xmath71 m .
the curve ( marked by o symbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
m - @xmath37 m , with a power - law size index @xmath112 and wavelength @xmath114 = 0.684 @xmath71 m .
the curve ( marked by o symbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
m - @xmath37 m , with a power - law size index @xmath81 and wavelength @xmath113 = 0.684 @xmath71 m .
the curve ( marked by o symbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
.list of certain porosity values with the corresponding simulated @xmath72 ( in @xmath115 ) , @xmath73 ( in @xmath115 ) values and the crossover angle ( in degrees ) between the positive and the negative polarization branches , from bell - shaped polarization curves for comet .
only those simulated values are tabulated which are close to the observed polarization values . [ cols="<,^,^,^,>",options="header " , ] m , with a power - law size index @xmath116 .
the curve ( marked by o symbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
m , with a power - law size index @xmath117 .
the curve ( marked by o symbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
m , with a power - law size index @xmath118 .
the curve ( marked by o symbols ) is the one drawn using the dda method , and the curve ( joined by + symbols ) is the one drawn using the mie theory . ]
the aim and main emphasis of the present work , was to develop a new generic algorithm able to generate realistic porous aggregates with tunable porosity for a given dust size distribution .
we focused on porous spheres , but generalization to other shapes ( e.g. porous ellipsoids ) is straightforward . using the ddscat light scattering code
, we showed that such dust particles can generate meaningful polarization values for comets , and hence provide a tool to study in a systematic way the effect of dust porosity on cometary light scattering phenomenon , such as polarization , _ color _ , etc . from the analysis of the simulated polarization curves
, we find that with the increase in porosity ( from @xmath119 to @xmath58 ) both the maximum and minimum polarization values change in a non - monotonic way . indeed , the negative polarization branch becomes deeper with the increase in porosity and then again it becomes shallower .
this phenomenon is true for both the wavelengths @xmath120 m and @xmath121 m and with the three values of power law index @xmath80 and @xmath41 considered here .
+ more specifically , we have successfully generated polarization values which match grossly with the observed polarization data for comet halley .
it appears that the matching with the observed polarization data of comet halley is better with @xmath122 porosity and @xmath123 radius - distribution power law index .
a sequel work is planned to find the optimum values of the two parameters @xmath124 which realize the least square fitting to the observed data . on the other hand , the simulated cometary _ color _
values show only weak dependence on the porosity values or on the scattering angles ( except when close to back - scattering ) . at the porosity @xmath122 and power law index @xmath125 , the simulated _ color _ value @xmath126 does not match with the average observed _ color _ value tabulated earlier by kolokolova et al ( 1987 ) which is @xmath127 .
this discrepancy may come from the underestimation of the upper limit of the particle radius or from the size dependence of single particle @xmath128 . generally , the simulated _ color _ values solely appear to be dependent on the power law index of dust size distribution function .
thus the dust _ color _ depends on dust particle size and this finding confirms a trend reported by kolokolova et al ( 1997 ) earlier . at last , we showed that the effective medium theory can only be used qualitatively to understand various features in the polarization or _ color _ behaviors .
if quantitative achievement is required , more detailed theory has to be used , such as the dda method applied to porous particles as generated by the algorithm given in the present paper .
r.b . thanks the assam university , silchar , for hospitality in the framework of ` global initiative for academic network ' ( gian ) programme from mhrd , april 2016 .
the authors aks and rv thank erasmus mundus - namaste programme for funds to do this collaborative work . 99 barabsi a.l . , vicsek t. , 1990 , j. phys.a , 23 , l729 bertini i. , thomas n. , barberi c. , 2007 , a&a , 461 , 351 bohren , c. f. , hofman , d. r. , 1983 . absorption and scattering of light by small particles .
wiley , new york .
brownlee d. e. , 1985 , ann .
earth planet .
, 13 , 147 das h.s . ,
d. paul , a. suklabaidya and a. k. sen,2011 , mnras 416,94 - 100 das h.s . , sen a. k. , 2011 , j. quant .
radiative transfer , 112 , 1833 demers a. , greene d. , hauser c. , irish w. , larson j. , shenker s. , sturgis h. , swinehart d. , terry d. in proceedings of the 6th annual acm symposium on principles of distributes computing ( podc 87 ) , 1987 .
p. 1 - 12
draine b.t .
, flatau p.j .
discrete - dipole approximation for scattering calculations .
j opt soc am a 1994;11:1491 - 1499 .
draine b.t . in : mishchenko m.i .
, hovenier , j.w . ,
travis , l.d . , editors .
light scattering by nonspherical particles , theory , measurements , and applications .
new york : academic press , 2000 .
p. 131 - 145
gupta r. , vaidya d. b. , bobbie j.s ; chylek petr , astrophysics & space science , 301 , issue 1 , pp .
21 - 31 , 2006 gustafson b.s . , kolokolova l. , 1999 , j. geophys .
res . , 104 , 31711 greenberg j. m. , & hage j. i. , 1990 , apj , 361 , 260 hadamcik , e . , sen , a.k .
, levasseur - regourd , a.c . , gupta , r .
, lasue , j .
, 2010.astron.astrophys.517,a86 .
hadamcik , e . ,
sen , a.k .
, levasseur - regourd , a.c . ,
gupta , r .
, lasue , j . , botet , r.,2013 , icarus222,774 - 785 .
hadamcik , e . ,
sen , a.k .
, levasseur - regourd , a.c . ,
roychoudhury , s . ,
lasue , j . , gupta , r. , botet , r.,2014 .
meteor.planet.sci.49(nr 1 ) , 36 - 44 . hall , j.s .
1949 , sci , 109 , 166 hiltner , w.a . 1949 , sci , 109 , 165 von humboldt , a.1858 , cosmos - a sketch of the physical description of the universe - . reprinted the johns hopkins university press , 1997 jiang , s.z . , editor .
focus on combustion research .
new york : nova science publishers , 2006 jullien , r. , botet , r. 1987 , aggregation and fractal aggregates , world scientific publishing , singapore kimura h. , 2001 , j. quant .
transfer , 70 , 581 kimura h. , kolokolova l. , mann i. , 2006 , a&a , 449 , 1243 kolokolova l. , kimura h. , kiselev n. , rosenbush v. , 2007 , a&a , 463,1189 kolokolova l , jockeres k , gustafson bo a s , and gunther l , journal of geophysical research , 2001 , vol .
e5 , 10,113 - 10,127 kolokolova l , jockers k , chernova g 1997 , icarus 126 , 351 - 361 maxwell j.c .
1873 nature viii , 437 - 441 maxwell - garnett j.c.1904 philos trans r soc a , 203 : 385 - 420 . meakin p. , 1983 ,
j. colloid interface sci .
, 96 , 415 meakin p. , 1984
a , 29 , 997 mishchenko , m. i. , & travis , l. d. 1994 , opt .
, 109 , 16 moreno , f. , muoz , o. , guirado , d. and vilaplana , r. , 2007 , jqsrt , 106 , 348 mukai t , mukai s , & kikuchi s , 1987 , astron .
astrophys . , 187 , 650 petrova e. v. , tishkovets v. p. , jockers k. , 2004 , sol . syst .
38 , 309 rotundi a. , and other 39 authors . , 2015 , science,347,6220 , p.3905 - 1 lord rayleigh . 1871 , philosophical magazine,41 , p.107 - 120 seizinger a. , kley w. , 2013 , astron & astrophys , a65 sen a .k , deshpande m r , joshi u c , rao n k , raveendran a v , 1991 , astron & astrophys , 242 , 496 shen y. , draine b. t. , johnson e. t. , 2008 , apj , 689 , 260 shen y. , draine b. t. , johnson e. t. , 2009 , apj , 696 , 2126 tishkovets v. p. , petrova e.v . , jockers k. , 2004 , j. quant .
transfer , 86 , 241 taflove a. , hagness s.c .
, computational electrodynamics : the finite - difference time - domain method , 3rd ed .
norwood : artech house , 2005 van de hulst h.c .
, light scattering by small particles , wiley , new york , 1957 vilaplanaa r. , moreno f. , molina a. , 2006 , jqsrt,100,415 vilaplana r. , luna r. , guirado , d. , 2011 , jqsrt,112 , 1838 vaidya d.b , gupta r , snow t.p , 2007 mnras 379 , 791 - 800 voshchinnikov n.v .
, videen g. , and henning t. , 2007 applied optic , vol 46 , issue 19 , 4065 wang c. y. , liu p. l. , bassingthwaighte j. b. , 1995 , j. phys.a math .
, 28 , 2141 waterman , p. c. 1971 , phys .
d , 3 , 825 wurm g. , blum j. , 1998 , icarus , 132 , 125 xing z. , hanner m. s. , 1997 , a&a , 324 , 805
if you want to present additional material which would interrupt the flow of the main paper , it can be placed in an appendix which appears after the list of references . | cosmic dusts are mostly responsible for polarization of the light that we observe from astrophysical objects .
they also lead to color - extinction , thermal re - emission and other scattering related phenomena .
dusts are made of small particles which are characterised by their size ( radius ) , composition ( matter ) , and structure ( morphology , including porosity ) . in the present work ,
we address the question of the role of the dust particle porosity on light polarization and color , using discrete dipole approximation ( dda ) light scattering code . to answer this question
, we developed an algorithm to generate solid particles of arbitrary values of porosity . in brief , the model considers a given homogeneous structure made of touching dipoles .
the dipoles are randomly removed one by one , such that the remaining structure remains connected .
we stop the removal process when the desired porosity is obtained
. then we study the optical properties of the porous particle .
that way , we show how the proper value of the porosity affects the polarization and color of the light scattered by these porous particles .
in addition to polarization , porosity has important effects on photometric color . considering an important application
, we emphasize the possible role of the porosity of the cometary dust particles on polarization and color of the light scattered by cometary coma . * the effect of porosity of dust particles on polarization and color with special reference to comets * @xmath0 _ department of physics , assam university , silchar-788001 , india + @xmath1 _
@xmath2 _ laboratoire de physique des solides umr8502 , universit paris - sud , universit paris - saclay , orsay 91405 , france + @xmath3 _ @xmath4 _ dpto .
de fsica aplicada , universidad politcnica de valencia , e.p.s.a .
alcoy 03801 , spain + @xmath5 _
@xmath6 _ department of physics , assam university , silchar-788001 , india + @xmath7 _
@xmath8 _ iucaa , ganeshkhinde , pune 411007 , india + @xmath9 _ date : 3 october 2016 * keywords : * dust : extinction- dust : polarization- dust : color- comet : general- porous- aggregate |
[ cols=">,^,^,^",options="header " , ] [ tab : predictions ] full mars @xcite tracking and showering monte carlo simulations were conducted for 6 gev and 24 gev protons incident on the target , returning predictions for the pion yield and energy deposition densities .
the detailed level of the mars simulations is illustrated by figure [ marsgeom ] , using the example of several 24 gev proton interactions in an inconel band .
figure [ phadron ] shows the corresponding yield and momentum spectra for all hadrons ; figure [ ppion ] gives more detailed information for the pions .
several scatter plots to illustrate the distribution in phase space of the produced pions are displayed in figure [ scatter ] .
the plots are seen to be relatively symmetric in the x and y coordinates , which indicates that any asymmetries due to the band tilt and elliptical beam spot are largely washed out by the large phase space volume occupied by the produced pions .
the yield per proton for positive and negative pions - plus - kaons - plus - muons at 70 cm downstream from the central intersection of the beam with the target was predicted for the kinetic energy range 32@xmath0232 mev that approximates the capture acceptance of the entire cooling channel .
note that the material in the flanges of the i - beam for the inconel and nickel targets was not included in the calculation ; its inclusion might result in a small change in the predicted yield .
table [ tab : predictions ] summarizes the yield and energy deposition results from the mars calculations .
it includes the several rows of derived results that assume the scenario , taken from section [ sec : protons ] , of @xmath1 captured pions .
these derived quantities are identified with a superscript `` 3.2 '' and include : the required number of protons per pulse , @xmath2 , the required total proton pulse energy , @xmath3 , the maximum localized energy deposition in the target material and corresponding temperature rise , @xmath4 and @xmath5 .
approximately 7% of the proton beam energy is deposited in the target .
detailed 3-dimensional maps of energy deposition densities were generated for input to the dynamic target stress calculations that are discussed in the following section .
24 gev protons .
this is a smaller bunch charge than would be typical for muon colliders ; the distribution of stress values will scale in approximate proportion to the bunch charge unless the material s fatigue strength is exceeded .
, height=240 ] 6 gev protons with transverse dimensions as given in table [ target_band_specs ] .
the time origin corresponds to the arrival of the proton pulse .
the stress values are shown for the position of maximum stress in all cases . ,
width=336 ] , but for an incident bunch of @xmath6 24 gev protons . , width=336 ] and [ nick3 ] , for 6 gev and 24 gev proton beams respectively , showing the close correspondence in the stress time development .
, width=336 ] , for @xmath6 24 gev protons on a nickel target , but extended to larger time values to show the dissipation of the shock stresses after multiple reflections from the band surfaces .
, width=336 ] , for @xmath6 24 gev protons on a nickel target but for both 50 ns and 100 ns time steps in the ansys simulation .
the reasonable agreement between the two curves suggests that the normal 100 ns step size is adequately short for approximate stress predictions .
, width=336 ] probably the most critical issue faced in solid - target design scenarios for pion production at neutrino factories or muon colliders is the survivability and long - term structural integrity of solid targets in the face of repeated shock heating . to investigate this , finite element computer simulations of the shock heating stresses
have been conducted using ansys , a commercial package that is widely used for stress and thermal calculations .
the target band geometry was discretised into a 3-dimensional mesh containing approximately 30 000 elements .
this was as fine as the computing capacity and memory allowed and was judged to be adequate for the accurate modeling of shock wave propagation .
the ansys simulations conservatively assumed that the deposited energy is all converted to an instantaneous local temperature rise .
the dynamic stress analyses were preceded by a transient thermal analysis to generate temperature profiles using as input the 3-dimensional energy deposition profiles previously generated by mars for the production assumption of @xmath1 total captured pions ( see the preceding section ) .
dynamic stress calculations were then performed both for a `` free edge '' band , i.e. , with no i - beam flanges , and with a `` fixed edge '' constraint , in which the edges of the band are constrained against displacement in both the radial and axial direction .
the `` free edge '' boundary condition is appropriate for the titanium alloy band ; the `` fixed edge '' model is considered likely to provide an improved approximation to the inconel and nickel bands with their i - beam flanges without requiring the extra computing capacity that would be needed to simulate the more complicated true geometry . the von mises stress (
i.e. , the deviation from the hydrostatic state of stress ) was found to be initially zero but to develop and fluctuate over time as the directional stresses relax or are reflected from material boundaries .
figure [ band_vonmises_initial ] gives an example snapshot of the predicted von mises stress distribution at one microsecond after the arrival of a proton pulse , and the remaining figures [ nick4 ] to [ nick6 ] show various aspects of the predicted stress at the position of maximum stress , respectively : the time development for 6 gev protons and for all three band material candidates ; the same for 24 gev protons ; superimposed plots for 6 gev and 24 gev protons and for the nickel band ; the stress development over a long enough time - span to see the attenuation of the stress levels ; and a check on the time step used in the ansys calculations .
table [ tab : predictions ] summarizes the ansys predictions for the maximum stress created at any time and any position in each of the band materials , @xmath7 .
these values were obtained by reading off from figures [ nick4 ] and [ nick3 ] and then scaling to the bunch charge for a total yield of @xmath1 captured pions .
the final row of table [ tab : predictions ] displays the percentage of the fatigue strength ( from table [ band_materials ] ) that this represents . for the inconel band ,
the calculated fraction of the fatigue strength that the band would be exposed to in this `` worst case '' proton bunch scenario , 53 - 69% , is either close to or slightly above what could be considered a safe operating margin for the target band .
a more definitive determination of the proton beam parameters that allow survivability and adequate safety margins for this target scenario could be provided by data from the ongoing bnl e951 targetry experiment @xcite , with planned stress tests for bunched 24 gev proton beams incident on several types of targets , including inconel 718 .
the inconel target may well be appropriate for some proton beam specifications at a muon collider , and it has already been shown @xcite to likely give a wide safety margin for the more relaxed beam parameters of neutrino factories . the titanium alloy was predicted to have a very conservative safety margin even for the assumed muon collider beam parameters : only 10 - 14% of the fatigue strength . although the yield is about 20% lower than for the other two candidate materials , target bands from titanium alloys look likely to survive with any proton bunch charges that might reasonably be contemplated for muon colliders .
finally , nickel targets are known to evade the predictions for fatigue strength limits , as already mentioned .
test beam experiments would be required to establish the suitability or otherwise of a nickel band production target for any particular muon collider scenario .
all of the above calculations apply for a circumferentially continuous band .
it remains to check the level of von mises stresses at the gaps between the eight welded band sections , although it is noted that the bnl g-2 target was deliberately segmented longitudinally in order to reduce the beam stresses .
for rotating band targets in muon colliders , additional periodic slots in the webbing may also be considered for thermal stress relief and eddy current reduction in rotating band targets for muon colliders .
in summary , the inconel rotating band target design appears to be a promising option for pion production targets at muon colliders . the design concept appears to be manageable from an engineering point of view , and initial simulations of target yields and stresses are encouraging for each of three candidate target materials : inconel 718 , titanium alloy 6al-4v grade 5 and nickel .
priorities for further evaluation of this target scenario include engineering designs of the components , optimization of the band geometry for pion yield and calibration of the target stress predictions to experimental targetry results .
we acknowledge helpful discussions with charles finfrock , george greene and charles pearson , all of bnl .
this work was performed under the auspices of the u.s .
department of energy under contract no .
de - ac02 - 98ch10886 . ,
b.j . king _
et al . _ ,
pac99 , ieee , pp .
3041 - 3 . , b.j .
king , nim a 451 ( 2000 ) pp .
335 - 343 , proc . icfa / ecfa workshop `` neutrino factories based on muon storage rings ( nufact99 ) '' , physics/0005007 . | a conceptual design is presented for a high power pion production target for muon colliders that is based on a rotating metal band .
three candidate materials are considered for the target band : inconel alloy 718 , titanium alloy 6al-4v ( grade 5 ) and nickel .
a pulsed proton beam tangentially intercepts a chord of the target band that is inside a 20 tesla tapered solenoidal magnetic pion capture channel similar to designs previously considered for muon colliders and neutrino factories .
the target band has a radius of 2.5 meters and is continuously rotated at approximately 1 m / s to carry heat away from the production region and into a water cooling tank .
the mechanical layout and cooling setup of the target are described , including the procedure for the routine replacement of the target band .
a rectangular band cross section is assumed , optionally with i - beam struts to enhance stiffness and minimize mechanical vibrations .
results are presented from realistic mars monte carlo computer simulations of the pion yield and energy deposition in the target and from ansys finite element calculations for the corresponding shock heating stresses .
the target scenario is predicted to perform satisfactorily and with conservative safety margins for multi - mw pulsed proton beams .
1.5 cm |
as always in astronomy , the determination of distances is a crucial and difficult problem . until recently , the distance scale of grbs was unknown by over 12 orders - of - magnitude . in 1997
, the discovery of optical and radio counterparts @xcite proved that at least the long - duration bursters were at cosmological distances with redshifts of @xmath3 .
the measurement of grb redshifts requires deep optical spectra , and to date only 24 redshifts are known for bursts with unknown selection effects . if grb distance indicators can be found that use only gamma - ray data ,
then we can measure the demographics and cosmology of large and well - understood samples of bursts .
two such luminosity ( and hence distance ) indicators have recently been proposed .
the first @xcite relates the burst luminosity ( @xmath4 ) with the spectral lag ( @xmath5 ) , which can be idealized as the time between peaks as recorded at high and low photon energies .
the second @xcite relates the luminosity with the variability ( @xmath6 ) , which is a specific measure of the spikiness of the burst light curve .
high luminosity bursts have short lags and spiky light curves , while low luminosity events have long lags and smooth light curves .
these two relations make grbs into standard candles , in the same sense as for cepheids and supernovae where an observed light curve property can yield the luminosity and then the distance . the two luminosity indicators were originally proposed and calibrated with 6 or 7 bursts .
the addition of 3 or 2 further bursts ( for a total of nine ) have fallen on the original relations , hence adding confidence in their utility .
further , if both the @xmath7 and @xmath8 relations are true , then there must be a particular @xmath9 relation , and this prediction has been strongly confirmed with an independent sample of 112 batse bursts @xcite .
finally , the very long lag bursts have been shown to be very low luminosity as demonstrated by their concentration to the local supergalactic plane @xcite .
these luminosity indicators have been used to identify specific bursts @xcite that are at redshifts of @xmath10 as well as to show that the star formation rate of the universe is rising steadily @xcite from @xmath11 to @xmath12 .
this _ letter _
reports on the construction of a hubble diagram ( a plot of luminosity distance , @xmath13 , versus redshift ) as a means of measuring the expansion history of our universe .
only nine grbs have the required information of red shift ( @xmath14 ) , peak flux ( @xmath15 ) , lag time ( @xmath5 ) , and variability ( @xmath6 ) . these data are collected in table 1 , along with the characteristic photon energy ( @xmath16 ) and the observed luminosity ( @xmath17 ) .
these nine bursts were all detected by batse with redshifts measured from optical spectra of either the afterglow or the host galaxy .
the highly unusual grb980425 ( associated with supernova sn1998bw ) is not included because it is likely to be qualitatively different from the classical grbs .
bursts with red shifts that were not recorded by batse can not ( yet ) have their observed parameters converted to energies and fluxes that are comparable with batse data .
simplistically , plots of @xmath17 versus @xmath5 and @xmath17 versus @xmath6 can calibrate the luminosity indicators , which then can yield luminosity distances to each burst for plotting on a hubble diagram and fitting to cosmological models . in practice
, there must be a simultaneous chi - square minimization for the luminosity calibrations and the cosmology , so as to avoid the effects of possible correlations between the indicators and distance .
if the earth is not along the central axis of the grb s jet , then various off - axis effects will change the observed properties .
for light viewed off - axis by angle @xmath18 from a jet moving at @xmath19 times the speed of light , the transverse doppler shift , redshift , and beaming will change the various observed properties by powers of @xmath20 .
the observed value of @xmath16 will scale as @xmath21 , @xmath5 will scale as @xmath22 , @xmath6 will scale as the inverse of a time and hence as @xmath21 , while @xmath4 will have a model dependent variation that scales roughly as @xmath23 . for structured jets @xcite
, the luminosity and bulk lorentz factor can be parameterized as being proportional to some positive power of @xmath21 . when viewed on - axis
, the bursts will presumably display fairly tight lag / luminosity and variability / luminosity relations .
these relations can be converted to similar relations involving observed off - axis quantities with an extra factor of @xmath21 to some power .
fortunately , the observed distribution of @xmath16 is remarkably narrow for bursts of a given brightness @xcite , and the removal of kinematic effects shows that the on - axis @xmath16 is virtually constant @xcite .
thus , @xmath16 should be proportional only to some power of @xmath21 , and this relation can be used to convert the @xmath21 dependency of the luminosity relations into a @xmath16 dependency .
essentially , we can use the observed @xmath16 value to get information about the off - axis angle and correct the observed @xmath5 and @xmath6 values to those which would be seen on - axis . in practice
, the correction factor can be taken as some power of @xmath24 .
( the division by 400 kev is to minimize correlations between the normalization constant and the exponent during the fits . )
the exponents of the correction factors will be free parameters which depend on the scenario and jet structure . with the off - axis correction
, the luminosities can be derived by fitting the relations @xmath25 and @xmath26 , where @xmath27 ^{e_{lag}}$ ] and @xmath28 ^{e_v}$ ] .
plots of @xmath17 versus @xmath29 and @xmath17 versus @xmath30 for the nine bursts with known redshifts ( see figure 1 ) have slopes of @xmath31 and @xmath32 , while the scatter in these plots will be minimized for the best values of @xmath33 and @xmath34 .
this gives @xmath35 , @xmath36 , @xmath37 , and @xmath38 .
thus : @xmath39^{-1.27 } , \ ] ] @xmath40^{1.57 } , \ ] ] where @xmath4 is in @xmath41 , @xmath5 is in seconds , @xmath6 is dimensionless , and @xmath16 is in kev . as independent checks , a plot of @xmath29 versus @xmath30 for 93 batse bursts will have a slope of @xmath42 , and the scatter will be minimized for the best values of @xmath33 and @xmath34 .
this gives @xmath43 , @xmath44 , and @xmath45 ; values which are consistent with the calibration from figure 1 .
the lag and the variability produce two independent luminosity values through eqs 1 and 2 .
these can be combined as weighted averages ( with the weights equal to the inverse square of the one - sigma errors ) to produce a combined luminosity ( @xmath46 ) and then @xmath13 , as given in table 1 . a plot of @xmath13 versus @xmath14 is a hubble diagram ( figure 2 ) .
equations 1 and 2 are optimal for studies of grbs where the cosmology is not in question . when the goal is to use the bursts as cosmological markers , care must be taken to avoid circular logic .
in particular , the previous paragraph assumed a specific cosmology ( @xmath47 for a flat universe and @xmath48 ) for calculating the @xmath13 values as a function of @xmath14 . to avoid this circularity ,
the fits must be performed with the cosmological parameters as additional fit parameters .
that is , the model being fit to the data will have cosmological plus grb parameters that are to be simultaneously determined by the minimization of the chi - square .
this will tie the cosmology conclusions into grb calibrations , but both will be well constrained when many bursts are available . a correct procedure to find the best value of @xmath49 in a flat universe is ( 1 ) to fix @xmath49 , ( 2 ) derive @xmath13 and @xmath17 for each burst for that cosmology , ( 3 ) fit @xmath17 as power laws of @xmath29 and @xmath30 , ( 4 ) use the best fits to derive @xmath46 and then @xmath13 for each burst , ( 5 ) calculate chi - square by comparing the model and observed @xmath13 values , ( 6 ) repeat steps 1 - 5 to identify the best fit @xmath49 value as that where chi - square is minimized , and ( 7 ) identify the one - sigma range where chi - square is within unity of its minimum value . by this procedure
, the lowest chi - square corresponds to @xmath50 with the one - sigma constraint @xmath51 .
the three - sigma limit is larger than unity , so this particular measure is not helpful for cosmology .
this new result is completely independent of , yet in agreement with , those from supernovae and other methods @xcite .
however , the current value for @xmath49 is not yet of high reliability because the number of degree of freedom in the fit are small and the slope of the hubble diagram is dominated by just two high redshift events .
while this new result does not seriously constrain current cosmology , it does act as a demonstration of principle and a sign for the future .
a comparison with the supernova - based hubble diagram is inevitable .
grbs have both advantages and disadvantages when compared to supernovae .
at least currently , supernovae have an advantage over grbs for the accuracy of the distance measurements for a single event .
well - observed nearby type ia supernovae have an rms scatter about their hubble diagram of 0.18 mag @xcite which translates into an uncertainty in the log of the distance of 0.04 .
grbs currently have an rms scatter about their calibration curves of 0.35 and 0.20 in the logs of lag and variability ; which translates into uncertainties in the log of the distance of 0.22 and 0.16 respectively .
these two independent measures can be combined to give an uncertainty of 0.13 in the log of the distance .
thus , individual supernovae are roughly three times more accurate as standard candles than are individual grbs .
the physics of gamma - ray emission from relativistic shocks in grbs is largely known ( but not how to create the magnetic fields ) although the explosion scenario is still uncertain . the basic physics of supernovae is well known ( although the type ia progenitors are still not identified ) .
supernovae can have their decline - rate versus luminosity calibration from low - redshift events , and this allows the elegance of being able to make the calibration substantially separate from the cosmology .
thus , supernovae are much better understood than are grbs , and this will be a substantial comfort for using them as standard candles .
nevertheless , it is unclear where this advantage for supernovae will pay off .
currently , the analyses of both supernova and grb hubble diagrams are entirely empirical with no contribution from theory , so a better theoretical understanding of supernovae has not helped in any specific manner .
supernova will have better known evolution effects ; yet @xcite show that evolution is no problem for supernovae and their argument applies identically to grbs .
supernovae have substantial problems with a higher dust density at high red shift @xcite and the possibility of grey dust @xcite .
gamma radiation suffers from no extinction . after a tremendous observational effort
, the current supernova hubble diagram @xcite extends only out to @xmath52 .
( sn1997ff at @xmath53 has very large uncertainties @xcite even without the large corrections for gravitational lensing @xcite and the uncertainty of the supernova type . )
the dedicated snap satellite @xcite , proposed for launch in 2008 , will go out to @xmath54 with exquisite accuracy in the light curves .
in contrast , the grb hubble diagram is already out to @xmath0 and likely it will be extended to @xmath1 @xcite or farther @xcite . with the launch of the swift satellite in 2003 , the grb hubble diagram will be available for the redshift ranges @xmath55 and @xmath56 roughly five years before snap can extend the supernova hubble diagram only to @xmath55 . in summary
, grbs are @xmath57 worse in accuracy than supernovae , but this is traded - off for the lack of problems from extinction , extension to @xmath10 , and an answer many years before snap .
i do not think that the grb hubble diagram will replace the supernova hubble diagram .
the reason is that _ both _ diagrams will have known and unknown systematic problems which will make results from any _ one _ method not conclusive .
what is needed is the concurrence of multiple independent methods .
so _ both _ grb and supernova hubble diagrams are needed for a confident result .
conceptually , the biggest question related to a construction of the grb hubble diagram is the possibility of luminosity evolution .
fortunately , there are two strong reasons that any such effects must be small .
first , theoretical explanations for the luminosity indicators @xcite tie the relations to simple physics ( involving relativistic effects operating on any light source , the delay in light travelling different paths , and the cooling of any body by radiation ) that are unrelated to properties ( such as metallicity ) that might evolve .
second , just as for the purely empirical supernova hubble diagram @xcite , any drift in population average properties @xcite is irrelevant as the calibrations will work for any individual event whether near or far .
that is , it does not matter whether population average properties ( like luminosity ) evolve because the distances are derived for individual events that are each correctly calibrated . within two years
, the grb situation will be greatly improved with the dedicated swift satellite @xcite currently scheduled for launch in september 2003 .
swift is expected to get @xmath58 bursts per year , of which @xmath59 will have directly measured redshifts .
the redshift will be difficult to measure for @xmath60 bursts , yet it is possible by locating the lyman break with broad - band photometry or spectroscopy extending into the near infrared . in its nominal three - year lifetime , we can get @xmath2 bursts with measured values for @xmath14 , @xmath15 , @xmath5 , @xmath6 , and @xmath16 .
this many bursts will allow for high accuracy calibration of the luminosity indicators .
the large number of bursts will also allow for the reduction of statistical errors ( dominated by the intrinsic scatter of the bursts ) by a factor of @xmath61 .
if the bursts are divided into six redshift bins , each bin will have an error of 0.03 in the log of the distance .
this accuracy will , for example , yield an uncertainty of @xmath62 in the derived value of @xmath49 for a flat universe , even with no improvements in the luminosity indicators .
so with swift , the prospect is that we can produce an accurate grb hubble diagram with @xmath2 bursts from @xmath63 by 2006 .
what can we learn from the grb hubble diagram ?
we can test the predicted shift of the universe from matter to dark - energy domination in the range @xmath64 .
we will also look for various predicted quintessence @xcite and non - standard @xcite effects in the range @xmath65 .
the typical sizes of these effects are 0.4 in the log of the distance around a redshift of @xmath66 . the current paradigm of cosmology ( the new inflation perhaps with quintessence ) is so new and untested that surprises could easily await in the unknown regime of @xmath67 .
aguirre , a. 1999 , , 525 , 583 brainard , j. j. et al .
1998 , 20th texas symposium , paris ( astro - ph/09904039 ) branch , d. , perlmutter , s. , baron , e. , & nugent , p. 2001 , astro - ph/0109070 bromm , v. & loeb , a. 2002 , , in press , ( astro - ph/0201400 ) costa , e. et al .
1997 , nature , 387 , 783 durrer , r. & novosyadlyj , b. 2001 , , 324 , 560 fenimore , e. e. , & ramirez - ruiz , e. 2000 , , submitted , ( astro - ph/0004176 ) frail , d. et al .
1997 , nature , 389 , 261 gehrels , n. a. 2000 , proc .
s.p.i.e , 4140 , 42 hogg , d. w. 1999 , astro - ph/9905116 ioka , k. & nakamura , t. 2001 , , 554 , l163 lamb , d. q. & reichart , d. e. 2000 , , 536 , 1 lewis , g. f. & ibata , r. a. 2001 , , 324 , l25 lloyd - ronning , n. , fryer , c. l. & ramirez - ruiz , e. 2000 , , in press , ( astro - ph/0108200 ) mallozzi , r. s. , et al .
1995 , , 454 , 597 mannheim , p. d. 2002 , astro - ph/0204202 moertsell , e. , gunnarsson , c. & goobar , a. 2001 , , 561 , 106 norris , j. p. 2002
, , submitted , ( astro - ph/0201503 ) norris , j. p. , marani , g. f. , & bonnell , j. t. 2000 , , 534 , 248 paciesas , w. s. , et al .
2001 , , 122 , 465 perlmutter , s. et al .
1999 , , 517 , 565 phillips , m. m. , lira , p. , suntzeff , n. b. , schommer , r. a. , hamuy , m. & maza , j. 1999 , , 118 , 1766 plaga , r. 2001 , , 370 , 351 riess , a. g. et al .
2001 , , 560 , 49 rossi , e. , lazzati , d. , & rees , m. j. 2002 , , in press , ( astro - ph/0112083 ) schaefer , b. e. 2001 , , submitted , ( astro - ph/0101462 ) schaefer , b. e. 2002 , , submitted schaefer , b. e. , deng , m. , & band , d. l. 2001 , , 563 , l123 snap collaboration 2002 , http://snap.lbl.gov/ totani , t. & kobayashi , c. 1999 , , 526 , l65 van paradijs , j. et al .
1997 , nature , 386 , 686 weller , j. & albrecht , a. 2001 , phys .
lett . , 86 , 1939 zhang , b. & mszros , p. 2002
, , in press , ( astro - ph/0112118 ) ccccccccc 970508 & 0.84 & @xmath68 & @xmath69 & @xmath70 & @xmath71 & @xmath72 & @xmath73 & @xmath74 + 970828 & 0.96 & @xmath75 & @xmath76 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 + 971214 & 3.41 & @xmath82 & @xmath83 & @xmath84 & @xmath85 & @xmath86 & @xmath87 & @xmath88 + 980703 & 0.97 & @xmath89 & @xmath90 & @xmath91 & @xmath92 & @xmath93 & @xmath94 & @xmath95 + 990123 & 1.60 & @xmath96 & @xmath97 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & @xmath102 + 990506 & 1.30 & @xmath103 & @xmath104 & @xmath105 & @xmath106 & @xmath107 & @xmath108 & @xmath109 + 990510 & 1.62 & @xmath110 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath116 + 991216 & 1.02 & @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 + 000131 & 4.5 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & @xmath128 & @xmath129 & @xmath130 + | gamma - ray bursts ( grbs ) are tremendous explosions visible across most of the universe , certainly out to redshifts of @xmath0 and likely out to @xmath1 .
recently , grbs have been found to have a roughly constant explosive energy as well as to have two luminosity indicators ( the spectral lag time and the variability ) that can be used to derive the burst s luminosity distance from the gamma - ray light curve alone .
there currently exists enough information to calibrate luminosity distances and independent redshifts for nine bursts . from these
, a grb hubble diagram can be constructed , where the observed shape of the curve provides a record of the expansion history of our universe .
the current nine burst diagram is sparse , yet formal limits can be placed on the mass density of a flat universe .
this first grb hubble diagram provides a proof of concept for a new technique in cosmology at very high redshifts . with the launch of the swift satellite in 2003 , we should get @xmath2 bursts to produce a hubble diagram impervious to all effects of dust extinction and out to redshifts impossible to reach by any other method . |
magnetic resonance ( mr ) spectroscopy , including nuclear magnetic resonance ( nmr ) @xcite and electron paramagnetic resonance ( epr),@xcite has been widely used to probe local structural and dynamic properties of a wide range of systems . however , because the conventional mr spectrometer is based on the detection of microwave and radio - frequency photons , whose energies are small compared to photon energies detected in single molecule fluorescence ( fl ) spectroscopy , mr spectroscopy inherently has much lower sensitivity . with a current typical spectrometer setup ,
@xmath0 electron spins and @xmath1 @xmath2h nuclear spins are required to obtain epr and nmr signals at room temperature , respectively . a nitrogen - vacancy ( nv ) center is a paramagnetic color center in diamond with unique electronic , spin , and optical properties including its stable fl signals , @xcite long decoherence time , @xcite and capability to initialize the spin states of nv centers by applying optical excitation and to readout the states by measuring the fl intensity .
@xcite epr of a single nv center is also observable using optically detected magnetic resonance ( odmr ) spectroscopy .
@xcite most interestingly , a new type of magnetometry using nv centers has been proposed .
@xcite magnetometers based on a nv center are actively being developed , and detection of small magnetic fields has been demonstrated .
@xcite moreover , detection of a single or a few electron spins surrounding a nv center has also been demonstrated using the double electron - electron resonance ( deer ) spectroscopy of a single nv center .
@xcite in this article , we discuss epr spectroscopy on a small ensemble of electron spins using a single nv center in a type - ib diamond .
although deer spectroscopy is a powerful method to determine the strength of a magnetic dipole field from a single or a few electron spins , it is challenging to estimate the number of the detected small ensemble spins from the observed signal when the detected spins are more than a few . here
we demonstrate the estimation of the number of detected spins by simulating the deer signal .
we first demonstrate the detection of a single nv center using fl autocorrelation measurement and odmr spectroscopy , and present the performance of rabi and spin echo measurements of the single nv center .
then , we employ deer spectroscopy to measure the epr spectrum of electron spins surrounding the nv center inside the diamond crystal .
the electron spins are identified as substitutional single nitrogen ( n ) impurities existing in diamond by the analysis of the observed epr spectrum .
finally , we employ a computer simulation to estimate the number of the n spins detected by the deer experiment .
3 @xmath3m@xmath4 . the color scheme for fl intensity is shown in the legend .
an isolated single nv center used for subsequent measurements is circled .
the inset shows autocorrelation measurement .
the anti - bunching signal at zero delay ( @xmath5(@xmath6=0)@xmath70.5 ) confirms that the fl signal is from a single quantum emitter .
( b ) cw odmr measurement of the single nv center when the external magnetic field of 35.7 mt was applied along the @xmath8 $ ] axis of the diamond crystal using a permanent magnet .
the reduction of fl intensity at the microwave frequency of 1.868 ghz corresponds to the @xmath9=0 @xmath10 @xmath11 transition of the nv center .
the solid line indicates a fit to the lorentzian function .
( c ) rabi oscillation measurement of the single nv center .
fl intensity oscillations were observed as @xmath12 was varied .
square dots connected with lines indicate the measurement and solid line indicates the fit .
the oscillation frequency ( @xmath13 ) of 7.4 @xmath14 0.1 mhz and the decay time ( @xmath15 ) of 0.77 @xmath14 0.04 @xmath3s were obtained from a fit to @xmath16 $ ] .
the inset shows the pulse sequence used in the measurement .
the schematic timings for excitation laser ( exc . ) , microwave ( mw ) , and fl measurement ( fl ) are shown .
( d ) spin echo measurement to extract @xmath17 of the nv center .
the lengths of @xmath18/2- and @xmath18-pulse used were 34 and 68 ns , respectively , which were determined from the rabi oscillation measurement .
fl intensity decay was monitored as a function of @xmath19 , and @xmath200.1 @xmath3s was obtained from a fit to @xmath21$].@xcite the inset shows the pulse sequence used in the measurement . for both pulsed odmr measurements , the microwave frequency of 1.868 ghz
was used to drive the @xmath22 transition of the nv center at @xmath23 mt , and laser initialization pulse ( init . ) of 2 @xmath3s , and laser read - out pulse ( ro ) and fl measurement pulse ( sig . ) of 300 ns were used .
also the fl intensity was normalized and re - scaled into the @xmath24 state population of the nv center .
, width=75 ] figure 1 shows the detection and coherent manipulation of a single nv center in a bulk diamond crystal using an odmr setup.@xcite the diamond used in the experiment is a synthetic type - ib single crystal ( @xmath25 mm@xmath26 , sumitomo electric inc . ) . for microwave excitations , two microwave synthesizers , a power combiner , and a 10 watt amplifier
are connected to a 20 @xmath3m - thin gold wire placed on a surface of the diamond sample .
first a fl image of the diamond was taken in order to map out fl signals in the diamond crystal ( see fig .
after choosing a suitable , well - isolated fl spot , we carried out the autocorrelation and continuous - wave ( cw ) odmr measurements in order to identify fl signals from a single nv center .
as shown in the inset of fig .
1a , the autocorrelation measurement of the chosen fl spot reveals the dip at zero delay which confirms the fl signals originated from a single quantum emitter .
in addition , cw odmr measurement of the selected single nv center was performed with application of the external magnetic field ( @xmath27 ) of 35.7 mt along the @xmath8 $ ] axis of the diamond lattice . as shown in fig .
1b , the reduction of the fl intensity was observed at the microwave frequency of 1.868 ghz corresponding to odmr signals of the @xmath22 transition of the nv center .
thus , the observations of autocorrelation and cw odmr signals confirm the successful identification of a single nv center .
next , we demonstrated coherent manipulations of the identified single nv center using pulsed odmr measurements . for all pulsed measurements ,
the nv center was first prepared in the @xmath24 state by applying a long initialization laser pulse , and the microwave pulse sequence was applied for the desired manipulation of the spin state of the nv center , then the final spin state of the nv center was determined by applying a read - out laser pulse and measuring the fl intensity . in addition , the fl signal intensity was mapped into the population of the nv s @xmath24 state ( @xmath28 ) using two reference signals each time the sequence was applied .
@xcite the two references were the maximum and minimum fl intensities corresponding to the @xmath24 and @xmath29 states of the nv center , respectively .
moreover , the pulse sequence was averaged on the order of @xmath30 times to obtain a single data point .
first , the rabi oscillation measurement was performed at @xmath23 mt and 1.868 ghz , which corresponds to the @xmath22 transition of the nv center . as the microwave pulse length ( @xmath12 )
was varied in the measurement , pronounced oscillations of @xmath28 was observed as shown in fig .
1c . by fitting the observed rabi oscillations to the sinusoidal function with the gaussian decay envelope , @xcite the lengths of @xmath31- and @xmath18-pulses for the single nv center
were determined to be 34 and 68 ns , respectively .
second , we carried out the spin echo measurement of the nv center to measure the spin decoherence time ( @xmath17 ) .
as shown in the inset of fig .
1d , the applied microwave pulse sequence consists of the conventional spin echo sequence , widely used in epr spectroscopy , @xcite and an additional @xmath31-pulse at the end of the sequence to convert the resultant coherence of the nv center into the population of the @xmath24 state.@xcite figure 1d shows the measured fl intensity as a function of free evolution time ( 2@xmath6 ) . by fitting the observed echo decay
to @xmath32,@xcite @xmath17 of the single nv center was determined as 1.2 @xmath14 0.1 @xmath3s .
-pulse in mw to rotate the surrounding spins .
( b ) the spectrum obtained implementing the deer sequence .
the spin echo intensity was recorded for fixed @xmath6 of 500 ns while sweeping the frequency of mw2 pulse .
the simulated epr spectrum is shown in the dotted line .
the simulation was done based on the spin hamiltonian of single substitutional nitrogen impurity ( n ) given in the main text .
( c ) coherent control of the n spins group corresponds to the peak at @xmath33 ghz ( see fig .
@xmath34 ns was used in the measurement .
oscillation of the spin echo intensity was observed as a function of the pulse length of mw2 ( @xmath35 ) , which corresponds to the rabi oscillations of n spins . by fitting the observed oscillation to a sinusoidal function
, we obtained the oscillation frequency of 5.5 @xmath14 0.3 mhz .
all the measurements were performed at 35.7 mt . , width=50 ] moreover , we performed the deer spectroscopy at @xmath23 mt and 1.868 ghz ( corresponding to the @xmath22 transition of the nv center ) to probe coupling between the single nv center and surrounding paramagnetic impurities in the diamond crystal .
figure 2a shows the pulse sequence used in the deer measurements,@xcite which was adopted from the field of epr.@xcite the deer sequence consists of the spin echo sequence for the nv center and an additional @xmath18-pulse at different microwave frequency ( denoted as mw2 ) . when the additional @xmath18-pulse is resonant with surrounding electron spins near the nv center , the magnetic moments of the surrounding spins are flipped , and this alters the magnetic dipole field experienced by the nv center from the surrounding spins during the second half of the spin echo sequence .
the alteration causes a shift in the larmor frequency of the nv center , which leads to different phase accumulation of the larmor precession during the second half of the sequence from the first half . as a result ,
the nv center suffers a phase shift in the echo signal and a reduction of fl intensity from the original spin echo signal is observed . as shown in fig . 2b , we monitored the spin echo intensity of the nv center at fixed @xmath36 ns as a function of the microwave frequency of mw2 ( @xmath37 ) , and observed clear intensity reductions at five frequencies ( 0.90 , 0.92 , 1.01 , 1.10 , and 1.12 ghz ) .
the observed deer spectrum corresponds to the resonance positions of single substitutional nitrogen ( n ) defects in diamonds .
the spin hamiltonian of n spins is given by @xmath38 where @xmath39 is the bohr magneton , @xmath40 is the external magnetic field , @xmath41 is the isotropic @xmath42-value , @xmath43 and @xmath44 are the electronic and nuclear spin operators , respectively , and @xmath45 ( @xmath46 mhz and @xmath47 mhz ) is the anisotropic hyperfine coupling tensor to @xmath48 nuclear spin .
@xcite because of the anisotropy in the hyperfine coupling and four different orientations of n spins , the spectrum of n spins consists of five lines with application of the magnetic field along the @xmath8 $ ] direction
. the simulated spectrum of n spins using the above hamiltonian is in a good agreement with the observed deer signal ( see fig .
2b ) . in the deer measurement , we also verified the length of @xmath18-pulse for the n spins ( @xmath35 ) by measuring rabi oscillations of the n spins .
figure 2c shows the spin echo intensity of the nv center as a function of @xmath35 .
the frequency of the mw2 pulse was set at 0.92 ghz . since @xmath6 was fixed , the oscillation we observed in the spin echo intensity of the nv center as a function of @xmath35 corresponds to the rabi oscillations of the n spins , which allowed us to choose the optimum pulse length of 90 ns for the @xmath18-pulse of the n peaks at 0.90 and 0.92 ghz . where @xmath49 nm is the lattice constant of diamond .
red sphere denotes the nv center placed at the origin of the lattice .
blue spheres represent randomly distributed n spins with 20 ppm concentration .
green spheres are a subset of n spins which are in @xmath8 $ ] orientation with @xmath50 .
n spins at or next to the origin , next to other n spins , and overlap to other n spins , were suppressed in the simulation .
( b - d ) intensity plots of @xmath51 versus @xmath52 for @xmath53 s , 10 ms , and 100 s , respectively , from @xmath54 spatial configurations .
the unit pixel of the plots are 0.01 in x - axis ( @xmath52 ) and 1 in y - axis ( @xmath51 ) , and the total intensity ( sum of all intensities in every pixel ) sums up to @xmath54 , the total number of spatial configurations simulated . the color scheme for intensity is shown in the legend .
vertical solid lines show the margins where deer peak intensity fluctuated for the outer right peak in fig .
2a . from the simulation
, we evaluated that approximately less than 45 n spins are responsible for the observed deer signal from n spins in @xmath8 $ ] orientation with @xmath50 for all @xmath55 values we considered .
the insets show the occurrence of spatial configurations of the simulated lattice that resulted in @xmath56 as a function of @xmath51 , which reveals that all the configurations that resulted in such @xmath52 have @xmath57 .
the number of such configurations were @xmath581200 , @xmath581300 , and @xmath59 for @xmath53 s , 10 ms , and 100 s , respectively , which translate to 4@xmath6013 @xmath61 out of total @xmath54 configurations . , width=80 ] finally , we performed a computer simulation to estimate the number of the n spins which is contributing to the observed deer signal at @xmath62 ghz by considering the effective magnetic dipole field at the nv center from the surrounding n spins ( @xmath63 ) .
the deer peak at @xmath62 corresponds to the @xmath64 transition of n spins oriented in the @xmath8 $ ] direction .
first step of the simulation was to generate a model configuration of the nv center and n spins in a diamond lattice , and this was done by placing the nv center at the origin of the diamond lattice and assigning the positions of n spins randomly in the lattice sites ( see fig .
the concentration of n spins was chosen to be 20 parts per million ( ppm ) carbon atoms based on the measured @xmath17 of the nv center ( 1.2 @xmath3s as shown in fig .
1d),@xcite and @xmath5818,000 n spins were placed in the simulated diamond lattice with a diameter of @xmath58200 nm .
next , we randomly chose the orientation of n spins where only a quarter of n spins were assigned in the @xmath8 $ ] direction and the rest of n spins were assigned in the other three directions , _
i.e. _ , @xmath65 $ ] , @xmath66 $ ] , and @xmath67 $ ] .
we then assigned the nuclear spin value ( @xmath68 ) of either @xmath69 , @xmath70 , or @xmath11 to all n spins with equal probability .
for the simulation of the deer signal of the nv center , we only considered the contribution coming from the n spins oriented in the @xmath8 $ ] direction with @xmath50 ( @xmath62 ghz in fig .
thus , @xmath63 was computed from only @xmath71-th of all n spins in the simulated lattice on average ( @xmath581500 n spins ) .
finally we assigned the electron spin value ( @xmath9 ) of either @xmath72 or @xmath73 with equal probability to the n spins .
@xmath63 is given by the sum of individual dipolar field from each n spin as , @xmath74 where @xmath75 is the dipolar field strength of @xmath76-th n spin at the nv center , @xmath77 is the number of the n spins oriented in the @xmath8 $ ] direction with @xmath50 , @xmath78 is the vacuum permeability , @xmath79 and @xmath80 are the magnitude and polar angle of the vector that connects the nv center and @xmath76-th n spin , respectively , and @xmath81 is the electron spin value of @xmath76-th n spin .
we then calculated the number of n spins ( @xmath82 ) out of @xmath77 to obtain more than 90@xmath61 of the @xmath63 by successively adding individual dipolar field term from each n spin in descending order of magnitude ( _ i.e. _ , when @xmath83 ) .
the deer intensity ( @xmath84 ) is calculated by considering the change in the effective dipolar field before and after the @xmath18-pulse of the n spins that causes a shift in the precession frequency of the nv center as , @xmath85 where @xmath86 is the @xmath42-value of the nv center,@xcite @xmath87 is the planck constant , and @xmath36 ns is the fixed value used in the deer measurement ( see the inset of fig .
it was assumed that @xmath18-flip of the n spins is instantaneous ( _ i.e. _ the effect of a finite width of the pulse was not considered ) . in addition , we took into account for the electron and nuclear spin relaxations of n spins during the deer measurement time ( @xmath58100 s with @xmath58@xmath30 averaging ) by statistically re - assigning the electron and nuclear spin values ( @xmath9 and @xmath68 ) according to the electron and nuclear longitudinal relaxation times ( @xmath88 and @xmath55 ) , respectively .
although @xmath88 has been investigated previously ( @xmath88@xmath5810 ms),@xcite @xmath55 has not been reported before to the best of our knowledge . in the simulation
, we used @xmath89 ms and considered @xmath55 to be 10 ms , 1 s and 100 s , therefore the assignment of @xmath68 happens @xmath90 times for @xmath53 s ( @xmath54 and 1 times for @xmath91 ms and 100 s , respectively ) and a total of @xmath54 iterations ( 100 s/@xmath88 ) occurs during the simulation .
then what we finally computed were the average of @xmath54 values of @xmath82 and @xmath84 ( _ i.e. _ , @xmath92 and @xmath93 ) .
in addition , in order to see the spatial configuration dependence on @xmath51 and @xmath52 , we repeated the procedure described above for @xmath54 spatial configurations and calculated @xmath54 values of @xmath51 and @xmath52 . moreover , including the effects of @xmath17 decay ( 1.2 @xmath3s as shown in fig .
1d ) , the detected deer signal in @xmath28 is given as @xmath94 where we re - scale @xmath95 to the spin echo intensity without @xmath18-pulse and @xmath96 to @xmath97 .
figure 3b - d shows the simulated result of @xmath51 as a function of the deer intensity in @xmath28 for @xmath53 s , 10 ms , and 100 s , respectively , from @xmath54 spatial configurations . as shown in fig .
3b , the number of the n spin that contributes to the deer signal ( @xmath51 ) depends on the detected deer intensity .
we first noted that , for all @xmath54 configurations we simulated , only a small portion ( @xmath9860 ) out of @xmath581500 n spins contributes to more than 90 @xmath61 of @xmath63 on average .
for example , in the case of homogeneously distributed spins , 60 spins with 20 ppm concentration will be located within @xmath5870 nm diameter . when the deer intensity is large ( @xmath99 ) , @xmath51 is smaller because @xmath63 in such spatial configurations is only dominated by a smaller ensemble of the n spin located in the vicinity of the nv center .
in addition , there was little observation of @xmath100 ( _ i.e. _ , @xmath101 because a small probability exists for the nv center to be coupled with a single or a few n spins and the resultant @xmath52 is a weighted sum of many oscillatory functions with different frequencies . on the other hand , when the deer intensity is small ( @xmath102 ) , @xmath51 is larger because the n spins in such spatial configurations spread uniformly and the n spins located farther away from the nv also contribute to @xmath63 .
the experimentally observed deer intensity was @xmath103 ( see fig .
the inset of fig .
3b shows a histogram of the occurrence of @xmath51 from the simulation that yielded @xmath104 - 0.65 .
the occurrence was in the range of 7 - 42 ( the mean of @xmath51 was 14 ) and 95@xmath61 of the occurrence was in 7 - 29 .
as shown in fig .
3c , there was little to no change in @xmath51 with @xmath91 ms ( the mean of @xmath51 was 15 spins and 95@xmath61 of the occurrence was in @xmath105 - 30 ) .
on the other hand , in the case of @xmath51 with @xmath106 s , slightly broader distribution was observed due to no nuclear spin relaxation during the simulation .
the mean of @xmath51 was 4 spins and 95@xmath61 of the occurrence was in 2 - 14 .
therefore the detected n spins in the deer measurement was estimated to be a few tens of spins .
in summary , we presented epr spectroscopy on a small ensemble of spins using a single nv center in diamond .
we first identified a single nv center using fl and odmr spectroscopy , then employed deer spectroscopy to detect epr signals of electron spins surrounding the nv center in diamond . from the epr spectrum
, we determined that the detected electron spins are substitutional single nitrogen impurities in diamond .
finally , by comparing the deer intensity with the simulated deer signal , we estimated the detected spins in the deer measurement to be a few tens of spins , thus the result demonstrated significant improvement of the epr sensitivity .
this work was supported by the national science foundation ( dmr-1508661 ) and the searle scholars program ( s.t . ) . | we report electron paramagnetic resonance ( epr ) spectroscopy of a small ensemble of spins using a single nitrogen - vacancy ( nv ) center in diamond .
we first identify a single nv center using fluorescence and optically detected magnetic resonance spectroscopy , then demonstrate double electron - electron resonance ( deer ) spectroscopy to perform epr detection of surrounding spins . through the epr spectrum analysis
, we determine that the detected spins are nitrogen impurities in diamond .
moreover we estimate the number of the detected spins to be a few tens of spins using simulation of the deer signal . |
extremely metal poor ( emp ) stars provide important clues to the chemical history of our galaxy , the role and type of early sn , the mode of star formation in the proto - milky way , and the formation of the galactic halo .
the sample of known emp stars is summarized by @xcite .
they compiled a list of the key properties of the 12 stars identified up to that time with [ fe / h ] @xmath4 dex , 7 of which are emp giants and subgiants within the range of @xmath5 between 4900 and 5650 k. our 0z project has the goal of increasing the sample of such stars through data mining of the hamburg / eso survey @xcite .
this is an objective prism survey from which it is possible to efficiently select qsos @xcite as well as a variety of interesting stellar objects , among them emp stars @xcite .
our 0z project has been systematically searching the database of the he s for this purpose over the past five years .
we have just published @xcite , and we present at this conference , a sample of new emp giants with @xmath6 k and [ fe / h ] @xmath4 dex which substantially increases the number of such stars known .
our 0z project has obtained moderate resolution follow up spectra of @xmath71550 candidate emp stars selected from the he s .
these follow up spectra were processed through the code described in @xcite , which is essentially identical to that used by the hk survey@xcite until recently ; the latest updates to the algorithm as used by the hk survey are described in @xcite .
the algorithm uses the strength of h@xmath8 and of the 3933 ca ii line to assign a metallicity to each star , which we denote as [ fe / h](hes ) .
( the segue survey uses an even more sophisticated algorithm taking advantage of the uniform wide wavelength coverage of the sdssii spectra and multi - color photometry ; a preliminary description is given in @xcite . )
we attempt to observe all stars with [ fe / h](hes ) @xmath9 dex as inferred from these moderate resolution follow up spectra at high resolution .
we present the results of detailed abundance analyses based on high resolution and high signal - to - noise spectra of eight extremely metal poor ( emp ) stars with [ fe / h ] @xmath4 dex , four of which are new . only stars with @xmath10 k
are included .
past work on emp stars , e.g. @xcite , @xcite and @xcite , has emphasized the small scatter at a fixed [ fe / h ] for trends of element abundance ratios [ x / fe ] .
the @xmath11 about the mean at each [ fe / h ] appears to be consistent with very little intrinsic scatter , just the observational errors contributing .
this applies only to c - normal metal - poor stars , and to elements between na and the end of the fe - peak .
the light elements li , be , c , n , and o in emp first ascent rgb stars may be affected by mixing of internally processed material , while emp agb stars may have na and al affected as well @xcite .
we established in @xcite that the neutron capture elements are decoupled from the fe - peak elements ; the site or conditions of their production , presumably via the @xmath12 or @xmath13process , must be different from that of the fe - peak elements .
the most interesting thing we have discovered recently is the existence of outliers in these relations among emp stars .
the most sensational case we have found , for which a brief description was given in @xcite with a full study in @xcite , is he1424@xmath20241 .
this star is the most metal poor star in our sample , with [ fe / h ] @xmath14 dex .
it has highly anomalous abundance ratios unlike those of any other known emp giant , with very low si , ca and ti relative to fe , and enhanced mn and co , again relative to fe .
1 and 2 illustrate the uniqueness of this star among other halo giants with regard to its ca and si abundance ratios with respect to fe .
only ( low ) upper limits for c and n can be derived from the non - detection of the ch and nh molecular bands .
si , ca and ti are formed via explosive @xmath3burning while mg , which is normal with respect to fe in he1424@xmath20241 , is formed via hydrostatic @xmath3burning .
a much smaller separation of abundance ratios ( @xmath15 dex ) for elements between these two types of @xmath3burning was seen by @xcite in the galactic bulge , but that of he1424@xmath20241 is much more extreme , and can not be explained with recent sn nucleosynthesis yields .
we suggest a peculiar snii produced the chemical inventory of this star .
he0132@xmath22429 , another sample star , has excesses of n and sc with respect to fe .
is shown for all of the candidate emp stars with hires spectra analyzed by the 0z project to date , including the present sample .
c - rich stars are not shown .
the solid horizontal line denotes the solar ratio .
the plot includes well studied galactic globular clusters , mostly from analyses by j. cohen and her collaborators , as well as samples of halo field stars from the sources indicated on the symbol key in the lower right of the figure .
note the highly anomalous position of he1424@xmath20241 , the only star with [ si / fe ] @xmath16 dex . ] . ]
the strong outliers in abundance ratios among the fe - peak elements in these c - normal stars , not found at somewhat higher metallicities ( [ fe / h ] @xmath17 dex ) , are definitely real .
the anomalies discussed here can not be due to observational or analysis errors .
there are two hires spectra for each of these two stars .
the effects are very large , and are measured relative to the many other emp stars we have analyzed in a homogeneous manner with identical codes and procedures .
they suggest that at such low metallicities we are beginning to see the anticipated and long sought stochastic effects of individual supernova events contributing to the fe - peak material within a single star .
another of our recent results ( see * ? ? ?
* ) comes from a detailed comparison of the analysis procedures adopted by our 0z project compared to those of the first stars vlt large project .
we find a systematic difference for [ fe / h ] of @xmath70.3 dex , our values always being higher , largely due to differences in adopted reddenings , temperature scales , and fe solar abundance between the two projects .
a preliminary metallicity distribution function for the he s was published in @xcite .
we examine here the likely validity of this result , focusing on the most extreme of the emp stars .
we ignore issues of incompleteness , which are currently being studied by t. schrck and n. christlieb , and which while small at [ fe / h ] @xmath18 dex , become quite large by @xmath19 dex .
our 0z project has obtained moderate resolution follow up spectra of @xmath71550 candidate emp stars selected from the he s .
we have attempted to observe all stars with [ fe / h](hes ) @xmath9 dex as inferred from these follow up spectra at high resolution .
we list below the 12 objects from the 0z project for which the resulting [ fe / h](hes ) was less than @xmath20 dex .
( the i d is truncated for stars whose analyses we have not yet published . ) since these are in principle stars of great interest , the follow up spectra were scrutinized with care .
three of the 12 were eliminated in that way ; two are m dwarfs and one is a qso .
each object that survived ( with one exception ) has now been observed at high resolution with either hires at keck or mike at magellan .
of the three objects predicted to have [ fe / h ] @xmath22 dex , none have [ fe / h](hires ) @xmath23 dex .
this leaves 8 of the 12 objects .
three of these are c - stars , two of them have [ fe / h](hires ) @xmath24 dex , the third is the only one that does not yet have a high resolution spectrum .
it is , however , an c - star with extremely strong ch , and therefore is very likely to have [ fe / h ] @xmath25 dex based on our earlier work described in @xcite . we assume for present purposes that this is the case ; a hires spectrum to be obtained
shortly should confirm this .
two of the stars listed in the table are rediscoveries of previously known emp stars found by the hk survey@xcite . in the end , after all this effort , we have found only two new stars with [ fe / h](hires ) @xmath21 dex , out of @xmath71550 follow up spectra .
genuine emp stars are very rare , and the emp tail of the he s metallicity distribution function is heavily contaminated with stars which because of observational or analysis errors have landed up there , as well as with c - stars whose metallicity has been underestimated ( see * ? ? ?
* ) and with some spurious objects .
algorithms currently in use with segue do not attempt to assign a metallicity based on h@xmath8 and the ca ii k line to very c - rich stars .
this is a big step forward .
however , the extreme low metallicity end of any mdf generated from he s follow up spectra is likely to be heavily contaminated with additional higher metallicity objects which through various issues , including observational errors , end up getting assigned a spurious very low [ fe / h](hes ) .
we now have hires observations in hand for essentially all of the 16 stars from our 0z project with @xmath26 [ fe / h](hes ) @xmath27 dex .
of the six analyzed thus far , only three have [ fe / h](hires ) @xmath23 dex .
we are in the process of analyzing the rest of these to see what fraction of those in the regime of [ fe / h](hes ) between @xmath20 and @xmath28 dex are genuine . in the meantime , we urge caution in looking at the low - metallicity tail of any preliminary mdf from the he s in which the metallicities of stars at [ fe / h ] @xmath29 dex are based on moderate - resolution spectra rather than high - resolution spectroscopy . based in part on observations obtained at the w.m .
keck observatory , which is operated jointly by the california institute of technology , the university of california , and the national aeronautics and space administration .
some of the data presented herein were obtained at the palomar observatory .
we are grateful to the many people who have worked to make the keck telescope and hires a reality and to operate and maintain the keck observatory .
j.g.c . is grateful to nsf grant ast-0507219 for partial support .
n.c . is a research fellow of the royal swedish academy of sciences supported by a grant from the knut and alice wallenberg foundation .
he also acknowledges financial support from deutsche forschungsgemeinschaft through grants ch 214/3 and re 353/44 .
we thank tim beers for providing us with [ fe / h](hes ) values using the most recent version of his codes .
spite , m. et al . , _
a&a _ * 430 * 655 ( 2005 ) j. g. cohen , n. christlieb , y. z. qian & g. j. wasserburg , _ apj _ * 588 * 1082 ( 2003 ) j. g. cohen , a. mcwilliam , n. christlieb , s. shectman , i. thompson , j. melendez , l. wisotzki , & e. reimers , _ apjl _ * 659 * l161 ( 2007 ) | we give an update on our 0z survey to find more extremely metal poor ( emp ) stars with [ fe / h ] @xmath0 dex through mining the database of the hamburg / eso survey .
we present the most extreme such stars we have found from @xmath1 moderate resolution follow up spectra .
one of these , he1424@xmath20241 , has highly anomalous abundance ratios not seen in any previously known halo giant , with very deficient si , moderately deficient ca and ti , highly enhanced mn and co , and low c , all with respect to fe .
we suggest a snii where the nucleosynthetic yield for explosive @xmath3burning nuclei was very low compared to that for the hydrostatic @xmath3burning element mg , which is normal in this star relative to fe . a second , less extreme ,
outlier star with high [ sc / fe ] has also been found .
we examine the extremely metal - poor tail of the he s metallicity distribution function ( mdf ) .
we suggest on the basis of comparison of our high resolution detailed abundance analyses with [ fe / h](hes ) for stars in our sample that the mdf inferred from follow up spectra of the he s sample of candidate emp stars is heavily contaminated for [ fe / h](hes ) @xmath0 dex ; many of the supposed emp stars below that metallicity are of substantially higher fe - metallicity , including most of the very c - rich stars , or are spurious objects .
address = california institute of technology address = uppsala university , sweden address = carnegie observatories of washington address = carnegie observatories of washington address = carnegie observatories of washington |
stars rarely form in isolation . in fact , it is well known that the vast majority of stars in the galaxy , and also in nearby galaxies , are found in groups ranging from small associations , containing some 100 m@xmath0 , to compact , old `` globular '' and young massive clusters .
the nearest examples of these latter objects include the galactic star - forming regions ngc 3603 and westerlund 1 , and the giant starburst region 30 doradus with its central star cluster r136 in the large magellanic cloud .
although the older galactic _ open _ clusters ( with ages of several gyr ) are undoubtedly gravitationally bound objects , their lower masses compared to the globular cluster population , and more diffuse structures make them much more vulnerable to disk ( and bulge ) shocking when they pass through the galactic disk ( or close to the bulge ) , thus leading to enhanced cluster evaporation .
these objects are therefore unlikely globular cluster progenitors .
it appears that the conditions for the formation of compact , massive star clusters that have the potential to eventually evolve into globular cluster - type objects by the time they reach a similar age
are currently not present in the galaxy , or at best to a very limited extent ( e.g. , westerlund 1 ; hanson 2003 ) .
the production of luminous , massive yet compact star clusters seems to be a hallmark of the most intense star - forming episodes , or starbursts .
the defining properties of young massive star clusters ( yscs ; with masses often significantly in excess of @xmath1 ) have been explored in intense starburst regions in several dozen galaxies , often involved in gravitational interactions of some sort ( e.g. , holtzman et al .
1992 , whitmore et al .
1993 , oconnell et al .
1994 , conti et al .
1996 , watson et al .
1996 , carlson et al .
1998 , de grijs et al .
2001 , 2003a , b , c , d , e ) .
an increasingly large body of observational evidence suggests that a large fraction of the star formation in starbursts actually takes place in the form of such concentrated clusters , rather than in small - scale star - forming `` pockets '' .
yscs are therefore important as benchmarks of cluster formation and evolution .
they are also important as tracers of the history of star formation of their host galaxies , their chemical evolution , the initial mass function ( imf ) , and other physical characteristics in starbursts . in a detailed study of the young star cluster population associated with the fading starburst region `` b '' in the prototype nearby starburst galaxy m82 ( de grijs et al .
2001 ) , we concluded that the last tidal encounter between m82 and its large neighbour spiral galaxy m81 , which occurred about @xmath2 myr ago ( brouillet et al .
1991 , yun 1999 ) had a major impact on what was probably an otherwise normal , quiescent , disk galaxy .
it caused a concentrated starburst , resulting in a pronounced peak in the clusters age distribution , roughly 1 gyr ago ( de grijs et al .
2001 , 2003c , parmentier et al .
the enhanced cluster formation decreased rapidly , within a few hundred myr of its peak .
however , general star formation activity continued in the galactic disk of m82 s region b , probably at a much lower rate , until @xmath3 myr ago .
the evidence for the decoupling between cluster and field star formation is consistent with the view that star cluster formation requires special conditions , such as large - scale gas flows , in addition to the presence of dense gas ( cf . ashman & zepf 1992 , elmegreen & efremov 1997 ) .
such conditions occur naturally in the extreme environments of gravitationally interacting galaxies . using optical observations of the `` mice '' and `` tadpole '' interacting galaxies ( ngc 4676 and ugc 10214 , respectively ) based on a subset of the early release observations obtained with the _ advanced camera for surveys _ on board the _ hubble space telescope ( hst ) _ and the novel technique of pixel - by - pixel analysis of their colour - colour and colour - magnitude diagrammes
, we deduced the systems star and star cluster formation histories ( de grijs et al .
2003e ) . in both of these interacting systems
we find several dozen yscs ( or , alternatively , compact star - forming regions ) , which overlap spatially with regions of active star formation in the galaxies tidal tails and spiral arms ( from a comparison with h@xmath4 observations that trace active star formation ; hibbard & van gorkom 1996 ) .
we estimate that the main gravitational interactions responsible for the formation of these clusters occurred @xmath5 myr ago .
in fact , we show that star cluster formation is a major mode of star formation in galactic interactions , with @xmath6% of the active star formation in encounters occurring in star clusters ( de grijs et al .
the tidal tail of the tadpole system is dominated by star forming regions , which contribute @xmath7% of the total flux in the _ hst _ _ i_-band filter ( decreasing to @xmath8% in the equivalent _ b_-band filter ) .
if the encounter occurs between unevenly matched , gas - rich galaxies then , as expected , the effects of the gravitational interaction are much more pronounced in the smaller galaxy .
for instance , when we compare the impact of the interaction as evidenced by star cluster formation between m82 ( de grijs et al .
2001 , 2003b , c ) and m81 ( chandar et al . 2001 ) , or the star cluster formation history in the `` whirlpool galaxy '' m51 ( bik et al . 2003 ) , which is currently in the process of merging with the smaller spiral galaxy ngc 5194 , the evidence for enhanced cluster formation in the larger galaxy is minimal if at all detectable .
the ngc 6745 system represents a remarkably violently star - forming interacting pair of unevenly matched galaxies .
the optical morphology of ngc 6745 , and in particular the locations of the numerous bright blue star - forming complexes and compact cluster candidates , suggest a recent tidal passage by the small northern companion galaxy ( ngc 6745c ; nomenclature from karachentsev et al .
1978 ) across the eastern edge of the main galaxy , ngc 6745a .
the high relative velocities of the two colliding galaxies likely caused ram pressure at the surface of contact between both galaxies , which in turn is responsible for the triggering of enhanced star and cluster formation , most notably in the interaction zone in between the two galaxies , ngc 6745b ( cf .
de grijs et al .
the smaller galaxy , however , does not show any significant enhanced cluster formation , which is most likely an indication that it contains very little gas . for the ngc 6745 young cluster system we derive a median age of @xmath9 myr .
based on the age distribution of the star clusters , and on the hi morphology of the interacting system , we confirm the interaction - induced enhanced cluster formation scenario once again .
ngc 6745 contains a significant population of high - mass clusters , with masses in the range @xmath10 .
these clusters do not have counterparts among the galactic globular clusters ( e.g. , mandushev et al .
1991 , pryor & meylan 1993 ) , but are similar to or exceed the spectroscopically confirmed mass estimates of the `` super star clusters '' ( sscs ) in m82 ( e.g. , m82 f and l ; see smith & gallagher 2001 ) and the antennae galaxies ( mengel et al .
we caution , however , that these massive ssc candidates may not be gravitationally bound objects , but more diffuse star forming regions or aggregates of multiple unresolved clusters instead .
nevertheless , we measure a very compact effective radius for the most massive object ( @xmath11 ) of only @xmath12 pc .
however , this object appears very elongated , or may in fact be a double cluster .
we should also keep in mind that this high mass estimate is a strong function of the ( low ) metallicity assumed ; if we had assumed ( higher ) solar metallicity for this object , the derived age would have been significantly smaller ( @xmath13 myr vs. @xmath14 gyr assumed in our study ) , and the mass could be smaller by a factor of @xmath15 .
even so , if we could confirm this mass estimate spectroscopically , either of the subcomponents would be the most massive cluster known to date , significantly exceeding cluster w3 in ngc 7252 , which has a mass of about @xmath16 , depending on the age , metallicity and imf assumed ( schweizer & seitzer 1998 ; maraston et al .
2001 , 2004 ) .
our detection of such massive sscs in ngc 6745 , which are mostly located in the intense interaction zone , supports the scenario that such objects form preferentially in the extreme environments of interacting galaxies .
the ( statistical ) derivation of galaxy formation and evolution scenarios using their star cluster systems as tracers is limited to the study of integrated cluster properties for all but the nearest galaxies , even at _ hst _ spatial resolution .
the question remains , therefore , whether or not at least a fraction of the yscs seen in abundance in extragalactic starbursts , are potentially the progenitors of globular cluster - type objects in their host galaxies .
if we could settle this issue convincingly , one way or the other , the implications of such a result would have profound and far - reaching implications for a wide range of astrophysical questions , including ( but not limited to ) our understanding of the process of galaxy formation and assembly , and the process and conditions required for star ( cluster ) formation . because of the lack of a statistically significant sample of similar nearby objects , however , we need to resort to either statistical arguments or to the painstaking approach of one - by - one studies of individual objects in more distant galaxies . with the ever increasing number of large - aperture ground - based telescopes equipped with state - of - the - art high - resolution spectroscopic detectors and the wealth of observational data provided by the _ hst _
, we may now be getting close to resolving this important issue .
it is of paramount importance , however , that theoretical developements go hand in hand with observational advances .
the present state - of - the - art teaches us that the sizes , luminosities , and in several cases spectroscopic mass estimates of most ( young ) extragalactic star cluster systems are fully consistent with the expected properties of young milky way - type globular cluster progenitors ( e.g. , meurer 1995 , van den bergh 1995 , ho & filippenko 1996a , b , schweizer & seitzer 1998 , de grijs et al .
2001 , 2003d ) .
for instance , for the young massive star cluster system in the centre of the nearby starburst spiral galaxy ngc 3310 , we find a median mass of @xmath17 ( de grijs et al .
2003d ) ; their mass distribution is characterised by a gaussian width of @xmath18 dex . in view of the uncertainties introduced by the poorly known lower - mass slope of the stellar imf ( @xmath19 ) ,
our median mass estimate of the ngc 3310 cluster system which was most likely formed in a ( possibly extended ) global burst of cluster formation @xmath20 yr ago is remarkably close to that of the galactic globular cluster system ( cf .
de grijs et al .
2003d ; fig .
1 ) . however , the postulated evolutionary connection between the recently formed massive star clusters in regions of violent star formation and starburst galaxies , and old globular clusters in the nearby universe is still a contentious issue .
the evolution and survivability of young clusters depend crucially on the stellar imf of their constituent stars ( cf .
smith & gallagher 2001 ) : if the imf is too shallow , i.e. , if the clusters are significantly depleted in low - mass stars compared to ( for instance ) the solar neighbourhood , they will disperse within a few orbital periods around their host galaxy s centre , and likely within about a gyr of their formation ( e.g. , smith & gallagher 2001 , mengel et al .
2002 ) .
ideally , one would need to obtain ( i ) high - resolution spectroscopy ( e.g. , with 8m - class ground - based telescopes ) of all clusters in a given cluster sample in order to obtain dynamical mass estimates ( assuming that the clusters are in full virial equilibrium ) , and ( ii ) high - resolution imaging ( e.g. , with the _ hst _ ) to measure their luminosities .
one could then estimate the mass - to - light ( m / l ) ratios for each cluster , and their ages from the spectra .
the final , crucial analysis would involve a direct comparison between the clusters locations in the m / l ratio vs. age diagramme with models of `` simple stellar populations '' governed by a variety of imf descriptions ( cf . smith & gallagher 2001 , mengel et al .
. however , individual young star cluster spectroscopy , feasible today with 8m - class telescopes for the nearest systems , is very time - consuming , since observations of large numbers of clusters are required to obtain statistically significant results .
instead , one of the most important and most widely used diagnostics , both to infer the star ( cluster ) formation history of a given galaxy , and to constrain scenarios for its expected future evolution , is the distribution of cluster luminosities , or alternatively their associated masses , commonly referred to as the cluster luminosity and mass functions ( clf , cmf ) , respectively .
starting with the seminal work by elson & fall ( 1985 ) on the young cluster system in the large magellanic cloud ( with ages @xmath21 yr ) , an ever increasing body of evidence seems to imply that the clf of yscs is well described by a power law down to the lowest luminosities ( and thus masses ) . on the other hand , for the old globular cluster systems in the local universe , with ages @xmath22 gyr ,
the clf shape is well established to be roughly gaussian ( whitmore et al .
1995 , harris 1996 , 2001 , harris et al .
this shape is almost universal , showing only a weak dependence on the metallicity and mass of the host galaxy ( e.g. , harris 1996 , 2001 , whitmore et al .
2002 ) .
this type of observational evidence has led to the popular but thus far mostly speculative theoretical prediction that not only a power - law , but _ any _ initial clf ( and cmf ) will be rapidly transformed into a gaussian distribution because of ( i ) stellar evolutionary fading of the lowest - luminosity ( and therefore lowest - mass ) clusters to below the detection limit ; and ( ii ) disruption of the low - mass clusters due both to interactions with the gravitational field of the host galaxy , and to cluster - internal two - body relaxation effects ( such as caused by star - star collisions and the resulting redistribution of mass inside the cluster ) leading to enhanced cluster evaporation ( e.g. , elmegreen & efremov 1997 , gnedin & ostriker 1997 , ostriker & gnedin 1997 , fall & zhang 2001 ) .
we recently reported the first discovery of an approximately gaussian clf ( and cmf ) for the star clusters in m82 b formed roughly simultaneously in a pronounced burst of cluster formation ( de grijs et al .
2003b , c ) .
this provides the very first sufficiently deep clf ( and cmf ) for a star cluster population at intermediate age ( of @xmath14 billion years ) , which thus serves as an important benchmark for theories of the evolution of star cluster systems ( but see goudfrooij , this volume ) .
the shape of the clf ( cmf ) of young cluster systems has recently attracted renewed theoretical and observational attention .
various authors have pointed out that for young star clusters exhibiting an age range , one must first correct their clf to a common age before a realistic assessment of their evolution can be achieved ( e.g. , meurer 1995 , fritze v .
alvensleben 1998 , 1999 , de grijs et al .
2001 , 2003b , c ) .
this is particularly important for young cluster systems exhibiting an age spread that is a significant fraction of the system s median age , because of the rapid evolution of the colours and luminosities of star clusters at young ages ( below @xmath14 gyr ) .
whether the observed power laws of the clf and cmf for young star cluster systems are intrinsic to the cluster population or artefacts due to the presence of an age spread in the cluster population which might mask a differently shaped underlying distribution is therefore a matter of debate ( see , e.g. , carlson et al . 1998 ,
zhang & fall 1999 , vesperini 2000 , 2001 ) . nevertheless , the clf shape and characteristic luminosity of the m82 b cluster system is nearly identical to that of the apparently universal clfs of the old globular cluster systems in the local universe ( e.g. , whitmore et al .
1995 , harris 1996 , 2001 , ashman & zepf 1998 , harris et al . 1998 ) .
this is likely to remain virtually unchanged for a hubble time , if the currently most popular cluster disruption models hold . with the very short characteristic cluster disruption time - scale governing m82 b ( de grijs et al . 2003c )
, its cluster mass distribution will evolve towards a higher characteristic mass scale than that of the galactic globular cluster system by the time it reaches a similar age .
thus , this evidence , combined with the similar cluster sizes ( de grijs et al .
2001 ) , lends strong support to a scenario in which the current m82 b cluster population will eventually evolve into a significantly depleted old milky way - type globular cluster system dominated by a small number of high - mass clusters .
however , they will likely be more metal - rich than the present - day old globular cluster systems . the connection between young or intermediate - age star cluster systems , as in m82 b , and old globular clusters lends support to the hierarchical galaxy formation scenario .
old globular clusters were once thought to have been formed at the time of , or before , galaxy formation , i.e. , during the first galaxy mergers and collisions .
however , we have now shown that the evolved clf of the compact star clusters in m82 b most likely to survive for a hubble time will probably resemble the high - mass wing of the `` universal '' old globular cluster systems in the local universe .
proto - globular cluster formation thus appears to be continuing until the present .
in summary , in this review i have shown that young , massive star clusters are the most significant end products of violent star - forming episodes ( starbursts ) triggered by galaxy collisions and gravitational interactions in general .
their contribution to the total luminosity induced by such extreme conditions dominates , by far , the overall energy output due to the gravitationally - induced star formation .
the general characteristics of these newly - formed clusters ( such as their ages , masses , luminosities , and sizes ) suggest that at least a fraction may eventually evolve into equal , or perhaps slightly more massive , counterparts of the abundant old globular cluster systems in the local universe , although they will likely be more metal rich than the present generation of globular clusters .
establishing whether or not such an evolutionary connection exists requires our detailed knowledge of not only the physics underlying the evolution of `` simple '' stellar populations ( i.e. , idealised model clusters ) , but also that of cluster disruption in the time - dependent gravitational potentials of interacting galaxies .
initial results seem to indicate that proto - globular clusters do indeed continue to form today , which would support hierarchical galaxy formation scenarios . settling this issue
conclusively will have far - reaching consequences for our understanding of the process of galaxy formation and assembly , and of star formation itself , both of which processes are as yet poorly understood .
ashman k.m . ,
zepf s.e . , 1992 ,
apj , 384 , 50 ashman k.m . ,
zepf s.e . , 1998 , globular cluster systems , cambridge university press bik a. , lamers h.j.g.l.m . , bastian n. , panagia n. , romaniello m. , 2003 , a&a , 397 , 473 brouillet n. , baudry a. , combes f. , kaufman m. , bash f. , 1991 , a&a , 242 , 35 carlson m.n . , et al . , 1998 , aj , 115 , 1778 chandar r. , ford h.c . , tsvetanov z. , 2001 , aj , 122 , 1330 conti p.s .
, leitherer c. , vacca w.d . , 1996 ,
apj , 461 , l87 de grijs r. , oconnell r.w . ,
gallagher j.s .
, 2001 , aj , 121 , 768 de grijs r. , anders p. , lynds r. , bastian n. , lamers h.j.g.l.m . , fritze v .
alvensleben u. , 2003a , mnras , 343 , 1285 de grijs r. , bastian n. , lamers , h.j.g.l.m .
, 2003b , apj , 583 , l17 de grijs r. , bastian n. , lamers h.j.g.l.m . , 2003c , mnras , 340 , 197 de grijs r. , fritze v .
alvensleben u. , anders p. , gallagher j.s . , bastian n. , taylor v.a . , windhorst r.a .
, 2003d , mnras , 342 , 259 de grijs r. , lee j.t . , mora herrera m.c . ,
fritze v .
alvensleben u. , anders p. , 2003e , new astron .
, 8 , 155 elmegreen b.g .
, efremov y.n . , 1997 , apj , 480 , 235 elson r.a.w . ,
fall s.m . , 1985 , pasp , 97 , 692 fall s.m . , zhang q. , 2001 , apj , 561 , 751 fritze
alvensleben u. , 1998 , a&a , 336 , 83 fritze
alvensleben u. , 1999 , a&a , 342 , l25 gnedin o.y .
, ostriker j.p . , 1997 , apj , 474 , 223 hanson m.m . , 2003 , apj , 597 , 957 harris w.e . , 1996 , aj , 112 , 1487 harris w.e . , 2001 , in : star clusters , saas - fee advanced course 28 , spinger - verlag , 223 harris w.e . ,
harris g.l.h .
, mclaughlin d.e . , 1998 ,
aj , 115 , 1801 hibbard j.e . ,
van gorkom j.h . , 1996 ,
aj , 111 , 655 ho l.c . , filippenko a.v . ,
1996a , apj , 466 , l83 ho l.c .
, filippenko a.v . , 1996b , apj , 472 , 600 holtzman j.a . , et al . , 1992 ,
aj , 103 , 691 karachentsev i.d . , karachentseva v.e . , shcherbanovskii a.l . , 1978 , pazh , 4 , 483 ( english transl . in soviet astr .
lett . , 4 , 261 ) mandushev g. , spassova n. , staneva a. , 1991 , a&a , 252 , 94 maraston c. , kissler - patig m. , brodie j.p .
, barmby p. , huchra j.p .
, 2001 , a&a , 370 , 176 maraston c. , bastian n. , saglia r.p .
, kissler - patig m. , schweizer f. , goudfrooij p. , 2004
, a&a , in press ( astro - ph/0311232 ) mengel s. , lehnert m.d . , thatte n. , genzel r. , 2002 , a&a , 383 , 137 meurer g.r . , 1995 , nat . , 375 , 742 oconnell r.w . ,
gallagher j.s . ,
hunter d.a . , 1994 ,
apj , 433 , 65 ostriker j.p . ,
gnedin o.y . , 1997 , apj , 487 , 667 parmentier g. , de grijs r. , gilmore g.f . , 2003 , mnras , 342 , 208 pryor c. , meylan g. , 1993 , in : structure and dynamics of globular clusters , djorgovski s.g . , meylan g. , eds . ,
san francisco : asp , p. 357
schweizer f. , seitzer p. , 1998
, aj , 116 , 2206 smith l.j . ,
gallagher j.s . , 2001 ,
mnras , 326 , 1027 van den bergh s. , 1995 , nat . , 374 , 215 vesperini e. , 2000 , mnras , 318 , 841 vesperini e. , 2001 , mnras , 322 , 247 watson a.m. , et al . , 1996 , aj , 112 , 534 whitmore b.c . , schweizer f. , kundu a. , miller b.w . , 2002 ,
aj , 124 , 147 whitmore b.c . , schweizer f. , leitherer c. , borne k. , robert c. , 1993 , aj , 106 , 1354 whitmore b.c .
, sparks w.b .
, lucas r.a . ,
macchetto f.d . ,
biretta j.a . , 1995 ,
apj , 454 , l73 yun m.s . , 1999 ,
iau symp . 186 : galaxy interactions at low and high redshift , barnes j.e .
, sanders d.b .
zhang q. , fall s.m .
, 1999 , apj , 527 , l81 | the currently available empirical evidence on the star formation processes in the extreme , high - pressure environments induced by galaxy encounters , mostly based on high - resolution _ hubble space telescope _
imaging observations , strongly suggests that star _ cluster _
formation is an important and perhaps even the dominant mode of star formation in the starburst events associated with galaxy interactions .
the production of young massive star clusters ( yscs ) seems to be a hallmark of intense star formation , particularly in interacting and starburst galaxies .
their sizes , luminosities , and mass estimates are entirely consistent with what is expected for young milky way - type globular clusters ( gcs ) .
yscs are important because of what they can tell us about gc formation and evolution ( e.g. , initial characteristics and early survival rates ) .
they are also of prime importance as probes of the formation and ( chemical ) evolution of their host galaxies , and of the initial mass function in the extreme environments required for cluster formation .
recent evidence lends support to the scenario that milky way - type gcs ( although more metal rich ) , which were once thought to be the oldest building blocks of galaxies , are still forming today . |
there are few concepts in classical physics that are equally fundamental as the conservation of angular momentum .
one formidable consequence of this law is the formation of gaseous accretion disks around a wide range of astrophysical objects . yet to explain observed luminosities based on the release of gravitational binding energy @xcite , a robust mechanism is required to circumvent the consequences of angular momentum conservation .
being sufficiently ionised , these discs harbor dynamically important magnetic fields , which render the disk unstable to a mechanism called magnetorotational instability ( mri , * ? ? ?
* ) , releasing energy from the differential rotation and converting it into turbulent motions . ultimately , these motions are what is powering the redistribution of angular momentum on large scales . in the absence of
externally imposed large - scale fields , a separate dynamo mechanism may be required to replenish sufficiently coherent fields to drive sustained mri turbulence @xcite . this requirement becomes particularly apparent in models that neglect the vertical structure of the disk . in this case
a convergence problem has been encountered , which has been attributed to the lack of an outer scale of the turbulence @xcite .
an alternative explanation has been suggested by , who point out the problem of resolving the radial fine - structure related to non - axisymmetric mri modes ( required to circumvent cowling s `` no dynamo '' theorem ) .
convergence can naturally be recovered when accounting for the vertical stratification of the disk @xcite . together with rotational anisotropy
, the introduced vertical inhomogeneity can produce a pseudo - scalar leading to a classical mean - field dynamo @xcite . in a previous paper ( * ? ? ? * hereafter ) , we have inferred mean - field closure parameters from isothermal stratified mri simulations ( also see * ? ? ? * ) and demonstrated that the `` butterfly '' pattern ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) typical for stratified mri can in fact be described in such a framework . in the current paper
, we aim to extend this line of work towards a more realistic thermodynamic treatment , including heating from turbulent dissipation and crude heat transport .
this new effort is largely initiated by a recent paper of @xcite , hereafter .
the authors make the intriguing suggestion that the treatment of the disk thermodynamics will have a strong effect on the dynamo and accordingly on the saturation level of the turbulence .
this is in contrast to the result of @xcite , who find that turbulent transport is not affected by the presence of convection . as a first step to improve
the realism of mri simulations compared to the commonly applied isothermal equation of state , assume thermal conduction with constant thermal diffusivity , @xmath0 .
they then identify a critical value @xmath1 below which the primary mode of vertical heat transport changes from being predominantly conductive to being dominated by turbulent convective motions .
one motivation of the present work is to scrutinize these models and at the same time obtain mean - field closure coefficients for the two suggested regimes , thereby confirming the assumption that the thermodynamic treatment of the disk affects underlying dynamo processes .
a second focus of our paper will be the authors choice of impenetrable vertical boundary conditions , which we find to have profound implications for the resulting vertical disk equilibrium .
the simulations presented in this paper extend the simulation of @xcite to include a more realistic treatment of thermodynamic processes as pioneered by @xcite .
simulations are carried out using the second - order accurate nirvana - iiicode @xcite , which has been supplemented with the hlld riemann solver @xcite for improved accuracy .
we solve the standard mhd equations in the shearing - box approximation @xcite employing the finite - volume implementation of the orbital advection scheme as described in @xcite ; for interpolation we use the fourier method by @xcite .
we here neglect explicit viscous or resistive dissipation terms but include an artificial mass diffusion term ( as described in * ? ? ?
* ) to circumvent time - step constraints due to low density regions in upper disk layers .
we remark that , owing to the total - variation - diminishing ( tvd ) nature of our numerical scheme and the total energy formulation , heating via dissipation of kinetic and magnetic energy at small scales is accounted for even in the absence of explicit ( or artificial ) dissipation terms
. written in a cartesian coordinate system ( @xmath2 ) and with respect to conserved variables @xmath3 , @xmath4 , and the _ total _ energy @xmath5 , with @xmath6 being the thermal energy density , the equations read : @xmath7 \partial_t(\rho{\mathbf v } ) + \nabla{\!\cdot\!}[\rho\mathbf{vv}+p^{\star}-\mathbf{bb } ] & = & \rho\ , [ -\nabla\phi + \mathbf{a}_{\rm i } ] \,,\\[6pt ] \label{eq : energy } \partial_t e + \nabla{\!\cdot\!}[(e + p^{\star}){\mathbf v}- ( { \mathbf v}{\!\cdot\!}{\mathbf b}){\mathbf b } ] & = & \rho { \mathbf v}{\!\cdot\ ! } [ -\nabla\phi + \mathbf{a}_{\rm i } ] \nonumber\\ & + & \nabla{\!\cdot\!}(k\,\nabla t ) \,,\\[6pt ] \partial_t { \mathbf b}-\nabla{\!\times\!}({\mathbf v}{\!\times\!}{\mathbf b } ) & = & 0 \,,\nonumber\\ \nabla{\!\cdot\!}{\mathbf b } & = & 0\ , , \label{eq : mhd}\end{aligned}\ ] ] with the total pressure given by @xmath8 , and a fixed external potential @xmath9 .
the inertial acceleration @xmath10 arises due to tidal and coriolis forces in the local hill system , rotating with a fixed @xmath11 , and where the shear - rate @xmath12 has a value of @xmath13 for keplerian rotation .
we furthermore assume an adiabatic equation of state , such that the gas pressure relates to the thermal energy density as @xmath14 , with the ratio of specific heats @xmath15 , as appropriate for a mono - atomic dilute gas .
finally , the temperature , @xmath16 , appearing in the conductive energy flux , is obtained via the ideal - gas law @xmath17 , where we chose units such that the factor @xmath18 relating to the gas constant disappears . following the approach taken by ,
we adopt a thermal conductivity , @xmath19 , in terms of a constant diffusivity coefficient @xmath0 , related via @xmath20 for the isothermal run , we do not evolve equation ( [ eq : energy ] ) but instead obtain the gas pressure via @xmath21 , with fixed temperature @xmath22 .
note that , in our units , @xmath23 differs from by a factor of two , owing the alternative definition of the initial hydrostatic equilibrium , which is @xmath24 in our case .
for all simulations , we adopt the same box size of @xmath25 at a numerical resolution of @xmath26 grid cells in the radial , azimuthal , and vertical direction , respectively .
this corresponds to a linear resolution of @xmath27 in all three space dimensions . as typical for shearing box simulations , we initialize the velocity field with the equilibrium solution @xmath28 , and adiabatically perturb the density and pressure by a white - noise of @xmath29 rms amplitude .
the magnetic configuration is of the zero - net - flux ( znf ) type with a basic radial variation @xmath30 to obtain a uniform transition into turbulence , we further scale the vertical field with a factor @xmath31 resulting in @xmath32 ( @xmath33 initially ) .. ] owing to the divergence constrain , this of course can only be done by introducing a corresponding radial field at the same time . in practice , we specify a suitable vector potential .
horizontal boundary conditions ( bcs ) are of the standard sheared - periodic type , and we correct the hydrodynamic fluxes to retain the conservation properties of the finite - volume scheme @xcite . for the vertical boundaries ,
we implement stress - free bcs ( i.e. , @xmath34 ) with two different cases for the treatment of the vertical velocity component @xmath35 , namely : ( i ) impenetrable , and ( ii ) allowing for outflow ( but preventing in - fall of material from outside the domain ) .
we will demonstrate that this distinction will have profound implications for the resulting density and temperature profiles within the box .
to counter - act the severe mass loss occurring in the case of open bcs , we continuously rescale the mass density , keeping the velocity and thermal energy density intact . unlike in earlier work
@xcite , which was adopting an isothermal equation of state , we here do not restore towards the initial profile , but simply rescale the current profile .
this is , of course , essential to allow for an evolution of the vertical disc structure , owing to heating via turbulent dissipation .
we remark that such a replenishing of material can be thought of as a natural consequence of radial mass transport within a global disk . as in
, we use @xmath36 as boundary condition for the magnetic field , and impose a constant temperature @xmath37 at the top and bottom surfaces of the disk .
the latter choice is motivated by the assumption that the upper disk layers are likely optically thin , and there exists a thermal equilibrium with their surroundings . as in previous work , we compute the density and thermal energy of the adjacent grid cells in the @xmath38 direction to be in hydrostatic equilibrium .
the main motivation of this paper is to reproduce , as closely as possible , the results of , where we then aim to establish mean - field dynamo effects for the contrasting cases of efficient versus inefficient thermal conduction , as studied there .
moreover , we shall begin to explore the impact of vertical boundary conditions , by studying the somewhat more realistic case allowing for a vertical outflow of material .
.overview of simulation parameters , and results .
[ tab : models ] [ cols="<,^,^,^,^,^ " , ] we have performed in total four simulations : the two main simulations ( ` m2 ' and ` m3 ' ) adopt thermal diffusivity of @xmath39 , and @xmath40 , respectively , representing the regimes of efficient and inefficient thermal transport ( at molecular level ) , respectively . for both these simulations , and for a third isothermal reference run ( model ` m1 ' ) , we adopt _ outflow _ boundary conditions . to assess the impact of the imposed vertical bcs ( see column `` vbc '' in table [ tab : models ] ) , and to make direct contact with previous work , we adopt a fourth model , ` m4 ' , with a value @xmath40 , and _
impenetrable _ boundaries at the top and bottom of the domain ( labeled `` wall '' in the following ) .
.,title="fig : " ] + we ran the different models for approximately 300 orbital times , @xmath41 ; note that use time units of @xmath42 instead .
all time averages are taken in the interval @xmath43 $ ] .
for the isothermal reference run , we obtain a time - averaged maxwell stress , @xmath44 of @xmath45 , which is comparable to the @xmath39 case with @xmath46 . in the case
@xmath47 with outflow boundaries we find a somewhat higher value of @xmath48 , which is however significantly lower than in the otherwise identical case with impenetrable boundaries with @xmath49 .
while stresses are indeed increased ( by approximately 30% ) at lower conductivity , clearly , the effect of the treatment of the boundaries is much more significant .
$ ] orbits ) vertical profiles of the average maxwell stress for the different models.,title="fig : " ] + due to the strong fluctuations , the relative amplitudes are best seen in figure [ fig : prof_maxw ] , where we plot time - averaged vertical profiles of @xmath50 normalized to the initial midplane gas pressure , @xmath51 .
compared to the isothermal case m1 , the thermally conductive model m2 has a very similar vertical structure .
the low conductivity model m3 , where turbulent overturning motions dominate , is not too different in terms of its vertical profile either .
markedly , model m4 , with impenetrable vertical bcs , shows maxima of @xmath50 around @xmath52 .
this is similar to the profiles shown in figure 8 of , but note that their model shows a strong peak within @xmath53 of the vertical boundaries , while we only observe a very thin boundary layer in our model . for : model m4 with @xmath40 and impenetrable boundaries ( top ) , model m3 with @xmath40 and outflow boundaries ( middle ) , and model m2 with @xmath39 and outflow ( lower panel ) .
+ , title="fig : " ] + for : model m4 with @xmath40 and impenetrable boundaries ( top ) , model m3 with @xmath40 and outflow boundaries ( middle ) , and model m2 with @xmath39 and outflow ( lower panel ) .
+ , title="fig : " ] + for : model m4 with @xmath40 and impenetrable boundaries ( top ) , model m3 with @xmath40 and outflow boundaries ( middle ) , and model m2 with @xmath39 and outflow ( lower panel ) .
+ , title="fig : " ] at the end of their result section , point - out that the spatio - temporal behavior of the dynamo is remarkably different between the @xmath39 , and @xmath40 cases ( see their figure 9 ) . if we compare to run m4 with impenetrable boundaries ( uppermost panel of figure [ fig : spctm ] )
, we indeed find a similarly irregular butterfly diagram .
argue that they find `` no evidence for cyclic activity or pattern propagation '' in the conductive regime .
in contrast to this , looking at figure [ fig : spctm ] , it appears that even model m4 at some level shows a propagating dynamo wave .
more importantly , models m2 , and m3 ( which presumably are in the conductive , and convective regimes , respectively ) show extremely similar dynamo patterns and cycle frequency ( middle and lower panels in figure [ fig : spctm ] ) .
this again suggests that the vertical boundaries have a profound effect on the mechanism of vertical heat transport and , as a consequence , on the dynamo .
$ ] orbits ) temperature profiles.,title="fig : " ] + to establish the fact that our model m4 is indeed comparable to the corresponding simulation of , we now look at vertical profiles of the gas density and temperature . for the latter , had found a peculiar `` tent '' shape ( see their figure 4 ) , i.e. a linear dependence of @xmath54 on @xmath38 , joined together at @xmath55 by a parabolic segment . in figure
[ fig : prof_t ] , we compile the @xmath56 profiles , and such a tent - like profile can indeed be seen for model m4 .
however , in the case of low thermal diffusivity , @xmath40 , we observe a much weaker deviation from the isothermal profile with outflow boundary conditions . , but for the average density.,title="fig : " ] + along with the tent - shaped temperature profile , found that heat transport from convection would erase the vertical density stratification and lead to a constant density near the disk midplane ( see their figure 5 ) .
we observe a very similar density profile for model m4 , which moreover shows strong density peaks at the domain boundary .
these peaks are a consequence of enforcing the hydrodynamic fluxes to be zero at the domain boundaries . unlike for impenetrable vertical boundaries ,
model m3 shows a more regular gaussian density profile . for models
m1-m3 , the width of the bell - shaped density profiles is consistent with the trend in temperature and reflects hydrostatic equilibrium .
apparently , such an equilibrium can not be obtained if solid - wall boundaries are applied .
( solid lines ) , and turbulent convective heat flux @xmath57 ( dashed lines ) for the three non - isothermal models.,title="fig : " ] + conjecture that when @xmath0 crosses a critical value of @xmath58 , the vertical heat transport changes from being mainly conductive to being predominantly convective .
this was illustrated by their figure 6 , where they compared the conductive heat flux @xmath59 with the mean turbulent heat flux @xmath60 where horizontal lines as usual indicate averages over the @xmath61 and @xmath62 directions . in figure
[ fig : heat_flux ] , we show the corresponding quantities for our non - isothermal runs : as in previous plots , model m4 agrees very well with the result of , but again differs significantly from the otherwise identical model m3 with outflow boundary conditions . compared to m4
, the net transport of heat is much reduced in the case of models m2 , and m3 . for clarity
we plot these curves separately in the inset of figure [ fig : heat_flux ] with magnified ordinate .
much as expected , model m2 is dominated by a positive ( i.e. outward ) conductive heat flux , whereas the turbulent flux is negligible . unlike model
m4 with solid - wall boundaries , the low - conductivity model , m3 , only shows a very moderate level of convective heat flux , and notably one of the opposite sign
. this negative flux appears to largely balance its positive conductive counterpart , implying low levels of net - conductive heat transport towards the disk surface and net - convective transport towards the midplane .
it is instructive to note that the heating predominantly occurs at @xmath63 , where the velocity and magnetic field fluctuations are highest , and not , as one might think , near the midplane .
one objective of this work was to test the dependence of the dynamo on the mechanism of vertical heat transport .
one way of doing this is to establish mean - field closure parameters for the new class of models with constant thermal diffusivity . like in previous work
, we utilize the test - field ( tf ) method @xcite to measure coefficients such as the @xmath64 effect , turbulent pumping , and eddy diffusivity .
the used method @xcite is `` quasi - kinematic '' in the sense that it has been found to remain valid into the non - kinematic regime in the absence of magnetic background fluctuations @xcite whether this covers mri is a topic of discussion .
we here only briefly recapitulate the general framework of mean - field mhd and refer the reader to for a more detailed description . a recent review about mean - field dynamos
can be found in @xcite . for the shearing - box approximation , due to its periodic character in the horizontal direction ,
there are no characteristic gradients expected in the radial or azimuthal directions .
the natural mean - fields are accordingly those , which only vary in the vertical direction .
with respect to the velocity @xmath65 , the mean - field induction equation reads @xmath66 \ , , \label{eq : mf_ind}\end{aligned}\ ] ] where we have ignored a contribution due to microscopic magnetic diffusivity .
note that the explicit @xmath61 dependence in the shear term , @xmath67 , drops out once the curl operation is applied .
furthermore , @xmath68 because of flux conservation in the periodic box . in this description ,
turbulence effects due to _ correlated _ velocity and magnetic field fluctuations are embodied in the mean electromotive force @xmath69 which is typically parametrized as @xcite : @xmath70 & & { \rm where\ ; } i , j \in \left\{x , y\right\},\ , k\!=\!z\ , .
\label{eq : closure}\end{aligned}\ ] ] given explicit knowledge of the rank - two @xmath64 and @xmath71 tensors , this closure allows to formulate the mean - field induction equation ( [ eq : mf_ind ] ) in terms of mean quantities alone , leading to the classical @xmath72 dynamo description , first applied to mri turbulence by @xcite . .
axis labels indicate curves plotted in dark ( @xmath73 ) or light ( @xmath74 ) colours , respectively . the mean vertical velocity , @xmath75 , in panel ( b ) , and the classical estimate for the turbulent diffusion , in panel ( c ) , are shown as dashed lines.,title="fig : " ] + tensor coefficients representing the closure parameters
are presented in figure [ fig : dyn_k120 ] for model m2 , with a high value @xmath39 of the thermal diffusivity .
the obtained profiles are largely similar to the corresponding curves of the isothermal model m1 ( not shown ) , which moreover agree , which generally appears to be poorly constrained by our simulations . ] with previous isothermal results reported in .
notably , for model m2 , the contrast between the disc midplane and the upper disc layers is somewhat less pronounced compared to the purely isothermal case .
this trend is continued when going to lower thermal diffusivity ( see figure [ fig : dyn_k004 ] below ) .
unlike reported in , we now find @xmath76 and @xmath77 to be predominantly of the _ same _ sign , which would argue in favor of a kinematic ( rather than magnetic ) origin of the effect .
note however the significant fluctuations in @xmath76 , which cast some doubt on whether this coefficient can be meaningfully determined in the presence of shear . on the other hand ,
a fluctuating @xmath64 should not be disregarded as a possible source of a mean - field dynamo @xcite .
before we proceed , we briefly discuss the remaining coefficients . in panel
( b ) of figure [ fig : dyn_k120 ] , we show the off - diagonal tensor elements of the @xmath64 tensor , which are dominantly symmetric , i.e. @xmath78 .
we remark that for the classical diamagnetic pumping effect , one would require _ anti_-symmetric parity .
the observed symmetry may however be interpreted as differential pumping , i.e. transporting radial and azimuthal field in opposite directions . for reference
, we plot the mean vertical velocity @xmath75 ( see dashed line ) , which additionally transports the mean field and hence leads to the characteristic acceleration in the butterfly diagram .
the diagonal parts of the @xmath71 tensor are shown in panel ( c ) , where we also plot the rms velocity fluctuation ( dashed line ) .
apart from the boundary layers , the turbulent diffusivity agrees well with the theoretical expectation @xmath79 that is , assuming a coherence time @xmath80 of the turbulence .
given the dominance of the azimuthal field , the coefficients @xmath81 , and @xmath82 are surprisingly isotropic ( unlike predicted by * ? ? ?
it is however interesting to note that while @xmath83 , shown in panel ( d ) of figure [ fig : dyn_k120 ] , is identical to the diagonal elements of the diffusivity tensor in panel ( c ) , its counterpart @xmath84 is much smaller . with negative shear and both coefficients positive , the dynamo based on the @xcite effect is decaying ( cf . the dispersion relation in * ? ?
* appendix b ) .
this does however not exclude the possibility that @xmath84 has an effect on the overall pattern propagation . , but for the model m3 with @xmath40 .
note that for the sake of direct comparison , axis ranges are kept fixed with respect to the previous figure.,title="fig : " ] + we now proceed to the corresponding coefficients for the case of inefficient thermal conduction . in figure
[ fig : dyn_k004 ] we accordingly show the @xmath64 and @xmath71 tensor components for model m3 .
we recall that in the quasi - isothermal case ( cf .
figure [ fig : dyn_k120 ] ) the @xmath76 and @xmath77 coefficients showed a trend to flatten and even reverse their slope near the midplane ( also cf .
figure 9 in * ? ? ?
this was reasoned to be related to a _ negative _
@xmath64 effect due to magnetic buoyancy .
in contrast , here the @xmath76 curve shows a more monotonic dependence on @xmath38 , indicating that such magnetic effects may be less pronounced in this case .
such a trend appears consistent with reduced magnetic buoyancy in the case of a stiffer effective equation of state . moving on to panel ( b )
, we note that @xmath85 is now suppressed and even shows a slight trend to change its sign indicating a possible significance of diamagnetic pumping in the adiabatic case . for the turbulent diffusivity plotted in panel ( c ) ,
we observe a significant deviation from the classical expectation ( dashed line ) , resulting in a nearly constant @xmath86 .
we conclude that the equation of state and the means by which energy is transported to the upper disk layers indeed have subtle effects on the inferred dynamo tensors . which of the differences seen between figures [ fig : dyn_k120 ] and [ fig : dyn_k004 ] is in the end responsible for the enhanced dynamo activity seen in model m3 , will require further careful study . for completeness , in figure [ fig : dyn_k004_wall ] , we also show the results for model m4 , where dynamo coefficients are larger by a factor of several .
this can be considered consistent with the much higher maxwell stresses observed in this case . unlike for model m3
, @xmath87 now shows a pronounced @xmath38 dependence , and @xmath77 , and @xmath88 are in fact negative , and accordingly positive for @xmath89 ( for @xmath90 ) as suggested by @xcite . , but for the model m4 with @xmath40 and impenetrable vertical boundaries . note the different axis ranges as compared to the previous two figures.,title="fig : " ] +
the primary goal of our work was to make contact with recent results by , and to complement their work with a direct measurement of mean - field dynamo effects via the tf method . given that we have used a very similar numerical method and applied identical parameters ( i.e. for model m4 ) , it should not surprise the reader that we can satisfactorily confirm all aspects of the corresponding simulation by .
minor discrepancies arise with respect to the boundary layers , which may be related to the detailed treatment of the hydrostatic equilibrium there .
accordingly , we can confirm their main result , namely that in the presence of _ impenetrable _ vertical boundary conditions
one observes a transition from a conductively dominated vertical heat transport to a state that is regulated by convective overturning motions .
this transition obviously depends on the value of the applied constant thermal diffusivity @xmath0 . like
reported in , in the @xmath40 case , we observe a flat density profile ( even with a slight minimum at the disk midplane ) , and a `` tent''-shaped temperature profile presumably established by convective heat transport as a result of rayleigh - taylor - type instability .
this case is also associated with a much increased dynamo activity ( by an order of magnitude in the tf coefficients , not shown here ) , resulting in an overall maxwell stress that is increased by a factor of three compared to the isothermal reference model . unlike speculated by ,
we however do not think that the enhanced dynamo activity is related to the magnetic boundary conditions .
this is despite the fact that such a connection indeed exists for the unstratified case @xcite , where the different magnetic boundary conditions serve to create an inhomogeneity in an otherwise translationally symmetric system .
we rather attribute the different dynamo regime to the overall different hydrodynamic state
which however appears largely influenced by the choice of impenetrable boundary conditions . a separate set of models ( m1-m3 ) with a more natural condition allowing the gas to flow out of the domain shows much less dramatic effects when going to the low thermal diffusivity regime
naturally , one arrives at moderately hotter disk interiors along with more spread - out , yet still gaussian density profiles .
dynamo tf coefficients are somewhat altered in this case , along with a roughly 30% higher turbulent maxwell stress . establishing a link @xcite between the turbulent transport coefficients in the momentum equation (
i.e. the maxwell and reynolds stresses ) and the induction equation ( i.e. the @xmath64 effect , turbulent diffusion , etc . ) will be key to understand magnetized accretion in a quantitative manner , and derive powerful closure models in the spirit of @xcite or @xcite . a possible direct extension of the existing models with varying amounts of thermal conductivity may be to cross - correlate @xmath91 with e.g. @xmath77 for various values of @xmath0 . such a connection has been suggested by @xcite and been derived in the quasi - linear regime by .
a complication in this endeavor however arises from the fact that the tf coefficients are likely measured in a magnetically affected , i.e. quenched state . given the dramatic effect of open versus closed vertical boundaries demonstrated in this paper
, it will be of prime interest to study the connection between the disk and the launching of a magnetically driven wind @xcite , including a possible influence of the wind on the disk dynamo .
the amount of recent work illustrates the importance of this issue . in terms of the vertical boundaries imposed on the temperature , including a transition into an optically - thin disk corona will be important .
then a radiative boundary condition consistent with black - body radiation can be applied . to conclude , we want to emphasize that , clearly , the presented simulations can only be regarded as a first step towards a realistic treatment of the disk thermodynamics .
ideally , full - blown radiative transfer should be employed , and simulations of radiation dominated accretion disks @xcite demonstrate that this has indeed become feasible .
the author acknowledges the anonymous referee for providing a well - informed report and wishes to thank axel brandenburg and gianluigi bodo for useful comments on an earlier draft of this manuscript .
this work used the nirvana - iiicode developed by udo ziegler at the leibniz institute for astrophysics ( aip ) .
computations were performed on resources provided by the swedish national infrastructure for computing ( snic ) at the pdc centre for high performance computing ( pdc - hpc ) . ,
a. , nordlund , . , stein , r. f. , & torkelsson , u. 1995 , in lecture notes in physics , berlin springer verlag , vol .
462 , small - scale structures in three - dimensional hydrodynamic and magnetohydrodynamic turbulence , ed .
m. meneguzzi , a. pouquet , & p .-
sulem , 385 | we investigate the saturation level of hydromagnetic turbulence driven by the magnetorotational instability in the case of vanishing net flux .
motivated by a recent paper of @xcite , we here focus on the case of a non - isothermal equation of state with constant thermal diffusivity .
the central aim of the paper is to complement the previous result with closure parameters for mean - field dynamo models , and to test the hypothesis that the dynamo is affected by the mode of heat transport .
we perform computer simulations of local shearing - box models of stratified accretion disks with approximate treatment of radiative heat transport , which is modeled via thermal conduction .
we study the effect of varying the ( constant ) thermal diffusivity , and apply different vertical boundary conditions . in the case of impenetrable vertical boundaries ,
we confirm the transition from mainly conductive to mainly convective vertical heat transport below a critical thermal diffusivity .
this transition is however much less dramatic when more natural outflow boundary conditions are applied .
similarly , the enhancement of magnetic activity in this case is less pronounced .
nevertheless , heating via turbulent dissipation determines the thermodynamic structure of accretion disks , and clearly affects the properties of the related dynamo .
this effect may however have been overestimated in previous work , and a careful study of the role played by boundaries will be required . |
in this paper we report on the serendipitous detection of a foreground edge - on galaxy seen in absorption against the x - ray core of a cooling flow galaxy cluster .
the cluster , abell 2029 , is a nearby ( @xmath0 ) cluster which contains a central cd galaxy whose diffuse light extends up to 850 kpc @xcite .
it has been ( re)classified by @xcite as richness class 4.4 .
x - ray studies of the cluster show that it is a very relaxed system , and is one of the most luminous galaxy clusters [ @xmath13 ergs s@xmath14 .
we have examined the archival @xmath9 image of the cluster to study the structure of the thermal intracluster medium ( icm ) in the core of the cluster and the interaction of the central radio source with the icm ( t. e. clarke et al .
, in preparation , hereafter paper ii ) .
while analyzing the @xmath9 image , we discovered a linear absorption feature roughly 15 south of the cluster core .
this feature is due to the galaxy uzc j151054.6 + 054313 ( hereafter uzc j151054 ) , which is a beautiful example of a late - type spiral galaxy seen ( nearly ) edge - on in the foreground of the dense cluster abell 2029 .
it has a distinctive bulge component and a rather thin disk with a disk - to - bulge ratio of @xmath15 , consistent with the parameters of an sb / sc galaxy ( figure [ fig : optic_xray]a ) .
@xcite measured redshifts of 31 galaxies toward abell 2029 and found uzc j151054 to be in the foreground of the cluster , at a redshift of @xmath1 , while the cluster is at a redshift of @xmath0 .
the linear feature is due to photoelectric absorption of soft x - rays in the disk of the spiral galaxy which is seen against the diffuse thermal intracluster medium . in [ obs ] , we discuss the x - ray , optical , and radio observations and reductions for this galaxy .
the optical and details of the system are given in
[ opt_prop ] and [ hi_prop ] , respectively .
[ x - ray_prop ] presents the x - ray absorption due to the foreground spiral , and in [ dis ] we discuss the results of these observations . throughout this paper ,
we adopt a luminosity distance of 90 mpc for uzc j151054 .
we express the physical properties in terms of this distance , and scale quantities with @xmath8 .
this distance is consistent with a pure hubble flow and wmap cosmology @xcite , which gives @xmath16=93.2 mpc .
this is also in agreement with the tully - fisher distance of @xmath17 83 mpc ( see [ hi_prop ] below ) . at a distance of 90 mpc , the scale is 0.44 kpc / arcsec .
the central region of abell 2029 was observed with the @xmath9 x - ray observatory on 2000 april 12 for a total time of 19.8 ksec ( obsid 891 ) .
the observations were centered on the back - illuminated acis - s3 chip .
we present a detailed description of the data reductions in paper ii and only provide a brief summary here .
the data were analyzed using ciao version 2.3 and caldb 2.18 .
the events were filtered for energy ( 0.310 kev ) and grade ( asca 0,2,3,4 , and 6 ) .
the background light - curve was determined from the s1 chip and showed no strong flares , and we removed only 128 seconds of data due to high background .
the background for the observation was determined from the blank sky background files collected by m. markevitch . for the purposes of studying the absorption due to the foreground galaxy ,
it is the soft x - ray background in the 0.31 kev band which is most important . the soft background , which is mainly galactic , varies by a factor of @xmath182 even at high galactic latitudes ( abell 2029 is at @xmath19 ) .
thus , it would be preferable to use an in - situ background from the observation .
unfortunately , the @xmath20 pspc image of abell 2029 @xcite shows that the thermal x - ray emission from the cluster covers the entire s3 chip , and even extends beyond the @xmath9 s1 chip ( the other backside - illuminated chip ) . as a result
, it is difficult to determine an in - situ background from the observation . for this absorption study
, it would also be useful to separate the background into particle and galactic foregrounds and cosmological x - ray background .
although the correct background is thus uncertain by a factor of about two , the dominant `` background '' in these measurement is actually the diffuse cluster emission from abell 2029 , and the results presented here are affected by less than 10% by this uncertainty in the real background .
the sense of the variation is that , if the background is higher than we have assumed ( the likely sense of the variation ) , we have underestimated the absorption by the spiral galaxy , albeit very slightly . figure [ fig : optic_xray]b , c shows gaussian smoothed ( @xmath21=1 ) images in the soft and hard x - ray bands of the region around the spiral galaxy uzc j151054 , both corrected for exposure and background .
the emission gradient seen from north to south in the image is due to the thermal icm of the core of the cluster abell 2029 .
detailed x - ray images of the galaxy cluster are presented in paper ii .
the x - ray image was compared to the two micron all sky survey ( 2mass ; * ? ? ?
* ) to check the positional accuracy .
there are three 2mass ids which are associated ( within 1 ) with x - ray sources , and comparison of positions shows no significant shift between the two coordinate systems .
@xmath2 , @xmath3 , @xmath4 , and @xmath5 band photometry has been obtained for a different project by one of us ( jmu ) in collaboration with s. p. boughn ( haverford ) and j. r. kuhn ( hawaii ) .
the observations were taken with the 0.9 m telescope on kitt peak national observatory on 1987 may 20 and 1988 may 10 ( @xmath4 ) , 1988 may 8 ( @xmath3 ) , and 1988 may 14 ( @xmath2 ) .
the no . 3 rca ccd was used for these observations mounted at the f:7.5 cassegrain focus , which resulted in square pixels of 0@xmath2286 on a side .
the seeing was @xmath23 .
the observations produced mosaics of abell 2029 which included the foreground spiral galaxy that we discuss in this paper ( see * ? ? ?
* for details of the observing and data reduction techniques ) .
@xmath5 band photometry of abell 2029 was obtained on 1998 april 19 using the t2ka camera mounted at the f:7.5 cassegrain focus , which resulted in square pixels of 0@xmath2268 on a side . throughout the run
the seeing was excellent , between 0@xmath227 and 0@xmath229 , which resulted in an effective seeing of @xmath242 due to the available pixel size .
the spatially - overlapping exposures were used to make a mosaic of about 35@xmath25 ( ra ) by 58@xmath25 ( dec ) again including the galaxy discussed here @xcite .
figure [ fig : optic_xray]a shows our galaxy immersed in the diffuse halo that pervades the center of abell 2029 and extends about 850 kpc from the cluster center @xcite .
the coordinate system of the @xmath5 band image of the spiral galaxy was tied to the 2mass position frame using five matches in the field .
a @xmath26 cm continuum image has been obtained from a deep hi survey of abell 2029 .
the nrao very large array ( vla ) was used in its c configuration to cover the frequency range of 1300.7 mhz to 1335.2 mhz appropriate to the redshift range of the cluster .
although the frequency range precluded the observation of at the redshift of the foreground spiral , it yields a good continuum image of the field , with a resolution of @xmath27 and an rms noise of @xmath28 @xmath29jy / beam .
figure [ fig : rad_cont ] shows that the spiral galaxy is undetected in the radio continuum ; we derive an integrated flux of @xmath30 @xmath29jy which leads to a @xmath31 upper limit of 200 @xmath29jy over the full area of the galaxy ( 270 arcsec@xmath32 )
. observations were obtained for us by jodie martin ( uva ) and john hibbard ( nrao ) in two 10 minute observations with the nrao green bank telescope ( gbt ) .
the observations were made with the spectral processor using two orthogonal polarizations with bandwidths of 10 mhz with a frequency resolution of 9.8 khz .
the elevation was @xmath33 degrees which caused a fair amount of radio frequency interference ( rfi ) pickup .
the data were ( twice ) hanning - smoothed off - line in order to minimize a ripple due to strong out - of - band rfi , giving a final resolution of @xmath34 .
the data were reduced using the dish package in aips++ . even in the presence of rfi ,
the spectral stability was quite adequate and we have only subtracted a linear baseline from the data . the calibration was referenced to 3c295 which was observed about 12 hours after the galaxy at about the same elevation .
we adopt a value for the flux of 3c295 of 22.46 jy at our observing frequency ( 1391 mhz ) from the @xcite scale which leads to the spectrum shown in figure [ fig : hi ] . from the scatter of the data in the line free channels
, we estimate the 1-@xmath21 error per channel for the two polarizations to be 3.5 mjy ( continuous line ) and 4.4 mjy ( dashed line ) .
the average redshift is 6438 ( optical definition , heliocentric ) in good agreement with the value of @xmath35 listed in the updated zwicky catalog @xcite , which was obtained by @xcite using the arecibo telescope .
correction to the frame of the cosmic microwave background radiation yields a redshift of c@xmath36 placing the galaxy at a distance of @xmath37 mpc assuming a pure hubble flow .
we shall use this distance to derive physical quantities below but express those values in units of @xmath8 where appropriate .
we have used the prescription in @xcite to determine the inclination and find a most - likely value of @xmath38 , assuming an intrinsic aspect ratio of 10:1 for the disk .
the lower limit is @xmath39 if we assume that the galaxy is thin .
we have adopted a value of @xmath40 .
the holmberg diameter is barely reached by the observations and has a value of @xmath41 @xmath8 kpc .
table [ tab : opt_phot ] shows the results of the photometry .
given the seeing differences we only determine the photometry for the center of the galaxy ( a @xmath42 box ) , the bulge , the disk , and the galaxy as a whole .
to define the regions used for photometry we took an average of the @xmath2 , @xmath3 , @xmath4 , and @xmath5 frames to make masks and used the same area for all bands in the photometry .
the values are corrected for foreground extinction using the prescription of @xcite ; but are not corrected for internal extinction ( see below ) .
the background was determined by first masking all sources visible in any of the four bands , then estimating the average and gradient due to the cd halo from the remaining pixels on either side of the nucleus both above and below the disk .
we estimate the photometry to be accurate to @xmath43 mag given the difficulty to estimate the local `` background '' levels in the presence of faint cluster galaxies and the large envelope of the cd galaxy that dominates abell 2029 .
similarly , we estimate the colors to be accurate to @xmath18 0.05 0.1 mag . the @xmath44 colors are consistent with those of scd galaxies given by @xcite with the @xmath45-corrections interpolated linearly to @xmath46 .
for example , the global color is @xmath47 when 0.82 ( 1.14 ) is the corresponding `` average '' color of an scd ( sbc ) galaxy .
the @xmath48 color is a bit blue ( 0.34 ) where 0.53 is expected for a typical scd galaxy ( 0.77 for an sbc ) , and could be affected by a contribution from as well .
the @xmath49 color is even bluer ( 1.17 when 1.37 is expected for an scd galaxy and 1.76 for an sbc ) , which might be an indication of recent star formation .
these colors are uncorrected for internal extinction which would , of course , make them even bluer .
there is a long - standing debate on the optical depth of ( late ) spiral galaxies ( e.g. , * ? ? ?
* and references therein ) .
the correction to a face - on orientation is usually parametrized in terms of the aspect ratio as @xmath50 where @xmath51 is the correction to the i band magnitude , @xmath52 and @xmath53 are the apparent semi - major and semi - minor axes of the galaxy , and @xmath54 is an empirical coefficient . because our galaxy is about 2 magnitudes dimmer than @xmath55 , @xmath56 @xcite leading to a correction of @xmath57 . for comparison , we have used the two relations offered by tully and collaborators @xcite to compute the same correction , which in our case lead to values of @xmath58 using the relation that depends on the absolute magnitude , and @xmath59 using the dependence on the velocity width .
given the statistical character of these prescriptions , we find their agreement quite encouraging .
adding a @xmath45-correction of @xmath60 magnitudes from @xcite leads to an @xmath5 band estimate of the luminosity of @xmath61 .
the expected stellar mass - to - light ratio of uzc j151054 is @xmath62 in solar units @xcite yielding an estimate of the total stellar mass of the galaxy of @xmath63 .
we have opted to use the average of both of the @xcite methods to similarly estimate a correction to the @xmath2 band photometry , which yields a value of @xmath64 .
this allows us to estimate an absolute magnitude of @xmath65 ( which includes a @xmath45-correction of @xmath66 from * ? ? ?
* ) corrected to face - on but not for any residual effects from scattering or internal extinction .
our ( final ) estimate of the @xmath2 band luminosity of this galaxy is thus @xmath67 .
the expected stellar mass - to - light ratio of uzc j151054 is @xmath68 in solar units @xcite .
thus , the total stellar mass of the galaxy is expected to be @xmath69 in good agreement with the estimate from the @xmath5 band photometry .
[ [ hi_prop ] ] the spectrum shows an asymptotic velocity of @xmath70 .
although the emission often extends beyond the optical emission , we derive a lower limit to the total mass of the galaxy by estimating the mass internal to the optical radius of 12 @xmath8 kpc as @xmath71 .
integration of the spectra yield estimates of the total fluxes of @xmath72 jy and @xmath73 jy for polarizations one and two , respectively , where the errors are derived from the scatter of the line - free channels and do not include calibration errors . as seen in figure [ fig : hi ] , the second polarization is corrupted by rfi , so that we shall ignore it henceforth ( although it does offer confirmation of the velocity extent of the ) . after our analysis was completed , we have received the spectrum from @xcite electronically .
their measured total flux is @xmath74 jy , which we have been told should be increased by perhaps 5% to account for pointing errors and should be accurate to about 15% ( m. haynes , private communication ) . given our short integration , we consider the agreement of the two derived total fluxes somewhat coincidental . nevertheless , we are encouraged to derive an mass from our spectrum and arrive at a value of @xmath75 which is accurate to @xmath18 10% .
thus , the galaxy shares in the common `` dark matter '' problem with ratios of @xmath76 , @xmath77 ( in solar units ) .
the total of the stellar and mass is approximately @xmath78 , which is still about five times smaller than the total mass from the rotation velocity .
thus , the dark matter mass of the galaxy is approximately @xmath79 , out to a radius of @xmath80 kpc .
we have also used the `` baryonic tully - fisher '' relation derived by @xcite to estimate the distance to uzc j151054 . given the uncertainty that recent star formation adds to the @xmath2 band determination of the stellar mass , we have used the stellar mass derived above from the @xmath5 band photometry ( @xmath81 ) .
equation 2 of @xcite yields an estimate of the mass in stars and gas of @xmath82 . correcting the mass for the he fraction yields a value of @xmath83 and a luminosity distance of @xmath84=83 mpc , in good agreement with the hubble flow distance given the uncertainties and the unknown peculiar velocity of uzc j151054 . using a value of @xmath85 appropriate for an 8 gyr sc galaxy with approximately continuous star formation @xcite , instead of our adopted value of 1.7 ( corresponding to a 12 gyr sc galaxy ) would lead to @xmath84=87 mpc .
examination of figure [ fig : optic_xray ] shows that the x - ray absorption region is roughly the size of the optical disk of uzc j151054 .
comparison of the hard ( 1.07.0 kev ) and soft ( 0.31.0 kev ) x - ray images indicates that the absorption is more prominent in the soft x - ray band as expected for photoelectric absorption .
the @xmath9 0.31.0 kev image contains 82 counts in the region of absorption , compared to 197 and 184 counts in the same size region above and below the absorption feature .
this corresponds to a @xmath86 detection of the foreground spiral galaxy absorption feature
. a careful comparison of the position angle of the optical galaxy and the x - ray absorption feature shows that the x - ray absorption may be slightly tilted ( @xmath87 ) relative to the galaxy disk ( figure [ fig : optic_xray ] ) .
such a misalignment could be the result of a warp in the hi disk relative to the optical disk in uzc j151054 .
the extent of the absorption region was determined by taking cuts through the x - ray images .
we used square regions of width 36 running roughly east to west ( pa=872 ) along the x - ray absorption feature and determined the counts in each region from the 0.31.0 kev @xmath9 image .
the box height is approximately the size of the optical spiral disk and the width was set small enough to allow several boxes to be placed along the absorption feature with sufficient counts to determine the extent of the absorption region . in figure
[ fig : counts ] we show a series of exposure and background corrected counts from 34 regions running east to west along the absorption feature .
figure [ fig : counts ] also shows the results from shifting the boxes up and down by one box height and determining the counts both above and below the absorption feature . in figure
[ fig : counts_perp ] we show a similar plot for a set of 31 regions of size 55 @xmath88 05 running north - south parallel to the disk . note that the psf of the @xmath9 acis - s3 detector at these positions and energies is mainly determined by the pixel size on the detector , and the 80% encircled energy radius is about 07 .
thus , the absorption from the disk is resolved perpendicular to the disk .
the limited number of counts in the absorption region preclude any detailed modeling of the spatial structure of the absorption .
the full width of the absorption along the disk is about 56 ( 25 kpc ) .
the dashed vertical lines in figure [ fig : counts ] show the approximate i - band optical extent of the galaxy .
the width of the absorption feature agrees fairly well with the optical width of the galaxy . in some ways
, this might be surprising , since many spiral galaxies have disks which extend beyond their optical disks .
on the other hand , the x - ray absorption at these energies is mainly due to heavy elements ( primarily oxygen ) , and the abundances of heavy elements are likely to be low outside of the optical disk . in the vertical direction ,
the width of the absorption disk is about 5 ( 2.2 kpc ) , after making a small correction for the psf of the @xmath9 acis - s3 .
the size and strength of the x - ray absorption can also be characterized by the equivalent width , or the width of a region with complete absorption and the same reduction in flux as that of the spiral galaxy absorption feature . using the 0.31.0 kev data , we determine an equivalent width of @xmath18 34 ( 15 kpc ) along the disk , and @xmath18 3 ( 1.3 kpc ) perpendicular to it . these values are smaller than the full width of the feature because the absorption is not complete . to analyze in more detail the absorption due to the spiral disk , we extracted spectra from the feature and a larger surrounding `` background '' area .
the spectra were binned to a minimum of 25 cts / bin to provide good statistics .
the region of the absorption feature and the surrounding thermal icm were fit in xspec with a wabs * zwabs * mekal model after applying a correction for the quantum efficiency degradation .
the mekal model represents the background cluster emission , while the zwabs is the absorption from the foreground spiral . the galactic column in wabs
was set to @xmath89 @xmath11 as determined from @xcite ) are less than 10% .
we also allowed the galactic column to be free in the fits and found a best fit value lower than the galactic value ( @xmath90 = @xmath91 @xmath11 ) .
this low galactic hydrogen column density is consistent with that found by @xcite for the outer regions of the cluster . using the lower galactic column for the fits results in less than a 5% change in the excess absorption column due to the foreground spiral galaxy . ] . in the fit to the absorption feature ,
the temperature and metallicity for the cluster emission were fixed at the value determined by the surrounding `` background '' region , but the excess absorption was allowed to vary .
the surrounding thermal icm is best fit with a temperature of 7.1 kev ( @xmath92/d.o.f .
= 141/175 ) , slightly lower than the deprojected temperature found by @xcite .
the model fits in xspec were taken for the energy range of 0.510.0 kev .
we extended the model fits down to 0.3 kev and found consistent results although there were significantly larger residuals at the lowest energies in the fits of the surrounding region . the best fit model ( @xmath92/d.o.f .
= 152/193 ) to the x - ray spectrum ( figure [ fig : xspec ] ) gives an excess absorbing column ( with 90% confidence interval ) of @xmath10 @xmath11 . at these columns ,
almost all of the absorption is due to heavy elements , particularly oxygen .
thus , this measurement is most directly interpretable as a measurement of the column density of oxygen ( in any form containing k - shell electrons ) of @xmath93 @xmath11 ( assuming the metallicity of * ? ? ?
the column density of hydrogen can thus be written as @xmath94^{-1 } $ ] @xmath11 , where @xmath95 is the oxygen abundance in units of the solar abundance .
this corresponds to a mass of @xmath96^{-1 } \ , d_{90}^2 \ , $ ] , assuming that the absorber is spread uniformly over the 556 by 36 region covered by the disk of the galaxy . of course
, it is likely that the interstellar gas and dust in uzc j151054 are concentrated to the center of the galaxy and the middle of the disk , and that the absorption is not uniform .
thus , it is likely that the column density and derived mass are really lower limits to the total amount of absorbing material .
we note that the hydrogen mass determined from the x - ray absorption ( assuming a uniform absorber and solar abundances ) is about a factor of 5 smaller than the atomic hydrogen mass from the 21 cm observations
. this may be due to a non - uniform distribution of the absorber , or due to a lower than solar metallicity in the interstellar medium ( ism ) in uzc j151054 .
the x - ray absorption feature appears to be limited vertically to the approximate width of the optical disk in uzc j151054 . in order to limit the absorption due to material in the halo of this galaxy
, we also extracted the spectra from regions which were 556 by 36 but shifted by 36 above and below the disk .
the spectra of the two regions above and below the disk were combined , and fit in the same way as the absorption from the disk .
the best - fit model had an excess absorption of @xmath97 @xmath11 ( @xmath92/d.o.f .
= 173/217 ) .
since the excess is consistent with zero at the 90% confidence level , we interpret this measurement as a 90% confidence upper limit of @xmath98 @xmath11 .
we have found a deep x - ray absorption feature due to a foreground spiral galaxy seen in projection against the core of the galaxy cluster abell 2029 . based on the 0.31.0 kev @xmath9 data , the x - ray deficit associated with the spiral disk of uzc j151054
is detected at @xmath99 8@xmath21 significance .
the x - ray absorption is strongest at low energies , as expected for photoelectric absorption . from the x - ray spectrum of the absorption feature , we find an absorbing column of @xmath10 @xmath11 , assuming solar abundances .
this corresponds to a mass of @xmath100 of hydrogen in the spiral disk , if the absorber is uniformly distributed over the region of the spiral disk . if the absorber is not uniform , as seems more likely , the required mass of the absorber is higher .
an analysis of the optical data shows that the galaxy colors are consistent with those of an scd galaxy , while the axial ratio gives an inclination of @xmath40 .
we fail to detect radio continuum emission from uzc j151054.6 + 054313 at a reasonably low level .
our limit of @xmath101jy is consistent with the lack of significant emission in the ir ; the galaxy is barely detectable in the 2mass survey images @xcite .
the detected x - ray absorption is smaller than that expected from the amount of neutral hydrogen detected in 21-cm emission from the disk of uzc j151054 .
this may indicate that the ism in uzc j151054 is not uniformly distributed over the absorption region , or that the metallicity is low .
the x - ray absorption measurements also provide an upper limit on any additional diffuse gas or dust in this galaxy .
we note that x - ray absorption is relatively insensitive to the physical state of the diffuse material .
any form of gas or dust at temperatures below @xmath102 k would produce absorption . at the columns of interest here ,
most of the absorption is due to the k - shell electrons in oxygen .
the measurements directly give the column density of oxygen , rather than hydrogen .
thus , the limits of the mass of the absorbing material depend inversely on the abundance of oxygen relative to hydrogen . the mass of the absorbing material we detect , @xmath103 , is much less than the total mass of the galaxy or the dark matter mass , @xmath104 .
thus , our measurements limit the possibility that the dark matter is diffuse baryonic gas .
if the dark matter is in the form of a roughly spherical halo , then our limits on the absorption outside the disk yield a limit on the mass of @xmath105 , which is about a factor of three smaller than the dark matter mass .
if the dark matter is in the disk and is uniformly distributed , then the limit on its mass is certainly less than the total absorbing mass determined for the disk of @xmath106 .
thus , the dark matter can not be diffuse gas , unless it either has very low metal abundances or is very inhomogeneous in its distribution . in recent years
, there have been a number of suggestions that dark matter in galaxies is indeed baryonic , and is due to dense clouds of gas ( e.g. , * ? ? ?
* ; * ? ? ?
this gas might either be located in a thick disk @xcite or in a more spherical halo @xcite .
however , in either model , the clouds of gas are rather small and dense , and have a small covering factor .
thus , the maximum optical depth for absorption across the galaxy is limited by this covering factor . as a result ,
our measurements for uzc j151054 do not refute or strongly constrain these theories .
finally , we note that a much longer ( 80 ksec ) _ chandra _ observation of the center of abell 2029 is planned for cycle 5 . among other aims
, this observation should allow the absorption feature due to uzc j151054 to be studied in more detail .
it would also be useful to have a 21 cm line image of the galaxy to compare the emission line with absorption from the interstellar medium .
we thank renzo sancisi for a very interesting conversation about baryonic dark matter in galaxies .
we thank steve boughn for useful comments on @xmath45-corrections , riccardo giovanelli and mort roberts for their insights on internal extinction in late - spiral galaxies .
we also thank jodie martin and john hibbard for donating the 20 minute observation with the gbt and collecting the data for us as well as for the initial reduction of the data with the dish package .
martha haynes kindly sent us a digitized copy of the arecibo spectrum of uzc j151054.6 + 054313 .
support for this work was provided by the national aeronautics and space administration through _ chandra _ award numbers go2 - 3159x and go2 - 3160x , issued by the _
chandra _ x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 39073 .
support for e. l. b. was provided by nasa through the _ chandra _ fellowship program , grant award number pf1 - 20017 , under nasa contract number nas8 - 39073 .
we have used the software packages aips and dish / aips++ of the nrao .
the national radio astronomy observatory is operated by associated universities , inc .
, under cooperative agreement with the national science foundation .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
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 .
, r. m. , et al .
2001 , explanatory supplement to the 2mass second incremental data release , the two micron all sky survey at ipac website : http://www.ipac.caltech.edu/2mass/releases/second/doc/explsup.html lccccc center & 36 & 18.70 & 18.21 & 17.70 & 17.31 + bulge & 58 & 18.55 & 18.05 & 17.59 & 17.20 + disk & 212 & 18.60 & 18.31 & 17.92 & 17.66 + galaxy & 270 & 17.82 & 17.42 & 16.99 & 16.65 + [ tab : opt_phot ] | we have detected an x - ray absorption feature against the core of the galaxy cluster abell 2029 ( @xmath0 ) which we identify with the foreground galaxy uzc j151054.6 + 054313 ( @xmath1 ) .
optical observations ( @xmath2 , @xmath3 , @xmath4 , and @xmath5 ) indicate that it is an scd galaxy seen nearly edge - on at an inclination of @xmath6 .
observations give a rotation velocity of 108 and an atomic hydrogen mass of @xmath7 , where @xmath8 is the distance to the galaxy in units of 90 mpc .
x - ray spectral fits to the @xmath9 absorption feature yield a hydrogen column density of @xmath10 @xmath11 assuming solar abundances .
if the absorber is uniformly distributed over the disk of the galaxy , the implied hydrogen mass is @xmath12 .
since the absorbing gas in the galaxy is probably concentrated to the center of the galaxy and the middle of the disk , this is a lower limit to the total hydrogen mass . on the other hand ,
the absorption measurements imply that the dark matter in uzc j151054.6 + 054313 is not distributed in a relatively uniform diffuse gas . |
it was proved by chatzidakis and pillay @xcite that if @xmath4 is a first order superstable theory , and the theory @xmath5 has a model companion @xmath6 , then @xmath6 is supersimple . throughout this paper
we refer to @xmath6 ( when it exists ) as the theory of models of @xmath4 equipped with a _
generic _ automorphism .
continuous first order logic is an extension of first order logic , introduced in @xcite as a formalism for a model theoretic treatment of metric structures ( see also @xcite for a general exposition of the model theory of metric structures ) .
it is a natural question to ask whether the theorem of chatzidakis and pillay generalises to continuous logic and metric structures .
the proof of chatzidakis and pillay would hold in metric structures if we used the classical definitions of superstability and supersimplicity literally ( namely , types do not fork over finite sets ) .
these definitions , however , are known to be too strong in metric structures , and need to be weakened somewhat in order to make sense .
for example , the theory of hilbert spaces has a countable language , is totally categorical and does not satisfy the classical definition of superstability .
the standard definition for @xmath1-stability and superstability for metric structures ( @xcite , and later @xcite ) comes from measuring the size of a type space not by its cardinality but by its density character in the metric induced on it from the structures .
a continuous theory is supersimple if for every @xmath7 , the @xmath8-neighbourhood of a type does not fork over a finite set of parameters , or equivalently , if ordinal lascar ranks corresponding to `` @xmath8-dividing '' exist .
a theory is superstable if and only if it is stable and supersimple .
similarly , @xmath1-stability is equivalent to the existence of ordinal @xmath8-morley ranks , which may be defined via a metric variant of the classical cantor - bendixson ranks ( see @xcite for a general study of such ranks ) . with these corrected definitions ,
the class of @xmath1-stable theories is rich with examples : hilbert spaces , probability algebras , @xmath9 banach lattices and so on .
furthermore , many classical results can be generalised .
for example , an @xmath1-stable theory has prime models over every set , an uncountably categorical theory in a countable language is @xmath1-stable , and so on ( see @xcite ) .
a somewhat more involved preservation result was shown by the first author @xcite , namely that the theory of lovely pairs of models of a supersimple ( respectively , superstable ) theory is again supersimple ( respectively , superstable ) . with superstability and supersimplicity defined as above , the question whether a superstable theory with a generic automorphism is supersimple arises again .
a specific instance of this question was asked by the second author and c. ward henson @xcite regarding the theory of probability algebras equipped with a generic automorphism .
it was answered negatively by the first author , showing that probability algebras with a generic automorphism are not superstable .
the proof appears in .
however , the notions of @xmath1-stability and/or superstability mentioned above might still be too strong : while they consider types of tuples up to arbitrarily small distance , one may further relax this and consider types also up to arbitrarily small perturbations of the entire language , or parts thereof .
this idea can be formalised with the theory of perturbations as developed in @xcite and somewhat restated in ( * ? ? ?
* section 4 ) .
we shall assume some familiarity with the second reference .
the goal of this paper is to study carefully two examples : hilbert spaces and probability algebras , both equipped with a generic automorphism .
the theory @xmath10 of atomless probability algebras and the theory @xmath11 of infinite dimensional hilbert spaces have some features in common .
they are @xmath1-stable , separably categorical over any finite set of parameters and types over sets are stationary .
it follows ( see @xcite ) that both @xmath12 and @xmath13 admit model companions @xmath0 and @xmath2 . in
we deal with the theory @xmath0 of hilbert spaces equipped with a generic automorphism .
we recall some if its properties from @xcite .
we use a corollary of the weyl - von neumann - berg theorem to show that @xmath0 is @xmath1-stable up to perturbation ( of the automorphism ) , and admits prime models up to perturbation over any set . unlike the arguments in @xcite ,
our arguments can be extended to a generic action of a finitely generated group of automorphisms ( i.e. , a generic unitary representation , see @xcite ) and even to hilbert spaces equipped with a generic action of a fixed finitely generated @xmath14-algebra .
this section also serves as a soft analogue for the main results of the other sections .
in we deal with the theory @xmath2 of probability algebras with a generic automorphism , first studied in @xcite .
specifically , we show that @xmath2 is @xmath1-stable up to perturbations of the automorphism .
it is an open question if @xmath2 admits prime models up to perturbations . in
we conclude with the first author s proof that without perturbation the theory @xmath2 is not superstable , showing that the results of are in some sense optimal .
let us consider a hilbert space @xmath15 and let @xmath16 denote the space of bounded linear operators on @xmath15 .
we recall that the _ operator norm _ of @xmath17 is @xmath18 .
we also recall the notions of the _ spectrum _ , _ punctual spectrum _ and _ essential spectrum _ of an operator @xmath17 : @xmath19 let @xmath15 be a hilbert space , @xmath20 .
we say that @xmath21 and @xmath22 are _ approximately unitarily equivalent _ if there is a sequence of unitary operators @xmath23 such that @xmath24 .
[ fct : wvnb ] let @xmath15 be a hilbert space and let @xmath20 be normal operators
. then @xmath21 and @xmath22 are approximately unitarily equivalent if and only if 1 .
@xmath25 2 .
@xmath26 for all @xmath27 in @xmath28 . when considering a hilbert space as a continuous structure we shall replace it with its unit ball , as described in @xcite .
we shall use the language @xmath29 , where @xmath30 and @xmath31 . notice that we can recover the norm as @xmath32 .
an axiomatisation for the class of ( unit balls of ) banach spaces in this language , excluding the symbol @xmath33 , appears in @xcite .
the symbol @xmath33 serves as a skolem function for the fullness axiom there , yielding a universal theory .
a banach space is a hilbert space if and only if the parallelogram identity holds , which is a universal condition as well : @xmath34 we obtain that the class of hilbert spaces is elementary , admitting a universal theory @xmath35 .
its model companion is @xmath11 , the theory of infinite dimensional hilbert spaces , obtained by adding the appropriate scheme of existential conditions .
it is easy to check that the theory @xmath35 has the amalgamation property , so @xmath11 eliminates quantifiers ( i.e. , it is the model completion of @xmath35 ) .
now let @xmath36 be a new unary function symbol and let @xmath37 be @xmath38 .
let @xmath12 be the theory @xmath39 .
since @xmath11 is @xmath1-stable and separably categorical even after naming finitely many constants , the theory @xmath12 admits a model companion @xmath0 ( see @xcite ) .
the universal part of @xmath12 is @xmath40 .
it is again relatively easy to check that @xmath41 has the amalgamation property . indeed , if @xmath42 for @xmath43 then we may write @xmath44 , where @xmath45 is the orthogonal direct sum , and then @xmath46 will do
. it thus follows that @xmath0 eliminates quantifiers as well .
[ prp : hilbertautspectrum ] let @xmath15 be a separable hilbert space and let @xmath36 be a unitary operator on @xmath15 .
then @xmath47 ( i.e. , @xmath48 is existentially closed as a model of @xmath12 ) if and only if @xmath49 .
clearly , if @xmath48 is existentially closed , then @xmath49 . on the other hand , assume that @xmath50 and that @xmath49 . passing to an elementary substructure , we may assume that @xmath15 is separable .
now let @xmath51 be separable and existentially closed .
since @xmath11 is separably categorical , we may assume that @xmath52 .
since @xmath53 , by there is a sequence @xmath54 of unitary operators on @xmath15 such that @xmath55 in norm .
it follows that if @xmath56 is a non - principal ultra - filter on @xmath57 then @xmath58 . on the other hand , @xmath59 for all @xmath60 .
thus @xmath61 , whereby @xmath62 .
henson and iovino observed that the theory @xmath0 is not @xmath1-stable ( or even small ) in the sense defined in the introduction .
indeed , let @xmath47 be @xmath63-saturated and for each @xmath64 let @xmath65 be a normal vector such that @xmath66
. then @xmath67 for @xmath68 .
thus the metric density character of @xmath69 is the continuum . on the other hand ,
it is shown in @xcite that @xmath0 is superstable .
let @xmath70 and @xmath71 denote the definable and algebraic closure ( in the real sort ) in models of @xmath0 .
we claim that if @xmath72 and @xmath73 , then @xmath74 , where @xmath75 is the definable closure of @xmath76 in the language @xmath77 .
indeed , let @xmath78 . then clearly @xmath79 . on the other hand , we may decompose @xmath80 , in which case @xmath81 , showing that @xmath82 .
we may similarly characterise non forking in models of @xmath0 . for @xmath47 and subsets @xmath83 , say that @xmath84 if @xmath85 for every @xmath86 . we leave it to the reader to check that @xmath87 satisfies the usual axioms of a stable notion of independence ( invariance , symmetry , transitivity , and so on ) , and therefore coincides with non forking .
let @xmath72 , @xmath88 .
then @xmath89 is stationary and @xmath90 is inter - definable with the set @xmath91 .
stationarity follows from the characterisation of independence ( and from quantifier elimination ) .
it is also clear that @xmath92 and @xmath93 , and since we already know that @xmath94 is stationary as well , we obtain @xmath95 . for the converse it suffices to show that for every @xmath96 , the projection @xmath97 belongs to the definable closure of any morley sequence in @xmath89 .
so let @xmath98 be such a morley sequence
. then @xmath99 for all @xmath96 and all @xmath100 , so @xmath101 forms an orthogonal sequence of bounded norm .
thus @xmath102 it follows that @xmath0 has weak elimination of imaginaries , namely that for every imaginary element @xmath103 there exists a real tuple ( possibly infinite ) @xmath76 such that @xmath104 , @xmath105 .
we now turn to perturbations of the automorphism in models of @xmath0 .
let @xmath106 for @xmath107 , and let @xmath108 .
we define an _
@xmath109-perturbation _ of @xmath51 to @xmath110 to be an isometric isomorphism of hilbert spaces @xmath111 which satisfies in addition @xmath112 the set of all @xmath109-perturbations will be denoted @xmath113 .
it is fairly immediate to verify that this indeed satisfies all the conditions stated in ( * ? ? ?
* theorem 4.4 ) , and therefore does indeed correspond to a perturbation system as defined there .
[ lem : spectralinclusion ] let @xmath42 be separable models of @xmath12 for @xmath114
. then we may write @xmath44 , and let us assume that @xmath115 and that @xmath116 has no isolated points .
then for every @xmath7 there is an isometric isomorphism @xmath117 , which fixes @xmath118 such that @xmath119 . under the assumptions we have @xmath120 .
we also assume that @xmath116 has no isolated points . therefore ,
if @xmath121 is isolated then its eigenspace in @xmath122 is entirely contained in @xmath123 , so the multiplicity ( possibly infinite ) of @xmath27 is the same for @xmath124 and for @xmath125 .
it follows that the hypotheses of hold , and we obtain @xmath126 such that @xmath127 . then @xmath128 will do .
[ thm : hilbertautaleph0stable ] the theory @xmath0 is @xmath1-stable up to perturbation of the automorphism .
let @xmath129 be separable , and let @xmath130 .
by we have @xmath131 , so @xmath132 by model completeness .
it will therefore be enough to show that every type over @xmath118 is realised , up to perturbation , in @xmath110 .
such a type can always be realised in a separable elementary extension @xmath133 . then @xmath134 and we may decompose the latter as @xmath135 .
notice that @xmath136 , so @xmath137 .
we may therefore apply , obtaining for every @xmath138 there an isometric isomorphism @xmath139 fixing @xmath118 such that @xmath140 .
we have thus shown that every type over @xmath118 is realised , up to arbitrarily small perturbation of the automorphism , in a fixed separable extension @xmath141 , as desired .
let @xmath142 be a finitely generated discrete group and let @xmath143 be the theory of hilbert spaces with a generic action of @xmath142 by automorphism ( see @xcite ) . using voiculescu s theorem @xcite in place of
, the same argument shows that the theory @xmath143 is @xmath1-stable up to perturbations of the automorphisms .
this can even be further extended to the theory @xmath144 of a generic presentation of a finitely generated @xmath14-algebra @xmath145 .
the theory @xmath0 has prime models up to perturbation over sets ( of real or imaginary elements ) .
by this we mean that for every set @xmath76 in a model of @xmath0 there exists a model @xmath110 , containing @xmath76 , such that if @xmath146 is any other model which contains @xmath76 then , up to arbitrarily small perturbation of @xmath147 to @xmath148 , we can embed @xmath110 elementarily in @xmath149 over @xmath76 .
we may assume that the set @xmath76 over which we seek a prime model is algebraically closed .
by weak elimination of imaginaries we may assume that @xmath76 is a real set , and we may further assume that @xmath150 .
it is therefore a hilbert subspace @xmath118 ( possibly finite dimensional ) on which @xmath151 is an automorphism .
moreover , since @xmath0 eliminates quantifiers , the type of @xmath118 is determined by the pair @xmath51 , and there is no need to consider the ambient structure .
if @xmath152 there is nothing to prove .
otherwise @xmath153 .
let @xmath154 be separable such that @xmath155 ( for example we may take @xmath156 in the lebesgue measure , with @xmath157 ) .
let @xmath130 .
clearly @xmath158 , so @xmath131 , and we shall prove that it is prime , up to perturbation of the automorphism , over @xmath118 .
so let @xmath159 and we may assume that @xmath160 is separable . as usual , we may decompose it as @xmath135 . since @xmath161
, we necessarily have @xmath162 , and since @xmath163 is moreover closed , it contains @xmath164 . by , for every @xmath7 there exists an isometric isomorphism @xmath117 fixing @xmath118 such that @xmath119 . by
we also have @xmath165 thus @xmath166 is as desired .
by a probability space we mean a triplet @xmath167 , where @xmath168 is a set , @xmath169 a @xmath170-algebra of subsets of @xmath168 , @xmath171 a @xmath170-additive positive measure on @xmath169 such that @xmath172 . a probability space @xmath167 is called _ atomless _ if for every @xmath173 there is @xmath174 such that @xmath175 .
we say that two elements @xmath176 determine the same _ event _ , and write @xmath177 if @xmath178 .
the relation @xmath179 is an equivalence relation and the collection of classes is denoted by @xmath180 and is called the _ measure algebra _ associated to @xmath167 .
operations such as unions , interesections and complements are well defined for events , as well as the measure .
the distance between two events @xmath181 is given by the measure of their symmetric difference .
this renders @xmath182 a complete metric space .
conversely , let @xmath183 be a boolean algebra and assume that @xmath184 is a complete metric on @xmath185 .
let @xmath186 be an abbreviation for @xmath187 and assume furthermore that @xmath188 , @xmath189 and @xmath190 .
then @xmath185 is the probability algebra associated to some probability space ( and we may moreover take that space to be the stone space of @xmath185 , equipped with the borel @xmath170-algebra ) . we may view probability algebras as continuous structures in the language @xmath191 ( the distance symbol is always implicit , and the measure can be recovered from it as above ) .
the class of probability algebras is elementary and admits a universal theory denoted @xmath192 .
its model completion is @xmath10 , the theory of atomless probability algebras .
it admits quantifier elimination , is @xmath1-categorical ( even over finitely many parameters ) and @xmath1-stable ( see @xcite ) .
let @xmath185 be a probability algebra .
an automorphism @xmath193 is said to be _ aperiodic _ if for every non - zero event @xmath194 and every @xmath195 there is a sub - event @xmath196 such that @xmath197 .
( in other words , the _ support _ of @xmath198 is @xmath199 for all @xmath200 . )
[ fct : rokhlin ] let @xmath185 be a probability algebra , @xmath193
. then @xmath36 is aperiodic if and only if , for every @xmath200 and every @xmath7 there is @xmath194 such that @xmath201 , @xmath202 ,
, @xmath203 are disjoint and @xmath204 .
now let @xmath205 where @xmath36 is a new unary function symbol .
let @xmath13 be the theory @xmath206 .
it was shown in @xcite that @xmath13 admits a model companion @xmath2 , consisting of @xmath13 together with axioms saying that @xmath36 is aperiodic . by the _ lebesgue space _ we mean the probability space @xmath207,\lambda)$ ] , where @xmath27 is the standard lebesgue measure .
the associated probability algebra @xmath208,\lambda)$ ] is the unique separable atomless probability algebra .
an _ automorphism _ @xmath36 of the lebesgue space is a measurable , measure - preserving bijection between measure one subsets of @xmath209 $ ] . 1 .
every automorphism of the probability algebra @xmath210 comes from an automorphism of the lebesgue space .
an automorphism @xmath36 of the lebesgue space induces an aperiodic automorphism on @xmath210 if and only if @xmath36 itself is aperiodic , namely if @xmath211\colon \tau^n(x)=x\ } = 0 $ ] .
let @xmath212 be a probability algebra .
we equip @xmath213 with the uniform convergence metric @xmath214 let @xmath215 for @xmath107 and let @xmath108 .
then an _ @xmath109-perturbation _ of @xmath216 to @xmath217 is an ( isometric ) isomorphism @xmath218 such that @xmath219 .
notice that this is essentially the same definition as for ( unit balls of ) hilbert space .
in particular , as in the hilbert space case , this definition satisfies the conditions of ( * ? ? ?
* theorem 4.4 ) and thereby comes from a perturbation system . let @xmath212 be a probability algebra , @xmath220 , @xmath221 .
we say that @xmath222 _ generate an @xmath223-partition ( of @xmath212 ) _ if @xmath224 are disjoint and @xmath225 .
an @xmath226-partition will simply be called an _ @xmath100-partition_. if @xmath227 , @xmath107 are probability algebras and @xmath228 is their free amalgam , we may identify @xmath229 with its image @xmath230 , and similarly @xmath231 .
in particular , if @xmath232 , @xmath107 then @xmath233 as well .
if in addition @xmath234 then @xmath235 is aperiodic , whereby so is @xmath236 , i.e. , @xmath237 . since @xmath2 is model complete we conclude that @xmath238 let @xmath239 be separable .
we say that @xmath240 is _ partitioned over @xmath212 _ if : 1 .
the probability algebra @xmath185 is isomorphic to @xmath241 over @xmath212 ( meaning that @xmath221 gets mapped to @xmath242 ) .
2 . under this isomorphism , for each @xmath243 there exists @xmath244 such that @xmath245 generate an @xmath100-partition , all of whose members belong to @xmath246 . [
lem : autextpart ] let @xmath239 be separable . then there exists a further elementary extension @xmath247 which is partitioned over @xmath212 .
first of all , we may assume that @xmath185 is atomless over @xmath212 .
indeed , @xmath248 is a separable elementary extension of @xmath240 and we may replace the latter with the former .
therefore we may assume that @xmath249 .
it is not difficult to construct an automorphism @xmath250 such that for each @xmath100 there is @xmath244 such that @xmath251 generate an @xmath100-partition .
let @xmath252 , so @xmath247 . on the other hand we have @xmath253 $ ] , and for every @xmath254 @xmath255 ) = 1 \otimes [ 1 \otimes \rho^k(c_n ) ] .
\end{gathered}\ ] ] now use the fact that @xmath256 to conclude . for a probability algebra @xmath257 and @xmath258 ,
let @xmath259 denote the ideal @xmath260 .
[ ntn : simplepartition ] for @xmath243 fix @xmath261 generating an @xmath100-partition such that in addition @xmath262 .
note that this determines @xmath263 up to isomorphism , and that it is _ not _ a model of @xmath2 .
[ lem : finiteorderautpert ] let @xmath239 , and assume that @xmath240 is partitioned over @xmath212 .
then for every @xmath243 there exists @xmath264 such that : 1 . @xmath265 .
2 . @xmath266 over @xmath212 , where @xmath267 are as in .
we may assume that @xmath268 and that this identification witnesses that @xmath240 is partitioned over @xmath212 .
therefore there exists @xmath269 are such that @xmath270 generate an @xmath100-partition all of whose members are in @xmath246 .
let @xmath271 for @xmath254 ( or , for that matter , for all @xmath272 ) .
let @xmath273 be as in .
since @xmath274 and @xmath275 , there is an isomorphism @xmath276 , which induces in turn @xmath277 .
we shall extend @xmath278 to an automorphism of @xmath185 as follows . for @xmath279
, observe that @xmath280 is a partition of @xmath281 , and @xmath282 .
we then define @xmath283 for each @xmath254 , @xmath284 restricts to an isomorphism @xmath285 , so @xmath286 .
in addition , if @xmath221 then : @xmath287 whereby @xmath288 .
thus @xmath284 acts as the identity on @xmath289 .
let @xmath290 .
for @xmath279 and for @xmath254 let @xmath291 , in which case @xmath292 and @xmath293 . assume now in addition that @xmath294 , i.e. , that @xmath295 . then in the expression @xmath296 , the instances of @xmath297 and @xmath284 can be expanded explicitly as follows : @xmath298 ( \tau_{\mathscr{a}}\otimes \rho_n ) \bigl [ ( \tau_{\mathscr{a}}\otimes \rho_n)^k \theta_1 \tau_{\mathscr{b}}^{-k } \bigr ] ( b_k ) \\ & = \tau_{\mathscr{b}}(b_k ) . \end{aligned}\ ] ] ( on the other hand , this fails for @xmath299 , since then @xmath297 expands to @xmath300 . )
it follows that @xmath301 ( since both events are sub - events of same measure of @xmath302 which has measure @xmath303 ) .
thus @xmath265 , as desired .
the theory @xmath2 is @xmath1-stable up to perturbations of the automorphism .
let @xmath304 .
by @xmath305 admits a partitioned extension @xmath240 .
we claim that up to perturbation , every type over @xmath212 is realised in @xmath240 .
indeed any such type is realised in some separable extension @xmath306 . by
again we may assume that @xmath307 is also partitioned over @xmath212 .
by any two partitioned extensions of @xmath212 admit an @xmath308 perturbation over @xmath212 to @xmath309 , and composing these we obtain an @xmath303-perturbation @xmath310 which fixes @xmath212 , and this for every @xmath243 .
the second author and henson @xcite asked whether the theory of probability algebras with a generic automorphism is superstable , as one might expect by analogy with a theorem of chatzidakis and pillay @xcite regarding generic automorphisms in classical first order logic . in this section
we present the first author s negative answer , which chronologically came before the results in , and to a large extent motivated them .
our aim is to show that the theory @xmath2 admits many types over small sets of parameters , and for this purpose it will suffice to show that there are many @xmath199-types over parameters which belong to the fixed algebra of the automorphism .
we therefore proceed in two steps , first characterising such types and then showing there are many of them . throughout
we let @xmath311 denote a monster model of @xmath2 , and let @xmath312 denote the fixed sub - algebra of @xmath56 under @xmath36 . consider a function
@xmath313 $ ] sending @xmath314 .
we call such a function _ shift invariant _ if @xmath315 we define @xmath316^{2^{<\omega}}$ ] to consist of all shift invariant mappings .
this is a closed subset of @xmath209^{2^{<\omega}}$ ] , and therefore compact .
let @xmath317 , and for @xmath60 , @xmath318 , let @xmath319 ( where @xmath320 , @xmath321 is the natural action of @xmath322 ) .
then @xmath323 is shift invariant , yielding a mapping @xmath324 sending @xmath325 .
this mapping @xmath326 is clearly continuous , and by quantifier elimination it is injective .
conversely , let @xmath327 , and let @xmath212 be any sufficiently homogeneous atomless probability algebra
. then one can find in @xmath212 a sequence of events @xmath328 such that @xmath329 , @xmath330 , and in general , for every @xmath100 and @xmath318 : @xmath331 this is indeed consistent by shift invariance .
moreover , shift invariance implies that for every @xmath332 and @xmath333 : @xmath334 ( by induction on @xmath335 ) .
it follows by quantifier elimination that the mapping @xmath336 is elementary and therefore extends to an automorphism @xmath337 , and we may embed @xmath305 in @xmath311 . in other words , for every @xmath327 we can find @xmath338 such that @xmath339 for all @xmath340 .
thus @xmath341 , showing that @xmath326 is bijective . since it is also continuous , from a compact to a hausdorff space , it is a homeomorphism . if @xmath342 is an arbitrary topological space we have @xmath343)^{2^{<\omega } }
= c\bigl ( y , [ 0,1]^{2^{<\omega } } \bigr)$ ] as sets . equipping @xmath344 with the compact - open topology and @xmath345 with the discrete topology these are homeomorphisms . ( the common topology can be given by a sub - basis , where a sub - basic open set is of the form @xmath346 ) \colon f[k\times\{s\ } ] \subseteq u\bigr\}$ ] , with @xmath347 compact , @xmath340 and @xmath348 $ ] is open ) .
we may define when a mapping @xmath349)^{2^{<\omega}}$ ] is shift invariant by as above , and let @xmath350 be the set of all such shift invariant functions .
it is then clear that @xmath351 . 1 .
the space @xmath353 admits a unique regular borel probability measure @xmath355 such that @xmath356 for all @xmath221 .
the probability algebra of @xmath357 is canonically isomorphic to @xmath212 , identifying the equivalence class of @xmath358 with the event @xmath201 .
the natural mapping @xmath359 is bijective . in other words ,
every equivalence class of bounded borel functions up to equality @xmath355-a.e . contains a unique continuous representative .
4 . let @xmath360 be a larger probability algebra .
then there exists a conditional expectation operation @xmath361\colon l_1(\widetilde { \mathscr{b } } ) \to l_1(\widetilde { \mathscr{a}})$ ] where @xmath362 $ ] is the unique function such that for all @xmath221 : @xmath363\,d\tilde \mu_{\mathscr{a}}. \end{gathered}\ ] ] see @xcite . the stone space associated to a boolean algebra
is discussed in 311 . the construction of the measure @xmath355 and the isomorphism between @xmath212 and the probability algebra of @xmath357 appears in 321j .
the identification of @xmath364 and @xmath365 can be found in 363 and 364k .
finally , conditional expectations are discussed in 365r .
now let @xmath366 , and let @xmath367 , say realised by @xmath368 . for @xmath318 ,
the type @xmath369 determines the mapping associating to every @xmath221 the measure @xmath370 . in other words , @xmath369 determines the function @xmath371 \in l_1(\widetilde { \mathscr{a}})$ ] .
since the essential range of @xmath372 lies in @xmath209 $ ] we have @xmath373 , and in fact @xmath374)$ ] .
let @xmath375)^{2^{<\omega}}$ ] , @xmath323 .
it is not difficult to see that @xmath376 satisfies , so identifying @xmath377)^{2^{<\omega}}$ ] with @xmath377^{2^{<\omega}})$ ] we actually have @xmath378 .
we have thus obtained a mapping @xmath379 .
again , it is injective by quantifier elimination and continuous , and a construction as above yields that it is surjective .
we have thus obtained a homeomorphism @xmath380 for a closed set @xmath381 and @xmath221 define @xmath382 \subseteq k \}.\end{gathered}\ ] ] it is not difficult to see that @xmath383 is closed in the compact - open topology . if @xmath384 is a partial type over @xmath385 and @xmath381 corresponds to the closed set @xmath386 \subseteq \operatorname{s}_1(apr_a)$ ] , then the closed set @xmath383 corresponds to @xmath387 \subseteq \operatorname{s}_1({\mathscr{a}})$ ] , where @xmath388 is a partial type over @xmath212 .
if @xmath389 holds we say that @xmath281 _ satisfies @xmath390 over @xmath201_. let @xmath391 denote the distance function between types over @xmath212 , and similarly @xmath392 .
it is fairly easy to verify that the distance mapping @xmath393 $ ] , i.e. , @xmath394 $ ] , is borel measurable ( but not continuous , since the theory is not @xmath1-categorical ) .
thus , if @xmath395 , then @xmath396 is a random variable from @xmath212 to @xmath209 $ ] , which we can integrate .
for every real @xmath404 , let @xmath405 be the type of one half of the circle on which @xmath36 acts by rotation by @xmath406 ( the measure on the circle being the lebesgue measure normalised to have total length @xmath199 ) .
if @xmath407 are irrational and linearly independent over the rationals then for every @xmath7 there exist @xmath408 such that @xmath409 .
if @xmath410 , @xmath411 then @xmath412 while @xmath413 .
it follows that @xmath414 , namely @xmath415 .
therefore @xmath416 ( it is not difficult to check that the distance between the types is in fact equal to @xmath417 ) .
this is a metric cantor - bendixson rank argument which applies more generally , saying that if @xmath4 is a non - small theory with a countable language then such a tree exists ( with @xmath430 possibly replaced with another positive constant ) . for an even more general statement of this fact see (
* propositions 3.16 and 3.19 ) .
define compacts subsets @xmath431 by induction on @xmath404 .
start with @xmath432 ; for @xmath404 limit , @xmath433 ; and given @xmath434 , obtain @xmath435 by removing from @xmath434 all points for which there is a relatively open neighbourhood of diameter @xmath436 .
since the language is countable , the topology on @xmath168 admits a countable base .
if we only take out basic open sets of diameter @xmath436 we still get the same sequence @xmath434 , and since the base is countable the sequence stabilises before @xmath63 .
let @xmath437 be an uncountable subset of types every pair of which have distance @xmath438 .
each set of diameter @xmath436 can contain at most one member of @xmath439 , so @xmath440 .
the topological space @xmath441 ( with the induced topology ) is @xmath442-perfect , meaning that every non - empty open subset has diameter @xmath438 .
let @xmath443 and @xmath444 .
then @xmath445 is closed , being the image of the closed set @xmath446 \subseteq \operatorname{s}_2(apr_a)$ ] under the projection @xmath447 ( a continuous mapping from a compact space to a hausdorff space is always closed ) .
thus @xmath448 is closed in @xmath449 .
we start with @xmath452 being the partial type defining @xmath441 .
it has the property that @xmath453^\circ \neq { \varnothing}$ ] .
assume now we have @xmath454 such that @xmath455^\circ \neq { \varnothing}$ ] .
the interior has diameter @xmath438 , so there are @xmath456^\circ$ ] such that @xmath457 .
thus @xmath458 , so they admit open neighbourhoods @xmath459^\circ$ ] and @xmath460^\circ$ ] such that @xmath461 .
we can then find smaller open neighbourhoods such that @xmath462 and @xmath463 .
letting @xmath464 be the partial type defining @xmath465 and @xmath466 the partial type defining @xmath467 we get : @xmath468 \supseteq [ \pi_{s{{^\frown}}0 } ] \supseteq [ \pi_{s{{^\frown}}0}]^\circ \neq { \varnothing}$ ] and @xmath468 \supseteq [ \pi_{s{{^\frown}}1 } ] \supseteq [ \pi_{s{{^\frown}}1}]^\circ \neq { \varnothing}$ ] .
finally , @xmath469 \times [ \pi_{s{{^\frown}}1 } ] ) \cap d_1 = { \varnothing}$ ] implies @xmath470 . for the other
, assume @xmath472 .
let @xmath473 be as in .
let @xmath474 be a sequence of independent events of measure @xmath442 , all fixed by @xmath36 , and let @xmath212 be the generated complete algebra . for @xmath475 and @xmath340
let @xmath476 , and let @xmath477 in other words , @xmath478 holds if and only if , for every @xmath340 , @xmath479 satisfies @xmath454 over @xmath480 .
it is easy to check that @xmath481 is consistent and implies @xmath482 for @xmath483 , so @xmath484 is consistent as well .
choose for each @xmath485 a complete type @xmath486 extending @xmath484 .
let @xmath487 , and let @xmath488 be such that @xmath489 . then over @xmath490 , @xmath491 takes only values in @xmath492 \subseteq \operatorname{s}_1(apr_a)$ ] , while @xmath493 only takes values in @xmath494 $ ] , and the opposite holds over @xmath495 .
thus @xmath496 over @xmath497 , which has measure @xmath417 .
we conclude that @xmath498 .
ita ben yaacov , alexander berenstein , c. ward henson , and alexander usvyatsov , _ model theory for metric structures _ , model theory with applications to algebra and analysis , volume 2 ( zo chatzidakis , dugald macpherson , anand pillay , and alex wilkie , eds . ) , london math society lecture note series , vol .
350 , cambridge university press , 2008 , pp . 315427 .
to3em , http://dx.doi.org/10.1007/s11813-008-0009-x[_topometric spaces and perturbations of metric structures _ ] , logic and analysis * 1 * ( 2008 ) , no . 34 , 235272 , http://arxiv.org/abs/0802.4458[arxiv:0802.4458 ] .
to3em , http://dx.doi.org/10.4115/jla.2009.1.1[_modular functionals and perturbations of nakano spaces _ ] , journal of logic and analysis * 1 * ( 2009 ) , no . 1 , 142 , http://arxiv.org/abs/0802.4285[arxiv:0802.4285 ] . ita ben yaacov and alexander usvyatsov , http://math.univ-lyon1.fr/~begnac/articles/cfo.pdf[_continuous first order logic and local stability _ ] , transactions of the american mathematical society , to appear , http://arxiv.org/abs/0801.4303[arxiv:0801.4303 ] .
jos iovino , _ stable banach spaces and banach space structures , i and ii _ , models , algebras , and proofs ( bogot , 1995 ) , lecture notes in pure and appl .
203 , dekker , new york , 1999 , pp . | we prove that @xmath0 , the theory of infinite dimensional hilbert spaces equipped with a generic automorphism , is @xmath1-stable up to perturbation of the automorphism , and admits prime models up to perturbation over any set .
similarly , @xmath2 , the theory of atomless probability algebras equipped with a generic automorphism is @xmath1-stable up to perturbation .
however , not allowing perturbation it is not even superstable . |
globular clusters ( gcs ) are generally considered to be good tracers of a galaxy s star formation history .
bimodal colour distributions are observed in many gc systems , suggesting that galaxies undergo multiple epochs of star and gc formation ( harris 2001 ) .
the exact nature of how these multiple star formation events occur is crucial to understanding galaxy formation .
ashman & zepf ( 1992 ) describe the production of metal - rich gcs during the merger of two gas - rich disc galaxies .
this scenario results in similar ages for the metal - rich gcs and the merger event . in this model
the blue , metal - poor gcs which belonged to the progenitor galaxies are universally old ( @xmath0 gyr ) .
forbes , brodie & grillmair ( 1997 ) propose a multi - phase collapse , during which metal - poor gcs form early during the pre - galaxy collapse phase and , at a later time , metal - rich gcs form out of more enriched gas at a similar time to the galaxy field stars .
this model implies a metal - rich population that is a few gyr younger than the metal - poor population .
the exact age difference may depend on mass and environment ( beasley 2002 ) . on the other hand , cote ,
marzke & west ( 1998 ) suggest the bimodality of gc colours in large ellipticals is due to the gradual accretion of metal - poor gcs from dwarfs , with metal - rich gcs being indigenous to the elliptical . in this picture , the metallicity of the metal - rich population is a function of the galaxy s luminosity and both sub - populations should be old .
these physical processes may all occur to some extent . to distinguish between the relative contributions of each process
, we need examples of elliptical galaxies covering a range of masses and in different environments . given that the above gc formation scenarios have different expectations concerning the ages of the metal - poor and metal - rich sub - populations , determining the age of individual gcs is a crucial step in testing these models .
there are several methods to measure gc properties such as age and metallicity from integrated spectra .
examples include the method of brodie & huchra ( 1990 ; hereafter bh90 ) , who present an empirical metallicity measure based on absorption line features .
another method is that of strader & brodie ( 2004 ; hereafter sb04 ) , based on a principal components analysis ( pca ) of 11 absorption line features .
methods involving spectral indices , predominantly the lick system , and their comparison to simple stellar population ( ssp ) models can break the age - metallicity degeneracy ( worthey 1994 ) .
these methods usually rely on an age sensitive index ( e.g. , h@xmath6 ) and a metallicity sensitive index ( e.g. , mgb ) to break the degeneracy . however , these methods suffer difficulties when individual indices are contaminated , due to cosmic rays , skyline subtraction or galactic emission .
another complication is non - solar @xmath4-abundance ratios which are difficult to determine from index - index plots and can affect the higher order balmer indices ( thomas , maraston & korn 2004 ) .
recently , proctor & sansom ( 2002 ; hereafter ps02 ) have used a multi - index @xmath7minimisation method to obtain ages , metallicities and abundance ratios for the stellar populations of late and early - type galaxies .
another advantage of using many indices is that it maximises the information used to break the age - metallicity degeneracy .
here we extend this method to extragalactic gcs for the first time ( see proctor , forbes & beasley 2004 for an application to galactic gcs ) .
hot luminous blue horizontal branch ( bhb ) stars have a significant effect on the measurement of several indices ( most notably the balmer indices , see lee , yoon & lee 2000 ) .
studies of galactic gcs show that those with similar ages and metallicities can have very different spectral contributions from their hb stars depending on whether the hb stars are hot ( blue ) or cool ( red ) . the way the hb is modelled has a significant impact on a ssp model ( maraston , greggio & thomas 2001a ) .
a colour - magnitude diagram is probably the best way to deduce hb morphology in a given gc , but such diagrams are limited to gcs in the very nearest galaxies .
thus information on hbs in extragalactic gcs must generally come from integrated spectra .
old metal - rich gcs with bhbs ( e.g. , ngc 6388 and ngc 6441 ; rich 1997 ) can falsely appear to be intermediate - age clusters due to this strengthening of the balmer lines ( schiavon 2004 ) , therefore , the ability to deduce hb morphology from integrated spectra is important . schiavon ( 2004 )
have recently introduced a method to detect anomalous hbs in gc spectra by differential comparisons of h@xmath8 , h@xmath9 , and h@xmath6 line strengths .
there is evidence for a significant number of proto - gcs in galaxies which are currently merging ( e.g. , the antennae ; whitmore & schweizer 1995 ) and very recent ( @xmath10500 myr ) mergers such as ngc 7252 ( maraston 2001b ; schweizer & seitzer 1998 ; miller 1997 ) and ngc 3921 ( schweizer , seitzer & brodie 2004 ; schweizer 1996 ) . the outstanding questions are whether a substantial fraction of the proto - gc sub - population formed in these mergers survives , and hence whether it can account for the gc systems of ellipticals ( see forbes , brodie & grillmair 1997 ) . an important link between ongoing mergers and `` old '' elliptical galaxies
is the intermediate aged ( 2 - 5 gyr ) merger remnants .
previous gc spectroscopic studies of merger remnant galaxies such as ngc 1316 ( goudfrooij 2001 ) and ngc 3610 ( strader 2003 ; strader , brodie & forbes 2004 ) have confirmed a small number of gcs with ages matching the time since the merger event ( @xmath11 gyr and @xmath1 gyr respectively ) .
however , the total number of intermediate - age gcs in these galaxies is currently poorly constrained spectroscopically . in this work
we examine the gc system of ngc 1052 , which is an excellent candidate for an elliptical that has undergone a minor merger event in the last few gyr .
ngc 1052 is an e4 galaxy , located in a small group at a distance of 18 mpc ( forbes , georgakakis & brodie 2001b ) . aside from containing an active nucleus ,
ngc 1052 displays several indications of a past merger or significant accretion event .
the gas angular momentum is higher than that of the stellar component , with the gas and stars having different rotation axes ( van gorkom 1986 , plana & boulesteix 1996 ) .
ngc 1052 also reveals hi tidal tails ( van gorkom 1986 ) , infalling hi gas onto the active nucleus ( van gorkom 1989 ) and dust lanes ( forbes , sparks & macchetto 1990 ) all signs of an accretion or merger .
van gorkom ( 1986 ) suggest that the hi observations could be explained by the accretion of a gas - rich dwarf , or other minor interaction , about 1 gyr ago . despite the evidence for a recent accretion event , ngc 1052 shows almost no optical disturbance ( it has a low fine structure value of @xmath12 = 1.78 , schweizer & seitzer 1992 ) and has a fundamental plane residual of + 0.07 ( prugniel & simien 1996 ) which is consistent with a normal galaxy on the fundamental plane ( forbes , ponman & brown 1998 ) .
we can further probe the evolutionary history of ngc 1052 by measuring the ages and metallicities of individual gcs in the galaxy .
the imaging of forbes ( 2001b ) revealed a bimodal gc system as is typical for elliptical galaxies .
the two sub - populations had a colour difference of @xmath13b [email protected] .
using worthey ( 1994 ) models , under the assumption that the blue gc sub - population is 15 gyr old and metal - poor with [ fe / h]=1.5 , forbes ( 2001b ) showed that any newly - formed red gcs ( assuming ages of @xmath152.5 gyr ) must have supersolar metallicities to fully account for the red sub - population .
in section [ sec_obs ] we will present our observations and data reduction methods .
measured indices and our analysis of these lick indices are given in section [ sec_inds ] . using the method of ps02 , metallicities , ages and abundance ratios are given in section [ sec_aam ] .
spectra of ngc 1052 itself are given in section [ sec_long ] and a brief analysis of gc kinematic data is in section [ sec_gal ] . finally we discuss the implications of our results in section [ sec_disc ] and present conclusions in section [ sec_conc ] .
spectra of gc candidates around ngc 1052 were obtained with the low resolution imaging spectrometer ( lris ; oke 1995 ) on the keck i telescope .
candidate selection , based on the keck imaging data of forbes ( 2001b ) , was designed to cover a wide range of potential gc colours ( i.e. 1.2 @xmath10 b
i @xmath10 2.5 , 0.6 @xmath10 v
i @xmath10 1.7 ) .
the properties of our candidates can be found in table 1 .
observations were obtained in 2003 january 25 - 27 with an integration time of @xmath16s = 8 hours for the slit - mask .
however , the signal - to - noise ratios of the combined spectra were improved by excluding the exposures made at the highest airmass . these suffered from signal - to - noise degradation , especially at bluer wavelengths . seeing ranged between 0.65@xmath17 and 0.75@xmath17 over the three nights . a 600 lines per mm grating blazed at 4000
was used for the blue side , resulting in an approximate wavelength range of 3300 - 5900 and a fwhm spectral resolution of @xmath18 .
table 1 presents observational data for the objects for which we obtained keck spectra .
figure [ n1052colour ] , a b
i colour magnitude diagram , shows that we have obtained spectra from both the blue and red sub - populations of gc candidates .
we have also generally sampled the most luminous gcs in ngc 1052 .
if any young ( @xmath19 few gyrs old ) gcs exist then they will tend to be brighter than the old population .
plotted in figure [ positions ] are the spatial positions of the confirmed gcs .
[ cols="<,^,^,^,^,^ , > , > " , ] [ tablefits ] lick indices were measured for the galaxy at each radius and the same offset to the lick system was applied as per the gc data .
denicolo ( 2005 ) presents indices , measured along the major axis of ngc 1052 , which are consistent with our observed values .
several indices show gradients with galactocentric radius . a sample of these are plotted in figure [ radial ] , showing some clear radial trends .
the metallicity and [ e / fe ] sensitive indices ( e.g. , cn@xmath20 and mg@xmath21 ) have a strong gradient towards higher values in the centre .
a similar trend is observed for other [ e / fe ] sensitive indices , e.g. , mgb .
the fe5270 line , which is primarily sensitive to iron metallicity [ fe / h ] , shows almost no gradient .
other fe lines such as fe4383 , fe4531 and fe5406 also show no gradient with radius .
these trends are also evident in the denicolo ( 2005 ) data .
the galaxy therefore posses strong gradients in [ e / fe ] sensitive indices , but no gradients in the [ fe / h ] sensitive indices .
the effect of central galactic emission is obvious in the h@xmath6 index plot , with large negative values at small radii .
the other balmer indices and fe5015 are similarly affected by emission .
due to the high total metallicity of the galaxy , we use the fe method from ps02 as opposed to the trager ( 2000a ) method used earlier to calculate the [ e / fe ] .
this is necessary because the isochrone shape is driven by [ z / h ] at low metallicity and [ fe / h ] at high metallicity ( see ps02 for a full explanation ) .
due to the high @xmath4element abundance ratios present in the galaxy , we extrapolate the tmk04 models to [ e / fe]=+0.9 .
figure [ mgbfe ] demonstrates the extremely high abundance ratios present in the central regions by showing a plot of @xmath10fe@xmath22 vs mgb , where @xmath10fe@xmath22=(fe5270+fe5335)/2 . to obtain reasonable fits to the tmk04 ssp models it was necessary to exclude several indices from all apertures during the fitting process .
galactic emission which increases towards the centre , forces the removal of all the balmer lines and fe5015 from the fitting process .
the index c4668 was also removed due to its systematic deviance from the best fit .
we also clipped individual deviant indices following a similar method to section 3.3 .
this resulted in a total of 12% of the remaining indices being removed ( mostly mg@xmath20 , ca4455 and fe4383 ) .
we were unable to obtain reasonable fits for the central aperture , probably due to strong galactic emission , and do not present the fitted parameters for it .
the values obtained for age , [ fe / h ] , [ e / fe ] and [ z / h ] are presented in table 5 and also plotted against radius in figure [ longall ] .
figure [ longall ] shows no strong radial gradient in either age or [ fe / h ] .
there is however a very strong gradient of decreasing [ e / fe ] with increasing radius . in the central arcsecond the estimated [ e / fe ] of @xmath19
+ 0.8 dex is unusually high .
using the kinematic information from the 16 gcs , we can estimate the mass enclosed within the radius of the gc system observed .
we use the projected mass estimator ( evans 2003 ) , assuming isotropy and an r@xmath23 distribution to derive a mass of 1.7@xmath24 within 19 kpc ( @xmath25 r@xmath26 ) .
the mass estimate error was calculated by bootstrapping the observed velocities and errors .
van gorkom ( 1986 ) used hi kinematics to measure a mass of @xmath27 within 23 kpc .
we have limited spatial coverage and too few gcs to make any comment on systematic rotation of the gc system . since our mass estimate is significantly higher than that for the hi kinematics , our assumption of isotropy may be incorrect .
for example , if the orbits had a rather strong tangental bias of @xmath28/@xmath29 = 2 , then our tracer mass estimate would be 30% too large ( evans et al .
to summarise , we have obtained spectra for 16 gcs which sample both the blue and red sub - populations in ngc 1052 .
ages , metallicities and abundance ratios for these gcs have been derived using the multi - line fitting technique of ps02 applied to the ssps of tmk04 .
we find that the blue gcs in ngc 1052 to be uniformly old @xmath3010 gyr and metal - poor ( with the exception of gc47 for which we strongly suspect an incorrect colour ) .
the red gcs are similarly old but metal - rich .
we do not find any young gcs associated with the recent merger event , even though we have sampled the expected colour and luminosity regime for young gcs .
we do , however find evidence for young _ galaxy _ stars associated with the merger event .
our luminosity - weighted age of @xmath1 gyr for the ngc 1052 stars is compatible with that estimated from the hi tidal tails and infalling hi gas of @xmath5 gyr ( van gorkom 1986 ) .
the high central value and strong gradient observed for [ e / fe ] in ngc 1052 is unusual for an elliptical galaxy .
for example , mehlert ( 2003 ) find negligible gradients in [ e / fe ] for 35 early - type galaxies in their sample of coma galaxies .
proctor ( 2002 ) using the same method as this work finds a maximum [ e / fe]@xmath31 + 0.4 and no significant [ e / fe ] gradient in the bulges of 28 late and early - type galaxies .
the @xmath4element abundance ratio for ngc 1052 is largely insensitive to the combination of indices we fit , unlike age and metallicity .
therefore , barring a major contamination of all the [ e / fe ] sensitive indices ( which must also affect the observations of denicolo 2005 ) we can confidently say that the central galaxy starlight is more @xmath4element enhanced than @xmath2 dex .
such @xmath4element enhancements and the deduced high central metallicities ( [ z / h]=+1.3 dex ) are far beyond current ssp models .
therefore , it is impossible to pin down exact values for the central galaxy field stars . at [ z / h]=+1.3
the total metallicity would be 20 times that of the sun , implying a metal content by mass of @xmath1940% .
this value is unrealistically high , which suggesting that our extrapolations beyond the current ssp models are unreliable .
van gorkom ( 1986 ) speculated that ngc 1052 had accreted a gas - rich dwarf galaxy about 1 gyr ago .
they find a total hi mass of @xmath32 which seems consistent with this idea .
such a small quantity of gas could also explain why we have detected no gcs forming in this merger ( even allowing for some cold gas to have been used up in field star formation and/or ionised )
. even a small burst by mass will give a young age for the galaxy central stars ( terlevich & forbes 2002 ) .
our results also indicate that the central merger - induced starburst involved very high metallicity and @xmath4element enhanced gas .
such high metallicity gas is not usually associated with dwarf galaxies but massive spirals .
the most obvious explanation for the high @xmath4element enhancement is a high central concentration type ii supernovae from the starburst .
inflows over any significant timescale are less likely , since no radial metallicity gradient is seen
. a spiral galaxy with a small bulge and on - going star formation in the disk could produce the young central ages observed without star formation from the merger itself being the dominant luminosity source . to better evaluate these possibilities
we need stellar population information for the galaxy out to at least one effective radius .
this will provide an age , metallicity and @xmath4element abundance ratio for the assumed old ( @xmath2210 gyr ) underlying population . with this information
the next step would be to apply a chemo - gas code in an attempt to reproduce the observed properties . in terms of the gc formation models described in the introduction ,
the detection of old red and blue gc sub - populations in ngc 1052 is generally consistent with the predictions of both forbes ( 1997 ) and cote ( 1998 ) .
the ages are not yet accurate enough to determine whether the red gcs are a few gyr younger than the blue ones as expected in forbes ( 1997 ) . on the other hand
, we find no evidence for gcs forming in the merger event , in contrast to the expectations of the merger model of ashman & zepf ( 1992 ) .
there are now several examples of galaxies with spectroscopically confirmed _ minor _ contributions to their gc systems from recently formed gcs , while the vast bulk of their gcs appear to be very old .
ngc 1399 , a central cluster galaxy which is old and has no photometric indication of intermediate - age gcs , may have a small number of @xmath192 gyr old gcs ( forbes 2001a ) .
photometry of gcs in the merger remnant ngc 3610 , which has a spectroscopic age of [email protected] gyrs ( denicolo 2005 ) , suggested the presence of an intermediate - age sub - population ( whitmore 2002 ) .
however , spectroscopy of some of these candidate intermediate - age gcs indicates that only a small portion of these clusters are in fact @xmath192 gyr old ( strader 2004 ) .
perhaps the best example of intermediate - age gcs in a merger remnant is ngc 1316 , it has a similar age ( @xmath193 gyrs ) and lies at a similar distance ( 22.9 mpc ) to ngc 1052 ( goudfrooij 2001 ) .
figure [ disc1316 ] shows a comparison of the colour magnitude diagrams for the spectroscopically confirmed gcs of these two galaxies . only the 3 brightest ngc 1316 gcs had high enough s / n to measure ages , and these are 2 magnitudes brighter in v than the brightest gcs for which we obtained spectra in ngc 1052 .
all three gcs were found to be @xmath193 gyrs old and solar metallicity ( goudfrooij 2001 ) .
the ngc 1316 gc colour distribution peaks at b
[email protected] , which lies between the peaks of the ngc 1052 gc colour distribution .
there is no significant population of ngc 1052 gcs in the same colour and magnitude parameter space as the young ngc 1316 gcs .
therefore we do not expect a similar sub - population of @xmath333 gyr gcs in ngc 1052 .
we have obtained low - resolution spectra for 19 globular cluster ( gc ) candidates associated with the merger remnant elliptical ngc 1052 . of this sample ,
16 are identified as @xmath34 gcs by their radial velocities . using these velocities
, we derive a virial mass of [email protected] x 10@xmath35 within a radius of @xmath196.5 r@xmath26 ( 19 kpc ) . if the orbits had a strong tangental bias then our tracer mass estimate would be 30% too large . using the multi - index @xmath7 fitting technique of proctor & sansom ( 2002 ) , and the simple stellar population models of thomas , maraston & korn ( 2004 ) we derive individual ages , metallicities and abundance ratios for the 16 confirmed gcs .
we find all of the gcs to be very old , i.e. @xmath36 gyr , with a range of metallicities .
the predicted colours , based on the derived ages and metallicities , agree well with the observed gc colours .
we find no evidence for young gcs associated with the likely minor merger event @xmath191 gyr ago .
we also obtained a long - slit spectrum covering the central @xmath37 of ngc 1052 .
no strong gradient in either age or metallicity was found .
however , a large abundance ratio gradient exists .
the stellar population in the central regions of ngc 1052 has a luminosity - weighted age of @xmath1 gyr with [ fe / h]@xmath2 and a very high @xmath4element abundance ratio of @xmath19 + 0.8 dex .
the recent central star formation episode was most likely induced by infalling gas associated with the recent merger .
thus , although ngc 1052 shows substantial evidence for a recent merger and an associated starburst , it appears that the merger did not induce the formation of many , if any , new gcs . as the formation of luminous star clusters appears to accompany most significant star formation events ,
the absence of young gcs and the high [ e / fe ] values in the center of the galaxy ( suggesting short star formation timescales ) may indicate that a relatively small amount of star formation occurred in the merger .
this interpretation is consistent with `` frosting '' models for the formation of early - type galaxies ( e.g. , trager 2000b ) .
we thank soeren larsen for help preparing the slit mask .
part of this research was funded by nsf grant ast-02 - 06139 the data presented herein were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
this research has made use of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , caltech , under contract with the national aeronautics and space administration .
df and rp thanks the arc for their financial support . | we have obtained keck spectra for 16 globular clusters ( gcs ) associated with the merger remnant elliptical ngc 1052 , as well as a long - slit spectrum of the galaxy .
we derive ages , metallicities and abundance ratios from simple stellar population models using the methods of proctor & sansom ( 2002 ) , applied to extragalactic gcs for the first time .
a number of gcs indicate the presence of strong blue horizontal branches which are not fully accounted for in the current stellar population models .
we find all of the gcs to be @xmath0 gyr old according to simple stellar populations , with a large range of metallicities . from the galaxy spectrum we find ngc 1052 to have a luminosity - weighted central age of @xmath1 gyr and metallicity of [ fe / h]@xmath2 .
no strong gradients in either age or metallicity were found to the maximum radius measured ( 0.3 r@xmath3 kpc ) .
however , we do find a strong radial gradient in @xmath4element abundance , which reaches a very high central value .
the young central starburst age is consistent with the age inferred from the hi tidal tails and infalling gas of @xmath5 gyr .
thus , although ngc 1052 shows substantial evidence for a recent merger and an associated starburst , it appears that the merger did not induce the formation of new gcs , perhaps suggesting that little recent star formation occurred .
this interpretation is consistent with `` frosting '' models for early - type galaxy formation .
globular clusters : general galaxies : individual : ngc 1052 galaxies : star clusters . |
the study of the critical properties of systems subject to quenched randomness is of great interest because of its relevance for many experimental systems .
this is a challenging research area which has been quite active in the last decades .
one has witnessed a growing interest in understanding the effect of randomness ( bond and site disorder ) , since harris showed that disorder is a relevant perturbation in a system undergoing a second order phase transition when the exponent @xmath2 of the specific heat in the pure system is positive.@xcite in this context the two dimensional ising model has attracted considerable interest , since the onsager singularity of the specific heat ( @xmath3 ) makes the harris criterion inconclusive ( for a review , see e.g. ref . ) .
renormalization group studies,@xcite supported by monte carlo simulations,@xcite established that homogeneous disorder is marginally irrelevant , i.e. it does not modify the critical behavior of the system , except for the appearance of logarithmic corrections . in accordance with the harris criterion , randomness is on the other hand a relevant perturbation for the @xmath4 and @xmath5 potts models .
furthermore , an infinitesimal amount of disorder was shown to turn the first order phase transition of the @xmath6 potts model into a second order one.@xcite numerical estimates of the critical exponents for an homogeneous disorder indeed showed a monotonic increase of the magnetization exponent @xmath7 with the value @xmath0 while the thermal exponent @xmath8 remains constant within numerical accuracy.@xcite thus for the two dimensional classical potts model , _ homogeneous _ disorder gives rise to several different universality classes dependent on the value of @xmath0 .
other systems with quenched randomness which have been investigated in the past years are quantum spin chains .
the interest in these models has grown considerably after a remarkable paper of fisher @xcite who derived many of the critical properties of the random transverse ising chain ( rtic ) using a real space renormalization - group scheme which is claimed to be asymptotically exact .
his predictions were checked numerically by several authors , using the mapping of the rtic onto a free fermion model,@xcite density - matrix renormalization group @xcite ( dmrg ) or quantum monte carlo.@xcite critical exponents calculated numerically were found in good agreement with fisher s results .
the renormalization scheme of fisher was also applied to the @xmath0-state disordered potts chain by senthil and majumdar;@xcite they found that the critical behavior does not depend on @xmath0 , therefore all critical exponents should be identical to those of the rtic obtained by fisher and corresponding to the case @xmath9 .
we recall that the disordered quantum potts chain is equivalent to a two dimensional classical potts model with _ correlated _ disorder ( analogous to the mccoy - wu model @xcite corresponding to @xmath9 ) , where the quantum formulation is obtained from the classical one in the extreme anisotropic limit.@xcite the conclusion of a unique universality class is thus different from what found in the classical potts model with homogeneous disorder discussed above . in this paper , we check the predictions of a unique universality class , independent of the number of states @xmath0 , by studying bulk and surface magnetization properties .
we consider both the regimes @xmath10 and @xmath6 , where the pure model exhibits a second and first - order transition , respectively .
we also report an investigation of the multiscaling behavior of the magnetization , which strongly supports the asymptotic expression for the probability distribution found by fisher in the rtic case .
the dynamical behavior is considered through the decay of the spin - spin correlation function along the time direction .
all numerical results are obtained by density matrix renormalization group techniques,@xcite applied to the two dimensional classical potts model with correlated disorder .
we use nishino s @xcite version of white s dmrg algorithm,@xcite where one renormalizes classical transfer matrices .
the exponents are calculated from finite size scaling analysis of finite lattice data for chains of lengths @xmath1 with exact enumeration over all possible ( @xmath11 ) disorder realizations , obtained from a binary distribution .
the justification for this type of approach will be presented in more detail in the text .
our numerical results are fully consistent with the conclusion of senthil and majumdar,@xcite i.e. the @xmath0-state potts model has critical properties which do not depend on @xmath0 .
moreover , dynamical properties which are not accessible by fisher s renormalization group , do not seem to depend on @xmath0 . the paper is organized as follows : in sec .
[ sec : mod ] , the model is presented and we give a justification for the choice of the numerical methods that are used .
section [ sec : exp ] is devoted to the estimates of the critical exponents .
multi - fractality is investigated in sec .
[ sec : multi ] .
the dynamical spin - spin correlation functions are studied in sec .
[ sec : corr ] .
finally , a brief summary and an outlook will be presented in sec .
[ sec : con ] . 0.2truecm
we consider a randomly layered ( see fig . [ fig5 ] ) two - dimensional classical @xmath0-state potts model on a square lattice .
the model is defined by the hamiltonian @xmath12 , \label{eq1}\ ] ] where the spins can take the values @xmath13 and the exchange couplings @xmath14 are randomly chosen according to the binary distribution @xmath15 .
\label{eq2}\ ] ] since the model is equivalent to the random quantum potts chain studied by senthil and majumdar,@xcite we use the quantum chain s vocabulary . in the case
@xmath9 , the system reduces to the rtic .
the particular arrangement of coupling constants , i.e. the fact that vertical couplings are identical to adjacent horizontal couplings to their right side ( see fig . [ fig5 ] ) , and that both couplings @xmath16 and @xmath17 have equal probabilities ( see eq .
( [ eq2 ] ) ) , makes the model self - dual .
the location of the critical point is known exactly:@xcite @xmath18 the strength of the disorder is monitored by the ratio @xmath19 , @xmath20 corresponding to the pure model .
although the dmrg method is capable of dealing with very large systems ( of few hundreds of sites ) with remarkable accuracy , some care has to be taken when working with disordered systems . the infinite lattice version of the algorithm is known to give incorrect results,@xcite especially at strong disorder ; it is _ essential _
@xcite when dealing with inhomogeneous systems to use the finite lattice dmrg algorithm @xcite which is designed to determine accurately properties of finite systems ( in our calculations we typically used three full sweeps through the lattice ) .
it is furthermore well known that due to the lack of self - averaging,@xcite an insufficient number of samples yields a wrong estimation of the average over randomness and might lead to typical values instead of the expected average quantities.@xcite as we aim to an accurate determination of critical exponents , we have chosen to restrict ourselves to relatively narrow strips ( @xmath1 which turns out to be large enough , since corrections to scaling will appear very small ) for which we were able to do an exact enumeration over all disorder realizations . as the system size is small compared to typical dmrg calculations a limited number of states kept ( @xmath21 ) is sufficient to obtain accurate results .
the setup of the calculation follows ref . , where the dmrg method was used to calculate exponents and magnetization profiles for the pure potts model .
the @xmath22-symmetry of the potts model has been broken by introducing fixed free boundary conditions : we fixed the spin at one edge of the system to a value @xmath23 , while at the other edge the spin is free .
this type of boundary conditions induces a finite magnetization in the system @xcite from which surface and bulk critical exponents can be calculated using finite size scaling analysis .
figure [ fig4 ] shows magnetization profiles ( defined by eq .
( [ eq4 ] ) ) of @xmath24 samples for the three - state potts model with @xmath25 and with disorder amplitude @xmath26 .
estimated error bars are much smaller than symbol sizes ; averaging over all disorder realizations produces the smooth profiles shown in fig .
[ fig1 ] . in this procedure
the only source of error comes from the truncation on the dmrg basis , this error is quite small even for @xmath27 states kept as we restricted ourselves to small systems . 0.2truecm
we used the standard order parameter of the potts model : @xmath28_{\rm av } = { q \left [ \langle\delta_{\sigma_{l , n},0}\rangle \right]_{\rm av } - 1\over q - 1 } , \label{eq4}\ ] ] where @xmath29 denotes the thermal average and @xmath30_{\rm av}$ ] the average over randomness .
standard finite size techniques have been used to estimate the bulk and surface critical exponents @xmath31_{\rm av}\sim l^{-x_b } , \quad [ m_l^s ] _
{ \rm av}\sim l^{- x_1 } , \label{eq5}\ ] ] where we have defined @xmath32_{\rm av } \equiv [ m_l(l/2 ) ] _ { \rm av}$ ] and @xmath33_{\rm av } \equiv [ m_l(l ) ] _ { \rm av}$ ] , which are the average magnetizations in the middle of the strip and at the free edge . from finite size data we calculated the approximants , defined as @xmath34_{\rm av } - \ln [ m_{l-1}^b ] _ { \rm av } \over \ln ( l+1)-\ln ( l-1 ) } \label{eq6}\ ] ] in the bulk case and similarly for the surface approximants . in the limit
@xmath35 the approximants scale towards the bulk and surface exponents @xmath36 and @xmath37 .
table [ table1 ] reports the finite size approximants for the surface exponents for various values of @xmath0 and of disorder ratios @xmath38 .
the extrapolation towards the thermodynamic limit was done using the bst algorithm.@xcite the boundary magnetization exponent , @xmath37 , exhibits a striking convergence for all values of @xmath0 and all disorder amplitudes , but for crossover effects discussed below .
it is clearly independent of @xmath0 , and in good agreement with the rtic value @xmath39 .
the best numerical results are obtained for @xmath9 , as the truncation error in the dmrg method is the smallest for this case , furthermore the surface exponent for the pure and disordered cases coincide for @xmath9 ( they are both equal to @xmath40 ) , therefore no crossover effects are to be expected . 2 we recall that the values of the surface exponents for the pure potts model are @xmath41 for @xmath4 , @xmath42 for @xmath5 and @xmath43 for @xmath44 .
the last exponent has also been recently calculated by dmrg methods @xcite ( the surface transition is continuous even if the bulk has a first order transition ) . in the case
@xmath45 , @xmath46 , one notices that the approximants approach the value @xmath40 in a non - monotonic fashion ( the same situation occurs with @xmath5 , @xmath26 ) .
this case corresponds to the weakest disorder considered and the non - monotonicity is most likely due to crossover effects from the exponent of the pure system ( @xmath47 ) .
it is not possible to extrapolate the data in the last column of table [ table1 ] with the bst method , since the largest size analyzed is still in the crossover region .
however we can give an estimate of the exponent using the analysis of the moments presented in section iv b ; from such an analysis we find 0.49(3 ) in good agreement with the value of the rtic , @xmath40 .
we also mention that calculations have been performed at @xmath45 , @xmath48 .
the approximants then do not suffer from so strong crossover effects , an observation which supports the explanation given above .
table [ table2 ] shows the approximants of the bulk exponent obtained from eq .
( [ eq6 ] ) for various @xmath0 and @xmath38 . as it can be seen
the approximants show a non - monotonic behavior , with oscillations which are stronger as the disorder ratio increases .
these oscillations , which were already observed in the rtic,@xcite are due to the choice of a binary distribution of exchange couplings which introduces an energy scale @xmath49 in the problem ; in other types of distributions , e.g. continuous ones , these oscillations are not present.@xcite lattice extrapolation techniques , as the bst method , are not able to deal with oscillating corrections to scaling .
we recall that the bst algorithm generates from an original series of @xmath50 approximants new series of @xmath51 , @xmath52 approximants at each iteration which are expected to have faster convergence towards the asymptotic value . in the bst analysis we found that even when starting from a monotonic series of data at weak disorder ratios , as in the cases @xmath9 , @xmath26 , the next series generated by the algorithm do not converge at all , but they show oscillating behavior .
therefore the bst method turns out to be useless in the analysis of the bulk exponent data . a better way to get a precise estimate of bulk exponent
is based on the analysis of moments , which will be explained in section [ sec : manal ] . in any case
the approximants of table [ table2 ] for different values of @xmath0 are rather close to each other , which supports again the idea of @xmath0-independence of their values .
we recall that for the pure potts model the bulk exponent is @xmath53 for @xmath54 and @xmath55 for @xmath4 .
for @xmath44 the bulk transition is first order .
fisher s prediction for the bulk exponent is:@xcite @xmath56 figure [ fig6 ] shows a log - log plot of bulk magnetization vs. system size .
the data correspond to different @xmath0 and @xmath38 and are all roughly parallel with slopes in the range @xmath57 indicated in the figure and obtained by a simple linear fit . in the scale of the figure oscillations
are not visible and the estimated slopes correspond to the average of the approximants of table [ table2 ] .
2 0.2truecm we mention that , surprisingly , we observed a weak dependence of the effective critical exponents with the amplitude of disorder @xmath38 in the range @xmath58 $ ] , while important cross - over effects with the strength of the disorder were observed , e.g. by picco,@xcite in the case of a two - dimensional potts model with a homogeneous disorder .
this is an evidence of the strongly attractive character of the random fixed point .
conformal invariance techniques are extremely accurate for studying second order phase transitions in isotropic pure systems .
conformal symmetry requires translation , rotation and scale invariance ; obviously a single disorder realization does not have such properties , but it is plausible that conformal symmetry is restored after averaging over different disorder configurations .
indeed in recent investigations , conformal mappings of correlation functions and order parameter profiles have been successfully applied in two - dimensional systems with homogeneous disorder.@xcite the case of correlated disorder , studied in this paper , is quite different as the system , even after disorder average , exhibits infinite anisotropy at the random fixed point ( see sec . [
sec : corr ] ) .
therefore conformal invariance should not hold . in spite of these restrictions ,
the magnetization profiles of the rtic with fixed - free boundary conditions have been quite satisfactorily fitted with the conformal expression,@xcite , obtained e.g. by burkhardt and xue @xcite @xmath59_{\rm av}={\cal a}(\pi / l)^{x_b } [ \sin\pi \zeta]^{-x_b } [ \cos\pi \zeta/2]^{x_1 } , \label{eq10}\ ] ] where @xmath60 .
the success of such an approach probably results from the geometrical constraints induced by the strip shape of the system combined to scaling arguments : in this sense , if eq .
( [ eq10 ] ) turns out to be valid also for the disordered potts chain , one should conclude that such a form of profile is not a strict consequence of conformal invariance , which does not hold in anisotropic critical systems,@xcite but follows from more general grounds .
0.2truecm in this section , we indeed ask the question whether eq .
( [ eq10 ] ) correctly describes the transverse magnetization profiles in the random potts chain .
senthil and majumdar @xcite indeed argued that the scaling functions should be independent of @xmath0 .
the quantity @xmath61 being such a scaling function , if eq .
( [ eq10 ] ) holds in the rtic chain when @xmath62 , it should also be valid in the same limit in the potts case .
unfortunately , in our case , the largest strip being @xmath63 , strong lattice effects make the fit to expression ( [ eq10 ] ) impossible in practice ( fig .
[ fig1 ] ) .
the same observation was reported for the rtic , but these finite - size effects vanish for the largest systems that can be reached using the free fermion mapping in the ising case . if eq .
( [ eq10 ] ) is valid , the quantity @xmath64_{\rm
av}- \ln [ m_l(l/4 ) ] _ { \rm av}\over \ln(\sqrt 2 - 1 ) } \label{eq11}\ ] ] should converge towards the surface exponent @xmath40 .
the results are shown in table [ table3 ] .
this method can not be considered as an exact determination of @xmath37 but it provides an indication of the possible validity of ( [ eq10 ] ) for this system .
the bulk exponent @xmath36 also follows from a three - point expression analogous to eq .
( [ eq11 ] ) and should be extracted from the magnetization profile , but our data are not accurate enough to allow this calculation .
according to fisher , the probability distribution of the surface magnetization @xmath65 for a strip of width @xmath66 is expected to behave asymptotically as : @xmath67 this expression was successfully used for the rtic by igli and rieger to rescale the probability distribution of the surface magnetization for several strips of different widths @xmath66 in ref . .
the scaling function @xmath68 was found to have a power - law decay in the rtic .
forgetting about the oscillations , the average behavior in fig .
[ fig2 ] is compatible with this observation .
the results of senthil and majumdar @xcite state that the scaling function @xmath68 is independent of @xmath0 in the thermodynamic limit . here
again , the widths of our strips are too small to give a smooth probability distribution , as shown in fig .
the figure shows that for the events which determine the average critical properties , i.e. those corresponding to the dominant contributions to the magnetization , the probability distribution is weakly dependent on @xmath0 , in accordance with universality of eq .
( [ eq12 ] ) . on the other hand , for the rare samples for which @xmath65 or @xmath69 ,
eq ( [ eq12 ] ) breaks down .
it is however possible to test expression ( [ eq12 ] ) by analyzing the moments of the distribution .
0.2truecm in this section we analyze moments of bulk and surface magnetization distributions . this analysis will help to avoid the problems related to the oscillations of the finite size estimates of the bulk exponent which
, as shown in the previous section , hindered a reliable extrapolation of such quantity .
moments analysis was recently used in studies of self - organized critical systems.@xcite it was shown that such method provides an accurate means of investigation of subtle aspects that seem to emerge for such models .
in our case we are interested in the scaling of the asymptotic behavior of the moments .
for instance , in the case of the surface or bulk magnetization we expect : @xmath70_{\rm av } \sim l^{-\chi ( \omega ) } .\ ] ] the exponent @xmath71 generalizes the definitions of eq.([eq6 ] ) , which correspond to @xmath72 .
if the system is self - averaging , we expect a linear dependence of @xmath73 with respect to @xmath74.@xcite at a fixed size @xmath66 , we consider the finite - size approximants of @xmath71 defined as usual by : @xmath75_{\rm av } - \ln { [ { ( m_{l-1 } ) ^\omega } ] _ { \rm av } } } { \ln ( l + 1 ) - \ln ( l - 1 ) } .\ ] ] plots of both quantities for @xmath76 and @xmath4 , @xmath77 are shown in fig .
the values at @xmath72 are those shown in the tables [ table1 ] and [ table2 ] .
the fact that @xmath78 is a consequence of the normalization of the probability distribution .
we notice that in both the surface and bulk case @xmath79 for large @xmath66 seems to be a constant independent of @xmath74 ; this constant equals the surface or bulk exponent .
this can be understood if one assumes the probability distribution given in eq .
( [ eq12 ] ) . as a consequence of this
one indeed obtains : @xmath80_{\rm av } } = \frac{1}{\sqrt{l } } \,\ , \int_0 ^ 1 d m_l^s \,\ , ( m_l^s)^\omega \,\ , \tilde{p } \left ( \frac { \ln m_l^s } { \sqrt l } \right ) \nonumber \\ = \frac{1 } { ( 1 + \omega ) \sqrt{l } } \int_0 ^ 1 d x \,\ , \tilde{p } \left ( \frac { \ln x } { ( 1 + \omega ) \sqrt{l } } \right ) , \label{scalgamma } \end{aligned}\ ] ] where we have used the change of variables @xmath81 .
we notice from this formula that @xmath74 does not modify the exponent of @xmath66 , but enters only as a prefactor of @xmath82 and consequently all the moments should have the same scaling dimension @xmath83 .
this is what observed in fig .
[ fig7 ] where the calculated @xmath79 does not seem to depend on @xmath74 , therefore our numerical results are consistent with a probability distribution of the form of eq .
( [ eq12 ] ) .
it is also important to stress that results would have been very different for a probability distribution with another scaling variable , e.g. @xmath84 .
we note that in eq .
( [ scalgamma ] ) , the last integral seems to be nothing but a normalization , but as we mentioned in the previous section the expression used for the probability distribution is only asymptotically valid for the dominant events , and thus the true normalization of the distribution differs from eq .
( [ scalgamma ] ) .
0.2truecm an interesting consequence of these observations is that one could calculate surface or bulk exponents from an arbitrary value of @xmath74 . from the point of view of the finite size scaling analysis we consider values of @xmath74 where the dependence on the system size @xmath66 is the smallest . for the data referring to the bulk exponent and shown in the figure
we see that around @xmath85 ( this point is indicated by a thick arrow in the figure ) finite size effects are very small , all @xmath79 seem to intersect into a single value ; for which we estimate @xmath86 in excellent agreement with the magnetic exponent obtained by fisher @xmath87 and also indicated by a thick dashed line in the figure .
table [ genscaling ] collects the values of bulk exponents calculated with this method . for all cases
studied we find a value of @xmath74 for which @xmath79 depends only very weakly on @xmath66 ; if the intersection of the @xmath79 for different @xmath66 is sharp ( as in the example shown in fig . [ fig7 ] ) one arrives to a very good estimate of the exponent .
we should also stress that , as the system size enters in eq .
( [ scalgamma ] ) together with @xmath74 in the form of a scaling variable @xmath88 .
this suggests that higher moments of the magnetizations should show weaker finite size corrections , as increasing @xmath74 has the same effect of enlarging the system size .
this is only partially true as eq .
( [ scalgamma ] ) describes only the leading scaling behavior in @xmath89 , and it does not include further corrections to scaling terms .
in this section , we deal with the dynamical properties of the random potts chain .
we thus focus here on correlation functions in the temporal direction , which are defined , in the quantum formalism , by : @xmath90_{\rm av } = \left[{\bra 0 \delta_{\sigma_i,0}\ e^{-{\cal h}t } \delta_{\sigma_i,0}\ket 0 \over \bra 0\ e^{-{\cal h}t}\ket 0}\right]_{\rm av } , \label{eq16}\ ] ] where @xmath91 denotes the ground state .
fisher s renormalization group does not provide any information about this quantity .
however , igli and rieger by means of scaling arguments derived the following asymptotic behavior : @xmath90_{\rm av}\sim ( \ln t)^{-2x_b}. \label{eq20}\ ] ] this asymptotic decay of the correlation functions is unusually slow with respect to usual power - law behavior expected at criticality .
this can be understood more easily in the classical version of the model and it is due to the fact that disorder along the transfer direction is strongly correlated @xcite ( layered disorder ) and consequently the model displays anisotropic critical behavior .
it means that at the critical point , distances measured in space and time directions are connected through power laws involving the dynamical exponent @xmath92 : @xmath93 .
the homogeneity assumption for the correlation functions thus becomes @xmath94_{\rm av}=b^{-2x_b}g(l / b , t / b^z ) , \label{eqgz}\ ] ] where @xmath95 is a rescaling factor .
the choice @xmath96 thus leads to the usual power - law decay @xmath94_{\rm av}\sim t^{-2x_b
/ z}g(l / t^{1/z } ) , \label{eqgzbis}\ ] ] while in the case of infinite anisotropy ( @xmath97 ) one should set @xmath98 and eq .
( [ eq20 ] ) follows . in the classical formulation of
the problem followed in this paper , the dynamical correlation function is defined by : @xmath99_{\rm av } = \left [ \frac { \langle v_0 | \delta_{\sigma_i , 0 } t^n \delta_{\sigma_i , 0 } | v_0 \rangle } { \lambda_0^n } \right]_{\rm av } , \label{classicalcorrelation}\ ] ] where @xmath100 and @xmath101 are the largest eigenvalue and the corresponding eigenvector of the transfer matrix @xmath102 . for the calculation , we used free boundary conditions on both edges ( the magnetization defined by eq .
( [ eq4 ] ) is thus zero in the whole strip ) and we restricted ourselves to the spin in the middle of the system . since large strip widths are needed in order to avoid finite size effects , average
are now performed with @xmath103 different samples . 0.2truecm figure [ fig8 ] shows the correlation function calculated by dmrg techniques following eq .
( [ classicalcorrelation ] ) for @xmath9 and @xmath48 .
we plot the logarithm of the correlation function vs. the double logarithm of the distance along the time direction ( called @xmath104 in fig .
[ fig5 ] ) ; the dashed line is the asymptotic slope predicted by eq .
( [ eq20 ] ) , where we have used fisher s exponent given in eq .
( [ eq7 ] ) .
this asymptotic behavior seems to be in agreement with our numerical data for the largest system size analyzed ( @xmath105 ) for which we took @xmath27 and @xmath106 disorder realizations .
0.2truecm figure [ fig8b ] shows the analogous plot of the time correlation function for @xmath4 and @xmath48 .
again for the largest system size investigated ( @xmath107 ) one notices good agreement with the expected asymptotic limit shown as a thick dashed line . differently from fig .
[ fig8 ] we notice here a crossing of the curves representing correlation functions for @xmath108 and @xmath107 .
this is we believe an effect of the dmrg approach which does not correctly describe the short distance correlation functions , since it does not reproduce the full transfer matrix spectrum . in the other hand , when the number of transfer matrix products becomes large in eq .
( [ classicalcorrelation ] ) the correlation function is determined by the extremal part of the spectrum of @xmath102 , which is believed to be well - reproduced by dmrg methods .
therefore we believe that the crossing of fig .
[ fig8b ] is due to the limits of the dmrg approximation , while results at larger @xmath104 should be more accurate
. in conclusion the analysis of the time correlation function for @xmath9 and @xmath4 also indicates that the universal dynamical properties of the disordered quantum potts chain are independent of @xmath0 .
we have provided numerical evidences supporting the conclusions of senthil and majumdar @xcite concerning the existence of a unique universality class for all finite value of the @xmath0-state potts model perturbed by a correlated disorder . for this purpose
, we have measured the bulk and surface critical exponents and found values fully compatible with those exactly calculated by fisher for the rtic .
we point out that a careful numerical study of the critical properties of disordered quantum chains requires an accurate average over randomness , since the most important source of error is due to the disorder average .
this effect becomes essential in some circumstances , possibly leading to wrong exponents , and may even dominate other contributions like finite - size effects . for that reason
, we generated _ all _ the disorder realizations for chains of relatively small lengths , using the fact that the finite - size corrections turned out to be quite small ( even compared to their importance in the corresponding pure models ) .
furthermore , we have checked the expression of the probability distribution by an analysis of its moments : since the disorder average was exact up to the accuracy of the dmrg method , this computation was allowed even for large moments .
the success of this approach is largely due the occurrence of an optimal moment for which the finite - size corrections appear to be extremely weak .
the method looks promising for further applications to disordered systems .
finally , we have also compared the expression of the dynamical correlation function with that of the rtic .
another interesting feature is the illustration , after the work of aizenman and wehr,@xcite that correlated disorder induces a second - order phase transition in the @xmath109state potts model with @xmath6 .
the universality class is however very robust , in contradistinction with the homogeneous two - dimensional disordered fixed point where the magnetic exponents continuously vary with @xmath0 .
we would like to thank r. couturier for help with the code parallelization and d. karevski for critical reading of the manuscript .
the computations were performed on the _ sp2 _ at the cnusc in montpellier under project no .
c990018 , and the _ power challenge array _ at the cch in nancy .
e.c . acknowledges grant from the ministre des affaires etrangres no 224679a .
w. selke , l.n .
shchur and a.l .
talapov , in _ annual reviews of computational physics _ , edited by d. stauffer ( world scientific , singapore , 1994 ) , vol 1 p. 17 .
m. picco , e - print cond - mat/9802092 .
h. rieger and n. kawashima , e - print cond - mat/9802104 .
b.m . mccoy and t.t .
rev.*176 * , 631 ( 1968 ) ; b.m .
mccoy and t.t .
wu , phys .
rev.*188 * , 982 ( 1969 ) ; b.m .
mccoy , phys .
* 188 * , 1014 ( 1969 ) .
kogut , rev .
phys . * 51 * , 659 ( 1979 ) .
s. r. white , , 2863 ( 1992 ) ; , 10345 ( 1993 ) .
i. peschel , x. wang , m. kaulke and k. hallberg ( eds . ) : _ density matrix renormalization : a new numerical method in physics _ , springer , 1999 .
similar remarks can be found in s. rapsch , u. schollwck , and w. zwenger , e - print cond - mat/9901080 .
we followed a different approach than in ref . , where a bulk magnetic field was added to induce a finite magnetization .
e - print cond - mat/9903270 . with another definition , namely @xmath110_{\rm av}\sim l^{-\omega\chi(\omega)}$ ]
, one would expect in case of self - averaging a unique exponent @xmath83 for all values of @xmath74 , i.e. all the moments would scale like the average . | we study the critical behavior of the random @xmath0-state potts quantum chain by density matrix renormalization techniques .
critical exponents are calculated by scaling analysis of finite lattice data of short chains ( @xmath1 ) averaging over all possible realizations of disorder configurations chosen according to a binary distribution .
our numerical results show that the critical properties of the model are independent of @xmath0 in agreement with a renormalization group analysis of senthil and majumdar ( , 3001 ( 1996 ) ) .
we show how an accurate analysis of moments of the distribution of magnetizations allows a precise determination of critical exponents , circumventing some problems related to binary disorder .
multiscaling properties of the model and dynamical correlation functions are also investigated .
# 1| # 1 # 1#1| # 1#2 # 1 . 2 |
knot floer homology is a powerful knot invariant constructed by ozsvth szab @xcite and rasmussen @xcite . in its basic form ,
the knot floer homology @xmath1 of a knot @xmath2 is a finite dimensional bigraded vector space over @xmath3 @xmath4 where @xmath5 is the maslov and @xmath6 is the alexander grading .
its graded euler characteristic @xmath7 is equal to the symmetrized alexander polynomial @xmath8 .
the knot floer homology enjoys the following symmetry extending that of the alexander polynomial .
d(k , i)=_d-2i(k ,-
i ) by the result of ozsvth szab @xcite , the maximal alexander grading @xmath6 , such that @xmath9 is the seifert genus @xmath10 of @xmath11 .
moreover , ghiggini showed for @xmath12 @xcite and yi ni in general @xcite , that the knot is fibered if and only if @xmath13 . a concordance invariant bounding from below the slice genus of the knot can also be extracted from knot floer homology @xcite . for torus knots the bound is sharp , providing a new proof of the milnor conjecture .
the first proof of the milnor conjecture was given by kronheimer and mrowka @xcite , then rasmussen @xcite proved it combinatorially by using khovanov homology @xcite .
knot floer homology was extended to links in @xcite .
the first combinatorial construction of the link floer homology was given in @xcite over @xmath14 and then in @xcite over @xmath15 .
both constructions use grid diagrams of links .
a _ grid diagram _ is a square grid on the plane with @xmath16 squares .
each square is decorated either with an @xmath17 , an @xmath18 , or nothing . moreover
, every row and every column contains exactly one @xmath17 and one @xmath18 . the number @xmath19 is called _ complexity _ of the diagram . following @xcite
, we denote the set of all @xmath18 s and @xmath17 s by @xmath20 and @xmath21 , respectively . given a grid diagram , we construct an oriented , planar link projection by drawing horizontal segments from the @xmath18 s to the @xmath17 s in each row , and vertical segments from the @xmath17 s to the @xmath18 s in each column .
we assume that at every intersection point the vertical segment overpasses the horizontal one .
this produces a planar _ rectangular diagram _
@xmath22 for an oriented link @xmath23 in @xmath24 .
any link in @xmath24 admits a rectangular diagram ( see e.g. @xcite ) .
an example is shown in figure [ diag ] . 8 in @xcite , @xcite
the grid lies on the torus , obtained by gluing the top most segment of the grid to the bottommost one and the leftmost segment to the right most one . in the torus ,
the horizontal and vertical segments of the grid become circles .
the mos complex is then generated by @xmath19tuples of intersection points between horizontal and vertical circles , such that exactly one point belongs to each horizontal ( or vertical ) circle .
the differential is defined as follows : @xmath25 where @xmath26 is the subset of generators that have @xmath27 points in common with @xmath28 . for @xmath29
, @xmath30 is the set of rectangles with vertices @xmath31 and @xmath32 , whose interior does not contain @xmath17 s and @xmath18 s or points among @xmath28 and @xmath33 .
moreover , a counterclockwise rotation along the arc of the horizontal oval , leads from the vertices in @xmath28 to the ones in @xmath33 .
the alexander grading is given by formula below , and the maslov grading by plus one .
the mos complex has @xmath34 generators .
this number greatly exceeds the rank of its homology . for the trefoil , for example
, the number of generators is 120 , while the rank of @xmath35 is 3 . in this paper
, we construct another combinatorial complex computing link floer homology , which has significantly less generators .
all knots with less than 6 crossings admit rectangular diagrams where all differentials in our complex are zero , and the rank of the homology group is equal to the number of generators .
our construction also uses rectangular diagrams .
given an oriented link @xmath23 in @xmath24 , let @xmath22 be its rectangular diagram in @xmath36 .
let us draw @xmath37 narrow short ovals around all but one horizontal and all but one vertical segments of the rectangular diagram @xmath22 in such a way , that the outside domain has at least one point among @xmath21 or @xmath20 .
we denote by @xmath38 the set of unordered @xmath39tuples of intersection points between the horizontal and vertical ovals , such that exactly one point belongs to each horizontal ( or vertical ) oval .
we assume throughout this paper that the ovals intersect transversely .
an example is shown in figure [ fig:5 - 2 ] .
8 a chain complex @xmath40 computing the hat version of link floer homology of @xmath23 over @xmath3 is defined as follows .
the generators are elements of @xmath38 .
the bigrading on @xmath38 can be constructed analogously to those in @xcite .
suppose @xmath41 is the number of components of @xmath23 .
then the alexander grading is a function @xmath42 , defined as follows .
first , we define a function @xmath43 . for a point @xmath44 ,
the @xmath45 component of @xmath46 is minus the winding number of the projection of the @xmath45 component of the oriented link around @xmath44 . in the grid diagram
, we have @xmath47 distinguished squares containing @xmath17 s or @xmath18 s .
let @xmath48 , be the vertices of these squares .
given @xmath49 , we set @xmath50 where here @xmath51 is the complexity of the @xmath45 component of @xmath23 ,
i.e. the number of horizontal segments belonging to this component .
the homological or maslov grading is a function @xmath52 defined as follows .
given two collections @xmath53 , @xmath54 of finitely many points in the plane , let @xmath55 be the number of pairs @xmath56 and @xmath57 with @xmath58 and @xmath59 .
let @xmath60 .
define [ mas ] m()=j(,)-2j ( , ) + j ( , ) .
8 to construct a differential @xmath61 we first need to consider the complex @xmath62 defined in the same way as @xmath40 but where the ovals are as long as @xmath63 grid .
an example is shown in figure [ long ] .
the differential @xmath64 where @xmath65 is the subset of generators that have @xmath27 points in common with @xmath28 and for @xmath66 , @xmath67 is the bigon with vertices @xmath31 and @xmath32 , whose interior does not contain @xmath17 s and @xmath18 s .
moreover , a counterclockwise rotation along the arc of the horizontal oval , leads from the vertex in @xmath28 to the one in @xmath33 .
note that @xmath62 coincides with the complex @xmath68 defined by ozsvth
szab in @xcite , where the vertical and horizontal ovals are identified with @xmath69 and @xmath70 curves , respectively , and @xmath21 and @xmath20 are extra basepoints . like the mos complex ,
this complex is combinatorial , since all domains suitable for the differential are either rectangles or bigons ( compare @xcite ) .
following @xcite , we will call elements of @xmath71 _ basepoints _ in what follows . to this complex
we further apply a simple lemma from homological algebra , that allows us to construct a homotopy equivalent complex @xmath40 with @xmath38 as a set of generators ( or to shorten the ovals @xmath69 and @xmath70 keeping track of the differential ) .
in section 1 , we describe an algorithm that for any domain connecting two generators decides whether it counts for the differential @xmath61 or not . furthermore , we distinguish a large class of domains that always count . in general
, however the count depends on the order of shortening of ovals , which replace the choice of a complex structure in the analytic setting .
let @xmath72 be the two dimensional bigraded vector space over @xmath14 spanned by one generator in alexander and maslov gradings zero and another one in maslov grading @xmath73 and alexander grading minus the @xmath6th basis vector .
the complex @xmath40 has much fewer generators than the mos complex ( compare section 3 ) .
recently , droz introduced signs in our construction @xcite and wrote a computer program realizing our algorithm over @xmath15 @xcite .
his program allows to determine the seifert genus and fiberedness of knots until 16 crossings and also to study the torsion part of knot floer homology .
the paper is organized as follows .
theorem [ main ] is proved in section [ proofs ] . in section [ ind ]
we introduce a big class of domains that always count for the differential . in the last section
we compute floer homology of @xmath74 knot and discuss further computations made by droz .
first of all , i wish to thank stephen bigelow for explaining to me his homological constructions of link invariants , which provided the original motivation for this project , and also for regular conversations and helpful hints .
i would like to express my gratitude to jean
marie droz for many interesting discussions and , especially , for writing a program computing the homology of the complex constructed in this paper .
i profited a lot from discussions with stephan wehrli .
special thanks go to ciprian manolescu for sharing his knowledge of heegaard floer homology and for his valuable suggestions after reading the preliminary version of this paper .
suppose @xmath22 is a rectangular diagram of complexity @xmath19 for an oriented link @xmath23 .
let @xmath75 be the complex generated over @xmath14 by @xmath39tuples of intersection points between horizontal and vertical ovals , such that exactly one point belongs to each horizontal ( or vertical ) oval as defined in the introduction .
the length of the ovals can be intermediate between long and short ones .
we also assume that the outside domain has at least one basepoint inside .
given @xmath76 , there is an oriented closed curve @xmath77 composed of arcs belonging to horizontal and vertical ovals , where each piece of a horizontal oval connects a point in @xmath28 to a point in @xmath33 ( and hence each piece of the vertical one goes from a point in @xmath33 to a point in @xmath28 ) . in @xmath78
, there exists an oriented ( immersed ) domain @xmath79 bounded by @xmath77 .
the points in @xmath28 and @xmath33 are called _ corners _ of @xmath79 .
let @xmath80 be the closures of the connected components of the complement of ovals in @xmath78 .
suppose that the orientation of @xmath80 is induced by the orientation of @xmath78 .
then we say that a domain @xmath81 connects two generators if for all @xmath6 , @xmath82 and @xmath22 is connected .
let @xmath83 be the set of all domains connecting two generators which contain neither corners nor points among @xmath21 and @xmath20 inside .
we define [ different ] : = _ m()=m()-1 _ d_,d m(d _ , ) , where @xmath84 is the maslov grading defined by ( [ mas ] ) and @xmath85 is a multiplicity of @xmath79 .
we set @xmath86 for rectangles and bigons without basepoints inside .
this defines the differential for the long oval complex . in general ,
@xmath87 can be defined by using the procedure of shortening of ovals described in the next section inductively . in particular , @xmath61 is defined in section 1.3 .
assume that one vertical and one horizontal oval used to define @xmath75 form a bigon with corners @xmath88 and @xmath89 without basepoints inside .
we further assume that a counterclockwise rotation along the arc of the horizontal oval , leads from @xmath88 to @xmath89 .
an example is shown on the left of figure [ fig : bigon ] .
suppose that @xmath90 is given by with known multiplicities .
let @xmath91 be a new complex obtained from @xmath75 as follows .
the set of generators @xmath92 is obtained from @xmath93 by removing all generators containing @xmath88 or @xmath89 .
the differential @xmath94 where @xmath95 and @xmath96 are the obvious inclusion and projection .
moreover , for any @xmath97 , @xmath98 is zero whenever @xmath99 , otherwise @xmath98 is obtained from @xmath28 by replacing @xmath89 by @xmath88 .
8 [ bigon ] @xmath91 is a chain complex homotopic to @xmath75 .
the differential @xmath100 can be given by for a new system of curves obtained by shortening of one oval as shown in figure [ fig : bigon ] .
for the proof of this proposition we will need the following lemma . [ sublemma ] whenever @xmath101 , we have @xmath102 we have to show that for @xmath28 with @xmath101 , @xmath103 with @xmath104 occurs in @xmath105 if and only if @xmath103 is obtained from @xmath28 by replacing @xmath88 by @xmath89 . indeed , if @xmath106 , then there exists a domain @xmath107 with @xmath108 .
let us construct its boundary @xmath109 .
we start at @xmath88 and go to @xmath89 along an arc of a horizontal oval .
no other points in @xmath28 or @xmath103 belong to this arc , since only one point of each generator is on the same oval .
there are two choices for this arc : the short or the long one .
the domain having the long arc as a part of its boundary contains basepoints inside and will not be counted .
analogously , going back from @xmath89 to @xmath88 we have to take the short arc of the vertical oval , otherwise the domain will not be in @xmath83 .
hence , @xmath110 contains a small bigon between @xmath88 and @xmath89 as a boundary component .
let us show that @xmath110 has only one boundary component .
clearly , there are no further boundary components inside of this bigon . on the other hand
, the bigon could not be an inner boundary component , hence in this case , its orientation would be reversed , and then @xmath111 would be in @xmath83 , but not @xmath110 . since @xmath83 contains connected domains only , we proved the claim .
let us define the maps @xmath112 and @xmath113 as follows .
@xmath114 here @xmath115 is the identity map .
it is not difficult to see , that @xmath116 is the identity map on @xmath117 since @xmath118 , @xmath119 and @xmath120 are zero maps .
the map @xmath121 is homotopic to the identity on @xmath122 , i.e. @xmath123 .
this is easy to see for generators in @xmath124 or for @xmath28 with @xmath101 .
let us assume @xmath125 , then @xmath126 , where @xmath127 is a linear combination of generators , such that any of them contains @xmath88 and @xmath128 is a linear combination of generators in @xmath124
. then @xmath129 by using @xmath130 and the previous claim . on the other hand ,
@xmath131 now using lemma [ sublemma ] one can easily show that the differential @xmath100 coincides with @xmath132 .
furthermore , using the homotopy to the identity , proved above , we derive @xmath133 .
this shows that @xmath134 is indeed a chain complex . to show that @xmath135 and @xmath136 are chain maps , i.e. @xmath137 and @xmath138
, we again use @xmath139 , which is a consequence of lemma [ sublemma ] . finally , we would like to show that the new differential @xmath140 can be realized by counting of maslov index one domains for a new system of curves .
we write @xmath141 if @xmath86 .
assume @xmath141 and @xmath28 , @xmath33 do not contain the corners of the bigon @xmath88 and @xmath89 .
for all such @xmath28 and @xmath33 , we also have @xmath142 , and they are connected by a maslov index one domain @xmath143 .
furthermore , assume @xmath144 do not contain @xmath88 and @xmath89 , then for any @xmath145 with @xmath125 , @xmath101 , such that @xmath146 , @xmath147 , and @xmath148 , @xmath149 occurs once in @xmath150 .
note that @xmath151 , since any domain connecting @xmath152 to @xmath149 contains either the bigon with negative orientation or basepoints , hence they do not count for the differential .
however , the new system of ovals contains a domain connecting @xmath152 to @xmath149 which is obtained from @xmath153 by shortening the oval .
this process is illustrated in figure [ poly ] , where @xmath152 and @xmath149 are given by black and white points respectively ; @xmath33 is obtained from @xmath152 by switching the black point on the dashed oval to @xmath89 and the upper black point to the white point on the same oval ; @xmath28 is obtained from @xmath33 by switching @xmath89 to @xmath88 .
8 it remains to show that the domain , obtained from @xmath153 by shortening the oval , belongs to @xmath83 .
the domain is obtained by connecting @xmath154 to @xmath155 by two arcs shown on the right of figure 4 , i.e. the domain is connected .
it has no basepoints inside , since @xmath154 and @xmath155 do not have them .
analogously , it can be written as @xmath156 with all @xmath82 .
finally , we show that our domain has no corners inside .
if @xmath157 is empty , it follows from the assumption that @xmath158 , i.e. have no corners inside .
if the intersection is not empty , then its boundary either contains no corners or at least two corners , one of them in @xmath159 ( see figure [ intersect ] ) .
the last is impossible since @xmath160 has no corners of @xmath28 and @xmath161 no corners of @xmath33 inside
. * note . *
the first statement of proposition [ bigon ] is a particular case of the gaussian elimination considered in e.g. ( * ? ? ?
* lemma 4.2 ) .
the complex @xmath162 is obtained from the complex with long ovals by applying lemma [ bigon ] several times , until the complex has @xmath38 as the set of generators .
this subsection aims to give a recursive definition of the multiplicity @xmath163 of @xmath149 in the differential @xmath164 for all @xmath165 . in general
, it will depend on the order in which the ovals were shortened .
let us fix this order .
if @xmath166 is a polygon we count it with multiplicity one .
for any other domain in @xmath83 we prolongate the last oval that was shortened to obtain this domain , and show whether in the resulting complex @xmath167 , one can find @xmath28 and @xmath33 with @xmath168 , @xmath146 and @xmath169 as in the proof of lemma [ bigon ] .
if there is an odd number of such @xmath28 and @xmath33 , then @xmath170 , otherwise the multiplicity is zero . to determine @xmath171 and @xmath172 in the new complex , we prolongate the next oval , and continue to do so until the domains in question are polygons which always count .
in fact , the algorithm terminates already when the domains in question are strongly indecomposable without bad components ( as defined in the next section ) since they all count for the differential ( cf .
theorem 2.1 below ) . it remains to compute the homology of our complex . by lemma [ bigon ] , the complex @xmath40 is homotopy equivalent to the complex with long ovals .
the homotopy preserves both gradings , since @xmath61 count only maslov index one domains without basepoints inside .
the complex @xmath62 coincides with the complex computing the hat version of link floer homology from the genus zero heegaard splitting of @xmath24 with extra basepoints ( see @xcite for summary ) .
the relative maslov and alexander of these two complexes are also the same .
moreover , the absolute alexander and maslov gradings in our complex are fixed in such a way , that @xmath173 this can be shown by comparing our and mos complexes ( see ( * ? ? ?
* theorem 4.2 ) for more details ) . hence proposition 2.4 in @xcite
computes the homology of our combinatorial complex .
to run the algorithm defining @xmath61 we have to decide at each step which domains count and which do not . in this section
we simplify the algorithm by selecting a large class of domains that always count for the differentials obtained from @xmath174 by applying lemma [ bigon ] several times .
let us start with some definitions .
let @xmath175 be the euler measure of a surface @xmath176 , which for any surface @xmath176 with @xmath177 acute right
angled corners , @xmath178 obtuse ones , and euler characteristic @xmath179 is equal to @xmath180 .
moreover , the euler measure is additive under disjoint union and gluing along boundaries . in ( * ? ? ?
* section 4 ) , lipshitz gave a formula computing the maslov index @xmath181 of @xmath79 as follows .
[ mas - domain ] m(d_,)=e(d _ , ) + n_+n _ , where @xmath182 .
the number @xmath183 is the local multiplicity of the domain at the corner @xmath88 , e.g. @xmath184 for an isolated corner , @xmath185 for an acute ( or @xmath186angled ) corner , @xmath187 for a straight ( or @xmath188angled ) corner or @xmath189 for an obtuse ( or @xmath190angled ) one . for a composition of two domains @xmath191
, we have @xmath192 .
a path in a domain starting at an obtuse or straight corner and following a horizontal or vertical oval until the boundary of the domain will be called a _
cut_. there are two cuts at any obtuse corner and one at any straight corner . a domain @xmath22 is called _ decomposable _ if it is a composition of maslov index zero and one domains ; any other maslov index 1 domain is called indecomposable .
a domain is called _ strongly indecomposable _ if the following conditions are satisfied : * it is indecomposable ; * no prolongations of ovals inside this domain destroy its indecomposability ; * the cuts do not intersect inside the domain .
an example of an indecomposable , but not strongly indecomposable domain is shown in figure [ notstrong ] .
8 the oval destroying the indecomposability is shown by the dashed line without any number .
[ notstrong ] in what follows any domain is assumed to belong to @xmath83 and to have maslov index one , i.e. our domains have no corners with negative multiplicities or with multiplicities bigger than @xmath193 ( since domains from @xmath83 have no corners inside ) .
we say that a cut touches a boundary component @xmath53 at an oval @xmath54 if either the end point of this cut belongs to @xmath54 or the cut leaves @xmath53 along @xmath54 .
we define the distance between two cuts touching @xmath53 to be odd , if one of them touches @xmath53 at a vertical oval and another one at a horizontal oval ; otherwise the distance is even .
a boundary component is called _ special _ if it is an oval with a common corner of two generators ( see figure [ domain ] right for an example ) , otherwise the component is non special .
let us call an inner boundary component _ bad _ , if it does not have obtuse corners .
[ count ] any strongly indecomposable domain without bad components counts for the differential . the proof is given in section 2.4 after the detailed analysis of the structure of indecomposable domains . * remark .
* in the long oval complex , @xmath174 coincides with the differential of link floer homology defined by counting of pseudo holomorphic discs .
the differential @xmath61 at least for some order of shortening can not be realized by such count .
this is because , according to the gromov compactness theorem the count of indecomposable domains is independent of the complex structure , i.e. each of them either counts or not for any complex structure . in our setting , e.g. for the domain in figure [ notstrong ] one can always choose an order of shortening in such a way that our count differs from the analytic one . indeed ,
if this domain is obtained by first shortening the dotted part of the oval labeled by 1 and then the dashed part , the domain does not count .
however , if we get it by shortening the part 2 , and then the dashed one , it counts . here we provide some technical results needed for the proof of theorem [ count ] .
a boundary component @xmath194 is called @xmath89connected with another component @xmath195 if for any point @xmath196 and @xmath197 disjoint from the corners , there exists a unique path without self intersections starting at @xmath89 and ending at @xmath198 , such that @xmath199 the path goes along cuts or @xmath200 , where the arcs of horizontal and vertical ovals alternate along the path ( an arc can consist of the union of a cut and some part of @xmath200 , as long as they are on the same oval ) ; @xmath201 the corners of the path ( i.e. intersection points of horizontal and vertical segments ) come alternatively from @xmath28 and @xmath202 , where @xmath202 contains @xmath33 and intersection points of cuts with @xmath200 ; @xmath203 the intersection of the path with a boundary component is neither a point nor the whole component .
@xmath204 the first corner belongs to @xmath28 ; as an example consider the domain shown in figure [ degen ] .
let us denote the left inner boundary component by @xmath205 and the right one by @xmath206 .
for any choice of a point @xmath89 on a vertical oval of @xmath205 ( disjoint from the corners ) , there is no path @xmath89connecting @xmath205 with @xmath206 .
on the other hand , if we choose @xmath89 on the horizontal oval of @xmath205 there are two such paths .
hence , @xmath205 and @xmath206 are not @xmath89connected .
8 [ new ] in a strongly indecomposable domain without bad components , any two boundary components are @xmath89connected .
let @xmath207 be the total number of boundary components in our domain .
the proof is by induction on @xmath207 .
assume first that our domain has no special components .
further , all corners of our domain have positive multiplicities not bigger than @xmath193 and our domain has maslov index one . since we have no bad components ,
implies that every inner boundary component has exactly one obtuse corner .
suppose @xmath208 .
if one of the cuts from the obtuse corner connects the inner boundary component with itself , the domain is decomposable ( see figure [ 02a ] @xmath209 ) . if it is not the case , then an easy check verifies the claim ( compare figure [ 02a ] @xmath210 ) . 8 assume the claim holds for @xmath211 .
suppose @xmath212 , and our domain is indecomposable .
let us denote by @xmath53 the @xmath19th component .
let us first assume that there are no cuts ending at @xmath53 . in this case the two cuts from the obtuse corner @xmath89connect @xmath53 with some other components which are all @xmath89connected by induction .
if there is a component connected with @xmath53 by two cuts , then it is @xmath89connected with the outside exactly in the case when @xmath53 has this property . to check this , it is sufficient to find a required path for two choices of @xmath89 ( before and after one corner ) on this component
. an example is shown in figure [ 0nc ] .
hence , when all cuts ending at @xmath53 come from components connected with @xmath53 by two cuts
( as in figure [ 0nc ] ) , then @xmath53 is @xmath89connected to the outside by the previous argument . 8 in the case , when @xmath213 is connected with @xmath53 by just one cut or @xmath53 and @xmath122 exchange their cuts , all components are again @xmath89connected to each other except when the following happens .
the path described in definition 2.2 after leaving @xmath53 ( along one of the cuts ) comes back to @xmath53 without visiting all other components .
since this path leaves and enter any component along cuts at odd distance ( compare figure [ 0na ] ) , and has no self intersections , it can be used to decompose the domain , which contradicts the assumption .
8 it remains to consider the case where @xmath79 has straight corners , or special components , i.e. ovals with a common corner of two generators .
we proceed by induction on the number of special components .
assume that we have one special component .
then the cut from the straight corner @xmath89connects this component with any other one ( otherwise not connected with the special one ) by the previous argument .
the obvious induction completes the proof .
the proof is again by induction on the number of boundary components @xmath207 in the domain . if @xmath214 , it is easy to see by prolonging ovals that any immersed polygon counts .
assume that for @xmath211 , the claim holds .
suppose our complex has a strongly indecomposable domain @xmath215 without bad components and @xmath212 .
examples with @xmath208 are drawn in figure [ domain ] .
8 let us stretch one oval in @xmath22 connecting two boundary components .
the result is a domain @xmath216 .
let @xmath88 and @xmath89 be the corners of the bigon , obtained after stretching .
the stretched oval connects @xmath89 to some boundary component , say @xmath53 . by lemma
[ new ] , @xmath89 can also be connected with @xmath53 by a unique path inside @xmath22 .
this path is not affected by the prolongation , since otherwise the domain would not be strongly indecomposable ( compare figure [ notstrong ] ) .
hence , @xmath216 can be represented as a union of two domains connecting some generators and having less boundary components .
the unique path connecting @xmath89 with @xmath53 leaves any boundary component along a cut .
moreover , the path has no self intersections .
therefore , @xmath216 is a union of two strongly indecomposable maslov index one domains without bad components and with positive multiplicities at the corners not bigger than @xmath193 .
these both domains count for the differential by the induction hypothesis .
we conclude that the domain @xmath22 also counts for the differential .
in this section we show how @xmath217 of small knots can be computed by hand and discuss the computer program written by droz .
figure [ newdiag ] shows a rectangular diagram for @xmath74 knot of complexity @xmath218 obtained from the original diagram in figure [ diag ] by cyclic permutations ( compare @xcite ) .
an advantage of this diagram is that there are no regions counted for the differential .
the alexander grading of a generator is given by the formula @xmath219 .
the maximal alexander grading is equal to one .
there are two generators in this grading shown by colored dots in figure [ newdiag ] .
both of them have maslov grading @xmath220 .
the homology of our complex is @xmath221 .
hence in alexander grading zero , we have 12 additional generators coming from the multiplication with @xmath222 .
note that our complex has 15 generators in alexander grading zero .
indeed , 12 of them can be obtained by moving one point of a generator in alexander grading one to the other side of the oval . in three cases , depicted by white dots
there are two possibilities to move a point .
this gives 3 additional generators .
note that these moves drop maslov index by one . to compute @xmath217 in the negative alexander gradings we use the symmetry ( [ sym ] ) .
finally , we derive that @xmath223 has rank two in the alexander
maslov bigrading @xmath224 , rank three in @xmath225 , and rank two in the bigrading @xmath226 . to compare , the alexander polynomial is @xmath227 . the knot @xmath74 is not fibered and its seifert genus is one . in @xcite ,
droz extend our construction over @xmath15 and wrote a computer program calculating the homology of the resulting complex @xcite .
the program is installed on the bar
natan s knot atlas . as a byproduct ,
his program generates rectangular diagrams of knots and links and allows to change them by cromwell
dynnikov moves .
the program can be used to determine seifert genus and fiberedness of knots until 16 crossings . according to droz s computations ,
the number of generators in our complex is significantly smaller than that in the mos complex .
moreover , for small knots , almost all domains suitable for the differential are embedded polygons , so they always count for the differential . for example , for knots admitting rectangular diagrams of complexity 10 , the number of generators in the mos complex is @xmath228 .
our complex has on average about @xmath229 generators among them about @xmath230 in the positive alexander gradings .
the knot @xmath231 admits a rectangular diagram of complexity 12 , where @xmath232 .
our complex has @xmath233 generators with @xmath234 of them in the positive alexander gradings . furthermore , droz s program produced examples of domains counted more than once over @xmath15 .
we do not know similar examples in the analytic setting . in figure 15 one such domain is shown .
this domain has a degenerate system of cuts and its count depends on the order of shortening of ovals .
one specific order gives multiplicity 2 for this domain .
d. bar - natan .
fast khovanov homology computations .
j. knot theory ramifications 16 ( 2007 ) 243255 j.m .
effective computation of knot floer homology .
acta math .
33 ( 2008 ) 471491 j.m .
python program at http://www.math.unizh.ch/assistenten/jdroz or at http://katlas.org/wiki/heegaard@xmath235floer@xmath235knot@xmath235homology . | we define a new combinatorial complex computing the hat version of link floer homology over @xmath0 , which turns out to be significantly smaller than the manolescu ozsvth
sarkar one . |
rotating spirals are a pervasive feature of two - dimensional excitable media , such as the belousov - zhabotinsky reaction @xcite .
figure [ fig : intro](a ) illustrates a spiral wave from a standard model of excitable media discussed below .
the wave character of the system is evident . as the spiral rotates ,
a periodic train of excitation is generated which propagates outward from the center , or core , of the spiral .
much of the historical study of excitable media has focused on the wave character of the problem , as illustrated by efforts to determine the selection of the spiral shape and rotation frequency @xcite .
however , it is now understood that these spiral waves also have particle - like properties .
this was first brought to the forefront by biktasheva and biktashev @xcite and has been developed in more recent years @xcite .
one of the more striking illustrations of a particle - like property is _ resonant drift _
@xcite , shown in fig .
[ fig : intro](b)-(d ) . resonant drift can occur spontaneously through instability , or due to spatial inhomogeneity , or as here , by means of resonant parametric forcing ( periodically varying the medium parameters in resonance with the spiral rotation frequency ) .
as is seen , the core of the spiral drifts along a straight line .
the speed is dictated by the forcing amplitude while the direction is set by the phase of the forcing , or equivalently the initial spiral orientation . and
@xmath0 , are indicated .
( d ) path of a drifting spiral in a square box .
the underlying spiral wave at one instant in time is shown faintly . in all cases
the plotted fields are the excitation variable , @xmath1 , of the reaction diffusion model .
details are given later in the text.,width=312 ] the trajectories of drifting spirals are unaffected by the domain boundaries ( or other spirals should they be present ) except on close approach , where often the result is a reflection of the drifting core @xcite , as illustrated in fig .
[ fig : intro](c ) . reflections are not specular the reflected angle @xmath0 is not in general equal to the incident angle @xmath2 .
when placed in a square box , the drift trajectory typically will ricochet off each boundary in such a way to eventually be attracted to a unique square path where @xmath3 , as shown in figure [ fig : intro](d ) .
( this is the more common case , but others are considered herein . ) the primary goal of this paper is firstly to determine accurately , through numerical simulations , the relationship between reflected and incident angles for some representative cases of spiral waves in excitable media , and secondly to explore the qualitative features of reflections in excitable media , particularly multiple reflections in square domains . while the numerical and theoretical study of reflecting trajectories was undertaken by biktashev and holden many years ago @xcite , much more extensive results are now possible and desirable , especially since phenomena strikingly similar to that seen in figs .
[ fig : intro](c ) and ( d ) have been observed in other macroscopic systems with both wave - like and particle - like properties @xcite .
our study is based on the standard barkley model describing a generic excitable medium @xcite . in the simplest form
the model is given by the reaction - diffusion equations @xmath4 where @xmath5 is the excitation field ( plotted in fig .
[ fig : intro ] ) and @xmath6 is the recovery field ; @xmath7 , @xmath8 , and @xmath9 are parameters . the parameters @xmath7 and @xmath8 collectively control the threshold for and duration of excitation while the parameter @xmath9 controls the excitability of the medium by setting the fast timescale of excitation relative to the timescale of recovery .
we consider two parameter regimes known commonly as the small - core and large - core regimes .
the small - core case is shown in fig .
[ fig : intro ] .
as the name implies , the core region of the spiral , where the medium remains unexcited over one rotation period , is small .
this is the more generic case for the barkley model and similar models and occupies a relatively large region of parameter space in which waves rotate periodically .
small - core spirals are found in the lower right part of the standard two - parameter phase diagram for the barkley model ( see figure 4 of ref . ) .
large - core spirals rotate around relatively large regions [ see fig . [
fig : large_core_intro](a ) discussed below ] .
such spirals occur in a narrow region of parameter space @xcite near the boundary for propagation failure .
the core size diverges to infinity near propagation failure .
parametric forcing is introduced through periodic variation in the excitability .
specifically we vary @xmath9 according to @xmath10 where @xmath11 and @xmath12 are the forcing amplitude and frequency .
the phase @xmath13 is used to control the direction of resonant drift . the forcing frequency
producing resonant drift will be close to the natural , unforced , spiral frequency .
however , due to nonlinearity , there is a change in the spiral rotation frequency under forcing and so @xmath12 must be adjusted with @xmath11 to produce resonant drift along a straight line .
we have studied reflections in two situations .
the first is reflection from a no - flux boundary .
this type of boundary condition corresponds to the wall of a container containing the medium .
we set the reflection boundary to be at @xmath14 and impose a homogeneous neumann boundary condition there : @xmath15 since there is no diffusion of the slow variable , no boundary condition is required on @xmath16 .
the medium does not exist for @xmath17 .
the second situation we have studied is reflections from a step change in excitability across a line _ within _ the medium .
we locate step change on the line @xmath14 .
we vary the threshold for excitation across this line by having the parameter @xmath8 vary according to @xmath18 unlike for the no - flux boundary , in this case waves may cross the line @xmath14 and so there is no boundary to wave propagation
. nevertheless , drifting spirals may reflect from this step change in the medium and we refer to this a _
step boundary_. the numerical methods for solving the reaction - diffusion equations are standard and are covered elsewhere @xcite .
some relevant computational details particular to this study of spiral reflections are as follows .
a converged spiral for the unforced system is used as the initial condition .
simulations are started with parametric forcing and the spiral drifts in a particular direction dictated by the phase @xmath13 in eq . .
the position of the spiral tip is sampled once per forcing period and from this the direction of drift , i.e. the incident angle @xmath2 , is determined by a least - squares fit over an appropriate range of drift ( after the initial spiral has equilibrated to a state of constant drift , both in speed and direction , but before the spiral core encounters a boundary ) . likewise , from a fit to the sampled tip path after the interaction with the boundary , we determine the reflected angle @xmath0 . by
varying @xmath13 we are able to scan over incident angles .
the simulations are carried out in a large rectangular domain with no - flux boundary conditions on all sides . for reflections from a neumann boundary , we simply direct waves to the computational domain boundary corresponding to @xmath14 .
we also study reflections more globally from all sides of a square domain with neumann boundary conditions , such as in fig .
[ fig : intro](d ) . in the study of reflections from the step boundary
, the computational domain extends past the step change in parameter .
we have run cases with the left computational boundary both at @xmath19 and @xmath20 and these are sufficiently far from @xmath14 that trajectory reflection is not affected by the computational domain boundary .
the dimensions of the rectangular computational domain are varied depending on the angle of incidence .
for @xmath21 we require a long domain in the @xmath22-direction , whereas for @xmath23 a much smaller domain may be used . in all cases
we use a grid spacing of @xmath24 .
the time step is varied to evenly divide the forcing period , but @xmath25 is typical . except where stated otherwise , the model parameters for the small - core case are : @xmath26 , and @xmath27 . for the large - core case they are @xmath28 , and @xmath27 . for the step boundary @xmath29 and @xmath30 .
different values of the forcing amplitude and period , @xmath11 and @xmath12 , are considered .
given the desire to measure incident and reflected angles precisely , we have required drift be along straight lines to high precision and in turn this has required high accuracy in the imposed forcing amplitude and period . in the appendix we report the exact values for the forcing parameters used in the quantitative incidence - reflection studies .
before presenting results from our study of reflections , it is important to be precise about the meaning of incident and reflected angles . as is standard ,
angles are measured with respect to the boundary normal .
this is illustrated in fig .
[ fig : intro](c ) .
what needs to be stressed here is that spirals have a chirality right or left handedness and this implies that we need to work with angles potentially in the range @xmath31 $ ] , rather than simply @xmath32 $ ] . specifically , we consider clockwise rotating spirals and define @xmath2 to be positive in the clockwise direction from the normal .
we define @xmath0 to be positive in the counterclockwise direction from the normal .
both @xmath2 and @xmath0 are positive in fig .
[ fig : intro](c ) and for specular reflections @xmath33 .
we begin with the small - core case already shown in fig .
[ fig : intro ] . figures [ fig : neumannbc ] and [ fig : threshbc ] illustrate the typical behavior we find in reflections from both types of boundaries . in both figures
the upper plot shows measured reflected angle @xmath0 as a function of incident angle @xmath2 over the full range of incident angles .
the lower plots show representative trajectories for specific incident angles indicated . here and throughout , the no - flux nature of the neumann boundary is indicated with shading ( @xmath14 is the at the rightmost edge of the shading ) , while the step in excitability at a step boundary is indicated with sharp lines .
all parameters are the same for the two cases ; they differ only in the type of boundary that trajectories reflect from . versus incident angle @xmath2 .
( b ) representative tip trajectories showing reflections at the incident angles marked with crosses in ( a ) .
the reflected angle is nearly constant for the full range of incident angles .
the forcing amplitude is @xmath34 .
, width=288 ] versus incident angle @xmath2 .
( b ) representative tip trajectories showing reflections at the incident angles marked with crosses in ( a ) .
the reflected angle is nearly constant for the full range of incident angles .
the forcing amplitude is @xmath34 .
, width=288 ] the reflections are far from specular .
this is particularly striking for @xmath35 where the incoming and outgoing trajectories lie on the same side of the normal .
the reflected angle is nearly constant , independent of the incident angle , except for incident angles close to @xmath36 .
there is a slight variation in the reflected angle , seen as undulation in the upper plots , but the amplitude of the variation is small .
one can also observe in the lower plots that the point of closest approach is also essentially independent of incident angle , except close to @xmath37 where the distance grows .
spiral trajectories come much closer to the step boundary than to the neumann boundary .
it is worth emphasizing that there is no effect of forcing phase in the results presented in figs .
[ fig : neumannbc ] and [ fig : threshbc ] . as the incident angle is scanned ,
the phase of the spiral as it comes into interaction with the boundaries will be different for different incident angles .
while this could have an effect on the reflected angle , we have verified that there is no such effect for the small - core cases we have studied , except at large forcing amplitudes near where spirals annihilate at the boundary ( discussed later ) . , including faint visualization of the @xmath1-field at a particular time instance , while lower plots show @xmath38 .
the reflected angle is nearly constant independently of incident angle , parameter values , and boundary type .
model parameters span a substantial range of the non - meandering small - core region : in ( a ) and ( b ) @xmath39 ; in ( c ) and ( d ) @xmath40 ; in ( e)-(h ) @xmath41 . cases ( g ) and ( h ) are step boundary , the others are all neumann boundaries . @xmath34 throughout .
, width=312 ] while we have not conducted detailed studies at other parameter values , we have explored the small - core region of parameter space .
figure [ fig : params ] shows representative results at distant points within the small - core region .
the figure indicates not only a qualitative robustness , but also a quantitative insensitivity to model parameter values throughout the small - core region . in each case
the upper plot shows @xmath42 while the lower plot shows @xmath38 .
the reflected angle varies by only a few degrees throughout all cases shown in the figure .
case ( a)-(b ) is close to the meander boundary while ( c)-(d ) is far from the meander boundary and corresponds to a very small core .
cases ( e)-(h ) are relatively large values of parameters @xmath7 and @xmath8 , both with neumann and step boundary conditions . in the step boundary case
, there is also the effect of @xmath43 to consider . across a number of representative incident angles
, we observed that as @xmath43 is incremented from @xmath44 up to @xmath45 , the closest approaches of the spiral tips occur further from the boundary .
we also find a slight reduction in the angle of reflection .
decrementing @xmath43 has the opposite effect .
however , if @xmath43 is too small then the repulsive effect at the boundary will be too small and spiral cores will cross the boundary .
we have examined the effect of forcing amplitude @xmath11 .
figures [ fig : neumvara ] and [ fig : threshvara ] show reflected angle as a function of incident angle for various values @xmath11 as indicated .
there is a decrease in the reflected angle with increasing forcing amplitude , or equivalently increasing drift speed .
generally there is also an increase in the oscillations seen in the dependence of reflected angle on incident angle .
the solid curves are from the biktashev - holden theory discussed in sec .
[ sec : bh ] . indicated .
solid curves are from biktashev - holden theory discussed in sec .
[ sec : bh ] .
, width=312 ] indicated .
solid curves are from biktashev - holden theory discussed in sec .
[ sec : bh ] .
, width=312 ] for sufficiently large forcing amplitudes small - core spirals may be annihilated as they drift into neumann boundaries . in such cases
no reflection occurs .
we have not investigated this in detail as it is outside the main focus of our study on reflections .
nevertheless , we have examined the effect of increasing the forcing amplitude through the point of annihilation for the case of a fixed incident angle @xmath46 .
the results are summarized in figure [ fig : annihil ] .
the reflected angle reaches a minimum for @xmath47 , and thereafter increases slightly , but does not vary by more than @xmath48 up to the amplitude where annihilation occurs , @xmath49 , as indicated in figure [ fig : annihil](a ) .
the forcing amplitude at which annihilation first occurs is rather large in that it corresponds to displacing the spiral considerably more than one unforced core diameter per forcing period .
figure [ fig : annihil](b ) shows tip trajectories on either side of the amplitude where annihilation occurs , while figure [ fig : annihil](c ) shows annihilation at much larger forcing amplitude .
we note that the exact amplitude at which annihilation first occurs depends slightly on the rotational phase of the spiral as it approaches the boundary .
( annihilation first occurs in the range @xmath50 depending on phase . )
likewise , the spiral phase can affect the reflected angle by nearly @xmath51 for @xmath52 .
the influence of phase is nevertheless small for the small - core spirals .
it is , however , more pronounced in the large - core case which we shall now discuss . up to the point of annihilation at a neumann boundary for small - core spirals .
the incident angle is fixed at @xmath46 .
( b ) tip trajectories a little below ( @xmath53 ) and a little above ( @xmath54 ) the forcing amplitude resulting in annihilation of the spiral at the boundary .
( c ) tip trajectory at @xmath55 showing annihilation at very large forcing amplitude .
, width=312 ] we now turn to the case where unforced spirals rotate around a relatively large core region of unexcited medium .
this case is illustrated in fig .
[ fig : large_core_intro](a ) where a rotating spiral wave and corresponding tip trajectory are shown in a region of space the same size as in fig .
[ fig : intro ] .
the larger tip orbit and unexcited core , as well as the longer spiral wavelength , in comparison with those of fig .
[ fig : intro](a ) are clearly evident .
while such spirals occupy a relatively narrow region of parameter space , they are nevertheless of some interest because asymptotic treatments have some success in this region @xcite and because this is nearly the same region of parameter space where wave - segments studies are performed @xcite . figure [ fig : large_core_intro](b ) shows a typical case of non - specular reflection for a large - core spiral compared with a small - core spiral forced at the same amplitude .
while many features are the same for the two cases , large - core spirals are found often to reflect at smaller @xmath56 and moreover , they can exhibit different qualitative phenomena .
space units [ same size as fig .
[ fig : intro](a ) ] .
( b ) resonant forcing and reflection for a large - core spiral shown in comparison to that of a small - core spiral .
the forcing amplitude is @xmath57 in both cases .
, width=312 ] figures [ fig : large_core ] and [ fig : large_core_paths ] summarize our findings for large - core spirals . reflected angle as a function of incident angle for three forcing amplitudes is shown in fig .
[ fig : large_core ] .
one sees the overall feature , as with the small - core case , that reflected angle is approximately constant over a large range of incident angles .
this is particularly true of low - amplitude forcing , @xmath58 . however , there are also considerable differences with the small - core case .
spirals are frequently annihilated at the boundary , [ fig .
[ fig : large_core_paths](d ) ] , over the range of incident angles indicated . for @xmath57 and @xmath59
trajectories glance from the boundary [ fig .
[ fig : large_core_paths](a ) ] . for @xmath60 and @xmath59 trajectories
become bound to the boundary .
[ see text and fig .
[ fig : large_core_paths](b ) . ]
wiggles are the effect of incident phase .
, width=312 ] at different forcing amplitudes .
( a ) with @xmath55 the trajectory glances from the boundary and moves off nearly parallel to it ( @xmath61 ) .
( b ) with lower amplitude @xmath62 , the trajectory becomes bound to the boundary .
( c ) with yet lower amplitude @xmath63 , the trajectory hugs the boundary for a while then leaves abruptly at an oblique angle ( @xmath64 ) . in
( d ) the incoming spiral with large forcing , @xmath55 , is annihilated at the boundary . in ( e )
the effect of phase is seen with two approaching trajectories shifted by half a core diameter .
otherwise the conditions are identical , @xmath57 .
the resulting reflected angles differ slightly .
, width=216 ] for large - core spirals the reflected angle increases with forcing amplitude .
this is opposite to what is found for small - core spirals in figs .
[ fig : neumvara ] and [ fig : threshvara ] .
moreover , the reflected angles are noticeably smaller than for the small - core case , as was already observed in fig . [
fig : large_core_intro](b ) .
we now focus in more detail on what happens in various circumstances .
the left portion of fig . [ fig : large_core ] indicates the different dynamics we observe , depending on forcing amplitude , at large negative incident angles ( @xmath59 ) , and figs .
[ fig : large_core_paths](a)-[fig : large_core_paths](c ) show representative trajectories with @xmath65 . at @xmath57 trajectories
glance off the boundary .
that is , they remain close for short while before moving off with a well defined large negative reflected angle .
the reflected angle is nearly constant at @xmath61 for incident angles @xmath59 . at @xmath60 , @xmath59 ,
trajectories become bound to the boundary and move parallel to it indefinitely . in fig .
[ fig : large_core_paths](c ) , with @xmath58 , one observes the trajectory moving along the boundary for a distance before abruptly leaving the boundary at a well - defined , relatively small positive reflected angle .
this behavior is not restricted to @xmath59 and is observed until @xmath66 .
in fact , this type of reflection is also observed for the other two forcing amplitudes studied for @xmath67 in a range above @xmath68 . for @xmath60
this occurs until @xmath67 is approximately @xmath69 , while for @xmath57 this is seen only until @xmath67 is about @xmath70 . at the higher
forcing amplitudes , as indicated for the case @xmath57 in fig .
[ fig : large_core ] , large - core spirals are frequently annihilated when they come into contact with the boundary .
figure [ fig : large_core_paths](d ) shows a typical example . whether or not a spiral is annihilated depends very much on the spiral phase on close approach to the boundary .
the points shown in fig . [ fig : large_core ] with @xmath57 are those where the trajectory reflected ; the absence of points indicates annihilation .
however , these results are for spirals all initiated a certain distance from the boundary .
changing that distance would affect the spiral phase at close approach and hence a different set of points would be obtained . nevertheless , the marked range of frequent annihilation is indicative of what occurs at this forcing amplitude . finally , we address the wiggles in the reflected angle curves in fig .
[ fig : large_core ] , most evident at large forcing amplitudes .
these wiggles are also due to the fact that the phase of spirals on close approach varies with incident angle .
figure [ fig : large_core_paths](e ) illustrates how the reflected angle depends on phase by showing two trajectories shifted by half a core diameter .
this shifts the spiral phase upon approach to the boundary and results in slightly different reflected angles . rather than eliminating these wiggles by averaging over various initial spiral distances
, we leave them in as an indication of the variability due to this effect . in general
, reflections of large - core spirals are much more sensitive to phase than reflections of small - core spirals , and one should understand that the data in fig .
[ fig : large_core ] will vary slightly if similar cases are run with spirals initiated at different distances from the boundary .
while we have not studied the step boundary in detail for large - core spirals , we have carried out a cursory investigation for such a boundary with @xmath71 . with the exception that there is no annihilation at the step boundary , we observe qualitatively similar behavior to that just presented for the neumann case .
most notably we find both glancing and bound trajectories .
many years ago biktashev and holden @xcite carried out a study very similar in spirit to that presented here .
moreover , they understood that a primary cause for the reflection from boundaries was the small changes in spiral rotation frequency occurring as spiral cores came into interaction with boundaries .
based on this they proposed an appealing simple model to describe spiral reflections .
the model is based on the assumption that both the instantaneous drift speed normal to the boundary and spiral rotation frequency are affected by interactions with a boundary , with the interactions decreasing exponentially with distance from the boundary .
while the actual interactions between spiral and boundaries are now known to be more complex ( see below ) , it is worth investigating what these simple assumptions give .
the beauty of the simple model is that it can be solved to obtain a relationship between reflected and incident angles , depending on only a single combination of phenomenological parameters .
( they called this combination @xmath72 , but we shall call it @xmath73
. they also used different definitions for incident and reflected angles . )
the model naturally predicts large ranges of approximately constant reflected angle depending on the value of @xmath73 .
what is nice is that while fitting the individual phenomenological parameters in their model would be difficult , it is also unnecessary . the value of @xmath73 can be selected to match the plateau value of @xmath56 observed in numerical simulations .
then the entire relationship between @xmath56 and @xmath67 from the theory is uniquely determined .
curves from the biktashev - holden theory are included in figs .
[ fig : neumvara ] and [ fig : threshvara ] .
while there are obvious limitations to the theory , it is nevertheless interesting to see that some of the features are reproduced just from simple considerations .
the theory would be expected to work best where the drift speed is small : low amplitude forcing . for
the large - core spirals the theory does not apply and so the corresponding curves are not shown in fig . [
fig : large_core ] .
we do not intend here to propose a more accurate theory for the reflection of drift trajectories .
it is worth emphasizing , however , that in recent years much has been understood about the interaction between spiral waves and various symmetry breaking inhomogeneities @xcite .
the key to this understanding has been response functions , which are adjoint fields corresponding to symmetry modes for spirals in homogeneous media @xcite . in principle
, one could now obtain rather accurate description of reflections using this approach , at least for cases in which the boundary could be treated as a weak perturbation of a homogeneous medium .
this is beyond the scope of the present study and we leave this for future research . as noted in the introduction , fig .
[ fig : intro](d ) , when placed within a square domain the trajectory of a drifting spiral will typically approach a square , reflecting from each domain boundary such that @xmath74 .
the reasons for this are simple ( see for example prati , _
et al._@xcite ) , but a brief analysis is useful , particularly for understanding when square orbits become unstable .
is the location , relative to the length of a side , of the @xmath75 reflection .
, width=240 ] figure [ fig : sketch ] shows the geometry of a consecutive pair of reflections in the case where the reflected angle is larger than @xmath76 . in this case the path will necessarily strike consecutive sides of the domain .
consider first the path in terms of angles and let @xmath77 and @xmath78 denote , respectively , the @xmath75 incident and reflected angles , starting from the initial reflection @xmath79 .
trivially , the geometry of the square domain dictates that @xmath80 .
then , if the trajectory approaches an attracting path with constant angles , @xmath81 , @xmath82 , it must be that this path satisfies @xmath83 .
that is , it must be a square or a rectangle . denoting the relationship between incident and reflected angle by @xmath84 ,
then a necessary condition for the square path to be attracting is that @xmath85 .
for the cases we have studied this is true since @xmath86 .
turning now to the points at which the path strikes the edge of the domain , we let @xmath87 denote the position of the @xmath75 reflection along a given side , relative to the length of a side .
one can easily see from the geometry that @xmath88 .
now , since @xmath89 , the fixed point @xmath90 is given by @xmath91 , or @xmath92 .
this corresponds to a square trajectory .
for example , from fig .
[ fig : neumvara ] with a forcing amplitude @xmath34 one can see that @xmath93 will necessarily be about @xmath94 , giving @xmath95 .
these are the values seen in the simulation in fig .
[ fig : intro](d ) . a necessary condition for this fixed point to be stable
is @xmath96 . for small - core spirals @xmath97 , so @xmath98 , and hence their square paths are stable . .
the reflected angle is considerably smaller than @xmath76 and the resulting trajectory bounces between opposite sides of the domain .
the spiral is shown faintly at one time instance .
( b ) @xmath99 .
the square trajectory is unstable .
for the first circuit around the nearly square path the full tip trajectory is plotted .
subsequently , for clarity only , the tip path sampled once per forcing period is shown .
the trajectory collapses towards the diagonal .
the final portion of the trajectory before the spiral approaches the corners is shown in bold .
the spiral undergoes a complicated reflection from the corner ( gray , dotted ) .
, width=312 ] while square trajectories occur for small - core spirals , for large - core spirals other trajectories are possible .
examples are shown in fig .
[ fig : large_core_other ] .
these occur when the reflected angle is smaller than @xmath76 .
( it is possible that @xmath100 might occur for small - core spirals in some regimes , although we have not observed them . )
when @xmath100 it is not necessarily the case that trajectories will strike consecutive sides of a square box .
this is seen in fig .
[ fig : large_core_other](a ) where the spiral in reflects between opposite sides of the domain .
the reflections satisfy @xmath101 .
the more interesting case is when @xmath56 is only slightly less than @xmath76 as is seen in fig .
[ fig : large_core_other](b ) .
the square trajectory is unstable . while @xmath85 and the angles converge quickly to @xmath102 , the equation @xmath103 exhibits growing period - two oscillations for @xmath104 slightly larger than 1 .
period - two oscillations in @xmath87 with @xmath102 correspond to approximately rectangular trajectories that approach a diagonal .
this ultimately leads the spiral into a corner of the domain where it may reflect in complicated manner .
the analysis just presented should not be viewed as a model for spiral trajectories .
rather it just shows what global dynamics can be deduced simply from a measured relationship between incident and reflected angles .
essentially this same analysis appears as part of a study of cavity solitions @xcite which also undergo non - specular reflections from walls and hence exhibit square orbits similar to figure [ fig : intro](d ) .
our simple analysis should be contrasted with the situation for drops bouncing on the surface of an oscillating liquid , so called _
walkers_. here physical models of the liquid surface and drop bounces account for many varied features of the system @xcite .
the corresponding theory for spiral waves would be that based on response functions in which many details of drift trajectories can be predicted @xcite , although as of yet not reflections from boundaries .
memory effects are important for walkers because bouncing drops interact with surface waves generated many oscillations in the past , and models necessarily take this into account @xcite .
however , path memory is absent from spiral waves in excitable media and this constitutes a significant difference between the two systems .
we have reported some quantitative and some qualitative features of resonant - drift trajectories in excitable media .
the main message is that reflections are far from specular the reflected angle generally depends only weakly on the incident angle and typically is nearly constant over a substantial range of incident angles ( particularly negative incident angles ) .
biktashev - holden theory @xcite accounts for some of the observed features , but a more detailed theory based on response functions @xcite is needed .
we have seen that the behavior of large - core spirals is more varied than that for small - core ones . rather than simply reflecting from a boundary
, large - core spirals may sometimes become bound to , or glance from , or be annihilated at a boundary , even at moderate forcing amplitudes .
finally we have considered what can occur as spirals undergo multiple reflections within a square domain , and in particular have shown that while small - core spirals are observed to meet the conditions of stable square trajectories , large - core spirals may fail to meet these conditions and exhibit more interesting dynamics .
we motivated this study with a broader discussion of macroscopic systems with wave - particle duality .
a large number of analogues to quantum mechanical systems have been reported for walkers on the surface of a vibrated liquid layer @xcite . as far as we are aware , this is less the case for the propagating wave segments studied by showalter _
et al._@xcite or the drifting spirals in excitable media considered here .
( we examined briefly small - core drift trajectories through a single slit , but did not observe diffraction - like behavior . ) nevertheless , for the reflection problem , spiral trajectories , propagating wave segments , cavity solitons , and walkers all share the feature of non - specular reflections @xcite and as a result these systems can show similar dynamics when undergoing multiple reflections within a bounded region @xcite .
it will be of interest to make further quantitative comparisons between these different systems in the future and to explore theoretical basis of this behavior .
in the appendix we report the exact values for the forcing parameters used in the detailed quantitative incidence - reflection studies , since obtaining high - precision values for resonant drift can be time consuming .
the values stated in the body of the paper are reported only to two significant figures .
46ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1371/journal.pone.0024388 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop ( ) | spiral waves in excitable media possess both wave - like and particle - like properties . when resonantly forced ( forced at the spiral rotation frequency )
spiral cores travel along straight trajectories , but may reflect from medium boundaries . here , numerical simulations are used to study reflections from two types of boundaries .
the first is a no - flux boundary which waves can not cross , while the second is a step change in the medium excitability which waves do cross . both small - core and large - core spirals are investigated .
the predominant feature in all cases is that the reflected angle varies very little with incident angle for large ranges of incident angles .
comparisons are made to the theory of biktashev and holden .
large - core spirals exhibit other phenomena such as binding to boundaries .
the dynamics of multiple reflections is briefly considered .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ wave - particle duality is typically associated with quantum mechanical systems . however , in recent years it has been observed that some macroscopic systems commonly studied in the context of pattern formation also exhibit wave - particle duality .
two systems in particular have attracted considerable attention in this regard : drops bouncing on the surface of a vibrated liquid layer @xcite and waves in chemical media @xcite .
the second case is the focus of this paper .
we explore non - specular reflections associated with spiral waves in excitable media reflections not of the waves themselves , but of the particle - like trajectories tied to these waves . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
the co - existence of an active galactic nucleus ( agn ) and young stars in the central region of seyfert galaxies is a widely known phenomenon , ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , supporting the so - called agn - starburst connection ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the above studies have pointed out that the main difference between the stellar population ( sp ) of active and non - active galaxies is an excess of intermediate age stars in the former .
in addition , near - infrared ( nir )
sp studies have revealed that the continuum is also dominated by the contribution of intermediate - age stellar population components ( spcs , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the nir ,
another component is commonly detected in the nuclear spectra of seyfert galaxies : unresolved hot dust emission @xcite .
thus , the study of the contribution of the sps and other components to the circumnuclear continuum of active galaxies is a fundamental key in the understanding of the nature of their central engine .
very recently , the use of integral field spectroscopy with adaptive optics at the gemini north telescope has allowed us to derive the contribution of distinct spcs to the nir spectra of active galaxies .
in addition , we have mapped spcs spatial distributions and performed the first two - dimensional ( 2d ) sp synthesis in the nir of the nuclear region of an active galaxy ( mrk1066 , * ? ? ? * ) . a spatial correlation between the intermediate age spc and a partial ring of low stellar velocity dispersions ( @xmath2 ) was found , supporting the interpretation that the low-@xmath2 structures commonly observed in the inner few hundreds parsecs of seyfert galaxies are due to colder regions with more recent star formation than the underlying bulge @xcite . using a different method modelling the equivalent width , supernova rate and mass - to - light ratio @xcite have quantified the star formation history in the centre of 9 nearby seyfert galaxies using their stars code .
they found that the ages of the stars which contribute most to the nir continuum lie in the range 1030myr , pointing out that these ages should be considered only as characteristic " , as they have not performed a proper spectral synthesis , suggesting that there may be simultaneously two or more sps that are not coeval @xcite .
as part of an ongoing project aimed at mapping the 2d age distribution of the nir sp in the inner few hundred parsecs of seyfert galaxies , we present the nir sp synthesis of the seyfert 2 galaxy mrk1157 ( ngc591 ) , an early type barred spiral galaxy located at a distance @xmath3mpc ( from nasa / ipac extragalactic database , ned ) , for which 1 corresponds to 296pc at the galaxy .
the near - ir gas excitation and kinematics as well as the stellar kinematics of the inner 450 pc of mrk1157 are discussed in riffel & storchi - bergmann ( _ in prep . _ ) using the same data used here .
this paper is structured as follows . in sec .
[ data ] we describe the observations , data reduction and cleaning procedures .
the spectral fitting procedures are discussed in sec .
[ synt ] . in sec .
[ results ] we present our results , which are discussed in sec . [ discussion ] .
the conclusions are presented in sec .
[ conclusions ] .
mrk1157 was observed with gemini nifs @xcite operating with the gemini north adaptive optics system altair in september / october 2009 under the programme gn-2009b - q-27 , following the standard object - sky - sky - object dither sequence , with off - source sky positions since the target is extended , and individual exposure times of 550s .
two sets of observations with six on - source individual exposures were obtained at different spectral ranges : the first at the j - band , centred at 1.25@xmath4 m and covering the spectral region from 1.14@xmath4 m to 1.36@xmath4 m , and the second at the k@xmath5-band , centred at 2.3@xmath4 m , covering the spectral range from 2.10@xmath6 m to 2.53@xmath6 m . at the j - band ,
the j_g5603 grating and zj_g0601 filter were used , resulting in a spectral resolution of @xmath7 , as obtained from the measurement of the full width at half maximum ( fwhm ) of arxe arc lamp lines .
the k@xmath5-band observations were obtained using the kl_g5607 grating and hk_g0603 filter and resulted in a spectral resolution of fwhm@xmath8 .
the data reduction was accomplished using tasks contained in the nifs package which is part of gemini iraf package , as well as generic iraf tasks .
the reduction procedure included trimming of the images , flat - fielding , sky subtraction , wavelength and s - distortion calibrations .
we have also removed the telluric bands and flux calibrated the frames by interpolating a black body function to the spectrum of the telluric standard star .
the angular resolution obtained from the fwhm of the spatial profile of the telluric standard star is 011@xmath9002 for the j - band and 012@xmath9002 for the k@xmath5-band , corresponding to [email protected] and [email protected] at the galaxy , respectively .
since the star has been observed for a shorter time than the galaxy , this value should be considered as a lower limit for the spatial resolution .
nevertheless , the performance of altair for larger integration times ( such those used for the galaxy ) in previous works by our group was found to be similar to the one reached in the observation of the star . in the case of ngc4151 ,
the spatial resolution of the galaxy is only 002 worse than those obtained from the fwhm of the brightness profile of the star @xcite . for each band ,
individual datacubes were created at an angular sampling of 005@xmath10005 and combined to a single datacube using the gemcombine iraf task .
more details about observations and data reduction can be found in riffel & storchi - bergmann ( _ in prep . _ ) . because of the complexity of the instruments , hyperspectral data frequently contains instrumental signatures that can not be removed with standard data reduction techniques , and which in many cases contaminates the science data .
the gemini nifs is no exception .
the wavelet principal component analysis ( wpca ) technique described below ( a full account of the technique will be addressed in a paper in preparation - ferrari et al .
2011 ) can separate the instrumental fingerprint and , after its removal , allow the analysis of the clean " data .
besides the relevant information , any set of data has some amount of redundant information and noise .
the goal of principal component analysis ( pca ) is to find a new basis on which the data the spectra from each spatial pixel - is expressed in a more meaningful way .
the basis vectors * the principal components * are obtained by searching for the maximum variance in the data .
there are as many vectors as the original variables , but usually the first few vectors contain most of the information of the data set : they provide a more concise description of the data .
often the physical content is restricted to few vectors and thus interpretation can be easier .
mathematically we proceed as follows . in a basis where the variables are all uncorrelated ( orthogonal )
their covariance or correlation matrix is diagonal .
we then find the principal components by diagonalizing the covariance or correlation matrix of the original data . in the case of covariance matrix , the * eigenvectors
* are the principal components and the * eigenvalues * are the variance associated with each of the eigenvectors . in the case of a hyperspectral cube , it is informative to measure the correlation between each principal component and the spectra in each spatial pixel : the * tomograms*. for example , tomogram 1 is the projection ( scalar product ) of eigenvector 1 and the spectra relative to each spatial pixel . for a full description on the method applied to astronomical data ,
see @xcite . the discrete wavelet transform ( dwt )
consists of describing a signal @xmath11 in terms of a smoothed component @xmath12 and detail wavelet coefficients @xmath13 @xcite : @xmath14 the maximum wavelet scale is @xmath15 . in the _ trous _
( i.e. with holes ) algorithm @xcite used in this work , a series of smoothed versions @xmath16 of @xmath11 are calculated by convolving @xmath11 with scaled versions of a low - pass filter @xmath17 . in this way
, the term @xmath18 is calculated by taking adjacent pixels from @xmath19 that are @xmath20 pixels far apart .
we proceed in this way for each of the scale @xmath21 . after this
, the discrete wavelet transform is obtained from the difference @xmath22 the wavelet coefficients @xmath13 now retain only details which are about @xmath20 in size , a multiresolution transform .
the maximum scale @xmath15 is mathematically arbitrary because , by virtue of eq .
( [ eq : dwt ] ) and ( [ eq : wdiff ] ) , reconstruction is complete for any @xmath15 .
physically , @xmath15 is chosen so that the spatial frequencies of interest are lower than @xmath23 .
the idea behind the use of the wavelet and the pca transforms together is to decompose the original signal both in the wavelet and in the pca space , select unwanted features and finally reconstruct the signal after eliminating these features .
the algorithm is schematically shown in figure [ fig : wpca ] .
the original datacube ( orange ) is first decomposed in wavelets scales @xmath24 .
then , the pca analysis is performed in each of the wavelet components , resulting in several eigenvectors for each scale . after removing those eigenvectors which contain unwanted structures ( represented by black dots in the figure ) , each of the wavelets components is restored with the pca reconstruction , @xmath25 , and then the resulting corrected wavelets @xmath26 are combined to form the resulting corrected cube ( green ) . )
space and then the pca analysis is performed . after identifying and correcting for unwanted effects ( instrumental signatures , fingerprints or noise , for example ) the cube is reconstructed in both pca and wavelet spaces . ]
figure [ pcangc591 ] shows an example of the wpca transform of mrk1157 datacube .
the columns are the wavelet decomposition of the original cube . in both fig.[fig : wpca ] and [ pcangc591 ] the dotted rectangle represents the same kind of information : in each wavelet scale a pca analysis is performed .
the original information is completely represented in the wavelet space , thus summing up all the components gives back the original data .
the wpca cleaning process is basically to identify the components ( in wavelet and pca space ) which contains mostly instrumental signatures ( i.e noise ) , then remove them and reconstruct the signal . in the case of mrk1157 observations ,
the wpca decomposition reveals that the @xmath27 tomograms are dominated by the instrumental signature ( horizontal and vertical stripes in fig .
[ pcangc591 ] ) . that instrumental fingerprint can also be identified in the eigenvectors as high frequency bumps ( not shown ) in the beginning and in the end of the spectra
the instrumental signature has a spatial and spectral behaviour .
thus , we remove @xmath27 and reconstruct the cube with @xmath28 .
we call the attention to the fact that we have analysed 5 wavelet scales and 20 principal componentes ( with 20 eigenvectors and 20 tomograms ; a total of 200 maps ) to identify the instrumental signature . for sake of brevity , in fig .
[ pcangc591 ] we only show the first 4 principal components , which account for @xmath2995% of the data variance . the _ clean _ datacube was used to perform the stellar population analysis .
the spatial resolution of the j and k@xmath30 bands are very similar , allowing us to combine both datacubes into a single datacube and then perform the spectral synthesis .
the combination was done using the _ scombine _ iraf task at a spectral sampling of 2 .
the astrometry was done using the peak of the continuum emission in both bands .
we have re - binned the resulting datacube to a spatial sampling of 01 in order to reach a signal - to - noise ratio ( s / n ) high enough to obtain a reliable fit of the stellar population .
the resulting s / n ranges from @xmath31 at the borders of the nifs field to up to 120 at the nucleus .
we used the _ wspectext _ iraf task to convert the individual spectra of the final datacube to 784 _ ascii _ files , used as input to the synthesis code .
the observation field is 28@xmath1028 , which corresponds to 830@xmath10830 pc@xmath32 at the galaxy .
our main goal in this section is to map in detail the nir spectral energy distribution ( sed ) components of the inner 400 pc of mrk1157 . for this purpose
we fit simultaneously the underlying continuum of the @xmath15 and @xmath33 bands applying the method described in @xcite and summarized below . in order to perform the stellar population synthesis we use the code @xcite .
fits the whole underlying spectrum , excluding emission lines and spurious data , mixing computational techniques originally developed for semi empirical population synthesis with ingredients from evolutionary synthesis models @xcite .
essentially , the code fits an observed spectrum , @xmath34 , with a combination , in different proportions , of @xmath35 single stellar populations ( ssps ) .
extinction is parametrised by the v - band extinction @xmath36 and modelled by as due to foreground dust . in the fits we use the ccm @xcite extinction law . in order to model a spectrum @xmath37
, the code solves the following equation : @xmath38 where * x * is the population vector , whose components @xmath39 , ( j = 1, .. ,n@xmath40 ) represent the fractional contribution of each ssp in the base to the total synthetic flux at @xmath41 .
@xmath42 is the spectrum of the @xmath43th ssp of the base of elements normalized at @xmath41 , the reddening term is represented by @xmath44 , @xmath45 is the synthetic flux at the normalisation wavelength , @xmath46 denotes the convolution operator and @xmath47 is the gaussian distribution used to model the line - of - sight stellar motions , which is centred at velocity @xmath48 with dispersion @xmath49 .
however , note that due to the low spectral resolution of eps models in the nir , the @xmath49 values can not be derived in a reliable way from the synthesis , thus , we do not use them . for details on @xmath49
see sec .
[ results ] .
the final fit is carried out minimizing the equation : @xmath50 ^ 2\ ] ] where emission lines and spurious features are excluded from the fit by fixing @xmath51=0 . in eq .
[ streq ] , the most important ingredient in stellar population synthesis is the base set , @xmath42 . as default base uses the ssps of @xcite
however , these ssps do not include the effect of the contribution of thermally pulsating asymptotic giant branch ( tp - agb ) stars , whose contribution is enhanced in the nir and crucial to model the stellar populations in this spectral region ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus , we update the base using the @xcite evolutionary population synthesis ( eps ) models as described in @xcite .
the base comprises ssps synthetic spectra covering 12 ages ( @xmath520.01,0.03 , 0.05 , 0.1 , 0.3 , 0.5 , 0.7 , 1 , 2 , 5 , 9 and 13 gyr ) and four metallicities ( @xmath530.02 , 0.5 , 1 , 2 z@xmath54 ) .
we also include black - body functions for temperatures in the range 700 - 1400k in steps of 100k @xcite and a power - law ( @xmath55 ) in order to account for possible contributions from dust emission ( @xmath56 ) and from a featureless continuum ( @xmath57 ) , respectively , at the nucleus ( e.g. * ? ? ? * ) .
the same spectral base was used to map the age distribution of the stellar population in the inner 300 pc of mrk1066 @xcite .
in fig . [ large ] ( top - left ) we show an optical image of mrk1157 taken with the wide field and planetary camera 2 ( wfpc2 ) at the hubble space telescope ( hst ) through the filter f606w @xcite . the 2.17@xmath4 m nifs datacube continuum image is also shown in fig .
[ large ] ( top - right ) . in order to illustrate the accuracy of our fits ,
we show in fig . [
large ] ( bottom ) sample spectra , obtained within 01@xmath1001 apertures for four distinct positions : the nucleus ( position n marked at top - left panel of fig . [ large ] ) and 04 west ( position a ) , 05 south - east ( position b ) and 08 north - east ( position c ) of the nucleus .
the synthetic spectra were overploted on the data as dotted lines .
as can be observed in this figure , the modelling of the stellar population reproduces very well the continuum / absorption spectra at all positions of mrk1157 .
following @xcite the observed and synthetic spectra were normalized at 2.12@xmath4 m , a region free of emission / absorption lines @xcite .
no contribution of the @xmath57 and/or @xmath56 components was necessary in order to reproduce the the nuclear continuum in mrk1157 .
similar result was obtained by @xcite when fitting the nuclear integrated spectrum of this source within a larger aperture ( r@xmath58200pc ) over the 0.82.4 spectral region .
we point out that the synthetic spectra was reddened using the a@xmath59 derived from the synthesis ( fig .
[ pop2 ] ) . following ( * ? ? ?
* see also riffel et al . , 2009 and cid - fernandes et al . 2004 ) , we have binned the contribution of the spcs ( @xmath39 ) into a reduced population vector with four age ranges : * young * ( @xmath60 : @xmath61 myr ) ; * young - intermediate * ( @xmath62 : @xmath63 gyr ) , * intermediate - old * ( @xmath64 : @xmath65 gyr ) and * old * ( @xmath66 : @xmath67 gyr ) . in fig .
[ pop1 ] , we show the spatial distribution of the percent flux contribution at 2.12@xmath4 m of the stars in each @xmath68 : while in regions farther than r@xmath69 0.8(240pc ) from the nucleus the contribution of the young - intermediate spcs reaches values of up to 100% , closer to the nucleus the contribution of this component is negligible . within this region
the stellar population is dominated by the old component(@xmath150% ) , closely surrounded ( r @xmath70 04 ) by an intermediate - old population ( also @xmath150% ) .
no signs of young stellar populations were detected close to the nucleus ; however , a significant contribution ( 25 - 60% ) is found in regions farther than @xmath7106 .
the light - fraction spc contributions depend of the normalization wavelength and thus the comparison with results from other spectral regions should be done with caution @xcite . however , a physical parameter which does not depend on the normalization point used in the fit is the stellar mass .
the mass - fraction of each population vector components is show in fig .
[ pop1 ] ( young : @xmath72 , young - intermediate : @xmath73 , intermediate - old : @xmath74 and old population : @xmath75 ) .
the maps of the mass - weighted spc follow a similar distribution to the light - weighted ones . however , in the former the contribution of the older ages is enhanced , particularly within 04 from the nucleus . besides the spc distributions , the outputs the average reddening of the stellar populations ( fig.[pop2 ] ) .
the highest values of @xmath76 ( we used @xmath77=3.1e(b - v ) ) are reached at the nucleus up to within @xmath6902 from it .
the goodness of the fit is measured in by the percent mean deviation : _ _ adev__=@xmath78 , where @xmath79 is the observed spectrum and @xmath80 is the fitted model @xcite .
the _ adev _ map for mrk1157 is shown in fig .
[ pop2 ] and presents values @xmath81% at most locations , indicating that the model reproduces very well the observed underlying spectra .
we point out that , in order to have a robust result on stellar population fitting with it is important to have a reliable flux calibration ( see manual ) . in other words
the fit depends on the overall shape of the observed spectrum from the @xmath15 to the @xmath33-band , and our nifs observations misses the @xmath82-band .
in order to verify the reliability of our flux calibration we have extracted a spectrum from our datacubes matching the aperture and position angle of our spex cross - dispersed spectrum @xcite .
note that the spex data were taken in the cross - dispersed mode , and thus are free from seeing and aperture effects ( see * ? ? ?
* for details ) .
the comparison between both spectra is shown in fig .
[ nifsxspex ] .
it is clear that our nifs spectra have a reliable flux calibration .
in fact the difference between the nifs and spex spectra is lower than 3% in the @xmath15 band . in addiditon
, we have performed a simulation varying the flux of the @xmath15 band by 20% ( from 90% up to 110% ) and the difference to the population vector components is , in mean , @xmath705% indicating that our results are robust .
in general , our results are similar to those obtained in the 2d mapping of the sp of mrk1066 @xcite as well as for previous nir studies using single aperture nuclear spectra @xcite . by mapping the stellar population in 2d we can analyse the spatial variations of the spcs in the inner few hundred parsecs of the galaxy .
further , the significance of these variations is enhanced by the comparison with the @xmath2 map presented in fig .
this map was obtained by fitting the k - band co absorption band - heads with the penalized pixel fitting ( ppxf ) method of @xcite using as stellar template spectra those of the gemini library of late spectral type stars observed with the gemini near - infrared spectrograph ( gnirs ) ifu and nifs @xcite .
more details on the stellar kinematics of the central region of mrk1157 can be found in riffel & storchi - bergmann(_in prep . _ ) . the @xmath2 map shows a partial ring of low-@xmath2 values ( @xmath1 50 - 60@xmath83 ) surrounding the nucleus at @xmath1 08 ( 230pc ) from it , immersed in higher @xmath2 values of the bulge stars ( @xmath1 100@xmath83 ) .
such rings are commonly observed in the central region of active galaxies and are due to kinematically colder regions with younger stars than the underlying bulge @xcite . the comparison between mrk1157 stellar population synthesis maps ( light- and mass - weighted ) and the @xmath2 map shows that the low @xmath2 ring is spatially correlated with the young - intermediate age spc , while the highest @xmath2s are associated with the old component .
interestingly , very similar results were found by @xcite for mrk1066 .
thus , the results found for mrk1157 and mrk1066 support the use of low stellar velocity dispersion as a tracer of young - intermediate age stars in the galaxy centre , thus confirming the interpretation of the above studies . the most compressed form to represent the spcs mixture of a galaxy is by its mean stellar age , which was defined by @xcite in two ways : the first is weighted by light fraction , @xmath84 and the second , weighted by the stellar mass , @xmath85 being both definitions limited by the age range used in our base , the former is more representative of the younger ages while the latter is enhanced by the old spc @xcite .
in order to compare our results with those derived using integrated spectra we have calculated the mean ages for the central region of mrk1157 .
these mean ages in light- and mass - fractions are @xmath86 = 8.48 @xmath87 = 8.65 , respectively .
note that these values were obtained taking the mean value over all the spectra in the map ( e.g. summing @xmath88 values over the map and dividing the result by the number of spectra in the map ) .
besides our study of mrk1066 @xcite , to date , the only previous 2d stellar population studies of active galaxies in the nir are those from the group of r. i. davies .
@xcite investigated the circumnuclear star formation in nine seyfert galaxies using the nir adaptive optics integral field spectrograph sinfoni at the very large telescope ( vlt ) .
based on measurements of the emission - line equivalent width , supernova rate and mass - to - light ratio , these authors found circumnuclear disks of typical diameters of tens of pc with a characteristic " age in the range 10300 myr .
such a characteristic " age can be associated with our mean age ( @xmath69 300 myr ) , thus our result for mrk1157 is in good agreement with those found by @xcite for their sample ( which does not include mrk1157 ) .
nevertheless , as pointed out by @xcite the methodology adopted here allowed us not only to obtain a characteristic " age , but to also map the spatial distribution of the spcs of different ages in the central region of a seyfert galaxy , on the basis of nir integral - field spectroscopy . moreover , our results are in reasonable agreement with those of @xcite where the authors use integrated long - slit spectroscopy ( considering the different apertures ) . in further support to the results of the synthesis , we found that the average reddening map derived for the stellar population ( fig.[pop2 ] ) is in close agreement with the one derived for the narrow - line region using emission - line ratios ( riffel & storchi - bergmann , _ in prep . _ ) , presenting a similar structure and a very similar average value to that found by ( * ? ? ? * e(b - v)@xmath10.63 ) for the sp , using nir integrated spectra .
as part of an ongoing project aimed at mapping the age distribution of the nir sp in the inner few hundred parsecs of seyfert galaxies , we present the spectral synthesis for the nuclear region of the seyfert 2 galaxy mrk1157 ( ngc591 ) within the inner @xmath69 400 pc at spatial resolution of @xmath135pc . we have used a cleaning method which allowed us to remove the noise effects and redundant data from the 1d extractions , thus allowing a more accurate analysis of the spcs . using the noise free cube we have mapped the distribution of stellar population components of different ages and of their average reddening .
the main conclusions of this work are : * the cleaning technique presented here has allowed us to perform the spectral synthesis in noisy spectra ( e.g. borders of the datacube ) in a more accurate way . *
the age of the dominant stellar population presents spatial variations : the flux and mass contributions within the inner @xmath1160 pc are dominated by old stars ( @xmath895gyr ) , while intermediate age stars ( @xmath90 gyr ) dominate in the circumnuclear region between 170 pc and 350pc . * as for mrk1066 @xcite , we found that there is a spatial correlation between the distribution of the intermediate age component and low stellar velocity dispersion values , which delineate a partial ring around the nucleus of mrk1157 .
similar structures have been found around other active nuclei , and our results for mrk1157 ( and mrk1066 ) reveals that these nuclear rings are formed by intermediate age stars . * no signatures of non - thermal and hot dust components are found in central region of mrk1157 .
we thank an anonymous referee for helpful suggestions .
the starlight project is supported by the brazilian agencies cnpq , capes and fapesp and by the france - brazil capes / cofecub program .
this work is based on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the science and technology facilities council ( united kingdom ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia e tecnologia ( brazil ) and south - eastcyt ( argentina ) .
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 .
this work has been partially supported by the brazilian institution cnpq .
heckman , t. m. , in ho l. ed , co - evolution of black holes and galaxies , from the carnegie observatories centennial symposia .
carnegie observatories astrophysics series .
cambridge university press , cambridge , pg . | we have used the gemini near - infrared integral field spectrograph ( nifs ) to map the age distribution of the stellar population in the inner 400pc of the seyfert 2 galaxy mrk1157 ( ngc591 ) , at a spatial resolution of 35pc .
we have performed wavelet and principal component analysis in the data in order to remove instrumental signatures . an old stellar population component ( age @xmath05gyr )
is dominant within the inner @xmath1130pc which we attribute to the galaxy bulge . beyond this region ,
up to the borders of the observation field , young to intermediate age components ( 0.10.7gyr ) dominate . as for mrk1066 , previously studied by us , we find a spatial correlation between this intermediate age component and a partial ring of low stellar velocity dispersions ( @xmath2 ) .
low-@xmath2 nuclear rings have been observed in other active galaxies and our results for mrk1157 and mrk1066 reveal that they are formed by intermediate age stars .
such age is consistent with a scenario in which the origin of the low-@xmath2 rings is a past event which triggered an inflow of gas and formed stars which still keep the colder kinematics of the gas from which they have formed .
no evidence for the presence of an unresolved featureless continuum and hot dust component as found in mrk1066 are found for mrk1157 .
[ firstpage ] galaxies : individual ( ngc591 ) galaxies : individual ( mrk1157 ) galaxies : seyfert galaxies : stellar content galaxies : near - infrared data analysis : wavelet transform
data analysis : principal component analysis |
recent shear flows studies ( @xcite , @xcite , @xcite ) have shown the importance of the early time dynamics , that in principle can lead to non - linear growth long before an exponential mode is dominant .
the recognition of the existence of an algebraic growth , due - among other reasons - to the non - orthogonality of the eigenfunctions ( @xcite ) and a possible resonance between orr - sommerfeld and squire solutions ( @xcite ) , recently promoted many contributions directed to study the early - period dynamics .
for fully bounded flows works by @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and for partially bounded flows works by @xcite , @xcite , @xcite , @xcite can be cited . as for free shear flows ,
the attention was first aimed in order to obtain closed - form solutions to the initial - value inviscid problem ( @xcite , @xcite ) and was successful by considering piecewise linear parallel basic flow profiles .
an interesting aspect observed in the intermediate periods is that the maximal amplification is generally associated to oblique disturbances , that , as a consequence , potentially can promote early transition , see e.g. @xcite .
in fact , these perturbations , which are asymptotically stable at all reynolds numbers , are the perturbations best exploiting the energy transient amplification . in this work
we consider as a prototype for free shear flow the two - dimensional wake past a bluff body .
the wake stability has been widely studied by means of modal analyses ( e.g. @xcite , @xcite , @xcite , @xcite ) . however , in this way only the asymptotic fate can be determined , regardless of the transient behavior and the underlying physical cause of any instability . in this work
we adopt the velocity - vorticity formulation to evaluate the general initial - value perturbative problem .
this method was proposed by criminale , drazin and co - authors in the years 1990 - 2000 ( @xcite , @xcite , @xcite , @xcite , @xcite ) . in synthesis , the variables are laplace - fourier transformed in the plane normal to the basic flow .
afterward , the resulting partial differential equations in time are integrated numerically .
this procedure allows for completely arbitrary initial expansions by using a known set of functions ( schauder basis in the @xmath2 space ) and yields the complete dynamics the early time transients and the asymptotic behavior ( up to many hundred time scales ) for any disturbance .
the long term dynamics would not , in fact , have been easily recovered by using the alternative method of the direct numerical integration of the linear equations since the integration over a range of time larger than a few dozen basic time scales is not feasible .
the base flow model that we employ includes the wake transversal velocity and thus the nonlinear and diffusive dynamics that are responsible for the growth of the flow and the associated mass entrainment .
we consider the first two order terms of the analytical navier - stokes expansion obtained by tordella and belan ( 2002 , 2003 ) ( @xcite , @xcite ) , see [ base flow ] .
in particular , we consider the longitudinal component of such an expansion solution , the problem is parameterized by @xmath3 , the longitudinal coordinate , and the reynolds number @xmath4 .
we use a complex wavenumber for the disturbance component aligned with the flow so that longitudinal spatially damped waves are represented .
it should be observed that a longitudinal spatial growth could not be considered physically admissible as an initial condition since the energy density of the initial perturbation would be infinite . in the context of the initial - value problem , this is an innovative feature adopted to introduce a possible spatial evolution ( damping ) of the perturbative wave in the longitudinal direction .
the perturbative equations are numerically integrated by the method of lines . the equations formulation and initial and boundary conditions are presented in [ formulation ] . the perturbation evolution is examined for the base flow configurations corresponding to the reynolds numbers of @xmath5 , and for a typical section , @xmath6 , of the intermediate region of the flow where the entrainment process is active .
a comparison with a base flow far field configuration , @xmath7 , is also proposed .
the normalized perturbation kinetic energy density is the physical quantity on which the transient growth is observed ( see [ transient ] ) . to determine the temporal asymptotics of the disturbance an equivalent of the modal temporal growth rate
is introduced ( see [ asymptotic ] ) . in the case of longitudinal disturbances ,
comparison with recent spatio - temporal multiscale orr - sommerfeld analysis ( @xcite , @xcite ) and with laboratory experimental data ( @xcite ) is carried out , see [ asymptotic ] . as noted , the initial conditions posed are arbitrary .
as far as the modal theory is concerned , the agreement is excellent both for the frequency , defined as the temporal derivative of the perturbation phase , and the temporal growth rate .
it is also in quantitative agreement with respect to the laboratory data .
the experiment shows couples of pulsation and wavelength of the cylinder vortex shedding that are close to those yielded by the ivp analysis when the wavenumber is that where the growth rate is maximum .
conclusions are given in [ conclusions ] .
the base flow is considered viscous and incompressible . to describe the two - dimensional growing wake flow , an expansion solution for the navier - stokes two - dimensional steady bluff body wake
( @xcite , @xcite ) has been used .
the @xmath8 coordinate is parallel to the free stream velocity , the @xmath9 coordinate is normal .
this approximated analytical navier - stokes solution incorporates the effects due to the full non linear convection as well as the streamwise and transverse diffusion .
the solution was obtained by matching an inner navier - stokes expansion in terms of the inverse of the longitudinal coordinate @xmath8 ( @xmath10 , @xmath11 ) with an outer navier - stokes expansion in terms of the inverse of the distance from the body . here
we take the first two orders ( @xmath12 ) of the inner longitudinal component of the velocity field as a first approximation of the primary flow . in the present formulation
the near - parallel hypothesis for the base flow , at a longitudinal position @xmath13 , is made .
the coordinate @xmath3 plays the role of parameter of the steady system together with the reynolds number .
the analytical expression for the profile of the longitudinal component is @xmath14 where @xmath15 is related to the drag coefficient ( @xmath16 ) and @xmath17 is an integration constant depending on the reynolds number .
as said in the introduction , this two terms representation is extracted from an analytical asymptotic expansion where the velocity vector and the pressure are determined to the forth order .
it should be observed that the transversal velocity component @xmath18 first appears at the third order @xmath19 , while the pressure only at the forth order @xmath20 . up to the second order ,
the field is thus parallel . beyond the second order
the analytical expression becomes much more complex , special functions as the confluent hypergeometric functions plays a role ( @xcite ) associated to the deviation from parallelism . by changing the @xmath3 values , the base flow profile ( [ u_wake_profile ] )
will locally approximate the behavior of the actual wake generated by the body . here ,
the region considered , if not otherwise specified , is fixed to a typical section , @xmath21 ( where @xmath22 is the spatial scale of the wake ) of the intermediate wake .
the term intermediate is used in the general sense as used by @xcite : intermediate asymptotics are self - similar or near - similar solutions of general problems , valid for times , and distances from boundaries , large enough for the influence of the fine details of the initial / or boundary conditions to disappear , but small enough that the system is far from the ultimate equilibrium state ... . the distance beyond which the intermediate region is assumed to begin varies from eight to four diameters @xmath22 for @xmath23 $ ] ( see @xcite , @xcite ) .
base flow configurations corresponding to a @xmath4 of @xmath5 are considered . in figure
[ fig : wake_profile ] a representation of the wake profile at differing longitudinal stations is shown . in the intermediate ( @xmath6 ) and far ( @xmath7 )
wake for different reynolds numbers , @xmath24 is the free stream velocity .
the diameter of the cylinder is out of scale ( three times ) with respect to the wake profiles.,width=340,height=302 ] by exciting the base flow with small arbitrary three - dimensional perturbations , the continuity and navier - stokes equations that describe the perturbed system are @xmath25 @xmath26 @xmath27 @xmath28 where ( @xmath29 , @xmath30 , @xmath31 ) and @xmath32 are the perturbation velocity components and pressure respectively .
the independent spatial variables @xmath33 and @xmath9 are defined from @xmath34 to @xmath35 , while @xmath8 is defined in the semispace occupied by the wake , from @xmath36 to @xmath35 .
all physical quantities are normalized with respect to the free stream velocity @xmath24 , the body scale @xmath22 and the density . by combining equations ( [ ns1 ] ) to ( [ ns3 ] ) to eliminate the pressure
, the linearized equations describing the perturbation dynamics become @xmath37 @xmath38 where @xmath39 is the transversal component of the perturbation vorticity field . by introducing the quantity @xmath40 , that is defined by @xmath41 we obtain three coupled equations ( [ os ] ) , ( [ squire ] ) and ( [ vel_vor ] ) . equations ( [ os ] ) and ( [ squire ] ) are the orr - sommerfeld and squire equations respectively , from the classical linear stability analysis for three - dimensional disturbances . from kinematics , the relation @xmath42 physically links together the perturbation vorticity components in the @xmath8 and @xmath33 directions ( @xmath43 and @xmath44 respectively ) and the perturbed velocity field . by combining equations ( [ os ] ) and ( [ vel_vor ] ) then @xmath45 which
, together with ( [ squire ] ) and ( [ vel_vor ] ) , fully describes the perturbed system in terms of vorticity .
this formulation is a classical one .
alternative classical formulations , as the velocity - pressure one , are in common use .
we chose this formulation because the vorticity transport and diffusion is the principal phenomenology for the dynamics of a wake system .
for piecewise linear profiles for @xmath46 analytical solutions can be found . for continuous profiles ,
the governing perturbative equations can not be analytically solved in general , but may assume a reduced form in the free shear case ( @xcite ) .
moreover , from the equations ( [ squire ] ) , ( [ vel_vor ] ) and ( [ os_gamma ] ) , it is clear that the interaction of the mean vorticity in @xmath33-direction ( @xmath47 ) with the perturbation strain rates in @xmath8 and @xmath33 directions ( @xmath48 and @xmath49 respectively ) proves to be a major source of any perturbation vorticity production .
the perturbation quantities are laplace and fourier decomposed in the @xmath8 and @xmath33 directions , respectively .
a complex wavenumber @xmath50 along the @xmath8 coordinate as well as a real wavenumber @xmath51 along the @xmath33 coordinate are introduced . in order to have a finite perturbation kinetic energy ,
the imaginary part @xmath52 of the complex longitudinal wavenumber can only assume non - negative values .
in so doing , we allow for perturbative waves that can spatially decay ( @xmath53 ) or remain constant in amplitude ( @xmath54 ) .
the perturbation quantities @xmath55 involved in the system dynamics are now indicated as @xmath56 , where @xmath57 indicates the laplace - fourier transform of a general dependent variable in the @xmath58 phase space and in the remaining independent variables @xmath9 and @xmath59 .
the governing partial differential equations are @xmath60\label{ivp2_fou2 } \\
\nonumber \frac{\partial \hat{\omega}_{y}}{\partial t } = & - & ( i k cos(\phi ) - \alpha_i ) u \hat{\omega}_{y } - i k sin(\phi ) \frac{du}{dy } \hat{v } \\ & + & \frac{1}{re } [ \frac{\partial^2 \hat{\omega}_y}{\partial y^2 } - ( k^2 - \alpha_i ^2 + 2 i k cos(\phi ) \alpha_i ) \hat{\omega}_y]\label{ivp2_fou3}\end{aligned}\ ] ] where @xmath61 is the angle of obliquity with respect to the @xmath8-@xmath9 physical plane , @xmath62 is the polar wavenumber and @xmath63 , @xmath64 are the wavenumbers in @xmath8 and @xmath33 directions respectively .
the imaginary part @xmath52 of the complex longitudinal wavenumber represents the spatial damping rate in the streamwise direction . in figure [ dt_perturbation_scheme ] the three - dimensional perturbative geometry scheme is depicted .
+ from equations ( [ ivp2_fou1])-([ivp2_fou3 ] ) , it can be noted that there can neither be advection nor production of vorticity in the lateral free stream .
the vorticity can only be diffused since only the diffusive terms remains in the limit when @xmath65 .
perturbation vorticity vanishes in the free stream , regardless if it is initially inserted there ( if inserted , vorticity is finally dissipated in time when @xmath66 ) .
this means that the velocity field is harmonic as @xmath65 .
governing equations ( [ ivp2_fou1 ] ) , ( [ ivp2_fou2 ] ) and ( [ ivp2_fou3 ] ) need proper initial and boundary conditions to be solved . among all solutions , those whose perturbation velocity field is bounded in the free stream are sought .
periodic initial conditions for @xmath67 can be cast in terms of a set of functions in the @xmath2 hilbert space , as @xmath68 for the symmetric and the asymmetric perturbations , respectively .
parameter @xmath69 is an oscillatory parameter for the shape function , while @xmath70 is a parameter which controls the distribution of the perturbation along @xmath9 ( by moving away , or bringing nearer , the perturbation maxima from the axis of the wake ) .
the trigonometrical system is a schauder basis in each space @xmath71 $ ] , for @xmath72 .
more specifically , the system @xmath73 , where @xmath74 , is the schauder basis for the space of square - integrable periodic functions with period @xmath75 .
this means that any element of the space @xmath2 , where the dependent variables are defined , can be written as an infinite linear combination of the elements of the basis .
the transversal vorticity @xmath76 is chosen initially equal to zero throughout the @xmath9 domain in order to ascertain which is the net contribution of three - dimensionality on the transversal vorticity generation and temporal evolution .
however , it can be demonstrated that the eventual introduction of an initial transversal vorticity does not actually affect the perturbation temporal evolution .
once initial and boundary conditions are properly set , the partial differential equations ( [ ivp2_fou1])-([ivp2_fou3 ] ) are numerically solved by the method of lines .
the spatial derivatives are centre differenced and the resulting system is then integrated in time by an adaptative multi - step method ( variable order adams - bashforth - moulton pece solver ) .
the transversal computational domain is large thirty body scales . by enlarging the computational domain to 50 and 100
body scales the results vary on the third and fourth significant digit , respectively .
one of the salient aspects of the initial - value problem is to observe the early transient evolution of various initial conditions . to this end
, a measure of the perturbation growth can be defined through the disturbance kinetic energy density in the plane @xmath77 ( see e.g @xcite,@xcite ) @xmath78 where @xmath79 is the extension of the spatial numerical domain .
the value @xmath80 is defined so that the numerical solutions are insensitive to further extensions of the computational domain size . here
, we take @xmath81 . the total kinetic energy can be obtained by integrating the energy density over all @xmath82 and @xmath51 . the amplification factor @xmath83
can be introduced in terms of the normalized energy density @xmath84 this quantity can effectively measure the growth of a disturbance of wavenumbers @xmath77 at the time @xmath59 , for a given initial condition at @xmath85 ( criminale _ et al .
_ 1997 ; lasseigne _ et al .
_ 1999 ) .
the temporal growth rate on the kinetic energy @xmath86 @xmath87 is introduced in order to evaluate both the early transient as well as the asymptotic behavior of the perturbations ( here it is the first moment of the perturbation which is assumed to asymptotically approach an exponential growth ) .
computations to evaluate the long time asymptotics are made by integrating the equations forward in time beyond the transient ( @xcite , @xcite ) until the temporal growth rate @xmath86 asymptotes to a constant value ( @xmath88 , where @xmath89 is of the order @xmath90 ) .
the angular frequency ( pulsation ) @xmath91 of the perturbation can be introduced by defining a local , in space and time , time phase @xmath92 of the complex wave at a fixed transversal station ( for example @xmath93 ) as @xmath94 and then computing the time derivative of the phase perturbation @xmath92 ( @xcite ) @xmath95 since @xmath92 is defined as the phase variation in time of the perturbative wave , it is reasonable to expect constant values of frequency , once the asymptotic state is reached .
we present a summary of the most significant transient behavior and asymptotic fate of the three - dimensional perturbations .
the temporal evolution is observed in the intermediate asymptotic region of the wake , which is the region where the spatial evolution is predominant .
it can be demonstrated that changing the number of oscillations @xmath69 and the parameter @xmath70 that controls the perturbation distribution along the @xmath9 direction can only extend or shorten the duration of the transient , while the ultimate state is not altered . more specifically ,
if the perturbation oscillates rapidly or is concentrated mainly outside the shear region of the basic flow , for a stable configuration , the final damping is accelerated while , for an unstable configuration , the asymptotic growth is delayed .
thus , these two parameters are not crucial , because their influence can be recognized a priori .
therefore , in the following we use the two reference values , @xmath96 and @xmath97 , and focus the attention mainly on parameters such as the obliquity , the symmetry , the value of the polar wavenumber and the spatial damping rate of the disturbance . in particular
, the polar wavenumber @xmath98 changes in a range of values reaching at maximum the order of magnitude @xmath99 , according to what is suggested by recent modal analyses ( @xcite , @xcite ) .
the order of magnitude of the spatial damping rate @xmath52 varies around the polar wavenumber value .
figure [ dt_asym_sym ] takes into account the influence , on the early time behavior , of the perturbation symmetry and of the wake region considered in the analysis , which is represented by the parameter @xmath3 .
all the configurations considered are asymptotically amplified , but the transients are different . , and ( c)-(d ) the perturbation vorticity components at @xmath93 as a function of time .
( a)-(c ) asymmetric initial condition , ( b)-(d ) symmetric initial condition .
intermediate ( @xmath100 , solid curves ) and far field ( @xmath101 , dashed curves ) wake configurations .
the periods @xmath102 are the periods of the modulation visible on @xmath103 , in the intermediate and far field , respectively .
the values of the vorticity component in part ( d ) have no physical meaning .
the plot simply shows that , on the contrary of the asymmetric case in part(c ) , the symmetric disturbance growth has a short transient after which it becomes homogeneous in time.,title="fig : " ] [ dt_asym_pi_4 ] , and ( c)-(d ) the perturbation vorticity components at @xmath93 as a function of time .
( a)-(c ) asymmetric initial condition , ( b)-(d ) symmetric initial condition .
intermediate ( @xmath100 , solid curves ) and far field ( @xmath101 , dashed curves ) wake configurations .
the periods @xmath102 are the periods of the modulation visible on @xmath103 , in the intermediate and far field , respectively .
the values of the vorticity component in part ( d ) have no physical meaning .
the plot simply shows that , on the contrary of the asymmetric case in part(c ) , the symmetric disturbance growth has a short transient after which it becomes homogeneous in time.,title="fig : " ] [ dt_sym_pi_4 ] , and ( c)-(d ) the perturbation vorticity components at @xmath93 as a function of time .
( a)-(c ) asymmetric initial condition , ( b)-(d ) symmetric initial condition .
intermediate ( @xmath100 , solid curves ) and far field ( @xmath101 , dashed curves ) wake configurations .
the periods @xmath102 are the periods of the modulation visible on @xmath103 , in the intermediate and far field , respectively .
the values of the vorticity component in part ( d ) have no physical meaning .
the plot simply shows that , on the contrary of the asymmetric case in part(c ) , the symmetric disturbance growth has a short transient after which it becomes homogeneous in time.,title="fig : " ] [ dt_asym_pi_4 ] , and ( c)-(d ) the perturbation vorticity components at @xmath93 as a function of time .
( a)-(c ) asymmetric initial condition , ( b)-(d ) symmetric initial condition .
intermediate ( @xmath100 , solid curves ) and far field ( @xmath101 , dashed curves ) wake configurations .
the periods @xmath102 are the periods of the modulation visible on @xmath103 , in the intermediate and far field , respectively .
the values of the vorticity component in part ( d ) have no physical meaning .
the plot simply shows that , on the contrary of the asymmetric case in part(c ) , the symmetric disturbance growth has a short transient after which it becomes homogeneous in time.,title="fig : " ] [ dt_sym_pi_4 ] the asymmetric cases ( a ) present , for both the intermediate position @xmath6 ( solid curve ) and the far field position @xmath104 ( dashed curve ) , two temporal evolutions . for @xmath100 a local maximum , followed by a minimum ,
is visible in the energy density , then the perturbation is slowly amplifying and the transient can be considered extinguished only after hundreds of time scales . for
@xmath101 these features are less marked .
it can be noted that the far field configuration ( @xmath7 ) has a faster growth than the intermediate field configuration ( @xmath105 ) up to @xmath106 . beyond this instant
the growth related to the intermediate configuration will prevail on that of the far field configuration . in the symmetric cases ( b ) the growths become monotone after few time scales ( @xmath107 ) and the perturbations quickly reach their asymptotic states ( around @xmath108 ) .
the intermediate field configuration ( @xmath105 , solid curve ) is always growing faster than the far field configuration ( @xmath7 , dashed curve ) .
this particular case shows a behavior that is generally observed in this analysis , that is , asymmetric conditions lead to transient evolutions that last longer than the corresponding symmetric ones , and demonstrates that the transient growth for a longitudinal station in the far wake can be faster than in the intermediate wake .
it should be noted that , even if the asymmetric perturbation leads to a much slower transient growth than that observed for the symmetric case , the growth rate become equal when the asymptotic states are reached ( see for example fig .
[ ivp_comparison_nmt_ivp_kfixed ] ) . the temporal window shown in fig.3 ( @xmath109 )
does not yet capture the asymptotic state of the asymmetric input .
however , we observed that further in time the amplification factor @xmath103 reaches the same order of magnitude of the symmetric perturbation .
the more noticeable results presented in fig .
3 are that the asymmetric growths in the early transient are much less rapid than the symmetric ones and that the function @xmath103 , in the case of asymmetric perturbations only , shows a modulation which is very evident in the first part of the transient , and which corresponds to a modulation in amplitude of the pulsation of the instability wave , see fig.[pulsation ] .
in fact , the pulsation varies : in the early transient it oscillates around a mean value with a regular period , which is the same visible on g , and with an amplitude which is growing until this value jumps to a new value around which oscillates in a damped way .
this second value is the asymptotic constant value .
this behavior is always observed in the case of asymmetric longitudinal or oblique instability waves .
instead , it is not shown by transversal ( @xmath110 ) waves or by symmetric waves , see fig .
4 , where , on the one hand , the asymptotic value , nearly equal to that of the asymmetric perturbation , is rapidly reached after a short monotone growth and where , on the other , the growth is many orders of magnitude faster , and as a consequence , a modulation would not be easily observable .
thus , we may comment here on the fact that two time scales are observed in the transient and long term behavior of longitudinal and oblique perturbations : namely , the periodicity associated to the average value of the pulsation in the early transient , clearly visible in the asymmetric case only , and the final asymptotic pulsation . the asymptotic value of the pulsation is higher than the initial one , typically is about @xmath111 times higher .
the period of the frequency modulation of the energy density @xmath103 is larger , nearly ( 1.4 - 1.7 ) higher , than the average period of the oscillating wave in the early transient , because @xmath103 is a square norm of the system solution .
thus the evolution of the system exhibits two periodic patterns at different frequencies : the first , of transient nature , and the other of asymptotic nature .
when the average damping of the energy density in the early transient is not strong and is then followed by a monotone asymptotic growth , the change of the frequency of the oscillation is evident , see fig.[pulsation ] .
this kind of behavior is often observed in the study of linear systems with an oscillating norm , a problem that naturally arises in the context of the linearized formulation of convection - dominated systems over finite length domains .
the occurrence of oscillating patterns in the energy evolution of the solutions is linked to the non - normal character of the linear operator which describes the system , see e.g. coppola and de luca ( 2006)@xcite . .
( a ) the amplification factor @xmath103 and ( b ) the temporal growth rate @xmath86 as functions of time .
asymmetric initial condition , @xmath112 ( solid curves ) , @xmath110 ( dashed curves ) .
( c)-(d ) the perturbation vorticity components as functions of time at @xmath93 , ( c ) @xmath113 and ( d ) @xmath114 .
the periods @xmath115 is the period of the modulation visible on @xmath103 , in the intermediate field .
, title="fig : " ] [ dt_g_2d3d ] .
( a ) the amplification factor @xmath103 and ( b ) the temporal growth rate @xmath86 as functions of time .
asymmetric initial condition , @xmath112 ( solid curves ) , @xmath110 ( dashed curves ) .
( c)-(d ) the perturbation vorticity components as functions of time at @xmath93 , ( c ) @xmath113 and ( d ) @xmath114 .
the periods @xmath115 is the period of the modulation visible on @xmath103 , in the intermediate field . ,
title="fig : " ] [ dt_r_2d3d ] . ( a ) the amplification factor @xmath103 and ( b ) the temporal growth rate @xmath86 as functions of time .
asymmetric initial condition , @xmath112 ( solid curves ) , @xmath110 ( dashed curves ) .
( c)-(d ) the perturbation vorticity components as functions of time at @xmath93 , ( c ) @xmath113 and ( d ) @xmath114 .
the periods @xmath115 is the period of the modulation visible on @xmath103 , in the intermediate field . ,
title="fig : " ] [ dt_g_2d3d ] . ( a ) the amplification factor @xmath103 and ( b ) the temporal growth rate @xmath86 as functions of time .
asymmetric initial condition , @xmath112 ( solid curves ) , @xmath110 ( dashed curves ) .
( c)-(d ) the perturbation vorticity components as functions of time at @xmath93 , ( c ) @xmath113 and ( d ) @xmath114 .
the periods @xmath115 is the period of the modulation visible on @xmath103 , in the intermediate field . , title="fig : " ] [ dt_r_2d3d ] figure [ dt_g_r_2d3d ] illustrates an interesting comparison between two - dimensional and three - dimensional waves ( note that a logarithmic scale on the ordinate is used in part ( a ) of fig .
[ dt_g_r_2d3d ] ) . the purely two - dimensional wave ( solid curve )
is rapidly reaching a first maximum of amplitude ( at about @xmath116 ) , then the perturbation decreases while oscillating and reaches a minimum around @xmath117 .
afterwards , the disturbance slowly grows up to @xmath118 , where an inflection point of the amplification factor @xmath103 occurs .
it should be noted that this behavior is controlled solely by the evolution of the @xmath119 , that is by the vorticity component present in the basic flow only .
then , the growth becomes faster and the perturbation is highly amplified in time .
the purely orthogonal perturbation ( dashed curve ) is instead immediately amplified .
the trend is monotone , and does not present visible fluctuations in time .
the initial growth is actually rapid and an inflection point of the amplification factor @xmath103 can be found around @xmath108 . beyond this point ,
the growth changes its velocity and becomes slower , but still destabilizing .
both cases have asymmetric initial conditions and are ultimately amplified . in agreement with squire theorem ,
the two - dimensional case turns out to be more unstable than the three - dimensional one , as the 2d asymptotically established exponential growth is more rapid than the 3d one ( see solid and dashed curves in fig .
[ dt_g_r_2d3d](a ) for @xmath120 ) .
however , it should be noted that , for an extended part of the transient ( up to about @xmath121 ) , the three - dimensional perturbation presents a larger growth than the two - dimensional one . .
( a ) the amplification factor @xmath103 and ( b ) the temporal growth rate @xmath86 as function of time .
symmetric initial condition , @xmath122.,title="fig : " ] [ g_k ] .
( a ) the amplification factor @xmath103 and ( b ) the temporal growth rate @xmath86 as function of time .
symmetric initial condition , @xmath122.,title="fig : " ] [ r_k ] .
( a ) the amplification factor @xmath103 as function of time and ( b ) the wave spatial evolution in the @xmath8 direction for @xmath123 .
asymmetric initial condition , @xmath124.,title="fig : " ] [ g_k ] .
( a ) the amplification factor @xmath103 as function of time and ( b ) the wave spatial evolution in the @xmath8 direction for @xmath123 .
asymmetric initial condition , @xmath124.,title="fig : " ] [ r_k ] fig .
[ g_r_k ] demonstrates that purely orthogonal three - dimensional unstable perturbations may become damped by increasing their wavenumber ( @xmath125 ) . in the case displayed in this figure , this happens when the wave number is increased beyond the value @xmath126 .
before the asymptotic stable states are reached , these configurations yield maxima of the energy density ( e.g. when @xmath127 at @xmath128 ) in the transients .
this trend is also typical of oblique and longitudinal waves , and it can be considered a universal feature in the context of the stability of near parallel shear flows . it should be noted that , in fig .
[ g_r_k ] , the perturbation is symmetric and again the amplitude modulation is not observed in the early transient , even in the asymptotically stable situations .
however , this example of transient behavior also contains a feature which is specific of orthogonal , both amplified or damped , and symmetric or asymmetric , perturbations , namely , the fact the most amplified component of the vorticity is the @xmath129 , see also fig.[dt_g_r_2d3d ] , part d. in fig .
[ g_r_alphai ] a significant phenomenon is observed for a longitudinal wave . by changing the order of magnitude of @xmath52
, it can be seen that perturbations that are more rapidly damped in space ( see , in fig .
[ g_r_alphai]b , the longitudinal spatial evolution of the wave ) yield a faster growth in time .
in fact , for nearly uniform waves in @xmath8 direction ( @xmath130 ) the configurations are asymptotically damped in time , while for increasing values of the spatial damping rate the perturbations are amplified in time ( note that a logarithmic scale is used on part ( a ) of the figure ) .
a possible general explanation is that the introduction in a physical system of a spatial concentration of kinetic energy is always destabilizing , hence the higher is the disturbance concentration the faster is the growth factor .
the dramatically increased growth rate of the symmetric mode with respect to the asymmetric mode can be understood from the induced velocity of the vorticity field . for simplicity , imagine that the wake consists of two parallel shear layers of opposite vorticity , as shown in figure 8(a ) .
further assume that the vorticity is discretized into a finite number of identical vortices .
suppose the upper shear layer is perturbed into a sinusoidal shape , as shown in figure 8(b ) .
the induced velocity at the crest of the sinusoid at point 1 due to the other vortices in the upper shear layer alone is up and to the right as indicated by the arrow , corresponding to the classical kelvin - helmholtz instability .
now consider the asymmetric mode , so that the lower shear layer is a sinusoid in phase with the upper one ( figure 8(c ) ) . the induced velocity at the crest of the lower shear layer at point 2 due to the other vortices in the lower shear layer alone is exactly the opposite of that at point 1 .
thus the growth in the asymmetric mode is due to only higher order gradients in the induced velocity field of one shear layer on the other .
in contrast , with a symmetric perturbation , the lower shear is the mirror image of the upper . in figure 8(d ) , the induced velocity at point 2 is necessarily the mirror image of that at point 1 .
both points move in concert to the right , without the low order velocity cancellation of the asymmetric mode .
thus the symmetric mode grows much faster .
this linear absolute instability takes place in the intermediate and far field and acts as a source of excitation for the pair of steady recirculating eddies in the lee of the cylinder .
the onset of a time periodic flow , a supercritical hopf bifurcation ( @xcite , @xcite ) indicates that in the end the vortices are shed alternatively from the separated streamlines above and below the cylinder forming the von karman vortex street .
the vortex street is the stable configuration after the bifurcation has taken place .
it has the symmetry of a traveling sinuous mode , which indicates asymmetry up - and - down of the cylinder .
thus , it can be observed that , also in the context of the vortex shedding , asymmetry shows higher stability properties with respect to symmetry .
figure [ ivp_comparison_nmt_ivp_kfixed ] presents a comparison between the initial - value problem and the asymptotic theory results represented by the zero order orr - sommerfeld problem ( @xcite ) in terms of temporal growth rate @xmath86 and pulsation @xmath91 .
temporal growth rate and @xmath131 pulsation .
comparison among the asymptotic results obtained by the ivp analysis ( circles : symmetric perturbation ; triangles : asymmetric perturbation ) and the normal mode analysis ( solid lines , see @xcite .
the asymptotic values for the ivp analysis are determined when the condition @xmath88 ( @xmath132 ) is satisfied . for the symmetric perturbations the asymptote
was reached before 50 time scales were elapsed , for the asymmetric perturbation the asymptote was determined after 500 time scales were elapsed.,title="fig : " ] [ dt_r35_full ] temporal growth rate and @xmath131 pulsation .
comparison among the asymptotic results obtained by the ivp analysis ( circles : symmetric perturbation ; triangles : asymmetric perturbation ) and the normal mode analysis ( solid lines , see @xcite .
the asymptotic values for the ivp analysis are determined when the condition @xmath88 ( @xmath132 ) is satisfied . for the symmetric perturbations the asymptote
was reached before 50 time scales were elapsed , for the asymmetric perturbation the asymptote was determined after 500 time scales were elapsed.,title="fig : " ] [ dt_om35_full ] in fig .
[ ivp_comparison_nmt_ivp_kfixed ] the imaginary part @xmath52 of the complex longitudinal wavenumber is fixed , and differing polar wavenumbers ( @xmath133 ) are considered .
for both the symmetric and asymmetric arbitrary disturbances here considered , a good agreement with the stability characteristics given by the multiscale near - parallel orr - sommerfeld theory can be observed .
however , it should be noted that the wave number corresponding to the maximum growth factor in the case of asymmetric perturbations is about 15% lower than that obtained in the case of symmetric perturbations and that obtained by the normal mode analysis .
when the perturbations are asymmetric , the transient is very long , of the order of hundreds time scales .
this difference can be due either to the fact that the true asymptote was not yet reached , or to the fact that the extent of the numerical errors in the integration of the equations is higher than that obtained in the case of symmetric transients , which last only a few dozen time scales .
note that this satisfactory agreement is observed by using arbitrary initial conditions in terms of elements of the trigonometrical schauder basis for the @xmath2 space , and not by considering as initial condition the most unstable waves given by the orr - sommerfeld dispersion relation .
moreover , a maximum of the perturbation energy ( in terms of @xmath86 ) is found around @xmath134 and confirmed by both the analyses . as shown in fig .
[ ivp_comparison_nmt_ivp_kfixed ] , we have also contrasted our results with the laboratory experimental results obtained in 1989 by williamson@xcite , who gave a quantitative determination of the strouhal number and wavelength of the vortex shedding oblique and parallel modes of a circular cylinder at low reynolds number .
the comparison is quantitatively good , because it shows that a wave number close to the wave number that theoretically has the maximum growth rate at @xmath135 ( see part a of fig . [ ivp_comparison_nmt_ivp_kfixed ] ) has a theoretically deduced frequency which is very close to the frequency measured in the laboratory . at this point ,
also the laboratory experimental uncertainty , globally of the order of a @xmath136 in an accurate measurement set up , should be introduced . the uncertainty associated to the laboratory method and to the theoretical model ( estimated through the difference between the position of the maximum growth rate showed by the two cases of asymmetric and symmetric perturbation ) overlaps , which confirms the quality of this comparison .
the same quantitative agreement is observed also at @xmath137 .
the three - dimensional stability analysis of the intermediate asymptotics of the 2d bluff - body viscous growing wake was considered as an initial - value problem .
the velocity - vorticity formulation was used .
the perturbative equations are laplace - fourier transformed in the plane normal to the growing basic flow .
the laplace transform allows for the use of a damped perturbation in the streamwise direction as initial condition . in this regard ,
the introduction of the imaginary part of the longitudinal wavenumber ( the spatial damping rate ) was done to explicitly include in the perturbation , which otherwise would have been longitudinally homogeneous , a degree of freedom associated to the spatial evolution of the system .
an important point is that the vorticity - velocity formulation , fourier - laplace transformation , allows ( over a reasonable lapse of computing time ) for the following of the temporal evolution over hundred of basic flow time scales and thus to observe very long transients .
such a limiting behaviour would not have been so easily reached by means of the direct numerical integration of the linearized governing equations of the motion .
two main transient scenarios have been observed in the region of the wake where the entrainment is present , the region in between @xmath100 diameters ( intermediate ) and @xmath101 diameters ( far field ) , for a @xmath4 number equal to @xmath138 and @xmath139 .
a long transient , where an initial growth smoothly levels off and is followed either by an ultimate damping or by a slow amplification for both oblique or 2d waves , and a short transient where the growth or the damping is monotone . the most important parameters affecting these configurations are the angle of obliquity , the symmetry , the polar wavenumber and the spatial damping rate . while the symmetry of the disturbance is remarkably influencing the transient behaviour leaving inalterate the asymptotic fate , a variation of the obliquity , of the polar wavenumber and of the spatial damping rate can significantly change the early trend as well as the final stability configuration .
interesting phenomena are observed .
a first one is that , in the case of asymmetric longitudinal or oblique perturbations , the system exhibits two periodic patterns , a first , of transient nature , and a second one of asymptotic nature .
a second phenomenon is that , in the case of an orthogonal perturbation , albeit always initially set equal to zero , the transversal vorticity component is the vorticity component which grows faster .
the third phenomenology is linked to the magnitude of the spatial damping rate .
perturbations that are more rapidly damped in space lead to a larger growth in time . for disturbances aligned with the flow
, the asymptotic behaviour is shown to be in excellent agreement with the zero order results of spatio - temporal multiscale modal analyses and with the laboratory determined frequency and wave length of the parallel vortex shedding at @xmath140 and @xmath139 .
it should be noted that the agreement between the initial value problem results and the normal mode theory is obtained not using as initial condition the most unstable wave given by the orr - sommerfeld dispersion relation at any section of the wake , but arbitrary initial conditions in terms of elements of the trigonometrical schauder basis for the @xmath2 space .
the authors would like to recognize robert breidenthal for his insightful interpretation of the perturbation dynamics in the study for wake vorticity .
his experience in this field was a significant contribution . | an initial - value problem ( ivp ) for arbitrary small three - dimensional vorticity perturbations imposed on a free shear flow is considered .
the viscous perturbation equations are then combined in terms of the vorticity and velocity , and are solved by means of a combined laplace - fourier transform in the plane normal to the basic flow .
the perturbations can be uniform or damped along the mean flow direction .
this treatment allows for a simplification of the governing equations such that it is possible to observe long transients , that can last hundreds time scales . this result would not be possible over an acceptable lapse of time by carrying out a direct numerical integration of the linearized navier - stokes equations .
the exploration is done with respect to physical inputs as the angle of obliquity , the symmetry of the perturbation and the streamwise damping rate .
the base flow is an intermediate section of the growing two - dimensional circular cylinder wake where the entrainment process is still active .
two reynolds numbers of the order of the critical value for the onset of the first instability are considered .
the early transient evolution offers very different scenarios for which we present a summary for particular cases .
for example , for amplified perturbations , we have observed two kinds of transients , namely ( 1 ) a monotone amplification and ( 2 ) a sequence of growth - decrease - final growth . in the latter case ,
if the initial condition is an asymmetric oblique or longitudinal perturbation , the transient clearly shows an initial oscillatory time scale .
that increases moving downstream , and is different from the asymptotic value .
two periodic temporal patterns are thus present in the system .
furthermore , the more a perturbation is longitudinally confined the more it is amplified in time .
the long - term behavior of two - dimensional disturbances shows excellent agreement with a recent two - dimensional spatio - temporal multiscale modal analysis and with laboratory data concerning the frequency and wave length of the parallel vortex shedding in the cylinder wake .
* keywords : * initial - value problem ( ivp ) - perturbation - early transient - spatially developing flows -instability - absolute - periodic temporal patterns - growth - three - dimensional . * corresponding author , e - mail : [email protected] * _ @xmath0 dipartimento di ingegneria aeronautica e spaziale , politecnico di torino , _
10129 torino , italy _ , _ _ @xmath1 department of applied mathematics , university of washington , seattle , _ wa 98195 - 2420 , usa _ _ |
the problems discussed in this paper are inspired by isogeometric analysis ( iga ) , which was developed in @xcite .
the core idea of isogeometric analysis is to use the spline based representation of cad models directly for the numerical analysis of partial differential equations ( pde ) . for a more detailed description of the isogeometric framework
we refer to @xcite .
one of the advantages of iga is the possibility to have discretization spaces of high order smoothness .
these spaces can then be used to directly solve high order pde problems . there exist several fourth ( and higher ) order problems of practical relevance . for their application in iga
see @xcite , as well as @xcite for kirchhoff - love shells , @xcite for the cahn - hilliard equation and @xcite for the navier - stokes - korteweg equation .
the spline based representation of the physical domains allows for high order smoothness within one b - spline patch .
however , most geometries of practical relevance can not be represented directly with one patch but have to be parametrized using a multi - patch approach .
it is not trivial to construct smooth function spaces over multi - patch domains .
many results for multi - patch domains can be derived from considerations on a single interface , hence in the following we will restrict ourselves to two - patch domains .
we are interested in @xmath0 isogeometric function spaces over two - patch domains .
more precisely , we compute the dimension and construct a stable basis for a special class of , so called , analysis - suitable @xmath3 parameterizations .
this class of geometries was first introduced in @xcite , where it was also shown that exactly the geometries of this class allow under certain assumptions @xmath4 isogeometric spaces with optimal approximation properties ( cf .
@xcite ) . note that an isogeometric function is @xmath0 if its graph surface is @xmath3 .
hence , the analysis - suitability condition is a restriction of the more general geometric continuity condition .
we refer to @xcite for the definition of geometric continuity . the existing literature about the construction of @xmath4-smooth isogeometric functions on two - patch ( and multi - patch ) domains can be roughly classified into two possible approaches .
the first one employs @xmath5-surface constructions around extraordinary vertices to obtain a set of @xmath4-smooth functions , see e.g. @xcite .
in contrast , the second approach studies the entire space of @xmath4-smooth isogeometric functions on any given two - patch ( or multi - patch ) parameterization and generate a basis of the corresponding @xmath4 isogeometric space . examples
are @xcite . in this paper
we follow the second approach , especially explored in @xcite .
we extend these results in two main directions .
first , these existing constructions and investigations are limited to piecewise bilinear domains .
our approach encloses the much wider class of analysis - suitable @xmath5 parameterizations , which contains the class of piecewise bilinear domains .
second , our construction works for non - uniform splines of arbitrary bidegree @xmath1 with @xmath2 and of arbitrary regularity @xmath6 with @xmath7 within the two single patches .
in contrast , the constructions @xcite are restricted to biquartic ( for special cases ) and to biquintic bzier elements and the constructions @xcite are restricted to bicubic and biquartic uniform splines of regularity @xmath8 .
further differences to @xcite are that our approach allows the construction of nested @xmath4 isogeometric spaces and that our basis functions are explicitly given , whereas the basis functions in @xcite are implicitly defined by means of minimal determining sets for the bzier coefficients .
similar to @xcite , the explicitly given basis functions possess a very small local support and are well conditioned .
moreover , the spline coefficients of our basis functions can be simply obtained by means of blossoming or fitting .
this could provide a simple implementation in existing iga libraries .
in addition , we present the study of the dimension of the resulting @xmath4 isogeometric spaces for all possible configurations of analysis - suitable @xmath5 two - patch parameterizations .
the remainder of the paper is organized as follows . in section [ sec : prel ] we describe some basic definitions and notations which are used throughout the paper
. section [ sec : c1-smooth ] recalls the concept of @xmath4-smooth isogeometric spaces over analysis - suitable @xmath5 two - patch parameterizations .
the dimension of these spaces , which depends on the considered two - patch parameterization , is analyzed in section [ sec : dimension ] .
then we present in section [ sec : basis ] an explicit construction of basis functions , and describe in section [ sec : matrices ] their resulting spline coefficients by means of blossoming ( and fitting ) .
finally , we conclude the paper in section [ sec : conclusion ] .
let @xmath9 be the interval @xmath10 $ ] or the unit square @xmath10 ^ 2 $ ] .
we denote by @xmath11 the ( tensor - product ) spline space of degree @xmath12 ( in each direction ) , which is defined on @xmath9 by choosing the open knot vector @xmath13 ( in each direction ) given by @xmath14 where @xmath15 , @xmath16 for all @xmath17 and @xmath18 is the number of different inner knots ( in each direction ) .
thereby @xmath6 describes the resulting @xmath19-continuity of the space @xmath11 at all inner knots .
the range for the regularity parameter @xmath6 is in general @xmath20 .
however , the focus here is on @xmath21 . of course
, in case of tensor - product splines , the knot vectors could be different in each direction .
moreover , the knot multiplicities could be different for every knot . to keep the presentation simple ,
we consider only the presented case .
the spline spaces @xmath22)$ ] and @xmath22 ^ 2)$ ] are spanned by the ( tensor - product ) b - splines @xmath23 and @xmath24 , @xmath25 , respectively .
each function @xmath26)$ ] and @xmath27^{2})$ ] possesses a b - spline representation @xmath28 and @xmath29 with spline control points @xmath30 and @xmath31 , respectively .
in addition , we consider the knot vectors @xmath32 and @xmath33 , which are obtained by inserting into the knot vector @xmath34 the knot @xmath35 with the index @xmath36 ( in case of @xmath37 ) and the knots @xmath35 and @xmath38 with the indices @xmath39 and @xmath40 ( in case of @xmath41 .
the resulting spline spaces are denoted by @xmath42 and @xmath43 , respectively , and are again @xmath19-smooth at all inner knots except at the knot @xmath35 or at the knots @xmath35 , @xmath38 , respectively , where the spaces are only @xmath44-smooth .
moreover , we denote by @xmath45 the space of ( tensor - product ) polynomials of degree @xmath12 on @xmath9 .
we consider a planar domain @xmath46 composed of two quadrilateral spline patches @xmath47 and @xmath48 , i.e. @xmath49 , which share a whole edge as common interface @xmath50 .
we assume that each patch @xmath51 , @xmath52 , is the image @xmath53^{2})$ ] of a regular , bijective geometry mapping @xmath54 ^ 2 \rightarrow \omega^{(s ) } , \quad { \boldsymbol{f}}^{(s ) } \in \mathcal{s}(\mathcal{t}^{p , r}_{k},[0,1]^2),\ ] ] with the spline representations @xmath55 for the sake of simplicity , we assume that the two patches @xmath56 and @xmath57 share the common interface at @xmath58.\ ] ] we denote the parameterization of the common curve at @xmath59 by @xmath60 \rightarrow { \mathbb r}^{2}$ ] and assume that @xmath61 .
the space of isogeometric functions on @xmath62 is given as @xmath63 ^ 2 ) , \mbox { } s \in \{l , r \ } \}.\ ] ] the graph surface @xmath64 of an isogeometric function @xmath65 consists of the two graph surface patches @xmath66^{2 } \rightarrow \omega^{(s ) } \times { \mathbb r } , \quad s \in \{l , r \},\ ] ] possessing the form @xmath67 where @xmath68 ^ 2)$ ] .
since the geometry mappings @xmath69 , @xmath70 , are given , an isogeometric function @xmath65 is determined by the two associated spline functions @xmath71 , @xmath70 , with the spline representations @xmath72 we are interested in @xmath4-smooth isogeometric functions @xmath73 .
we assume @xmath74-smoothness condition of @xmath75 , that results in @xmath76 for @xmath77 $ ] .
we denote the function along the common interface by @xmath78 .
let us consider the space @xmath79 of @xmath4-smooth isogeometric functions on @xmath62 , i.e. @xmath80 in more detail .
an isogeometric function @xmath65 belongs to @xmath79 if and only if the two graph surface patches @xmath81 and @xmath82 possess a well defined tangent plane along the common interface @xmath83 , compare @xcite .
this is equivalent to the condition that there exist functions @xmath84 \rightarrow { \mathbb r}$ ] such that for all @xmath77 $ ] @xmath85 and @xmath86 see @xcite .
the above described equivalent conditions are called geometric continuity of order @xmath87 or @xmath5-smoothness ( cf .
@xcite ) .
note that the first two equations , i.e. @xmath88 uniquely determine the functions @xmath89 , @xmath90 and @xmath91 up to a common function @xmath92 \rightarrow { \mathbb r}$ ] ( with @xmath93 ) by @xmath94 where @xmath95 and @xmath96 see e.g. @xcite . therefore , an isogeometric function @xmath65 belongs to @xmath79 if and only if the equation @xmath97 is satisfied for all @xmath77 $ ] . equations and
lead to linear constraints on the spline control points @xmath98 of @xmath71 , @xmath99 , with indices @xmath100 belonging to the index space @xmath101 moreover , we denote by @xmath102 the index space formed by all spline control points @xmath98 , @xmath70 , i.e. @xmath103 furthermore , for functions @xmath104 and @xmath91 satisfying equations there exist non - unique functions @xmath105 \rightarrow { \mathbb r}$ ] such that @xmath106 see e.g. @xcite
. note that for generic patches @xmath107 ^ 2)$ ] the functions @xmath104 and @xmath91 fulfill @xmath108)$ ] as well as @xmath109)$ ] . for special configurations
the degree may be lower and the regularity may be higher . motivated by @xcite , we restrict in the following the considered geometry mappings @xmath56 and @xmath57 to geometry mappings possessing linear functions @xmath89 , @xmath90 , @xmath110 , @xmath111 , as stated in the following definition .
[ def : asg1 ] the two - patch geometry parameterization @xmath56 and @xmath57 is analysis - suitable @xmath5 ( as @xmath5 ) if there exist @xmath89 , @xmath90 , @xmath110 , @xmath112)$ ] , with @xmath89 and @xmath90 relatively prime ( i.e. , @xmath113 ) satisfying equations and . the condition on @xmath113 was not used in @xcite but is not restrictive : if @xmath114 one can replace @xmath89 , @xmath90 and @xmath91 by @xmath115 , @xmath116 and @xmath117 , respectively .
obviously , the polynomial @xmath91 is divisible by @xmath118 , which can be seen best in . with @xmath113 ,
the functions @xmath89 , @xmath90 and @xmath91 are uniquely determined up to a common constant .
it was shown in @xcite when the functions @xmath119 and @xmath120 , @xmath70 , are assumed to be of higher degree or even to be spline functions along the interface , then the polynomial representation along the interface is reduced to lower degrees . hence
, these spaces do not guarantee optimal approximation order .
for this more general case the investigation of a dimension formula or of a basis construction should be possible in similar way as presented in the following sections , but is beyond the scope of this paper .
we consider as @xmath5 geometries @xmath121 ^ 2)$ ] with the corresponding functions @xmath89 , @xmath122)$ ] and @xmath123)$ ] .
let @xmath124 be the maximum degree of the functions @xmath119 , i.e. @xmath125 here we consider the actual degree of the functions . since @xmath126)$ ] , we obtain either @xmath127 or @xmath128 .
let @xmath129 be defined as the number of different inner knots where the function @xmath91 possesses the value zero , i.e. @xmath130 since @xmath123)$ ] , @xmath131 .
depending on @xmath91 , we define a new knot vector @xmath132 .
first , if @xmath133 then @xmath134 . otherwise , assuming @xmath135 , we have three cases : if @xmath136 , we set @xmath137 , if @xmath138 or @xmath139 , we set @xmath140 or @xmath141 , respectively , where @xmath142 , and possibly @xmath143 with @xmath40 , are the indices of @xmath144 and @xmath38 , which are roots of @xmath91 .
we are interested in the dimension of the isogeometric space @xmath79 .
clearly , the space @xmath79 is the direct sum of the two subspaces @xmath145 and @xmath146 which implies that @xmath147 in @xcite , the functions of @xmath148 and @xmath149 have been called basis functions of the first and second kind , respectively .
we first state the dimension of @xmath148 .
the dimension of @xmath148 is equal to @xmath150 clearly , @xmath151 is equal to the number of control points @xmath152 , @xmath70 , possessing indices @xmath153 , i.e. @xmath154 . to analyze the dimension of @xmath149 , we need additional tools describing the situation at @xmath59 . some of these tools have been ( similarly ) introduced in @xcite . consider the transversal vector @xmath155 defined on @xmath59 such that @xmath156 observe that @xmath157 ( which is equivalent to ) and therefore we simply set @xmath158 in addition , we consider the space of traces and transversal derivatives on @xmath59 and its pullback , which are given by @xmath159 and @xmath160 respectively .
the transversal vector @xmath161 , @xmath162 and @xmath163 depend only on the choice of @xmath110 and @xmath111 , since @xmath89 and @xmath90 are now uniquely determined by the geometry mappings @xmath56 and @xmath57 .
associated to @xmath161 , we consider the transversal derivative of @xmath75 with respect to @xmath161 on @xmath59 , i.e. @xmath164 .
clearly , for @xmath165 the function @xmath166 is a spline function .
more precisely , we have : [ lem : space_g0 ] if @xmath167 , then @xmath168)$ ] .
analyzing equation , we observe that @xmath168)$ ] , since @xmath169)$ ] .
the following lemma ensures that for @xmath170 the function @xmath171 is a spline function , too .
[ lem : space_g1 ] if @xmath167 , then @xmath172)$ ] .
recall that @xmath173 .
an isogeometric function @xmath75 belong to the space @xmath149 if and only if @xmath174 for all @xmath77 $ ] .
since @xmath175 we obtain that @xmath165 if and only if @xmath176 for all @xmath77 $ ] .
( condition is exactly the same as condition by substituting @xmath91 via . )
recall that @xmath177 .
therefore , by dividing equation by @xmath89 , we see that @xmath178)$ ] .
analogously we can show that @xmath179)$ ] and obtain that @xmath180)$ ] .
there is a one - to - one correspondence between trace and transversal derivative at @xmath59 , and @xmath181 .
[ prop : representation ] for any @xmath182 there exists a unique @xmath167 such that @xmath183 , given , for @xmath70 , by @xmath184 recall the notation @xmath185 .
equation is equivalent to @xmath186 by taylor expansion of @xmath187 and using equation , we obtain @xmath188 for @xmath189
. two useful corollaries of proposition [ prop : representation ] follow below . both hold for any possible choice of @xmath110 and @xmath111 .
[ cor : space - dimensions - at - gamma ] the dimension of @xmath149 is equal to the dimension of @xmath163 .
[ cor : characterization - of - c1-functions ] it holds that @xmath190 if and only if : @xmath191)\label{eq : c1-char - g0}\\ g_{1}&\in & \mathcal{s}({\mathcal{t}^{p - d_{\alpha},r-1}_{k}},[0,1 ] ) \label{eq : c1-char - g1}\\ \alpha^{(s ) } g_{1 } + \beta^{(s)}g_{0}^\prime & \in & \mathcal{s}({\mathcal{t}^{p , r}_{k}},[0,1 ] ) , \text { for $ s \in \{l , r \}$.}\label{eq : c1-char - mixed } \end{aligned}\ ] ] the statement follows from lemma [ lem : space_g0][lem : space_g1 ] and in proposition [ prop : representation ] . to investigate the dimension of @xmath163
, we consider the spaces @xmath192 and @xmath193 clearly , @xmath194 .
the following two lemmas state the dimension of @xmath195 and @xmath196 .
it holds that @xmath197)\ ] ] and consequently @xmath198 thanks to corollary [ cor : characterization - of - c1-functions ] , we need to show that for all @xmath199)$ ] we can construct @xmath200 such that and holds . in case of @xmath201 , and also when @xmath202 with @xmath136 , we can simply set @xmath203 . in the rest of the proof
, we always assume @xmath202 . in case of @xmath138 ,
we choose @xmath204)$ ] , where @xmath144 is a root of @xmath91 .
we need to prove , that is , to show that @xmath205 is @xmath19-smooth , for @xmath70 .
we only need to check the @xmath6-th derivative of , for @xmath70 , which is @xmath206 that is continuous in @xmath10 \setminus \tau_{\ell } $ ] for the regularity of @xmath207 and continuous when @xmath208 since the second addendum vanishes in the limit . in case of @xmath139 , we choose @xmath209)$ ] , where @xmath210)$ ] and @xmath211)$ ] are non - unique functions such that @xmath212 , and @xmath213 with the indices @xmath214 and @xmath40 are the two roots of @xmath91 . as before ,
we need to prove , that is , to show that @xmath215 is @xmath19-smooth , for @xmath70 .
we only need to check again that the @xmath6-th derivative of , for @xmath70 , is continuous in @xmath10 $ ] , which can be done analogous to the case @xmath216 .
the dimension of @xmath217 follows directly from the definition of the spline space .
this concludes the proof .
it holds that @xmath218 and consequently @xmath219 thanks to corollary [ cor : characterization - of - c1-functions ] , @xmath220 if and only if @xmath221 . the dimension of @xmath222 follows directly from the definition of the spline space .
this leads to the following result .
the dimension of @xmath149 is equal to @xmath223 finally , we obtain : the dimension of @xmath79 is equal to @xmath224 our results are in agreement with those in @xcite , where the special case of bilinear geometry mappings @xmath56 and @xmath57 with @xmath225 , @xmath226 , @xmath227 and @xmath8 was considered .
we present an explicit basis construction for the space @xmath79 for as @xmath5 two - patch geometries .
our basis will consist of a basis for the space @xmath148 and of a basis for the space @xmath149 .
the functions in @xmath148 are not influenced by the interface @xmath59 .
hence , a basis of @xmath148 can be constructed in a straightforward way from the standard basis on a single patch .
consider for @xmath228 , @xmath229 , the isogeometric functions @xmath230 and @xmath231 determined by @xmath232 then the collection of isogeometric functions @xmath233 forms a basis of the space @xmath148 .
note that these basis functions do not depend on the geometry mappings @xmath56 and @xmath57 ( and therefore do not depend on @xmath89 , @xmath90 and @xmath91 , too ) .
this is in contrast to the basis functions of @xmath149 , see section [ subsec : basis_second ] .
we present a construction of a basis of the space @xmath149 .
thereby , the resulting basis functions depend on @xmath89 , @xmath90 , @xmath110 and @xmath111 , and hence on the geometry mappings @xmath56 and @xmath57 . the idea is to generate a basis of @xmath149 by means of a basis of @xmath195 and of a basis of @xmath196 ( technically by means of a basis of @xmath163 ) , since proposition [ prop : representation ] provides an explicit representation for the desired basis functions of @xmath149 , given by @xmath234 for @xmath70 . more precisely , the construction works as follows : 1 .
we select that pair of functions @xmath110 and @xmath111 , such that holds and which minimizes the term @xmath235 ) } + ||\beta^{(r ) } ||^{2}_{l_{2}([0,1])}.\ ] ] in case of @xmath236 , we have @xmath237 .
2 . let @xmath238)$ ] . depending on @xmath91 , we first choose for the space @xmath239)$ ] a basis @xmath240 , @xmath241 , and then for each pair @xmath242 , @xmath243 , the function @xmath244 as follows : * _ case @xmath245 _ : the functions @xmath240 , @xmath241 , are the b - splines of @xmath239)$ ] , and @xmath246 .
* _ case @xmath247 @xmath216 : _ let @xmath144 with the index @xmath248 be the root of @xmath91 .
the functions @xmath240 , @xmath249 , are the b - splines of @xmath250)$ ] and the function @xmath251 is one of the b - splines of @xmath252)$ ] with the property @xmath253 .
the function @xmath244 is given by @xmath254 * _ case @xmath247 @xmath255 : _ let @xmath256 with the indices @xmath257 and @xmath40 be the two roots of @xmath91 . the functions @xmath240 , @xmath258 , are the b - splines of @xmath250)$ ] , the function @xmath259 is one of the b - splines of @xmath252)$ ] with the property @xmath260 , and the function @xmath251 is one of the b - splines of @xmath261)$ ] with the property @xmath262 .
the function @xmath244 is given by @xmath263 + then each pair @xmath242 , @xmath264 , determines a basis function @xmath265 via representation , i.e. @xmath266 for @xmath70 . in case of @xmath267 , the representation of @xmath265 simplifies to @xmath268 except for the index @xmath269 if @xmath216 and for the index @xmath270 or @xmath269 if @xmath255 . in case of @xmath271 , the representation of @xmath265 even simplifies to @xmath272 3 .
let @xmath273)$ ] , and let @xmath274 , @xmath275 , be the b - spline basis functions of the space @xmath276)$ ] . for each pair @xmath277 , @xmath275 , a basis function @xmath278 is defined via representation , i.e. @xmath279 then the collection of isogeometric functions @xmath280 forms a basis of the space @xmath149 .
the basis functions @xmath265 and @xmath278 have small local supports which are comparable with the supports of the basis functions considered in @xcite , which were constructed for the special case of bilinear geometry mappings @xmath56 and @xmath57 with @xmath225 , @xmath226 , @xmath227 , @xmath8 and @xmath281 for @xmath282 .
our selection of the functions @xmath110 and @xmath111 , as described above , is of course only one possibility .
it would be even possible to choose for each function @xmath265 a different pair of functions @xmath110 and @xmath111 , if desired .
in addition , in case of @xmath202 and @xmath216 , the choice of the functions @xmath110 and @xmath111 satisfying @xmath283 , where @xmath144 with the index @xmath284 is a root of @xmath91 , would also lead for this case to the simplified representation for all functions @xmath265 .
[ ex : functions ] we consider the as @xmath5 two - patch geometry @xmath56 , @xmath285^{2})$ ] , which is shown in fig .
[ fig : ex_functions ] and is given by the control points @xmath286
@xmath287 @xmath288 @xmath289 and @xmath290 @xmath291
@xmath292 @xmath293 the corresponding functions @xmath89 , @xmath90 and @xmath91 are given by means of equations - and selecting the function @xmath294 as @xmath295 which leads to @xmath296 respectively .
the minimization of leads to @xmath297 clearly , we have @xmath298^{2})$ ] with @xmath299 , and we obtain for these selections of the geometry mappings @xmath56 and @xmath57 that @xmath300 below , let @xmath301 be the uniform knot vector @xmath302 fig . [ fig : ex_functions ] shows the graphs of the resulting basis functions @xmath265 and @xmath278 of the space @xmath149 for @xmath303 . [ cols="^,^,^ " , ]
we represent the spline functions @xmath304 , @xmath70 , for the previous constructed isogeometric basis functions @xmath75 of the space @xmath149 as a linear combination of the tensor - product b - splines @xmath305 .
thereby , these linear factors ( i.e. the b - spline coefficients of the spline functions @xmath71 with respect to the space @xmath22^{2})$ ] ) will be described by means of blossoming . for the sake of simplicity ( especially with respect to notation ) , we will restrict ourselves below to the case @xmath133 or @xmath226 .
note that our framework could be also extended to the remaining cases .
we give a short overview of the concept of blossoming . for more detail
we refer to e.g. @xcite .
given a univariate spline function @xmath306 with the spline representation , there exists a uniquely defined function @xmath307 , called the blossom of @xmath308 , possessing the following properties : * @xmath309 is symmetric , * @xmath309 is multi - affine , and * @xmath310 .
these properties imply ( by the so called dual function property , see @xcite ) @xmath311 and therefore fully determine the blossom , since the value @xmath312 for arbitrary values @xmath313 can be computed by recursively using the convex combinations @xmath314 for @xmath315 , @xmath316 and @xmath317 the concept of blossoming provides a simple way to perform knot insertion , to differentiate a spline function and to multiply two spline functions .
given the spline function @xmath26)$ ] with the blossom @xmath309 .
representing @xmath308 as a spline function in the space @xmath318)$ ] , the corresponding spline control points @xmath319 are given by @xmath320 the spline control points @xmath321 of the derivative of @xmath308 , i.e. @xmath322)$ ] , can be computed as follows : @xmath323 for @xmath324 .
given further the spline function @xmath325)$ ] with the blossom @xmath326 .
let @xmath327 and @xmath328 .
then the spline control points @xmath329 of the product @xmath330)$ ] are given by @xmath331 where the summation runs over all possibilities to split the set @xmath332 into the two disjoint subsets @xmath333 and @xmath334 .
let @xmath335 , @xmath336 and @xmath337 .
we denote by @xmath338 , @xmath339 , and @xmath340 , @xmath341 , for @xmath70 the spline functions @xmath342 , and @xmath343 , respectively .
in addition , let @xmath344 be the vector of functions , where @xmath345 as well as @xmath346 and let @xmath347 be the vector of functions , where @xmath348 as well as @xmath349 then there exists a matrix @xmath350 such that @xmath351 the matrix @xmath352 for @xmath70 has a block structure , resulting in an equation of the form @xmath353 where @xmath354 , @xmath355 and @xmath356 .
for each row the single entries of the matrix @xmath352 , @xmath70 , provide the b - spline coefficients of the corresponding spline function @xmath338 or @xmath340 with respect to the spline space @xmath22^{2})$ ] .
the following lemma provides the entries of the single matrices @xmath357 , @xmath358 and @xmath359 : [ lem : matrices ] we denote by @xmath360 the blossom of the b - spline @xmath361 . for a linear function
@xmath362 \rightarrow { \mathbb r}$ ] with the bzier representation @xmath363 , \quad w_{0},w_{1 } \in { \mathbb r},\ ] ] we define the matrix @xmath364 given by @xmath365 the matrices @xmath357 , @xmath358 and @xmath359 , compare , are given as follows ( depending on @xmath119 , @xmath120 or @xmath91 ) : * let @xmath271 .
then we have @xmath366 * let @xmath226 . then we have @xmath367 where the entries of the matrices @xmath368 and @xmath369 are given by @xmath370 and @xmath371 respectively . here , @xmath372 is the identity matrix of dimension @xmath373 .
the results follow directly from the concept of blossoming as presented in subsection [ subsec : blossoming ] .
the matrices @xmath357 , @xmath358 and @xmath359 , see , are sparse matrices ( compare example [ ex : matrixp3a1 ] ) . a further possibility to construct the matrices @xmath357 , @xmath358 and @xmath359 , see , is the use of the concept of fitting .
thereby , the @xmath374-th row of the matrices @xmath357 , @xmath358 and @xmath359 , denoted by @xmath375 , @xmath376 and @xmath377 , respectively , are computed by minimizing the terms @xmath378 @xmath379 and @xmath380 respectively , where @xmath381 are the greville abscissa of the b - splines @xmath382 , @xmath383 , of the spline space @xmath384)$ ] .
[ ex : matrixp3a1 ] let @xmath135 and @xmath226 , @xmath385 , @xmath8 , @xmath386 and @xmath281 for @xmath282 .
let @xmath362 \rightarrow { \mathbb r}$ ] be a linear function with the bzier representation .
then the rows @xmath387 , @xmath388 and @xmath389 of the matrices @xmath390 , @xmath391 and @xmath392 from lemma [ lem : matrices ] are given as follows : @xmath393 and @xmath394
we have studied the spaces of @xmath4-smooth isogeometric functions over a special class of two - patch geometries , so - called analysis - suitable @xmath5 ( as @xmath5 ) two - patch parameterizations ( cf .
this class of two - patch geometries is of particular interest , since exactly these geometries allow under certain assumptions @xmath4 isogeometric spaces with optimal approximation properties , see @xcite . more precisely , we have computed the dimension of these @xmath4 spaces and have presented an explicit basis construction .
the resulting basis functions are well conditioned , have small local supports and their spline coefficients can be simply computed by means of blossoming or fitting .
note that the constructed basis interpolates traces and transversal derivatives at the interface .
hence , the basis functions may be negative .
in fact , the functions interpolating the transversal derivative are by construction positive on one side of the interface and negative on the other .
the presented basis can be transformed easily to obtain locally supported basis functions which sum up to one .
however , it is unclear whether or not a non - negative , local partition of unity exists for all as @xmath5 parameterizations
. this will be of interest for future research .
one issue that remains to be studied is the flexibility of as @xmath3 geometries over general multi - patch domains .
the basis construction over two - patch geometries can be applied also to multi - patch configurations except for the basis functions around vertices , where modifications might be necessary .
we are confident that the presented basis representation can be extended to the multi - patch case and used for fitting procedures , that approximate any given geometry with an as @xmath3 parameterization .
the developed basis provides a simple representation that can be implemented in existing iga libraries .
thus , our @xmath4-smooth functions may be used to discretize different fourth - order partial differential equations .
it may be of interest for future studies to perform such simulations and analyze their properties as well as to investigate the class of as @xmath5 parameterizations for volumetric two - patch and multi - patch domains .
the authors wish to thank the anonymous reviewers for their comments that helped to improve the paper .
the first two authors ( m. kapl and g. sangalli ) were partially supported by the european research council through the fp7 erc consolidator grant n.616563 higeom , and by the italian miur through the prin `` metodologie innovative nella modellistica differenziale numerica '' . this support is gratefully acknowledged .
f. auricchio , l. beiro da veiga , a. buffa , c. lovadina , a. reali , and g. sangalli . a fully `` locking - free '' isogeometric approach for plane linear elasticity problems : a stream function formulation . , 197(1):160172 , 2007 .
j. kiendl , y. bazilevs , m .- c .
hsu , r. wchner , and k .- u .
the bending strip method for isogeometric analysis of kirchhoff - love shell structures comprised of multiple patches .
, 199(35):24032416 , 2010 . | we study the dimension and construct a basis for @xmath0-smooth isogeometric function spaces over two - patch domains . in this context ,
an isogeometric function is a function defined on a b - spline domain , whose graph surface also has a b - spline representation . we consider constructions along one interface between two patches .
we restrict ourselves to a special case of planar b - spline patches of bidegree @xmath1 with @xmath2 , so - called analysis - suitable @xmath3 geometries , which are derived from a specific geometric continuity condition .
this class of two - patch geometries is exactly the one which allows , under certain additional assumptions , @xmath0 isogeometric spaces with optimal approximation properties ( cf .
@xcite ) . such spaces are of interest when solving numerically fourth - order pde problems , such as the biharmonic equation , using the isogeometric method .
in particular , we analyze the dimension of the @xmath0-smooth isogeometric space and present an explicit representation for a basis of this space .
both the dimension of the space and the basis functions along the common interface depend on the considered two - patch parameterization .
such an explicit , geometry dependent basis construction is important for an efficient implementation of the isogeometric method .
the stability of the constructed basis is numerically confirmed for an example configuration .
isogeometric analysis , analysis - suitable @xmath3 geometries , @xmath0 smooth isogeometric functions , geometric continuity |
in this paper , all graphs considered are simple , finite and undirected .
we refer to the book @xcite for undefined notation and terminology in graph theory .
a path in an edge - colored graph is a _ proper path _ if any two adjacent edges differ in color .
an edge - colored graph is _ proper connected _ if any two vertices of the graph are connected by a proper path of the graph . for a connected graph @xmath3 , the _ proper connection number _ of @xmath3 , denoted by @xmath5 ,
is defined as the smallest number of colors required to make @xmath3 proper connected .
note that @xmath17 if and only if @xmath3 is a complete graph .
the concept of @xmath5 was first introduced by borozan et al .
@xcite and has been well - studied recently .
we refer the reader to @xcite for more details . as a natural counterpart of the concept of proper connection ,
the concept of proper vertex connection was introduced by the authors @xcite .
a path in a vertex - colored graph is a _ vertex - proper path _ if any two internal adjacent vertices on the path differ in color .
a vertex - colored graph is _ proper vertex connected _ if any two vertices of the graph are connected by a vertex - proper path of the graph . for a connected graph @xmath3 ,
the _ proper vertex connection number _ of @xmath3 , denoted by @xmath6 , is defined as the smallest number of colors required to make @xmath3 proper vertex connected .
especially , set @xmath18 for a complete graph @xmath3 .
moreover , we have @xmath19 if @xmath3 is a noncomplete graph .
actually , the concepts of the proper connection and proper vertex connection were motivated from the concepts of the rainbow connection and rainbow vertex connection . for details about them
we refer to a book @xcite and a survey paper @xcite . here
we only state the concept of the total rainbow connection of graphs , which was introduced by liu et al . @xcite and also studied in @xcite .
a graph is _ total - colored _ if all the edges and vertices of the graph are colored .
a path in a total - colored graph is a _ total rainbow path _ if all the edges and internal vertices on the path differ in color .
a total - colored graph is _ total rainbow connected _ if any two vertices of the graph are connected by a total rainbow path of the graph . for a connected graph @xmath3 ,
the _ total rainbow connection number _ of @xmath3 , denoted by @xmath7 , is defined as the smallest number of colors required to make @xmath3 total rainbow connected .
motivated by the concept of the total rainbow connection , now for the proper connection and proper vertex connection we introduce the concept of the total proper connection .
a path in a total - colored graph is a _ total proper path _
if @xmath0 any two adjacent edges on the path differ in color , @xmath1 any two internal adjacent vertices on the path differ in color , and @xmath2 any internal vertex of the path differs in color from its incident edges on the path .
a total - colored graph is _ total proper connected _ if any two vertices of the graph are connected by a total proper path of the graph . for a connected graph @xmath3 , the _ total proper connection number _ of @xmath3 , denoted by @xmath4 ,
is defined as the smallest number of colors required to make @xmath3 total proper connected .
it is easy to obtain that @xmath20 if and only if @xmath3 is a complete graph , and @xmath21 if @xmath3 is not complete .
moreover , @xmath22 we can also extend the definition of the total proper connection to that of the total proper @xmath23-connection @xmath24 in a similar way as the definitions of the proper @xmath23-connection @xmath25 , proper vertex @xmath23-connection @xmath26 and total rainbow @xmath23-connection @xmath27 , which were introduced by borozan et al . in @xcite , the present authors in @xcite and liu et al . in @xcite , respectively .
however , one can see that when @xmath23 is larger very little have been known .
almost all known results are on the case for @xmath28 .
so , in this paper we only focus our attention on the total proper connection @xmath4 of graphs , i.e. , @xmath24 for the case @xmath28 .
the rest of this paper is organized as follows : in section @xmath9 , we mainly determine the value of @xmath4 for some special graphs , and moreover , we present some preliminary results . in section @xmath29 , we obtain that @xmath8 for any @xmath9-connected graph @xmath3 and give examples to show that the upper bound @xmath10 is sharp . for general graphs
, we also obtain an upper bound for @xmath4 . in section @xmath10
, we prove that @xmath11 for a connected graph @xmath3 with order @xmath12 and minimum degree @xmath13 . in section @xmath30
, we compare @xmath4 with @xmath6 and @xmath5 , respectively , and obtain that @xmath14 for any nontrivial connected graph @xmath3 , and that @xmath4 and @xmath5 can differ by @xmath15 for @xmath16 .
in this section , we present some preliminary results on the total proper connection number and determine the value of @xmath4 when @xmath3 is a nontrivial tree , a complete bipartite graph and a complete multipartite graph .
[ pro1 ] if @xmath3 is a nontrivial connected graph and @xmath31 is a connected spanning subgraph of @xmath3 , then @xmath32 . in particular , @xmath33 for every spanning tree @xmath34 of @xmath3 .
[ pro2 ] let @xmath3 be a connected graph of order @xmath35 that contains a bridge . if @xmath36 is the maximum number of bridges incident with a single vertex in @xmath3 , then @xmath37 .
let @xmath38 denote the maximum degree of a connected graph @xmath3 .
we have the following . [ thm1 ] if @xmath34 is a tree of order @xmath35 , then @xmath39 .
since each edge in @xmath34 is a bridge , we have @xmath40 by proposition [ pro2 ] .
now we just need to show that @xmath41 .
let @xmath42 be the vertex with maximum degree @xmath43 and @xmath44 denote its neighborhood .
take the vertex @xmath42 as the root of @xmath34 .
define a total - coloring @xmath45 of @xmath34 with @xmath46 colors in the following way : let @xmath47 be a vertex in @xmath34 . if @xmath48 , color @xmath0 @xmath42 and its incident edges with distinct colors from @xmath49 , and @xmath1 @xmath50 with the color from @xmath51 for @xmath52 .
if @xmath53 , there exists a father of @xmath47 , say @xmath54 .
let @xmath55 denote the neighborhood of @xmath47 .
color the edges @xmath56 with distinct colors from @xmath57 , and the vertex @xmath58 with the color from @xmath59 for @xmath60 .
for any two vertices @xmath61 and @xmath62 in @xmath34 , let @xmath63 be a path from @xmath64 to @xmath42 , where @xmath65 . next we shall show that there is a total proper path @xmath66 between @xmath61 and @xmath62 .
if @xmath67 and @xmath68 are edge - disjoint , then @xmath69 ; otherwise , we walk from @xmath61 along @xmath67 to the earliest common vertex , say @xmath70 , and then switch to @xmath68 and walk to @xmath62 , i.e. , @xmath71 .
thus , @xmath41 , and therefore , @xmath39 .
the consequence below is immediate from proposition [ pro1 ] and theorem [ thm1 ] .
[ cor1 ] for a nontrivial connected graph @xmath3 , @xmath72 a _ hamiltonian path _ in a graph @xmath3 is a path containing every vertex of @xmath3 and a graph having a hamiltonian path is a _
traceable graph_. we get the following result . [ cor2 ] if @xmath3 is a traceable graph that is not complete , then @xmath73 .
let @xmath74 denote a complete bipartite graph , where @xmath75 .
clearly , @xmath76 and @xmath77 if @xmath78 . for @xmath79
, we have the result below .
[ thm12 ] for @xmath80 ,
we have @xmath81 . let the bipartition of @xmath74 be @xmath82 and @xmath83 , where @xmath84 and @xmath85 .
since @xmath74 is not complete , it suffices to show that @xmath86 .
now we divide our discussion into two cases . * case 1 . * @xmath87 .
we first give a total - coloring of @xmath74 with @xmath29 colors .
color @xmath88 the vertex @xmath89 and the edge @xmath90 with color @xmath91 , @xmath1 the vertex @xmath92 and the edge @xmath93 with color @xmath9 , and @xmath2 all the other edges and vertices with color @xmath29 .
then we show that there is a total proper path @xmath66 between any two vertices @xmath94 of @xmath74 .
it is clear that @xmath47 and @xmath42 are total proper connected by an edge if they belong to different parts of the bipartition .
next we consider that @xmath47 and @xmath42 are in the same part of the bipartition . for @xmath95
, we have @xmath96 . for @xmath97 , if one of them is @xmath98 , then @xmath99 ; otherwise , @xmath100 . * case 2 . * @xmath101 .
similarly , we first give a total - coloring of @xmath74 with @xmath29 colors .
color the vertices and edges of the cycle @xmath102 starting from @xmath89 in turn with the colors @xmath103 . for @xmath104 and @xmath105 ,
color @xmath0 @xmath106 with color @xmath91 , @xmath1 @xmath107 with color @xmath9 , and @xmath2 all the other edges and vertices with color @xmath29 .
now we show that there is a total proper path @xmath66 between any two vertices @xmath108 of @xmath74 .
it is clear that @xmath47 and @xmath42 are total proper connected by an edge if they belong to different parts of the bipartition . for @xmath109
, we have @xmath110 . for @xmath111
, we have @xmath112 .
it can be checked that @xmath47 and @xmath42 are total proper connected in all other cases .
therefore , the proof is complete .
since any complete multipartite graph has a spanning complete bipartite subgraph , we obtain the following corollary .
[ cor3 ] if @xmath3 is a complete multipartite graph that is neither a complete graph nor a tree , then @xmath73 .
in this section , we first prove that @xmath8 for any @xmath9-connected graph @xmath3 .
also we show that this upper bound is sharp by presenting a family of a @xmath9-connected graphs .
finally , we state an upper bound of @xmath4 for general graphs .
given a colored path @xmath113 between any two vertices @xmath98 and @xmath114 , we denote by @xmath115 the color of the first edge in the path , i.e. , @xmath116 , and by @xmath117 the last color , i.e. , @xmath118 .
moreover , let @xmath119 be the color of the first internal vertex in the path , i.e. , @xmath120 , and @xmath121 be the last color , i.e. , @xmath122 . if @xmath66 is just the edge @xmath123 , then @xmath124 , and @xmath125 .
[ defn1 ] let @xmath45 be a total - coloring of @xmath3 that makes @xmath3 total proper connected .
we say that @xmath3 has the strong property if for any pair of vertices @xmath126 , there exist two total proper paths @xmath127 between them ( not necessarily disjoint ) such that @xmath88 @xmath128 and @xmath129 for @xmath130 , and @xmath131 both @xmath132 and @xmath133 are @xmath29-sets .
let @xmath3 be a connected graph and @xmath31 be a spanning subgraph of @xmath3 .
we say that @xmath31 is a _ spanning minimally @xmath9-connected subgraph _ of @xmath3 if the removal of any edge from @xmath31 would leave @xmath31 @xmath91-connected .
[ thm2 ] let @xmath3 be a @xmath9-connected graph .
then @xmath8 and there exists a total - coloring of @xmath3 with @xmath10 colors such that @xmath3 has the strong property .
let @xmath134 be a spanning minimally @xmath9-connected subgraph of @xmath3 .
we apply induction on the number of ears in an ear - decomposition of @xmath134 .
the base case is that @xmath134 is simply a cycle @xmath135 .
obviously , @xmath136 and @xmath137 for @xmath138 .
next define a total - coloring @xmath45 of @xmath139 with @xmath10 colors by @xmath140 and @xmath141 clearly , the total - coloring @xmath45 makes @xmath134 have the strong property . in an ear - decomposition of @xmath134 ,
let @xmath66 be the last ear with at least one internal vertex since @xmath134 is assumed to be minimally @xmath9-connected . and denote by @xmath142 the graph after removal of the internal vertices of @xmath66 .
let @xmath47 and @xmath42 be the vertices of @xmath143 and then @xmath144 . by induction hypothesis , there exists a total - coloring of @xmath142 with @xmath10 colors such that @xmath142 is total proper connected with the strong property .
we give such a total - coloring to @xmath142 .
then there exist two total proper paths @xmath67 and @xmath68 from @xmath47 to @xmath42 such that @xmath88 @xmath128 and @xmath129 for @xmath130 , and @xmath131 both @xmath132 and @xmath133 are @xmath29-sets .
let @xmath145 .
color the edge @xmath146 with the color from @xmath147 , and then total - properly color @xmath66 from @xmath47 to @xmath42 so that @xmath148 and @xmath149 . if @xmath150 , it will become clear that this is the easier case , and so we consider the case that @xmath151 in the following . without loss of generality ,
suppose that @xmath152 .
we will show that @xmath134 is total proper connected with the strong property under this coloring .
for any two vertices of @xmath142 , there exist two total proper paths connecting them with the strong property by induction hypothesis . since @xmath153 forms a total proper connected cycle , any two vertices in this cycle also have the desired paths .
assume that @xmath154 and @xmath155 .
next we will show that there are two total proper paths from @xmath156 to @xmath70 with the strong property .
since @xmath157 , there exist two total proper paths @xmath158 and @xmath159 starting at @xmath70 and ending at @xmath47 with the strong property .
analogously , there exist two total proper paths @xmath160 and @xmath161 starting at @xmath70 and ending at @xmath42 with the strong property .
since these paths have the strong property , suppose that @xmath162 and @xmath163 are total proper paths .
if @xmath164 , then @xmath165 and @xmath166 are the desired pair of paths .
thus , assume that @xmath167
. then there exists a total proper walk @xmath168 for some @xmath65 ( suppose @xmath169 ) .
if @xmath170 is a path , then @xmath170 and @xmath171 are the desired two paths .
otherwise , let @xmath172 denote the vertex closest to @xmath70 on @xmath161 which is in @xmath173 .
now consider the path @xmath174 .
if @xmath175 is a total proper path , then @xmath175 and @xmath176 are the desired two paths , and so we suppose that @xmath177 . since @xmath67 and @xmath161 are total proper paths , @xmath178 , @xmath179 and @xmath180 .
then @xmath181 . let @xmath182 and @xmath183 .
obviously , @xmath184 and @xmath185 are two total proper paths .
note that @xmath186 .
thus , @xmath184 and @xmath185 have the strong property .
since @xmath187 by proposition [ pro1 ] , we have @xmath8 and there exists a total - coloring of @xmath3 with 4 colors such that @xmath3 has the strong property .
this completes the proof of theorem [ thm2 ] . in order to show that the bound obtained in theorem [ thm2 ]
is sharp , we give a family of @xmath9-connected graphs @xmath3 with @xmath188 ( see figure [ fig.1 . ] ) .
[ pro3 ] let @xmath3 be the graph obtained from an even cycle by adding two ears which are as long as their interrupting segments respectively , such that each segment has @xmath189 ( @xmath190 ) edges
. then @xmath188 .
[ 0.8]-connected graph with @xmath188.,title="fig : " ] before proving proposition [ pro3 ] , we give the following fact .
[ fact1 ] let @xmath135 . if there exists a total - coloring of @xmath139 with three colors such that there are two total proper paths @xmath191 and @xmath192 where @xmath193 and @xmath194 , then @xmath195 .
_ proof of proposition 3 : _ since @xmath8 by theorem [ thm2 ] , we just need to prove that @xmath196 .
assume that there is a total - coloring of @xmath3 with @xmath29 colors such that @xmath3 is total proper connected .
label the segments and some vertices of @xmath3 as in figure [ fig.1 .
] , where @xmath197 and @xmath198 are the neighbours of the vertex @xmath199 for @xmath200 .
firstly , we shall show that the segments @xmath201 and @xmath202 are two total proper paths .
if one of them is not , say @xmath201 , then there is no total proper path in @xmath201 from @xmath92 to @xmath203 or from @xmath204 to @xmath205 ( say from @xmath92 to @xmath203 ) .
hence there exists a total proper path through @xmath202 connecting @xmath92 and @xmath203 , suppose @xmath206 ( this assumption , as opposed to using any of @xmath207 or @xmath208 , does not lose any generality ) .
next we consider the total proper path between @xmath89 and @xmath209 .
then there must exist a total proper path using the segments @xmath210 or @xmath211 . if there is a total proper path @xmath212 , then @xmath213 .
thus the total proper path between @xmath214 and @xmath209 is unique , i.e. , @xmath215 , and then @xmath216 .
however , we can not find a total proper path from @xmath203 to @xmath217 , a contradiction .
if there is a total proper path @xmath218 , then @xmath216 .
thus the total proper path connecting @xmath203 and @xmath217 is unique , i.e. , @xmath219 .
then @xmath220 and @xmath221 are two total proper paths in @xmath222 which is an even cycle of length @xmath223 , which contradicts fact [ fact1 ] .
hence there is no total proper path from @xmath89 to @xmath209 , a contradiction .
therefore , the segments @xmath201 and @xmath202 are two total proper paths . secondly , we will show that at least one of @xmath224 or @xmath207 must be total proper ( and similarly , at least one of @xmath225 or @xmath208 ) .
suppose both @xmath224 and @xmath207 are not total proper .
then @xmath89 and @xmath92 are total proper connected by a path through @xmath225 or @xmath208 , say @xmath226 .
however , we can not find a total proper path connecting @xmath89 and @xmath209 in a similar discussion above , which is impossible .
thus , suppose @xmath224 and @xmath225 are total proper without loss of generality . finally , we know that at least one of the paths @xmath227 and @xmath228 must be not total proper by fact [ fact1 ] .
as we have shown , the only place which we can not go through is at the intersections , and so assume that the path @xmath229 is not total proper , where @xmath230 and @xmath231 . in the following , we consider the total proper path @xmath66 from @xmath232 to @xmath233 and divide our discussion into two cases : * case 1 . *
@xmath66 is @xmath234 or @xmath235 . between @xmath232 and @xmath209
, there must exist a total proper path @xmath67 . if @xmath67 is @xmath236 , then @xmath237 .
hence , there is only one total proper path @xmath215 from @xmath214 to @xmath209 ; otherwise @xmath226 and @xmath238 are two total proper paths in the cycle @xmath239 for a contradiction .
then it follows that @xmath240 .
similarly , we can deduce that there is no total proper path connecting @xmath203 and @xmath217 , which is impossible . if @xmath67 is @xmath241 , then @xmath240 . in a similar discussion
, we obtain that the total proper path from @xmath203 to @xmath217 is unique , i.e. , @xmath219
. then @xmath220 and @xmath221 are two total proper paths in @xmath222 , a contradiction . if @xmath67 is @xmath242 or @xmath243 , then @xmath226 and @xmath238 are two total proper paths in @xmath239 , which again contradicts fact [ fact1 ] . * case 2 .
* @xmath66 is @xmath244 .
consider the total proper path @xmath68 from @xmath232 to @xmath245 , where @xmath246 .
if @xmath68 is @xmath247 , then we can prove that this subcase could not happen in a similar way as case 1 . if @xmath68 is @xmath248 , then @xmath249 .
from @xmath205 to @xmath250 , there is only one total proper path @xmath251 since we can not go through @xmath247 .
however , @xmath252 and @xmath253 are two total proper paths in @xmath254 for a contradiction .
the proof is thus complete .
remember that for a @xmath9-connected graph @xmath3 , we have that the proper connection number @xmath255 ; see @xcite .
but , if we consider a @xmath9-connected bipartite graph @xmath3 , then we have that @xmath256 .
that means that the bipartite property can lower down the number of color by 1 .
however , from proposition 3 we see that the bipartite property can not play a role in general to lower down the number of colors for the total proper connection number , since the graphs in proposition 3 are bipartite but their total proper connection numbers reach the upper bound @xmath10 . finally , we prove an upper bound of @xmath4 for general graphs .
[ thm3 ] let @xmath3 be a connected graph and @xmath257 denote the maximum degree of a vertex which is an endpoint of a bridge in @xmath3
. then @xmath258 if @xmath259 and @xmath260 otherwise . in order to prove theorem [ thm3 ] , we need a lemma below .
let @xmath261 denote the set of colors presented on the vertex @xmath42 and edges incident to @xmath42 .
[ lem1 ] let @xmath31 be a graph obtained from a block @xmath262 with @xmath263 by adding @xmath264 nontrivial blocks and @xmath265 pendant edges at @xmath50 for @xmath266 .
consider @xmath257 as the maximum degree of a vertex which is an endpoint of a bridge in @xmath31 .
then @xmath267 .
let @xmath268 and @xmath269 .
we give a total - coloring @xmath45 of @xmath31 using @xmath224 as follows .
* step 1 . *
if @xmath262 is a trivial block , then we give a total - coloring with 3 colors to @xmath262 such that @xmath270 are different from each other ; otherwise , we give a total - coloring with 4 colors to @xmath262 that makes it have the strong property by theorem [ thm2 ] .
let @xmath271 modulo @xmath23 for @xmath266 .
* step 2 . * for @xmath266 , if @xmath272 , then we give a total - coloring with 4 colors from @xmath273 to each uncolored nontrivial block at @xmath50 , denoted by @xmath274 , that makes each of them have the strong property by theorem [ thm2 ] ; afterwards if @xmath275 , then color @xmath276 uncolored pendant edges at @xmath50 , denoted by @xmath277 , with distinct colors from @xmath278 and then color each pendant vertex @xmath279 using @xmath280 for @xmath281 .
next we show that @xmath31 is total proper connected under the coloring @xmath45 .
if @xmath262 is a nontrivial block , then each pair of the vertices in @xmath262 has two total proper paths between them with the strong property .
it will become clear that this is the easier case so we consider the case that @xmath262 is a trivial block .
let @xmath47 and @xmath282 be two vertices of @xmath31 .
it is obvious that there exists a total proper path connecting them if both belong to the same block .
suppose that @xmath272 and @xmath275 for @xmath130 .
if @xmath283 and @xmath284 , then there exist two paths @xmath158 and @xmath159 from @xmath47 to @xmath98 with the strong property . we know that @xmath285 is a total proper path for some @xmath286 ( suppose @xmath287 )
similarly , there exists a total proper path @xmath288 from @xmath282 to @xmath98 where @xmath289 is a total proper path connecting @xmath282 and @xmath290 .
thus , we can find a total proper path @xmath291 between @xmath47 and @xmath282 . if @xmath292 and @xmath293 , then @xmath294 is a total proper path under the coloring @xmath45 .
for the other cases , it can be checked that there exists a total proper path connecting @xmath47 and @xmath282 in a similar way .
therefore , @xmath295 .
now we are ready to prove theorem [ thm3 ] .
_ proof of theorem 4 : _ let @xmath296 be the blocks of @xmath3 and @xmath297 denote the block graph of @xmath3 with vertex set @xmath298 .
now , we consider a breadth - first search tree ( bfs - tree ) @xmath34 of @xmath297 with root @xmath299 and suppose that the blocks have an order @xmath296 .
let @xmath300 and @xmath269
. we will give a total - coloring @xmath45 using @xmath224 in the following .
we give a total - coloring to @xmath299 and its neighbor blocks of @xmath3 in a similar way as in lemma [ lem1 ] .
then we can get that @xmath3 is total proper connected if there are no more blocks in @xmath3 . hence , suppose that there are uncolored blocks in @xmath3 .
we extend our coloring from @xmath299 in a breadth first search way until there is no more blocks in @xmath3 , i.e. , if @xmath301 has uncolored neighbor blocks , we give a total - coloring to its uncolored neighbor blocks of @xmath3 in a similar way as * step 2 * ; otherwise , consider @xmath302 .
now we prove that @xmath3 is total proper connected .
let @xmath47 and @xmath282 be two vertices in @xmath3 .
it is obvious that there exists a total proper path between them if both belong to the same block .
suppose that @xmath303 and @xmath304 .
let @xmath66 denote the path from @xmath301 to @xmath305 in the bfs - tree @xmath34 .
then we can find a total proper path from @xmath47 to @xmath282 traversing the blocks on @xmath66 under the coloring @xmath45 .
therefore , @xmath258 if @xmath259 and @xmath260 otherwise .
in this section , we prove the following result concerning the minimum degree .
[ thm7 ] let @xmath3 be a connected graph of order @xmath12 with minimum degree @xmath13 , then @xmath11
. given a graph @xmath3 , a set @xmath306 is called a _ two - step dominating set _ of @xmath3 if every vertex in @xmath3 which is not dominated by @xmath202 has a neighbor that is dominated by @xmath202 .
moreover , a two - step dominating set @xmath202 is called a _ two - way two - step dominating set _ if @xmath307 every pendant vertex of @xmath3 is included in @xmath202 , and @xmath308 every vertex in @xmath309 has at least two neighbors in @xmath310 , where @xmath311 denotes the set of all vertices at distance exactly @xmath23 from @xmath202 .
further , if @xmath312 $ ] is connected , @xmath202 is called a _ connected two - way two - step dominating set _ of @xmath3 .
[ lem2]@xcite every connected graph @xmath3 of order @xmath138 and minimum degree @xmath13 has a connected two - way two - step dominating set @xmath202 of size at most @xmath313 .
_ proof of theorem 5 : _ the proof goes similarly as that of the main result in @xcite by li et al .
we are given a connected graph @xmath3 of order @xmath12 with minimum degree @xmath13 .
the assertion can be easily verified for @xmath314 and so suppose @xmath315 .
let @xmath202 denote a connected two - way two - step dominating set of @xmath3 and @xmath316 .
then we have @xmath317 by lemma [ lem2 ] .
let @xmath318 is a neighbor of @xmath156 in @xmath319 for @xmath320 and @xmath321 is a neighbor of @xmath70 in @xmath322 for @xmath323 .
* for each vertex @xmath323 , its neighbors in @xmath310 has at least one common neighbor in @xmath202 , i.e. , @xmath324 .
[ 0.7 ] of @xmath3.,title="fig : " ] we consider a spanning subgraph @xmath325 $ ] of @xmath3 ( see figure [ fig.2 . ] , where @xmath326 denotes the union of @xmath327 graphs each of which is isomorphic to the graph @xmath0 and similarly for @xmath328 and @xmath329 ) .
next , we give a total - coloring @xmath45 to @xmath330 using @xmath331 . for
the edges and vertices of @xmath312 $ ] , let @xmath34 be a spanning tree of @xmath312 $ ] .
then by theorem 1 , @xmath34 can be total - colored using @xmath332 such that for each edge @xmath333 , the colors of @xmath108 and @xmath334 are different from each other .
we color @xmath34 in such a way and the edges of @xmath312\backslash t$ ] with any used colors ( denote this coloring of @xmath312 $ ] by @xmath335 ) . for the other edges and vertices in @xmath330 ,
color them as depicted in figure [ fig.2 . ] .
since each pair of vertices @xmath336 has a total proper path @xmath66 connecting them such that @xmath337 and @xmath338 , it suffices to show that @xmath330 is total proper connected in the assumption that @xmath339 .
take any two vertices @xmath47 and @xmath42 in @xmath340 .
if @xmath341 , then @xmath47 has a neighbor @xmath54 in @xmath310 and similarly @xmath42 has a neighbor @xmath342 in @xmath310 .
hence , if @xmath343 , @xmath344 is a total proper path ; otherwise , @xmath345 is a total proper path where @xmath346 is another neighbor of @xmath47 in @xmath310 .
it is easy to check that @xmath47 and @xmath42 are total proper connected in all other cases
. * case 2 .
* there exists one vertex @xmath323 whose neighbors in @xmath310 has no common neighbors in @xmath202 , i.e. , @xmath347 . [ 0.7 ] of @xmath3.,title="fig : " ] we consider a spanning subgraph @xmath330 of @xmath3 ( see figure [ fig.3 . ] ) .
next , we give a total - coloring @xmath45 to @xmath330 using @xmath331 . for the edges and vertices in
@xmath312 $ ] , we use the total - coloring @xmath335 as in case 1 . for any vertex @xmath348 , color @xmath349 with color 1 and @xmath350 with color 2 where @xmath351 and @xmath352 .
and then color @xmath353 with color 2 and @xmath354 with color 1 where @xmath355 and @xmath356 .
for any vertex @xmath357 , color the edges incident to @xmath42 as depicted in figure [ fig.3 . ] .
moreover , we color the vertices of @xmath310 with color 3 and the vertices of @xmath309 with color 4 .
now we show that @xmath330 is total proper connected .
take any two vertices @xmath47 and @xmath42 in @xmath340 . if @xmath341 , there exist two paths @xmath358 and @xmath359 connecting to @xmath202 , where @xmath360 and @xmath361 .
thus , if @xmath362 , @xmath47 and @xmath42 are total proper connected by a path @xmath363 where @xmath66 is a total proper path from @xmath346 to @xmath364 in @xmath312 $ ] ; otherwise , there exists a total proper path connecting @xmath47 and @xmath42 in a similar discussion as case 1 .
it can be checked that @xmath47 and @xmath42 are total proper connected in all other cases .
therefore , we have @xmath365 by proposition [ pro1 ] and lemma [ lem2 ] .
let @xmath3 be a nontrivial connected graph .
recall that @xmath366 .
the question we may ask is , how tight are the inequalities @xmath367 and @xmath368 ? by ( * ? ? ?
* proposition 1 and theorem 1 ) , we have that @xmath88 @xmath18 if and only if @xmath3 is a complete graph , @xmath131 @xmath369 if and only if @xmath370 , and @xmath371 @xmath372 if and only if @xmath373 .
note that @xmath20 if and only if @xmath3 is a complete graph , and @xmath21 if @xmath3 is not complete .
thus , it follows that @xmath14 .
next we consider the tightness of the inequality @xmath367 .
observe that @xmath374 if and only if @xmath3 is a complete graph .
proposition [ pro4 ] below shows that there exists an example graph @xmath3 such that @xmath375 .
it can be verified that @xmath377 by ( * ? ? ?
* proposition 3 ) .
thus , it suffices to show that @xmath378 by ineq.@xmath379 .
a @xmath29-coloring of edges and vertices of @xmath3 is shown in figure [ fig.4 . ] to make @xmath3 total proper connected .
hence , we have that @xmath375 .
now we consider the difference between @xmath4 and @xmath5 . if @xmath34 is a tree of order @xmath35 , then @xmath383 by (
* proposition 2.3 ) and @xmath39 by theorem [ thm1 ] .
hence , @xmath384 .
moreover , there exists an example graph depicted as in proposition [ pro3 ] such that @xmath385 since @xmath188 and @xmath256 ( we give a @xmath9-edge - coloring of @xmath3 by coloring alternately the edges of the segments @xmath386 and the cycle @xmath387 ) . however , we have not found any graph @xmath3 such that @xmath4 and @xmath5 can differ by @xmath388 .
thus , we pose the following problem . | a graph is said to be _ total - colored _ if all the edges and the vertices of the graph is colored .
a path in a total - colored graph is a _ total proper path _
if @xmath0 any two adjacent edges on the path differ in color , @xmath1 any two internal adjacent vertices on the path differ in color , and @xmath2 any internal vertex of the path differs in color from its incident edges on the path .
a total - colored graph is called _ total - proper connected _ if any two vertices of the graph are connected by a total proper path of the graph . for a connected graph @xmath3 , the _ total proper connection number _ of @xmath3 , denoted by @xmath4 ,
is defined as the smallest number of colors required to make @xmath3 total - proper connected .
these concepts are inspired by the concepts of proper connection number @xmath5 , proper vertex connection number @xmath6 and total rainbow connection number @xmath7 of a connected graph @xmath3 . in this paper
, we first determine the value of the total proper connection number @xmath4 for some special graphs @xmath3 .
secondly , we obtain that @xmath8 for any @xmath9-connected graph @xmath3 and give examples to show that the upper bound @xmath10 is sharp . for general graphs
, we also obtain an upper bound for @xmath4 .
furthermore , we prove that @xmath11 for a connected graph @xmath3 with order @xmath12 and minimum degree @xmath13 . finally , we compare @xmath4 with @xmath6 and @xmath5 , respectively , and obtain that @xmath14 for any nontrivial connected graph @xmath3 , and that @xmath4 and @xmath5 can differ by @xmath15 for @xmath16 .
* keywords * : total - colored graph , total proper connection , dominating set * ams subject classification 2010 * : 05c15 , 05c40 , 05c69 , 05c75 . |
although it is now generally recognised that the environment experienced by the galaxies during their whole lifetime plays a role as important as the initial conditions of their formation , there are still many open questions . in order to define what is the amplitude and dispersion of a given galaxy property that can be ascribed to `` nature '' , a well characterised reference sample of isolated galaxies
is needed . to be statistically useful the sample must be large enough to allow us to assess environmental effects both as a function of morphological type and luminosity . the motivation of the amiga ( * a*nalysis of the interstellar * m*edium of * i*solated * ga*laxies http://amiga.iaa.es ) project is to identify such a sample of isolated galaxies .
the amiga project adopted the catalogue of isolated galaxies ( cig : * ? ? ?
* ) as a starting point .
the strength of the cig involves its size ( 1050 galaxies ) and its selection with a strong isolation criterion .
previous papers in this series included : 1 ) optical characterisation including derivation of the optical luminosity function , 2 ) morphological revision using poss - ii ( and sdss overlap ) and type - specific olf analysis and 3 ) mid- and far - infrared properties using the iras database .
studies of the radio continuum , , co and h@xmath1 emission properties have also been carried along with a study of the small agn population found in the sample .
this work focuses on a reassessment of the isolation degree for all galaxies in the cig with recession velocities @xmath2 km s@xmath3 .
by separating stars from galaxies ( based on poss - i plates ) , we were able to create a catalogue of more than 53000 neighbour galaxies , lying around the 950 cig galaxies .
the catalogues are available in electronic form at the cds @xcite or from http://amiga.iaa.es . a typical star / galaxy separation parameter plane from a poss - i e image ( cig 714 ) is shown in fig .
[ fig : paramspace714 ] .
the galaxies have a lower surface brightness than the stars and in the @xmath4(area ) vs. magnitude plane , the two classes of objects fall in different loci .
) parameters for the cig and the comparison samples .
the cig galaxies are represented by grey pluses .
the ktg are depicted by green triangles , the hcg by magenta squares and the aco by red dots .
the mean values of each sample are shown by horizontal and vertical lines , following the same colour code .
the catalogues are available at the cds @xcite . ] using the @xmath553000 neighbour galaxies , we calculated continuous parameters of isolation for 950 galaxies in the cig .
we used the local number density and the tidal force estimation to precisely describe the environment of each of the 950 cig galaxies .
the local number density of the neighbour galaxies is calculated by focusing on the vicinity of the isolated galaxy candidates , where the principal perturbers should lie .
an estimation of the local number density , @xmath6 , is found by considering the distance to the @xmath7 nearest neighbour . to probe a local region around the central galaxy , we consider @xmath8 equal to 5 , or less if there are not enough neighbours in the field : @xmath9 with the volume @xmath10 , where @xmath11 ( in @xmath12 ) is the projected distance to the @xmath7 nearest neighbour . in order to provide an estimation of the degree of isolation taking also into account the masses of the neighbours
, we calculated the tidal strength affecting the primary cig galaxies .
the tidal strength per unit mass produced by a neighbour is proportional to @xmath13 , where @xmath14 is the mass of the neighbour , and @xmath15 is its distance from the centre of the primary ( @xmath16 is the extension of the object ) .
we approximated @xmath15 by the projected separation , @xmath17 , at the distance of the cig galaxy .
accordingly , the estimator @xmath18 is defined as the ratio between the tidal force and binding force : figure [ fig : cigktghcgaco ] shows the comparison between the local number density and the tidal strength estimation for the cig and the other catalogues sampling denser environments .
these two isolation parameters are complementary and allow us to have a clear picture of the environment of each cig galaxy considered .
we compared the level of isolation of the galaxies in the cig with galaxies in denser environments : 41 triplets ( ktg * ? ? ? * ) , 34 compact groups ( hcg * ? ? ? * ) and 15 clusters ( aco * ? ? ?
* ) were selected for the comparison .
the two isolation parameters are very well suited to discriminate the isolated galaxies from galaxies more affected by their environments .
a more complete information can be found principally in and .
abell , g. o. 1958 , , 3 , 211 espada , d. , bosma , a. , verdes - montenegro , l. , athanassoula , e. , leon , s. , sulentic , j. , & yun , m. s. 2005 , , 442 , 455 hickson , p. 1982
, , 255 , 382 karachentseva , v. e. 1973 , astrofizicheskie issledovaniia izvestiya spetsialnoj astrofizicheskoj observatorii , 8 , 3 karachentseva , v. e. , karachentsev , i. d. , & shcherbanovskii , a. l. 1979 , astrofizicheskie issledovaniia izvestiya spetsialnoj astrofizicheskoj observatorii , 11 , 3 leon , s. , et al .
2008 , , 485 , 475 lisenfeld , u. , et al .
2007 , , 462 , 507 sabater , j. , leon , s. , verdes - montenegro , l. , lisenfeld , u. , sulentic , j. , & verley , s. 2008 , , 486 , 73 sulentic , j. w. , et al .
2006 , , 449 , 937 verdes - montenegro , l. , sulentic , j. , lisenfeld , u. , leon , s. , espada , d. , garcia , e. , sabater , j. , & verley , s. 2005 , , 436 , 443 verley , s. , et al . 2007a , , 470 , 505 verley , s. , et al .
2007b , , 472 , 121 verley , s. , combes , f. , verdes - montenegro , l. , bergond , g. , & leon , s. 2007c , , 474 , 43 verley , s. , et al .
2007d , vizier online data catalog , 347 , 505 verley , s. , et al . 2007e , vizier online data catalog , 347 , 20121 | the role of the environment on galaxy evolution is still not fully understood . in order to quantify and set limits on the role of nurture one must identify and study a sample of isolated galaxies .
the amiga project `` analysis of the interstellar medium of isolated galaxies '' is doing a multi - wavelength study of a large sample of isolated galaxies in order to examine their interstellar medium and star formation activity .
we processed data for 950 galaxies from the catalogue of isolated galaxies ( cig , karachentseva 1973 ) and evaluated their isolation using an automated star - galaxy classification procedure ( down to @xmath0 ) on large digitised poss - i fields surrounding each isolated galaxy ( within a projected radius of at least 0.5 mpc ) .
we defined , compared and discussed various criteria to quantify the degree of isolation for these galaxies : e.g. karachentseva s revised criterion , local surface density computations , estimation of the external tidal force affecting each isolated galaxy .
we found galaxies violating karachentseva s original criterion , and we defined various subsamples of galaxies according to their degree of isolation .
additionally , we sought for the redshifts of the primary and companion galaxies to access the radial dimension .
we also applied our pipeline to triplets , compact groups and clusters and interpret the isolated galaxy population in light of these control samples . |
the rise of exoplanet characterization since the initial discovery of the transit of planet hd 209458b a decade ago charb00,henry00 has been truly stunning .
the initial steps in atmospheric characterization of hot jupiters came after the realization that the transmission spectra of transiting planets would be diagnostic of the temperature and chemical mixing ratios in these atmospheres ss00,brown01,hubbard01 .
as with perhaps all subfields of astronomy and planetary science , the return on the investment in these myriad observations is amplified when the data sets are compared to models that aim to simulate the conditions in the atmospheres of these planets .
the model atmospheres , whether 1d or 3d , aim to predict the temperature structure , chemical mixing ratios , and wavelength dependent opacity , as a function of height .
a comparison of transmission spectra data with models can enable constraints on the mixing ratios of atomic and molecular absorbers [ e.g.][]charb02,fortney03,tinetti07,swain08,sing09b , madhu09,desert09,burrows10 .
of course it then directly follows that constraints on these atmospheres are _ model dependent _ , as it is likely that the choices that one makes in constructing a model atmosphere affect the calculated spectrum . in this exciting field ,
the great difficulty in obtaining high signal - to - noise observations , along with the somewhat unconstrained nature of current atmosphere models , can make interpretation difficult .
this was again typified by the recent works of stevenson10 and beaulieu10 , who , based on separate data reductions and model fits , disagree on the probable mixing ratio of methane in the atmosphere of gj 436b , which we dicuss below . at the forefront of exoplanet characterization ,
detections of lower mass objects have expanded the reservoir of constraints to parameter space beyond hot jupiters , broadening the scope of planetary characterization .
the premier hot neptune , gj 436b a 22.6 @xmath6 planet orbiting an m2.5 star , was detected via radial velocity reflex motion by butler04 , and later revisited by maness07 .
the first photometric detection of transits were obtained by gillon07a , providing the missing link needed to determine the mass and in turn the bulk density . gillon07b and deming07 refined system parameters with a _
transit lightcurve at 8 @xmath3 m , and deming07 and demory07 reported a detection of the planet s secondary eclipse at 8 @xmath3 m as well .
a recent attempt at transit characterization was pont09 , who probed the 1.4 @xmath3 m water band using nicmos on board the _ hubble space telescope_. this was the first attempt at a multi - wavelength transmission spectrum obtained for gj 436b .
ground based efforts have yielded h- and k - band radius measurements alonso08,caceres09 , in addition to radius measurements from epoxi obtained by ballard09 in the 0.35 - 1.0 @xmath3 m range . using irac on - board _ spitzer _
, beaulieu10 very recently obtained transit depth measurements at 3.6 , 4.5 and 8 @xmath3 m . in order to study
the dayside of the planet stevenson10 obtained secondary eclipse measurements in six _
bandpasses , from 3.6 to 24 @xmath3 m .
they and madhu10 interpret their results as providing evidence of thermochemical disequilibrium in gj 436b s dayside atmosphere .
the fits by madhu10 to the stevenson10 observations are quite interesting in the mixing ratios of ch@xmath1 , h@xmath0o , co , and co@xmath0 that they require .
they postulate a metal - rich atmosphere , in line with our understanding of uranus and neptune . however , compared to equilibrium chemistry , ch@xmath1 is strongly depleted , while h@xmath0o is also depleted , and co and co@xmath0 are strongly enhanced .
higher order hydrocarbon molecules , e.g. , c@xmath0h@xmath7 or hcn , were not considered in their fits , even though these molecules are the first products resulting from methane destruction via photolysis moses05 or reduction zahnle09b , and are strong absorbers in the near and mid infrared .
when suggesting that a low ch@xmath1 abundance may be due to the photolysis and/or vertical mixing , these higher - order hydrocarbon molecules should be included .
alternatively , beaulieu10 interpret the stevenson10 data ( some re - reduced by their team ) , as well as primary transit data , as potentially indicating a methane - rich atmosphere with a temperature inversion .
this only further strengthens the points that differences between data reduction methods and models clearly impacts conclusions .
the coming of the _ james webb space telescope _ will bring the next advancement in understanding exoplanet atmospheres and exoplanetary atmosphere modeling , as it will help to reduce the under - constrained nature of current models . for a planet like gj 436b where there are already signs from photometry that the atmospheric chemistry is complex
, the spectral capabilities of _ jwst _ will enhance our understanding , such as the identification of more complex chemical composition regimes and temperature structures .
understanding the integration times and applicable wavelength bands needed to resolve spectral features of particular atmospheres is key for planning future science initiatives with _
gj 436b is a particularly interesting planet because it is the transiting neptune - class object that orbits the brightest parent star .
this means it may long remain the best studied extrasolar neptune - class object .
given the poor signal - to - noise of transmission spectroscopy obtained with _
pont09 , and the uncertainties that arise from using a small number of wide photometric bands strung together as spectra , it may well fall to _ jwst _ to enable robust atmospheric characterization .
deriving accurate transit depths can be a complex task .
problems with _ spitzer _
irac primary transit observations include the issue of variable stellar fluxes between transit visits , due to starspots . in the most common situation where a planet does not occult spots directly ,
the unocculted parts of the stellar surface appear less bright , due to the spots .
compared to a stellar surface that lacks spots , one would derive a deeper transit depth [ see , e.g.][]pont08,agol10 .
there is also the difficulty of extracting the transit depth with great accuracy , which yields different groups to find different transit depths , with the same data sets [ e.g.,][]beaulieu08 , desert09 .
however , given that the interpretation of transmission spectroscopy is in principle less sensitive to the atmospheric pressure - temperature profile than day - side emission spectroscopy , transmission spectroscopy may lead to the most robust determinations of atmospheric abundances . here
we examine the transmission spectrum signatures of a variety of atmospheric chemistries , including those from thermochemical equilibrium calculations , those favored by stevenson10 , and those that include abundant higher order hydrocarbons not considered by these authors , guided by results from zahnle09b .
we combine these models with a detailed simulation of the nirspec and miri low resolution spectrograph instruments that will be aboard _ jwst _ , for a realistic simulation of what further knowledge we may gain for this important planet . in 2 , we make simple comparisons between our models , analytic relations , and other published work , in order to establish the robustness of derived atmospheric parameters , and to validate our code .
we find large differences between our work and that of g. tinetti and collaborators . in 3 ,
we describe model transmission spectra for gj 436b while investigating various chemistries . in
4 we present _ jwst _ simulations of gj 436b transmission spectra , and discuss the possibilities for identifying important atmospheric signatures in the near future . in 5
we discuss implications of our transmission spectra models as well as address future endeavors for atmospheric characterization via theoretical work .
we model the transmission spectrum of planets using a descendent of the code first described in hubbard01 . here
we ignore the effects of refraction and a glow of photons around the planet s limb due to rayleigh scattering , both of which hubbard01 found to be negligible for close - in planets . later works using this code included fortney03 , which investigated simple two - dimensional models of the atmosphere of hd 209458b , but included one planet - wide _ p - t _ profile , with 2d changes in the atmospheric opacities .
fortney05c examined the possible effects of cloud opacity for the slant viewing geometry appropriate for transits .
we refer the reader to fortney10 , which described the code in some detail , and extended our treatment to 3d planetary atmosphere models . in either 1d or 3d , atmospheric _
p - t _ profiles are lain atop an opaque atmosphere at a reference pressure of either 1 or 10 bar . the radius at this pressure level is adjusted to yield the best fit to observations . along 1000 light ray paths through the atmosphere parallel to the star - planet - observer axis ,
the local atmospheric density and opacity are each typically sampled at 1000 points along each ray for absorption , the wavelength dependent cross - section is calculated based on contributions from a variety of atoms and molecules .
the abundances can be based either on local chemical equilibrium at a given atmospheric _
point lodders02,lodders06,lodders09 , or they can be arbitrary . the rayleigh scattering cross - section is described in fortney10 , and in practice for our h@xmath0/he dominated atmospheres , we find a cross - section of 1.645e-24 @xmath8 molecule@xmath9 at 450 nm , and scale by @xmath10 at other wavelengths . here
we define the wavelength - dependent transit radius as the radius where the total slant optical depth reaches 0.56 , following the results of lecavelier08b[]hubbard01,fortney03 we used the profile of the slant optical depth vs. radius , set atop an opaque circle , to generate synthetic images from which we calculated the amount of stellar flux blocked by the planet .
we then defined a larger opaque circle whose cross - sectional area blocked this same amount of light as the planetary model
the radius of this larger circle yielded our transit radius .
more recently we have chosen the tau=0.56 level as the transit radius , simply for computation ease , but we find fine agreement between the two methods , as did lecavelier08b .
generally , for simple tests we find excellent agreement between our recent work and that of t. barman [ e.g.][]barman07 , who use a method similar to our current one , as well as with e. miller - ricci kempton [ e.g.][]millerricci09a .
] , who use a method similar to our former one .
opacities generally used in the field of exoplanet atmospheres are discussed in detail in sharp07 and freedman08 .
freedman08 outline the opacities that we use in modeling the atmospheres of hot jupiters , other jupiter - class planets , neptune - class planets , and brown dwarfs [ e.g.][]fortney08a , fortney08b , saumon06,cushing08 .
we will not repeat the discussions in that paper , but we will touch on the issues of water opacity and higher - order hydrocarbons in turn .
it is well established theoretically burrows97,marley99,ss00,sudar00,barman01 that water vapor opacity is the dominant infrared opacity source in warm giant planet atmospheres .
there is also inescapable observational evidence that this is true for brown dwarfs [ e.g.][]kirkpatrick05 , and this clearly appears to be true for hot jupiters as well [ e.g.][]swain08,grillmair08 .
comprehensive _ ab initio _ calculations of line lists of hundreds of millions of lines for h@xmath0o have been tabulated by , for instance , partridge97 and by barber06 .
both tabulations are widely used .
one of us ( r. s. freedman ) has done extensive comparisons of these two particular line lists at the temperatures of interest for planets and brown dwarfs , those below 2500 k , and these differences are described in freedman08 at being `` slight . '' in the detailed fits of m.s .
marley and collaborators to l- and t - type brown dwarfs , there is no hint that the partridge97 database is insufficient to match the high signal - to - noise medium - resolution nir and mid - ir spectra that have been achieved for scores of objects [ e.g.][]cushing08,stephens09 . in
we show calculated absorption cross - sections at 1500 k and 1 mbar .
clearly the partridge97 and barber06 line lists are nearly identical in this pressure / temperature regime .
if transmission spectra calculated by two different atmosphere codes differ ( see [ others ] ) , the choice between these two water line lists can not be an important contributing factor .
we can validate the predictions of the transmission spectrum code by turning to previous work .
in particular , lecavelier08a have shown that the relation between absorption cross - section , mixing ratio , atmospheric temperature structure , and transit radius can be treated analytically .
there is a particuarly straighforward relation for the wavelength - dependent transit radius for an atmosphere that obeys a few simple constraints .
these constraints are an isothermal temperature structure , a constant gravitational acceleration with height , and an opacity cross section @xmath11 that varies as @xmath12 where @xmath11 and @xmath13 are the wavelength dependent cross - section , and a reference cross - section , respectively , and @xmath14 and @xmath15 are the wavelength and a reference wavelength , respectively . given these constraints ,
the planet s radius can be written as @xmath16 where @xmath17 is the transit radius , @xmath14 is the wavelength , @xmath18 is boltzmann s constant , @xmath19 is the temperature , @xmath3 is the mean molecular mass , @xmath20 is the surface gravity , and @xmath21 is the scale height . for a pure rayleigh scattering atmosphere , @xmath22 . in figure [ rayl ]
we present an isothermal model at @xmath23k , @xmath24 m s@xmath25 , and @xmath26 , with all opacity turned off , save rayleigh scattering .
we find a model slope that is within 1% of the analytic relation from eq .
( [ eqdr ] ) , which we regard as excellent agreement .
in [ opac ] we described our implementation of the water vapor line list of partridge97 .
as a further test of the transmission code we can isolate specific wavelength regions where water opacity closely obeys the @xmath27-relation from eq . ( [ eqalpha ] ) . in particular , in figure [ waterfit]_a _ we show the absorption cross - section vs. wavelength at 1500 k and 10 mbar .
we have over - plotted fits for @xmath27 in three spectral regions .
the bluest and reddest wavelength ranges have a large negative slope , while the middle wavelength range has a positive slope .
if our transmission spectrum model is working correctly , we should be able to match the transit radius slope of @xmath28 from equation ( [ eqdr ] ) , for an atmosphere with a constant gravity and scale height , with water vapor being the only opacity source .
we choose 1500 k and the surface gravity of hd 209458b , 980 cm s@xmath25 .
this model is plotted in figure [ waterfit]_b_. one can readily see that over the three defined wavelength ranges that our model matches the analytic relation .
our choice of an isothermal hd 209458b - like model was based on models presented by g. tinetti and collaborators in a recent paper by beaulieu09 .
these authors used _ spitzer _
irac observations to measure the transit depth in 4 bandpasses from 3 to 10 @xmath3 m .
their nice model fit , compared to the data ( shown in their figure 10 ) , allowed the authors to assert that water vapor was the main absorber in that atmosphere . this may well be true .
however , as shown in fortney10 , our own hd 209458b models were not able to reproduce the large _ variation _ in absorption depths .
as discussed in fortney10 , we are generally unable to match the much larger variation in transit radius of the models of tinetti and collaborators [ e.g.][]tinetti07,tinetti10 .
although our two groups use different water opacity databases ( tinetii et al . use the bt2 list ) it does not appear that can be a contributing factor . to help sort out this issue
, we became interested in simple tests .
figure 9 of beaulieu09 additionally shows transmission spectra for isothermal model atmospheres of hd 209458b at 1500 , 2000 , and 2500 k , with opacity due only to water vapor ( with a mixing ratio of @xmath29 , the same value we use here ) .
we compare our 1500 k model to that of beaulieu09 , as well as the analytic relations , in figure [ compare ] .
( the model from beaulieu09 was obtained using a data extraction software package . )
the differences are large . here
we are able to match the analytic relations , while the model from beaulieu09 can not .
it is not immediately clear what causes these dramatic differences between the two models .
the large differences remain at 2000 k and 2500 k as well . as shown in figure [ water ] , differences in water opacity databases can not be a culprit .
differences in abundances can generally not be a reason either , as a higher ( lower ) water abundance would move the transit radius up ( down ) at all wavelengths .
also , that issue was eliminated for the simple test presented here .
we are left in the position of identifying what we believe is a problem , with tinetti et al .
models , but we are not in a position to speculate as to its cause .
we only dwell on this issue at length because it is the match of models to observations that allows for the identification of absorption features , and the determination of the mixing ratios of specific components .
we are left to doubt the validity of the tinetti transmission models , and the derivations of atmospheric abundances from some of the papers in which those models were used .
we further wish to stress that at this time we have only compared to transmission spectrum models , and not to day - side emission spectrum models .
certainly the comparison between models is an area in need of future work .
an interesting avenue would be to model the transmission spectrum of solar system planets , such as earth palle09 or saturn nicholson06 , which we will pursue in the near future .
our description of our methods complete , we can now turn our application to gj 436b .
gj 436b is a relatively small planet with a bulk density similar to neptune gillon07a , torres08 .
it may be mostly composed of fluid water , but a layer of h - he dominated atmosphere is clearly needed to account for the observed radius gillon07a , adams08,nettelmann10 .
gj 436b is one of the least - irradiated transiting planets , which makes its atmosphere cooler than many other well - studied planets .
based on thermochemical equilibrium models , the low temperatures suggests the dominant carbon - bearing molecule in the gaseous envelope is methane spiegel10a , madhu10,lewis10 .
the recent stevenson10 secondary eclipse measurements from warm _ spitzer _
, however , have been interpreted by these authors as suggesting otherwise . an atmosphere who s
carbon chemistry is methane - dominate would tend to yield a small flux ratio in the 3.6 @xmath3 m band , with more flux in the 4.5 @xmath3 m band .
however , stevenson10 report a strong detection at 3.6 @xmath3 m , and non - detection at 4.5 @xmath3 m .
this could be indicative of extensive methane depletion , which would allow one to probe deeply , to hotter gas , in the 3.6 @xmath3 m band , while a large mixing ratio for co and co@xmath0 , which both absorb strongly in the 4.5 @xmath3 m band , could suppress flux in this bandpass stevenson10 , madhu10 .
large co and co@xmath0 abundances have been shown to be indicators of high metallicity lodders02,visscher06,zahnle09a . in the favored scenarios ,
the co / ch@xmath1 and co@xmath0/ch@xmath1 mixing ratios are enhanced due to vertical mixing from hotter , ch@xmath1-poor gas below , along with the photochemical destruction of ch@xmath1 by incident uv photons .
if relatively abundant ch@xmath1 is indeed destroyed , this will give rise to a whole host of higher - order hydrocarbons , which is well understood for our solar system s giant planets [ e.g.][]moses05 .
detailed chemical models , including photochemistry and vertical mixing , were applied to cool transiting planets by zahnle09b , which predicted the formation of abundant c@xmath0h@xmath0 , c@xmath0h@xmath1 , c@xmath0h@xmath2 , and hcn .
more recently , the theory of a methane - poor atmosphere for gj 436b was challenged by beaulieu10 , who , with a combination of modestly different eclipse depths at 3.6 and 4.5 @xmath3 m , compute emission and transmission spectra models that allow abundant methane . since
a goal of the work is to explore the prospects for broad - wavelength - coverage spectra with _
jwst _ , we choose to explore a diverse set of model atmospheres .
our choices are guided by predictions from equilibrium chemistry , nonequilibrium chemistry , and fits to published spectra of the planet .
cross sections for the main molecules expected from thermochemical equilibrium are shown in figure [ eq_x ] .
there are certainly a rich number of molecular bands , particularly at wavelengths blueward of 5 @xmath3 m , where nircam and nirspec will be sensitive . in figure
[ double ] , we present gj 436b model transmission spectra , using these cross sections .
one model is for a 30@xmath30 solar metallicity atmosphere in thermochemical equilibrium ( black ) , as well as two models ( red and blue ) with abundances taken directly from the `` red model '' and `` blue model '' in table 2 of stevenson10 .
these mixing ratios are somewhat depleted in water , strongly depleted in methane , and co and co@xmath0-rich compared to our equilibrium calculations . in all cases , we include the atomic sodium and potassium abundances derived from the equilibrium model . for the three chemistry cases , we use two different _ p t _ profiles . in the upper panel ,
a hotter dayside average " profile ( named `` 2 pi '' ) is used , which simulates inefficient day night energy redistribution .
the bottom panel uses a cooler planet - wide average profile ( named `` 4 pi '' ) [ e.g.,][]fortney05 .
the two _ p t _ profiles are plotted in figure [ pt ] .
these models ( which we will refer to as models ` a ' through ` f ' ) , as well as models discussed in the following sections are described in table [ table1 ] . at the current time we do not investigate time - variable temperature structure or abundances .
the only published model of the atmospheric dynamics of gj 436b show little variability lewis10 , and there is no observational evidence as yet .
for the two _ p t _ profile cases , we adjust the 10 bar radius of all models so that the radii align in the optical .
the main chemical difference between the hotter and cooler profiles is that the hotter model yields a larger mixing ratio of na and k , yielding stronger features in the optical .
a structural difference is that the larger scale height in the warmer model leads to modestly larger changes in radius as a function of wavelength . compared to our equilibrium model ,
the two models from stevenson10 ( models ` c ' through ` f ' in table [ table1 ] ) show considerably smaller radii in the near and mid - infrared , which is predominantly due to the smaller mixing ratio of water .
therefore , a clear probe of the water mixing ratio is the near - ir radius , compared to the optical radius , where water vapor is much less important . as expected
the differences between the equilibrium model and the red / blue models are greatest in wavelengths where methane and water are the dominant absorbers , and smallest where co and co@xmath0 are the dominant absorbers .
stevenson10 and madhu10 suggested that there could be evidence for a very low ch@xmath1 mixing ratio in the atmosphere of gj 436b , due to a combination of vertical mixing and photolysis of ch@xmath1 .
however , they did not investigate how the methane - derived nonequilibrium chemical products may affect the spectra of the planet s atmosphere .
recently , zahnle09b have investigated nonequilibrium carbon for isothermal `` warm jupiter '' atmospheres ( @xmath31 k ) at a range of metallicities .
these temperatures are similar to those suggested for gj 436b .
zahnle09b find that methane is sustained at higher regions in the atmosphere , water is more stable , and oh and h@xmath0 quickly combine to form h@xmath0o and h. this effectively increases the c to o ratio leading to increased abundances of molecules such as hcn , c@xmath0h@xmath0 , c@xmath0h@xmath1 , and c@xmath0h@xmath2 . given the uncertainties in modeling this chemistry , and the wide range of atmosphere models that are consistent with the gj 436b data to date , our next aim is to explore transmission spectra with a range of nonequilibrium chemical products , guided by the results of zahnle09b .
the cross sections for these first generation products of methane , which are strong absorbers in the infrared , can be found in figure [ photo_x ] .
data for c@xmath0h@xmath0 , c@xmath0h@xmath1 and c@xmath0h@xmath2 , which are likely incomplete , are from the hitran database , while that for hcn is from the calculations of harris08 .
the cross - sections are similar in magnitude to those of the equilibrium chemistry products ( figure [ eq_x ] ) .
hcn and c@xmath0h@xmath0 have prominent features at 1.5 , 3.3 , 7 , and 13 @xmath3 m , c@xmath0h@xmath1 has a significant impact to absorption at 9.5 @xmath3 m , but c@xmath0h@xmath2 has very little effect , with most of its opacity residing between 10 and 13 @xmath3 m . in figure [ photo ]
we explore transmission spectra including the opacities with these nonequilibrium products .
we begin with model ` e ' in from table [ table1 ] , also shown as the red model from the top panel of figure [ double ] .
this model uses a dayside average _ p
t _ profile with the same best fit mixing ratios as the `` red model '' from table 2 of stevenson10 .
in addition to the original parameters of model ` e ' , we include absorption from hcn , c@xmath0h@xmath0 , c@xmath0h@xmath1 and c@xmath0h@xmath2 , with mixing ratios of @xmath32 , @xmath33 , @xmath34 and @xmath35 respectively , shown as the cyan model in figure [ photo ] ( model ` g ' from table [ table1 ] ) .
we derive these chemical abundances based on zahnle09b , where they depict mixing ratios as a function of height and @xmath36 ( eddy diffusion coefficient ) for an atmosphere at 1000 k. the mixing ratios chosen are similar to what is expected for an atmosphere that is vigorously mixing ( @xmath37 ) .
it is worthwhile to explore some variations on this model .
one variation removes the blanketing effects of hydrogen cyanide ( hcn ) , shown in purple ( model ` h ' from table [ table1 ] ) .
this model illustrates the difference in the width of the features at 3.3 , 7 and 13 @xmath3 m , where hcn and c@xmath0h@xmath0 have overlapping opacity . when the features are thick , they are dominated by hcn opacity .
when we remove hcn , we see thinner features , indicating that c@xmath0h@xmath0 dominates over hcn .
this characteristic may allow further constraints on the mixing ratios of cool transiting planet atmospheres . in green
is a model with these nonequilibrium products absent at pressures below 10 mbar ( model ` i ' from table [ table1 ] ) .
the green model illustrates the condition of modest vertical mixing , with @xmath36 on the order of @xmath38 @xmath8 s@xmath9 , similar to that favored by madhu10 . in this case ,
nonequilibrium products are not stable higher in the atmosphere than the @xmath39 10 mbar pressure level zahnle09b .
the main absorption features in the infrared will be weakened , as the green model confirms .
the strength of these features could be constraints on vertical mixing and @xmath36 .
absorption from these nonequilibrium products may mask features that would otherwise show an under - abundance of water in the transmission spectrum , as figure [ photo ] shows .
the red model , which generally shows smaller radii in the near- and mid - ir than in the optical , instead shows significantly larger radii when the nonequilibrium products are introduced .
spectroscopy , rather than photometry , will be key towards disentangling the effects of various molecules on the transmission spectrum . in figure
[ data ] , we explore model fits to recent _ spitzer _ irac data for 3.6 , 4.5 , and 8 @xmath3 m beaulieu10 , as well as data from epoxi in the 0.35 - 1.0 @xmath3 m range ballard09 , _ hst nicmos _ in the 1.1 - 1.9 @xmath3 m range pont09 , ground based h - band alonso08 , and ground based k - band caceres09 .
we plot our models in high resolution , and as band - averages , where appropriate .
looking at the optical and near infrared , the best match is to the model that uses abundances from equilibrium chemistry , shown in grey ( model ` a ' from table [ table1 ] ) . in the mid - infrared ,
for the _ spitzer _ irac data we see the same trend that we found in fortney10 : in comparison to beaulieu08 data for hd 189733b and to beaulieu09 data for hd 209458b , our models can not match the larger _ amplitude _ of the features the implied dramatic change in absorption depth as a function of wavelength .
the models of g. tinetti and collaborators , used in the beaulieu10 paper , do fit the observation reasonably well .
however , as discussed in 2.3 , we find that the tinetti et al . models overestimate the amplitude of absorption features .
it is clear that our models do not agree with the large peak to trough variation of spectra required by beaulieu10 to fit their data . if the methods employed by beaulieu08,beaulieu09,beaulieu10 are correct , and the error bars are not underestimated ,
the results for all of these planets imply that dramatic revisions to models of these atmospheres are needed .
given the uncertainties in the reduction of irac transit data , it may be up to _ jwst _ to confirm speculations about the molecules present at the terminator of gj 436b .
we now evaluate the observability of the differences in the gj 436b models by comparing them through the eyes of _ jwst_. we have developed a code that simulates _ jwst _ spectra by computing the number of photons detected using a model of the host star , a transmission model of the planet , and estimates of the total efficiency ( detected electrons per incident photon ) at each wavelength for the various _ jwst _ dispersive spectroscopic modes .
noise is also modeled and added to the simulated spectra .
the star gj 436 is relatively bright over the @xmath40 @xmath3 m spectral region , and the transmission models predict that gj 436b will have absorption features from many species over this wavelength range . therefore we illustrate the model similarities and differences with simulations of _ jwst _ observations using the nirspec @xmath41 spectroscopic mode over this spectral range .
the double - pass caf@xmath42 prism used in this mode provides spectroscopic resolution varying from @xmath43 at @xmath44 @xmath3 m to @xmath45 at @xmath46 @xmath3 m . we approximate this with a fourth order polynomial fit over the @xmath40 @xmath3 m range , and we assume that the prism has total transmission efficiency of 0.81 after two passes .
we estimate that the optical efficiency of nirspec s 14 reflective surfaces @xcite is approximately 0.58 over @xmath47 @xmath3 m , consistent with the values calculated by the nirspec team ( p. jakobsen , private communication 2003 ) and assumed by @xcite after removing the grating blaze function .
we adopt a quantum efficiency of 0.75 across the entire @xmath48 @xmath3 m spectral range , consistent with the nirspec detector requirements @xcite .
the telescope is estimated to have total reflectivity of 0.9 across this wavelength range .
total efficiency was modeled to decrease linearly by a factor of 2.0 as wavelength decreases from 1.0 to 0.7 @xmath3 m , driven mostly by reduction in reflectivity in the 14 reflective nirspec surfaces . like @xcite , we assume there will be no losses from the 16 wide entrance slit and that the only significant noise sources are photon noise and systematic noise due to small guiding errors during exposures
. photon noise is simulated by adding poisson noise appropriate for the number of detected photo - electrons in each resolution bin .
we adopt the systematic noise value of @xmath49 estimated by @xcite .
we do assume that this noise is gaussian in its distribution although @xcite found that it was somewhat non - gaussian .
even with high precision jwst instruments , we will suffer systematic noise at these modest but significant levels . in the `` 1-(in - transit / star ) '' computation , any additional natural or instrumental noise occurring at frequencies greater than the inverse of the transit observation period will impact the extracted spectrum .
this simulation program was coded in c , and it uses the public domain ranlib package for simulating photon noise and gaussian systematic noise . a high fidelity stellar model of the gj 436 host star was not readily available
, so we used a model of gj 411 which has m2 v spectral type , similar to the m2.5 v type of gj 436 . using our simulation code , we re - binned the @xcite @xmath50 model of gj 411 to the instrumental resolution of the _ jwst _ nirspec prism at each wavelength interval over @xmath51 @xmath3 m .
next , our code computed the number of stellar photons from this binned flux , reducing it by the ratio of the squared model planet radius divided by the squared stellar radius ( assumed to be @xmath52 cm ) at each wavelength .
we used a distance of 10.2 pc to gj 436 and an integration time of 1800 s for these calculations .
this integration time is @xmath53 shorter than the 2740 s duration of the transit @xcite .
we used the resultant simulated in - transit spectrum and the simulated stellar spectrum of equal integration time to compute the absorption depth at each wavelength , 1 - ( in - transit / star ) .
this is plotted for gj 436b models in figure [ jwst ] .
models ` a ' , ` c ' , and ` e ' ( shown in black , blue , and red respectively ) are plotted in the top panel of figure [ jwst ] . these are _ jwst _ simulations of the models in the top panel of figure [ double ] . models ` e ' , ` g ' , ` h ' , and ` i ' are plotted in the bottom panel in red , purple , cyan and green respectively , and are _ jwst _ simulations of the models in figure [ photo ] .
the absorption features are labeled here for clarity . in particular , in the top panel the differences between the water rich model ( black ) and water poor models ( red and blue ) are readily apparent , as is the strong co@xmath0 feature at 4.3 @xmath3 m and co feature at 4.5 @xmath3 m . in the bottom panel ,
the absorption features due to nonequilibrium hcn and c@xmath0h@xmath0 are clearly apparent , as is ch@xmath1 absorption from 3 - 4 @xmath3 m .
the prospects for detailed characterization of this planet , and others with _ jwst _ , is good .
given the high brightness of gj 436b ( @xmath54 or @xmath55 mag ) , it is likely that its observation will require use of a detector subarray that is smaller in the dispersion direction than the @xmath56 pixel length of the complete @xmath57 @xmath3 m @xmath5 spectrum @xcite . therefore , we find that acquiring the entire spectrum shown at the signal - to - noise in figure [ jwst ] may require 2 or 3 transits
. these observations may be best acquired in the higher resolution @xmath50 mode for stars as bright as gj 436 . at @xmath58 @xmath3
m miri low resolution spectrograph ( lrs ; r @xmath39 100 ) observations were also simulated for these models in a similar fashion using details of lrs models and actual measured performance .
these simulations are shown in figure [ ir ] and are another independent way of looking at the effects of nonequilibrium chemistry in the atmosphere of gj 436b . in the top panel ( models ` a ' ,
` c ' and ` e ' are shown as black , blue , and red respectively ) , water is the main opacity source in this wavelength range .
the bottom panel ( showing models ` e ' , ` g ' , ` h ' , and ` i ' as red , purple , cyan , and green respectively ) shows the clear distinction of hcn and c@xmath0h@xmath0 between the different models shown at 7 @xmath3 m .
absorption from c@xmath0h@xmath1 is shown at 9.5 @xmath3 m as well .
observations using miri will be the only way to probe the 9.5 @xmath3 m feature created by the presence of c@xmath0h@xmath1 .
the simulations were also made for a total integration time of 30 m in transit and 30 m on the star .
the flux of gj 436 is less than 1 jy over this wavelength range , faint enough for its entire @xmath58 @xmath3 m spectrum to be acquired simultaneously , which is an advantage miri will have over nirspec observations that require multiple transits to obtain the full spectrum .
the noise included in the spectral simulations and shown in figures [ jwst ] and [ ir ] are likely lower limits to the actual noise recorded in _
jwst _ spectra .
common - mode low frequency noises due to the observatory will likely be eliminated by the differential measurement of the star and planet .
all data for the simulated spectra shown are obtained on time scales of 2 3 eclipses , so any global variation in the star on longer time scales ( i.e. , subsequent transits or eclipses ) will also be removed .
variation of the planet between transits will be recorded , and this could , in principle , limit the usefulness of co - adding multiple spectra .
variations in the _ jwst _ instrumentation , the host star , or the planet on time scales shorter than 2 3 transit or eclipse events will appear in the data as systematic or random noise .
nevertheless , we expect that _ jwst _ should be able to obtain high signal - to - noise exoplanet spectra given that _
spitzer _ was able to obtain signal - to - noise approaching @xmath59 ( e.g. , * ? ? ? * and references therein ) with its older detectors , about a factor of 2 less than we predict for the systematic noise of _ jwst_.
we have explored possible transmission spectra of gj 436b .
we have investigated equilibrium chemistry cases and have compared them to mixing ratios derived to fit recent secondary eclipse measurements in the near infrared .
the differences in these mixing ratio regimes have a significant effect when comparing the corresponding transmission spectra
. the lack of water in some cases will be easily detected by _ jwst _ assuming the spectrum is not dominated by absorption from nonequilibrium chemical products of methane .
for this reason , we have made the first attempt to include opacity from higher order hydrocarbons for close - in giant planet atmosphere models .
specifically , we have included absorption from hcn , c@xmath0h@xmath0 , c@xmath0h@xmath1 , and c@xmath0h@xmath2 .
the special condition that favors production of these molecules the most is the relatively low temperatures found in this planet .
_ hst _ and _ spitzer _ may be limited in the capabilities needed to observe this planet s transmission spectrum . the pont09 near - ir spectrum from _ hubble
_ suffered from instrumental systematics that could not be overcome .
reduction of _ spitzer _ irac transit data suffers from the problem that different groups reducing the same data sets achieve different results , with errors bars that do not overlap beaulieu08,desert09 .
transit depths observed at different epochs allow for stellar vairability to complicate the interpretation .
these observations are _ extremely challenging _
, such that many results for gj 436b , and many of the transiting planets , must be regarded as provisional .
we have shown that _ jwst _ should be able to make important breakthroughs in answering questions about the chemical mixing ratios in the planet s atmosphere .
specifically , the presence of higher order hydrocarbons in the transmission spectrum of this planet would reinforce the recent claim of non - detection of methane in the emission spectrum .
we also may be able to constrain the composition and structure at the planetary limb .
if hydrocarbon chemistry is present , it will be possible to constrain the abundances of these species .
for example , hcn and c@xmath0h@xmath0 have strong absorption overlapping at 1.5 , 3.3 , 7 , and 13 @xmath3 m .
the thickness of the feature in each wavelength region will tell us which molecule is more abundant in the atmosphere .
when there is minimal hcn , the signature of c@xmath0h@xmath0 in these wavelength regions will be characteristically thinner .
this phenomena is depicted for _ jwst _ s nirspec in the lower panel of figure [ jwst ] for the 1.5 and 3.3 @xmath3 m features as well as for the miri lrs in the lower panel of figure [ ir ] at 7 @xmath3 m .
the strength of the features in the infrared produced by these hydrocarbons may also be instrumental in constraining information regarding atmospheric mixing .
weak features may indicate a small eddy diffusion coefficient ( @xmath36 ) , pointing towards weak vertical mixing in the planet s atmosphere , with these molecules forming at higher pressure levels . in order to produce accurate model spectra of an exoplanet , including nonequilibrium chemical products ( due to mixing and photochemistry ) a self consistent code
is needed .
this would ensure that the molecular mixing ratios , and their impact on opacities , are consistent with the atmospheric _
t _ profile .
this alone is difficult , but should be addressed in the future . in principle , all of this should be done in a three - dimensional model .
perhaps someday exquisite observations will warrant such a treatment .
clearly _ jwst _ will be an effective tool for reducing the under - constrained nature of current models , yet the brightness of gj 436 may be a limitation . the nirspec prism mode will be able to observe a larger spectral range at high signal - to - noise with fewer transits for planets orbiting stars that are relatively dimmer than gj 436 .
this issue however , does not pertain to obtaining spectra in the 5 - 10 @xmath3 m range using the miri lrs .
this study is the first in a series of papers that will investigate models of transmission and emission spectra of transiting planets , convolved with our realistic model of _ jwst _ observations . for a complementary approach over a large phase space of _ jwst _ exoplanet observations ,
see belu10 .
the model dependent nature of our understanding of the composition and structure of exoplanet atmospheres ensures that it will always be useful to refine atmosphere models .
thus developing , scrutinizing and refining these models brings us closer to the true nature of these planets . in
2 we validated our model against analytic relations , but found that a similar tinetti et al .
model could not match these same relations .
recent work on including nonequilibrium and photochemical products liang04,zahnle09a , zahnle09b will yield better predictions of chemical mixing ratios .
transmission spectrum models that incorporate the full 3d nature of the planetary atmosphere , which is particularly imporant at the terminator region fortney10,burrows10 will yield a more accurate understanding of the temperature structure and chemical abundances
. the conditions of gj 436b coupled with
_ jwst _ will provide the chance to explore the possibility of a compositionally complicated atmosphere in great detail , expanding our knowledge of planets as unique to their particular circumstance .
we thank jean - michel dsert , david sing , david spiegel , alain lecavelier des etangs , mark marley , heather knutson , and adam burrows for useful discussions , and eliza kempton , travis barman , and nikku madhusudhan for discussions and the sharing of model results .
j. j. f. and m. s. acknowledge the support of the _ spitzer _ theory program and a university affiliated research center ( uarc ) aligned research program ( arp ) grant .
uarc is a partnership between university of california , santa cruz and nasa ames research center .
g. acknowledges support from the jwst nircam instrument , nasa wbs 411672.05.05.02.02 .
, s. , christiansen , j. l. , charbonneau , d. , deming , d. , holman , m. j. , fabrycky , d. , ahearn , m. f. , wellnitz , d. d. , barry , r. k. , kuchner , m. j. , livengood , t. a. , hewagama , t. , sunshine , j. m. , hampton , d. l. , lisse , c. m. , seager , s. , & veverka , j. f. 2010 , , 716 , 1047 , j. , tinetti , g. , kipping , d. m. , ribas , i. , barber , r. j. , y - k .
cho , j. , polichtchouk , i. , tennyson , j. , yurchenko , s. n. , griffith , c. a. , batista , v. , waldmann , i. , miller , s. , carey , s. , mousis , o. , fossey , s. j. , & aylward , a. 2010 , a&a submitted , arxiv:1007.0324 , j. p. , kipping , d. m. , batista , v. , tinetti , g. , ribas , i. , carey , s. , noriega - crespo , j. a. , griffith , c. a. , campanella , g. , dong , s. , tennyson , j. , barber , r. j. , deroo , p. , fossey , s. j. , liang , d. , swain , m. r. , yung , y. , & allard , n. 2009 , mnras in press , arxiv:0909.0185 , b. j. , alexander , d. , brambora , c. k. , derro , r. , engler , c. , fox , o. , garrison , m. b. , henegar , g. , hill , r. j. , johnson , t. , lindler , d. j. , manthripragada , s. s. , marshall , c. , mott , b. , parr , t. m. , roher , w. d. , shakoorzadeh , k. b. , smith , m. , waczynski , a. , wen , y. , wilson , d. , xia - serafino , w. , cabelli , c. , cheng , e. , garnett , j. , loose , m. , zandian , m. , zino , j. , ellis , t. , howe , b. , jurado , m. , lee , g. , nieznanski , j. , wallis , p. , york , j. , regan , m. w. , bagnasco , g. , bker , t. , de marchi , g. , ferruit , p. , jakobsen , p. , & strada , p. 2007 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
6690 , society of photo - optical instrumentation engineers ( spie ) conference series , m. , holota , w. , posselt , w. , koehler , j. , melf , m. , bagnasco , g. , & marenaci , p. 2005
, in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
5904 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
j. b. heaney & l. g. burriesci , 185198 across three wavelength ranges .
see equations ( [ eqalpha ] ) and ( [ eqdr ] ) for details .
vertical dotted lines show the wavelength range of validity where we expect the analytic relation to hold .
( _ b _ ) the resulting planetary radius vs. ln(wavelength ) of the model . across the three wavelength ranges , the fit to the linear radius vs. ln(wavelength )
is very good .
again , vertical dotted lines show the range of wavelength validity.[waterfit],title="fig:",scaledwidth=75.0% ] + across three wavelength ranges
. see equations ( [ eqalpha ] ) and ( [ eqdr ] ) for details .
vertical dotted lines show the wavelength range of validity where we expect the analytic relation to hold .
( _ b _ ) the resulting planetary radius vs. ln(wavelength ) of the model . across the three wavelength ranges , the fit to the linear radius vs. ln(wavelength )
is very good .
again , vertical dotted lines show the range of wavelength validity.[waterfit],title="fig:",scaledwidth=75.0% ] | we explore the transmission spectrum of the neptune - class exoplanet gj 436b , including the possibility that its atmospheric opacity is dominated by a variety of nonequilibrium chemical products .
we also validate our transmission code by demonstrating close agreement with analytic models that use only rayleigh scattering or water vapor opacity .
we find broad disagreement with radius variations predicted by another published model . for gj 436b , the relative coolness of the planet s atmosphere ,
along with its implied high metallicity , may make it dissimilar in character compared to hot jupiters . "
some recent observational and modeling efforts suggest low relative abundances of h@xmath0o and ch@xmath1 present in gj 436b s atmosphere , compared to calculations from equilibrium chemistry .
we include these characteristics in our models and examine the effects of absorption from methane - derived higher order hydrocarbons . to our knowledge , the effects of these nonequilibrium chemical products on the spectra of close - in giant planets has not previously been investigated . significant absorption from hcn and c@xmath0h@xmath0
are found throughout the infrared , while c@xmath0h@xmath1 and c@xmath0h@xmath2 are less easily seen .
we perform detailed simulations of _ jwst _ observations , including all likely noise sources , and find that we will be able to constrain chemical abundance regimes from this planet s transmission spectrum .
for instance , the width of the features at 1.5 , 3.3 , and 7 @xmath3 m indicates the amount of hcn versus c@xmath0h@xmath0 present
. the nirspec prism mode will be useful due to its large spectral range and the relatively large number of photo - electrons recorded per spectral resolution element .
however , extremely bright host stars like gj 436 may be better observed with a higher spectroscopic resolution mode in order to avoid detector saturation .
we find that observations with the miri low resolution spectrograph should also have high signal - to - noise in the @xmath4 @xmath3 m range due to the brightness of the star and the relatively low spectral resolution ( @xmath5 ) of this mode . |
the hosts of high redshift quasars and radio galaxies provide an important insight into the formation of the most massive galaxies at high redshifts .
in particular , they can be used to study the origin of the black hole mass bulge mass / velocity dispersion relation ( magorrian et al .
1998 , gebhardt et al .
2000 ; ferrarese et al . 2000 ) and its possible explanation in terms of agn feedback on star formation in the host galaxy ( silk & rees 1998 ; king 2003 ) .
dust obscured quasars play a key role in such models , representing the intermediate stage after the black hole has begun accreting gas from its surroundings , but before agn driven winds have cleared gas and dust from the line of sight to the quasar ( fabian 1999 ; li et al .
2006 ) , in a manner similar to that proposed by sanders et al .
( 1988 ) . in order to test these models , it is particularly important to build up samples of dust - obscured quasars , especially at high redshifts , where an increased amount of quasar obscuration is thought to be due to dust in the host galaxy ( e.g. treister et al .
2009 ) , rather than by a nuclear torus as in the orientation - based unified scheme for local seyfert and radio galaxies / quasars ( antonucci 1993 ) . over the past few years
, we have been constructing a sample of mid - infrared selected , dust obscured quasars , building on the work of lacy et al .
( 2007a ) .
this sample includes @xmath5 objects with redshifts past the peak in number density of the normal quasar population at @xmath6 ( richards et al .
2006 ) .
obscured quasars also have an useful practical advantage over normal quasars , namely that the obscuration of the quasar nucleus by dust allows the host galaxy of the quasar to be characterized much more easily . only in these obscured quasars
is it possible to obtain accurate estimates of host galaxy stellar luminosities and galaxy morphologies ( zakamska et al .
2006 ; lacy et al .
2007b ) .
despite their importance , obscured quasars remain a relatively poorly studied population .
large numbers have been found in the sloan digital sky survey ( reyes et al .
2008 ) , but only through mid - infrared selection has it proved possible to find more than a handful of these objects at @xmath7 , where the [ oiii ] emission line is redshifted out of the rest - frame optical band ( martnez - sansigre et a. 2005 ; lacy et al . 2007a ; polleta et al . 2006 ; 2009 ) .
existing observations of high-@xmath4 dust obscured quasars do tend to support a scenario in which a significant fraction of the dusty quasar population is reddened by dust in the host galaxy , rather than a nuclear torus ( martnez - sansigre et al .
2006 ; lacy et al .
2007b ) , but key unknowns in the observational studies of these objects remain .
these include the masses of the molecular gas reservoirs in these objects , and the stellar luminosities of the hosts . from work that has been performed on similar objects , including @xmath6 ultraluminous infrared galaxies ( ulirgs ) selected at 24@xmath8 m with _ spitzer _ ( yan et al .
2010 ) , many of which are agn - dominated , and on a few other type-2 quasars at @xmath9 ( martnez - sansigre et al.2009 ; polletta et al .
2011 ) , it seems that molecular gas is present in such objects , but how the molecular gas content relates to other host galaxy properties is unclear .
we were thus motivated to study the co emission from luminous obscured quasars , since it traces the molecular gas content of the galaxy , and hence the reservoir of molecular gas available for star formation .
in particular , the low-@xmath10 transitions provide the best tracer for the total gas mass ( papadopoulos & ivison 2002 ) .
the low critical density of these transitions allows them to be seen against the cosmic microwave background in gas with density as low as @xmath11 . to do this
, we use the new capabilities of the extended very large array ( evla ) .
the vla has a large collecting area , but has hitherto been restricted through not having receivers in the 30 - 40ghz range , or a correlator capable of correlating a wide bandwidth .
the evla , with ka - band receivers and the wide - band widar correlator provides the opportunity to study these objects in the lowest co ( 1 - 0 ) transition .
when fully commissoned , the widar correlator will be able to correlate an 8ghz bandwidth .
it will thus be capable of measuring redshifts for galaxies whose distances are relatively poorly constrained ( for example , galaxies with only photometric redshifts ) , and for galaxies whose emission features in the optical / near - infrared are obscured by dust . for this initial investigation
, we used the restricted correlator bandwidth available through the open shared risk observation ( osro ) program and objects with well - constrained redshifts with relatively narrow ( @xmath12kms@xmath13 ) emission lines . the combination of these data with newly released archival data from the _ herschel _ telescope allows us to study the gas and dust content of these objects for the first time .
we assume a cosmology with @xmath14 and @xmath15 .
our objects were selected in the mid - infrared using the selection technique of lacy et al .
( 2004 ) , which finds candidate agn and quasars , both obscured and unobscured , by virtue of their red mid - infrared continua .
our obscured quasars comprise both dust - reddened type-1 quasars ( objects with red continua , but broad permitted lines ) , and type-2 quasars , selected from the _ spitzer _ swire survey ( lonsdale et al .
2003 ) and the extragalactic first look survey ( xfls ) ( lacy et al . 2005 ; fadda et al .
2006 ) .
optical spectra of the first 77 objects in this mid - infrared selected sample were published in lacy et al .
( 2007a ) . a further @xmath16 spectra
have now been obtained ( lacy et al .
2012 , in preparation ) . from this list we picked out a pilot sample of two obscured quasars for follow - up with the evla in osro mode , according to the following criteria : @xmath17 , @xmath18 ( to ensure good ras for night time summer observing ) and optical spectroscopic redshifts from multiple emission lines ( figure 1 ) .
unlike most other samples of quasars selected for co follow - up , our objects were not pre - selected for bright far - infrared emission .
the high redshifts of our sources mean that the available optical spectra sample the rest - frame uv . our working definition of a type-2 quasar as an object which lacks broad permitted lines in the rest - frame optical . as only xfls171419.97 + 602724.8 has a rest - frame optical spectrum ( figure 2 ) , we are only sure of this one as being a type-2 quasar by our definition
. however , the spectral energy distribution ( sed ) of sw161117.30 + 541759.2 suggests it is dominated by starlight in the rest - frame optical .
the spectra are discussed in more detail in section 3 .
xfls171419.97 + 602724.8 has had its optical spectrum previously published in lacy et al .
( 2007a ) ( but reproduced here for ease of comparison with the near - infared spectrum presented here ) .
the spectrum of sw161117.30 + 541759.2 was taken using hectospec on the mmt during june 2008 , we used the 270 groove per mm grating , resulting in a spectral resolution of @xmath19 nm . approximately 130 fibers were placed on candidate agn in each setting .
the field was observed for for @xmath20s .
.redshifts and photometry [ cols="<,^,^",options="header " , ] notes : @xmath21 is the fraction of starlight in the model in rest - frame @xmath22-band , @xmath23 is the optical depth in @xmath24-band and @xmath25 is the fractional contribution of the young stellar population in the dual stellar population model .
the limit on @xmath26 is calculated assuming the quasar is accreting at the eddington rate .
the best spectroscopic analogues with similar redshifts as these objects are high redshift radio galaxies ( mccarthy 1993 ) , and type-2 quasars detected in x - ray surveys ( e.g. norman et al . 2002 ; silverman et al .
the ly@xmath27 lines from the radio galaxies are , however typically much stronger relative to the uv emission lines , consistent with the large haloes of ionized gas around these objects being affected by the presence of the radio jets ( reuland et al .
these objects also have similar uv spectra to those of low redshift seyfert-2 galaxies ( e.g. kriss et al .
1992 ; kraemer et al .
1994 ) , and agn selected through lyman - break techniques ( hainline et al .
thus , by analogy with known agn both at high and low redshift we feel secure in assuming that the presence of high ionization , high equivalent width uv emission lines is a reliable tracer of agn activity .
few co ( 1 - 0 ) measurements of normal quasars at similar redshifts have been published .
riechers et al .
( 2006 ) measured co(1 - 0 ) of three @xmath28 quasars using the greenbank and effelsberg telescopes .
their luminosities are significantly higher than those measured for dusty quasars in this paper ( @xmath29 ) , but the quasars are intrinsically more luminous , and were pre - selected for having ultraluminous infrared emission , so form a biased sample . similarly , coppin et al .
( 2008 ) detected co emission in 5/6 optically - luminous , submm detected quasars at @xmath30 , with luminosities of @xmath31 . at higher redshifts , wang et al .
( 2010 ) find slightly lower co luminosities , @xmath32 , in a sample of eight @xmath33 quasars , again pre - selected to have submm continuum detections . at present
it is therefore difficult to say whether the dusty quasars in this paper have significantly more or less molecular gas than their unobscured counterparts , since ours are unbiased towards bright far - infrared luminosities ( but see also klamer et al .
this situation is likely to change rapidly , however , as the evla and alma are fully commissioned over the next 2 - 3 years .
[ [ dusty - quasars - and - the - evolutionary - paradigms - for - massive - galaxies - at - high - redshifts ] ] dusty quasars and the evolutionary paradigms for massive galaxies at high redshifts ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ our understanding of galaxy evolution would be greatly enhanced if we could compare the gas content , stellar masses , star formation rates and space densities in unbiassed samples of different types of massive , high redshift galaxies , in order to constrain evolutionary scenarios .
this would include smgs , warm ulirgs and obscured quasars selected in the mid - infrared or hard x - ray , radio galaxies , normal quasars and , if they exist , the quiescent equivalents of each class .
selection effects on current samples of these populations are strong , however , making it difficult to perform a definitive comparison . in our quasar hosts , there is a selection against objects with low stellar mass due to the requirement for detection in all irac bands , but no pre - selection for high dust masses and star formation rates based on mm / submm continuum detections . in the samples of _ spitzer _ selected ulirgs of yan et al .
( 2010 ) and type-2 quasars of martnez - sansigre et al.(2009 ) , a continuum detection at 1.2 mm ( i.e. a high star formation rate and dust mass ) was a requirement for follow - up observations for co. the smgs with co ( 1 - 0 ) measurements are also restricted to a fairly small sample at present , though this situation is likely to change in the next few months as evla commissioning progresses ( e.g. ivison et al .
nevertheless , we can attempt to piece together a consistent picture with what is currently available .
figure 6 shows our quasar host galaxies compared to other high redshift objects ( including other type-2 quasars ) in the co - luminosity stellar luminosity ( rest - frame @xmath22-band absolute magnitude ) plane .
assuming similar conversions between co luminosity and co mass , and stellar luminosity and stellar mass for the different populations shown here , we can begin to constrain some evolutionary scenarios between these types of objects .
the type-2 quasars vary considerably in co mass .
the two @xmath34 objects from polletta et al .
( 2011 ) have gas masses comparable to smgs , but the @xmath35 objects from this paper and the @xmath36 and @xmath37 objects from martinez - sansigre et al . ( 2009 ) have detections or limits significantly lower than this . with the exception of xfls171419.97 + 602724.8 ,
the type-2 quasars tend to have similar stellar masses to the smgs . making the assumption of conversion of gas to stars with 100% efficiency
, trajectories can be drawn on figure 6 ( such as the large magenta arrow ) to indicate how a galaxy might evolve with time .
( although such a scenario is highly simplistic as feedback effects will prevent the complete conversion of gas to stars , more gas is likely to accumulate too , both from expelled gas cooling and infalling , and from accretion of fresh material . )
this simple scenario would indicate that smgs and co - rich type-2 quasars could evolve into moderately - high stellar mass galaxies , comparable in mass to the hosts of normal type-2 quasars .
however , unless they are able to accrete significantly more gas than is currently present in their molecular reservoirs ( or merge with other massive galaxies ) , they would not be able to evolve into our most massive host galaxy , xfls171419.97 + 602724.8
. it may be significant that , although this object is extremely luminous in the infrared , it has a high dust temperature , perhaps indicating a more evolved system overall , with agn heating dominating the infrared sed .
the @xmath6 ulirgs seem to be a distinct population of relatively low stellar mass objects that could evolve into some of the lower mass ( and lower luminosity ) type-2 quasars .
similarly , the bx and bzk galaxies also look to be distinct from the majority of the smgs and type-2 quasars .
clearly , however , these is a need for larger samples of high redshift galaxies with better controlled selection effects .
these can now be provided using the new generation of micro - jansky depth , @xmath38deg@xmath39 near - infrared surveys such as the _
extragalactic representative volume survey ( mauduit et al . 2011 ) in combination with far - infrared surveys with herschel ( e.g. oliver et al
. 2010 ) and follow - up with evla and the atacama large mm / submm array ( alma ) .
we thank chris carilli for helpful advice on evla observations , hana schumacher for providing a more accurate estimate of the co ( 1 - 0 ) luminosity for one of the high redshift radio - intermediate type-2 quasars , and the referee for comments that significantly improved the manuscript .
the national radio astronomy observatory is a facility of the national science foundation , operated under cooperative agreement by associated universities , inc .
observations reported here were obtained at the mmt observatory , a joint facility of the smithsonian institution and the university of arizona .
mmt time was granted by noao through the telescope system instrumentation program ( tsip ) .
tsip is funded by nsf .
ml and ap were visiting astronomers at the infrared telescope facility , which is operated by the university of hawaii under cooperative agreement no .
nnx-08ae38a with national aeronautics and space administration , science mission directorate , planetary astronomy program .
the nasa / ipac extragalactic database ( ned ) and infrared science archive ( irsa ) are operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
_ herschel _ is an esa space observatory with science instruments provided by european - led principal investigator consortia with significant participation from nasa .
adelman - mccarthy , j. et al .
2008 , apjs , 175 , 297 alexander , d.m .
2008 , aj , 135 , 1968 antonucci , r. 1993 , ara&a , 31 , 473 bell , e.f .
2003 , apj , 586 , 794 bennert , v.n . , auger , m.w .
, treu , t. , woo , j .- h . & malkan , m.a . 2011 , apj letters , submitted ( arxiv:1102.1975 ) benson , a.j . ,
frenk , c.s . ,
lacey , c.g .
c.m . & cole , s. 2003 , apj , 599 , 38 borys , c. , smail , i. , chapman , s.c . ,
blain , a.w . ,
alexander , d.m .
& ivison , r.j .
2005 , apj , 635 , 853 bruzual , g. & charlot , s. 2003 , mnras , 344 , 1000 casey , c.m .
2009 , mnras , 399 , 121 ciliegi , p. et al . 1999 ,
mnras , 302 , 222 condon , j.j .
2003 , aj , 125 , 2411 coppin , k.e.k .
2008 , mnras , 389 , 45 coppin , k. et al . 2010a , apj , 713 , 503 coppin , k.e.k .
2010b , mnras , 407 , l103 croton , d.j .
2006 , mnras , 365 , 11 daddi , e. et al .
2010 , apj , 713 , 686 dasyra , k.m .
2009 , apj , 701 , 1123 de breuck , c. et al .
2010 , apj , 725 , 36 emonts , b.h.c .
2011 , mnras , submitted ( arxiv:1103.5806 ) fadda , d. , januzzi , b.t .
, ford , a. , storrie - lombardi , l.j .
2004 , aj , 128 fabian , a.c . 1999 ,
mnras , 308 , l39 farrah , d. et al .
2008 , apj , 677 , 957 ferrarese , l. & merritt , d. 2000 , apj 539 l9 frayer , d.t .
aj , 131 , 250 garn , t. , green , d.a . , hales , s.e.g . , riley , j.m .
& alexander , p. 2007
, mnras , 376 , 1251 garn , t. , green , d.a . , riley , j.m .
& alexander , p. 2008 , mnras , 383 , 75 gonzlez - solares et al .
2011 , in preparation greve , t.r .
2004 , a&a , 107 , 99 greve , t.r .
2005 , mnras , 359 , 1165 hainline , l.j . ,
blain , a.w .
, smail , i. , alexander , d.m .
, armus , l. , champan , s.c .
& ivison , r.j .
2010a , arxiv:1006.0238 hainline , k.n . ,
shapley , a.e . , greene , j.e . & steidel c.c .
2010b , arxiv:1012.0075 healey , s.e .
2007 , apjs , 171 , 61 hildebrand , r.h . 1983 , qjras , 24 , 267 hiner , k.d . ,
canalizo , g. , lacy , m. , sajina , a. , armus , l. , ridgway , s. & storrie - lombardi , l. 2009 , apj , 706 , 508 ivison , r.j . , smail , i. , papadopoulos , p.p .
, wold , i. , richard , j. , swinbank , a.m. , knieb , j .-
p . & owen , f.n .
2010 , mnras , 404 , 198 ivison , r.j . ,
papadopoulos , p.p .
, smail , i. , greve , t.r . ,
thomson , a.p .
, xilouris , e.m . & chapman , s.c .
2011 , mnras , 412 , 1913 kennicutt , r.c .
1998 , ara&a , 36 , 189 kewley , l.j .
2006 , mnras , 372 , 961 king , a. 2003 , apj , 596 , l27 aj , 136 , 1866 klamer , i. et al .
2005 , apj , 621 , l1 kraemer , s.b . , wu , c .- c . , crenshaw , d.m . & harrington , j.p . 1994 , apj , 435 , 171 kriss , g.a . , davidsen , a.f . ,
blair , w.p .
, ferguson , h.c .
& long , k.s . 1992 ,
apj , 394 , l37 kovcs , a. , chapman , s.c . ,
dowell , c.d . ,
blain , a.w .
, ivison , r.j . ,
smail , i. & phillips t.g .
2006 , apj , 650 , 592 lacy , m. et al .
2004 , apjs , 154 , 166 lacy , m. et al .
2005 , apjs , 161 , 41 lacy , m. , petric , a.o .
, sajina , a. , canalizo , g. , storrie - lombardi , l.j . ,
armus , l. , fadda , d. & marleau , f.r .
2007a , aj , 133 , 186 lacy , m. , sajina , a. , petric , a.o . ,
seymour , n. , canalizo , g. , ridgway , s.e . ,
armus , l. & storrie - lombardi , l.j . 2007b , apj , 669 , l61 lacy , m. et al .
2012 , in preparation li , y. et al .
2007 , apj , 665 , 187 leyshon , g. & eales , s.a .
1998 , mnras , 295 , 10 leyshon , g. , dunlop j.s . & eales , s.a . 1999 ,
mnras , 309 , 474 lonsdale , c.j .
2003 , pasp , 115 , 897 lowe , s.r .
2007 , a&a , 474 , 1093 maraston , c. 2005 , mnras , 362 , 799 marconi , a. & hunt , l.k .
2003 , apj , 589 , l21 mauduit , j.c .
2011 , in preparation marconi , a. , axon , d.j . ,
maiolino , r. , nagao , t. , pastorini , g. , pietrini , p. , robinson , a. & torricelli , g. 2008 , apj , 678 , 693 martnez - sansigre et al .
2005 , nat , 436 , 666 martnez - sansigre et al .
2006 , mnras , 370 , 1479 martnez - sansigre , a. & taylor , a. 2009 , apj , 692 , 964 martnez - sansigre , a. et al .
2009 , apj , 706 , 184 menndez - delmestre , k. et al .
2009 , apj , 699 , 667 nesvadba , n.p.h .
2010 , a&a , 525 , a43 norman , c. et al .
2002 , apj , 571 , 218 oliver , s.j . et al .
2010 , a&a , 518 , l21 papadopoulos , p.p . & ivison , r.j .
2002 , apj , 564 , l9 papovich , c. et al .
2006 , aj , 132 , 231 patnaik , a.r . ,
browne , i.w.a . , wilkinson , p.n .
& wrobel , j.m .
1992 , mnras , 254 , 655 pei , y.c .
1992 , apj , 395 , 130 persson , s.e . , murphy , d.c . ,
krzeminski , w. , roth , m. & rieke , m.j .
1998 , aj , 116 .
2475 polletta , m.c .
2006 , apj , 642 , 673 polletta , m. et al .
2009 , a&a , 492 , 81 polletta , m. et al .
2011 , a&a , in press ( arxiv:1107.2129 ) rayner , j.t .
2003 , pasp , 115 , 362 reuland , m. et al .
2007 , aj , 133 , 2607 reyes , r. et al .
2008 , aj , 136 , 2373 richards , g.t .
2006 , aj , 131 , 2766 riechers , d.a .
2006 , 650 , 604 riechers , d.a .
2010 , 720 , l131 robert , c. , leitherer , c. & heckman , t.m .
1993 , apj , 418 .
749 sajina , a. , scott , d. , dennefeld , m. , dole , h. , lacy , m. & lagache , g. 2006 , mnras , 369 , 939 sanders , d.b . , soifer , b.t .
, elias , j.h . , madore , b.f .
, matthews , k. , neugebauer , g. & scoville , n.z . 1988 , apj , 328 , 35 seymour , n. et al .
2007 , apjs , 171 , 353 shapley , a.e . ,
steidel , c.c . ,
pettini , m. & adelberger , k.l .
2003 , apj , 588 , 65 sibthorpe , b. , ferlet , m. , bendo , g. , papageorgiou , a. and the spire icc 2010,``spire beam model release note , version 1 '' silk , j. & rees , m.j .
1998 , a&a , 331 , l1 silverman , j.d .
2010 , apjs , 191 , 124 stevens , j.a . ,
amure , m. & gear , w.k .
2005 , mnras , 357 , 361 swinbank , a.m. et al .
2006 , mnras , 371 , 465 swinbank , a.m. et al .
2010 , nat , 464 , 733 tacconi , l.j .
2010 , nat , 463 , 781 wang , r. et al .
2010 , apj , 714 , 699 yan , l. et al .
2010 , apj , 714 , 100 wardlow , j.l .
2010 , mnras , submitted ( arxiv:2006.2137 ) wu , y. et al .
2011 , apj , 723 , 895 zakamska , n.l . et al .
2006 , aj , 132 , 1496 | we present optical through radio observations of the host galaxies of two dust obscured , luminous quasars selected in the mid - infrared , at @xmath0 and @xmath1 , including a search for co emission .
our limits on the co luminosities are consistent with these objects having masses of molecular gas @xmath2 , several times less than those of luminous submillimeter - detected galaxies ( smgs ) at comparable redshifts .
their near - infrared spectral energy distributions ( seds ) , however , imply that these galaxies have high stellar masses ( @xmath3 ) the relatively small reservoirs of molecular gas and low dust masses are consistent with them being relatively mature systems at high-@xmath4 . |
one of the early triumphs of quantum mechanics has been to enable one to calculate the discrete energy spectrum and the corresponding stationary states - eigenstates of the quantum hamiltonian - of atoms and molecules , neglecting their interactions with the quantized electromagnetic field .
however , if these interactions are taken into account , stationary states corresponding to discrete energies , save for the groundstate , are absent .
the data of atomic and molecular spectroscopy can be interpreted in terms of the decay of metastable states with energies close to the discrete energies , or eigenvalues , of the non - interacting atoms or molecules .
the decay of these states is accompanied by emission of photons with nearly discrete energies equal to differences between energies of stationary states ; ( bohr s frequency condition ) .
these metastable states are called `` ( quantum ) resonances '' .
their analysis is the subject of this paper : we further develop some key ingredients of the mathematical theory of resonances for the standard model of `` non - relativistic quantum electrodynamics '' ( qed ) and for nelson s model of electrons interacting with quantized ( longitudinal lattice ) vibrations , i.e. , phonons . due to the interactions of electrons with massless field quanta - photons or phonons - the standard techniques to analyze quantum - mechanical resonances developed during the past thirty or so years ( see , e.g. , and references therein )
can not be applied to realistic models of atoms or molecules .
our goal , in this paper , is to modify these techniques to cover the present models . before introducing the models we explain the resonance problem in general terms .
let @xmath0 be a quantum hamiltonian , where @xmath1 is a real parameter called the coupling constant .
assume there is a one - parameter family of unitary transformations , @xmath2 with @xmath3 , s.t .
the family @xmath4 has an analytic continuation in @xmath5 to a disc @xmath6 in the complex plane .
we call such an analytic continuation a _ complex deformation _ of @xmath0 .
we note that , while the essential spectrum of @xmath7 usually changes dramatically under such an analytic continuation , the eigenvalues are locally independent of @xmath8 for suitably chosen deformation transformations @xmath9 , at least when they are isolated . moreover , the real eigenvalues of @xmath7 coincide with the eigenvalues of @xmath0 . the complex eigenvalues of @xmath10 are called the ( quantum @xmath11)resonance eigenvalues - or just resonance eigenvalues - of the hamiltonian @xmath0 . the transformations most commonly used is the group of dilatations of positions and momenta ( see below ) , and the corresponding resonances are sometimes called `` dilatation resonances '' . it is plausible from our definition that resonances - at least for weakly coupled systems ( @xmath1 small ) - are closely related to eigenvalues of @xmath12 but what is their physical significance ?
physically , one thinks of quantum resonances as long - lived metastable states or as `` bumps '' in the scattering cross - section as a function of energy .
the energies and life - times of metastable states are given by the bumps centers and the inverse of the bumps widths .
a known approach to establish such properties is as follows .
let @xmath13 denote the dense linear subspace of @xmath11entire vectors , i.e. , vectors @xmath14 for which the family @xmath15 has an analytic continuation to the entire complex plane .
for such vectors one has the `` combes formula '' @xmath16 if we continue the r.h.s .
analytically , first in @xmath5 and then in @xmath17 then we see that matrix elements , @xmath18 of the resolvent , for @xmath19 and @xmath20 have an analytic continuation in @xmath21 across the essential spectrum of @xmath0 to the `` second riemann sheet '' whenever the resolvent set of the operator @xmath22 contains a part of this essential spectrum . ] .
clearly , eigenvalues of @xmath22 in the lower complex half - plane , @xmath23 , are poles of this analytic continuation , provided these eigenvalues are isolated . in the latter case ,
the metastability property can be established ( at least , for weakly coupled systems ) by using the relation - via the fourier transform - between the propagator and the resolvent , contour deformation and cauchy s theorem ( see @xcite ) .
the `` bumpiness '' of the cross - section can be connected to the resonance poles .
the real and imaginary parts of the resonance eigenvalues give the energy and the rate of decay , or the reciprocal life - time , of the metastable state .
the situation described above is exactly the one encountered in quantum mechanics . in non - relativistic qed and phonon models ,
the resonance eigenvalues are _ not isolated _ ; more precisely , a branch of essential spectrum is attached to every complex eigenvalue of the deformed hamiltonian @xmath7 .
this is due to the fact that photons and phonons are _ massless_. as a result , establishing the property of metastability and the pole structure of the resolvent ( and the related bumpiness of the cross - section ) becomes a challenge . in this paper , we prove , for non - relativistic qed and nelson s model , the metastability property of resonances and characterize them in terms of poles of a meromorphic continuation of the matrix elements of the resolvent on a dense set of vectors .
next , we introduce the models considered in this paper . the hamiltonian of the qed model is defined as @xmath24 where @xmath25 @xmath26 denotes the momentum of the @xmath27 particle and @xmath28 its mass , and @xmath29 is the potential energy of the particle system .
furthermore , @xmath30 denotes the quantized vector potential @xmath31 where @xmath32 @xmath33 is an ultraviolet cut - off that vanishes sufficiently fast at infinity , and @xmath34 are two transverse polarization vectors , i.e. , orthonormal vectors in @xmath35 satisfying @xmath36 moreover , @xmath37 is the photon ( quantized electromagnetic field ) hamiltonian defined as @xmath38 where @xmath39 the operator - valued distributions @xmath40 and @xmath41 are annihilation and creation operators acting on the symmetric fock space @xmath42 over @xmath43 .
they obey the canonical commutation relations @xmath44
= \left [ a_{\lambda}(k ) , a_{\lambda'}(k ' ) \right ] = 0 \quad,\quad \left [ a_{\lambda}(k ) , a_{\lambda'}^*(k ' ) \right ] = \delta ( k - k')\delta _ { \lambda , \lambda'},\ ] ] and @xmath45 where @xmath46 is the vacuum vector . the qed hamiltonian @xmath47 acts on the hilbert space @xmath48 , where @xmath49 is the hilbert space for @xmath50 electrons , e.g. @xmath51 ( neglecting permutation symmetry ) . in ( [ eq : nrqed ] ) ,
zeeman terms coupling the magnetic moments of the electrons to the magnetic field are neglected .
nelson s model describes non - relativistic particles without spin interacting with a scalar , massless boson field .
the hamiltonian of the model acts on @xmath48 , where @xmath42 is the symmetric fock space over @xmath52 and is given by @xmath53 here , @xmath54 denotes an @xmath50-particle schrdinger operator on @xmath49 .
we assume that its spectrum , @xmath55 consists of a sequence of discrete eigenvalues , @xmath56 below some real number @xmath57 called the _ ionization threshold_. for @xmath58 in @xmath59 , we denote by @xmath60 and @xmath61 the usual phonon creation and annihilation operators on @xmath42 .
they are operator - valued distributions obeying the canonical commutation relations @xmath62
= \left [ a(k ) , a(k ' ) \right ] = 0 \quad,\quad \left [ a(k ) , a^*(k ' ) \right ] = \delta ( k - k').\ ] ] the operator associated with the energy of the free boson field , @xmath37 , is given by the expression , except that the operators @xmath60 and @xmath61 now are scalar creation and annihilation operators as given above .
the interaction @xmath63 in @xmath64 is assumed to be of the form @xmath65 where @xmath66 dk.\ ] ] as above , the function @xmath67 denotes an ultraviolet cut - off , and the parameter @xmath68 is assumed to be positive .
next , we state our assumptions on the potential and the ultraviolet cut - off @xmath69 in particular concerning analyticity under dilatations .
* the potential @xmath29 is dilatation analytic , i.e. , the vector - function @xmath70 has an analytic continuation to a small complex disc @xmath71 for some @xmath72 an example of a dilatation - analytic potential @xmath73 is the coulomb potential for @xmath50 electrons and one fixed nucleus located at the origin . for a molecule in the born - oppenheimer approximation ,
the potential @xmath29 is not dilatation - analytic . in this case
, one has to use a more general notion of distortion analyticity ( see @xcite ) , which can be easily accommodated in our analysis . *
the function @xmath33 is dilatation analytic , i.e. , @xmath74 has an analytic continuation from @xmath75 to the disc @xmath6 . for instance
, we can choose @xmath76 for some fixed , arbitrarily large ultraviolet cut - off @xmath77 let @xmath0 denote either @xmath47 or @xmath78 . to define quantum resonances for the hamiltonian @xmath79 we use the dilatations of electron positions and of photon momenta : @xmath80 where @xmath5 is a real parameter .
such dilatations are represented by the one - parameter group of unitary operators , @xmath81 on the total hilbert space @xmath82 of the system .
this is one of the most important examples of the deformation groups mentioned above on page 25 . ] . following the general prescription , we define , for @xmath83 the family of unitarily equivalent hamiltonians @xmath84 by the above assumptions on @xmath73 and @xmath69 the family @xmath7 can be analytically extended , as a type - a family in the sense of kato , to all @xmath5 belonging to the disc @xmath6 in the complex plane , where @xmath85 is as in assumptions ( a ) and ( b ) .
the deformation resonances are now defined as complex eigenvalues of @xmath86 let @xmath87 .
we consider the eigenvalues @xmath88 of @xmath89 , or of @xmath90 with @xmath91 by the renormalization group analysis in @xcite , we know that , as the interaction between the non - relativistic particles and the field is turned on , these eigenvalues turn into resonances @xmath92 with @xmath93 and these resonances are @xmath94independent ; ( see also @xcite for a somewhat different model ) .
our goal is to investigate the properties of these resonances , as described above . to simplify our presentation
, we assume that @xmath88 is non - degenerate , and we denote by @xmath95 the normalized , unperturbed eigenstate associated with @xmath88 .
we also assume that * _ fermi s golden rule _ ( @xcite ) holds .
this condition implies that @xmath96 for some positive constant @xmath97 see for example @xcite .
the main results of this paper are summarized in the following theorems .
[ main_theorem ] let @xmath0 be either @xmath47 or @xmath78 .
given @xmath98 and @xmath99 as above , and under assumptions ( a)-(c ) formulated above , there exists some @xmath100 such that , for all @xmath101 and times @xmath102 @xmath103 where @xmath104 with @xmath105 appearing in ( [ eq : formfactor ] ) for the nelson model , and @xmath106 for qed .
we expect that our approach extends to situations where fermi s golden rule condition fails , as long as @xmath107 , and that we can improve the exponent of @xmath1 in the error term by using an initial state that is a better approximation of the `` resonance state '' ; see section 3 .
the analysis below , together with theorem 3.3 in @xcite , gives an adiabatic theorem for quantum resonances in non - relativistic qed .
theorem [ main_theorem ] estimates the survival probability , @xmath108 , of the state @xmath109 .
let @xmath110 and @xmath111 .
theorem [ main_theorem ] implies that @xmath112^{1/2},\ ] ] which is @xmath113 for @xmath114 .
this property is what we call the _ `` metastability '' _ of the resonance associated with the resonance eigenvalue @xmath99 .
there is a dense linear subspace @xmath13 of vectors with the property that , for every @xmath20 the family @xmath115 of vectors has an analytic extension in @xmath5 to the entire complex plane , with @xmath116 for any @xmath117 vectors in @xmath118 are called dilatation - entire vectors . next , for @xmath119 and @xmath120
we define domains @xmath121 let @xmath0 be either @xmath47 or @xmath78 .
let conditions ( a ) , ( b ) and ( c ) be satisfied , and let @xmath122 be an eigenvalue of @xmath89 .
then there are a constant @xmath123 and a dense set @xmath124 s.t .
, for @xmath125 and for all @xmath126 the function @xmath127 has an analytic continuation in @xmath21 from the upper half - plane , across a neighbourhood of @xmath128 , into the domain @xmath129 , for some @xmath130 and @xmath131 , and this continuation satisfies the relations @xmath132 with @xmath133 and @xmath134 are quadratic forms on the domain @xmath135 .
since we can rotate the essential spectrum of @xmath136 in @xmath137 using dilatation analyticity , if @xmath138 is large enough , we expect that @xmath139 can be analytically continued in @xmath21 from the upper half - plane into a neighbourhood of @xmath99 that is larger than @xmath140 given in theorem [ thm : resonpoles ] . in this case
the quadratic form @xmath141 would also depend on the homotopy class of the path along which @xmath142 is analytically continued from the upper half - plane to the point @xmath143 in the vicinity of @xmath144 for an operator @xmath145 on the one - particle space @xmath146 we denote by @xmath147 its `` lifting '' to the fock space @xmath148 ( second quantization ) .
the set @xmath149 in theorem [ thm : resonpoles ] can be chosen explicitly as @xmath150 where @xmath151 is the projection onto the vacuum @xmath152 in @xmath153 , for the nelson model . in this case
@xmath154 for qed , we define @xmath155 since @xmath156 , the set @xmath149 is dense in @xmath118 .
the main difficulty in the proofs of our main results comes from the fact that the unperturbed eigenvalue @xmath88 is the threshold of a branch of continuous spectrum . to overcome this difficulty
, we introduce an infrared cut - off that opens a gap in the spectrum of @xmath157 and we control the error introduced by opening such a gap using the fact that the interaction between the electrons and the photons or phonons vanishes sufficiently fast at low photon / phonon energies ( see @xcite and eqn .
below ) .
our paper is organized as follows . in sections
[ sec : hamiltonians]-[sec : ham ] we prove theorem [ main_theorem ] for the nelson hamiltonian , @xmath78 . in section
[ sec : qedproof ] we extend this proof to the qed hamiltonian , @xmath47 .
theorem [ thm : resonpoles ] is proven in section [ sec : polesproof ] . as we were completing this paper
, there appeared an e - print @xcite where lower and upper bounds for the lifetime of the metastable states considered in this paper are established by somewhat different techniques .
* acknowledgements*. j.fr . and i.m.s .
would like to thank m. griesemer for many useful discussions on related problems .
is grateful to i.m.s . and w.a.s .
for hospitality at the university of toronto and i.m.s . , and i.m.s . and w.a.s .
are grateful to j.fr . for hospitality at ethz .
let @xmath158 be the hamiltonian defined in @xmath64 .
we begin this section with a discussion of the dilatation deformation @xmath7 of @xmath0 defined in the introduction , eqn .
as was already mentioned above , by the above assumptions on @xmath73 and @xmath69 the family @xmath7 can be analytically extended to all @xmath5 belonging to a disc @xmath6 in the complex plane .
the relation @xmath159 holds for real @xmath5 and extends by analyticity to @xmath160 a direct computation gives @xmath161 where @xmath162 and @xmath163 . note that @xmath164 , with @xmath165 we now introduce an infra - red cut - off hamiltonian @xmath166 where @xmath167 , and @xmath168 . here
@xmath169 is an infrared cut - off function that we can choose , for instance , as @xmath170 .
we also define @xmath171 where @xmath172 .
we then have that @xmath173 we denote by @xmath174 and @xmath175 the symmetric fock spaces over @xmath176 and @xmath177 respectively .
it is well - known that there exists a unitary operator @xmath178 that maps @xmath179 to @xmath180 , so that @xmath181 here , @xmath182 acts on @xmath183 and is defined by @xmath184 the operators @xmath185 and @xmath186 denote the restrictions of @xmath37 to @xmath187 and @xmath175 respectively .
the unitary operator @xmath178 will be sometimes dropped in the sequel if no confusion may arise .
we note the following estimate that will often be used in this paper : @xmath188^{-1 } \right \| \leq \mathrm{c } g \sigma^{1/2 + \mu},\ ] ] where @xmath189 is a positive constant , and @xmath190 we now consider an unperturbed isolated eigenvalue @xmath88 of @xmath191 to simplify our analysis , we assume that @xmath88 is non - degenerate .
let @xmath192 which is positive .
it is shown in @xcite that , as the perturbation @xmath63 is turned on , the eigenvalue @xmath88 turns into a resonance @xmath99 of @xmath193 in other words , for @xmath194 with @xmath195 there exists a non - degenerate eigenvalue @xmath99 of @xmath7 _ not _ depending on @xmath8 with @xmath196 , @xmath197 , and , if fermi s golden rule condition holds , @xmath198 for some positive constant @xmath199 similarly , the operator @xmath182 has an eigenvalue @xmath200 bifurcating from the eigenvalue @xmath128 of @xmath201 having the same properties as @xmath92 with the important exception that @xmath202 depends on @xmath203 the reason for this is that @xmath204 furthermore , we have the crucial property ( see proposition [ theorem_hgeqsigma ] ) that the eigenvalue @xmath200 of @xmath205 is isolated from the rest of the spectrum of @xmath182 .
more precisely , @xmath206 for some positive constant @xmath207 independent of @xmath208 it is tempting to treat @xmath7 as a perturbation of @xmath182 .
however , we have to take care of the difference between @xmath99 and @xmath209 .
in order to deal with this problem , we `` renormalize '' the unperturbed part @xmath210 by setting @xmath211 here @xmath212 denotes the spectral projection onto the eigenspace associated with the eigenvalue @xmath200 of @xmath205 . as in @xmath213 , we have the representation @xmath214 where we have set @xmath215 by @xmath216 , we see that @xmath99 is a non - degenerate eigenvalue of @xmath217 . in proposition [ lambda_diff_estimate ] we will show that there exists a positive constant @xmath207 such that @xmath218 and that the operator @xmath219 still has a gap of order @xmath220 around @xmath144 then the decomposition @xmath221 is replaced by @xmath222 where @xmath223 let @xmath224 denote one of the operators @xmath0 , @xmath7 , @xmath210 or @xmath182 .
we write its resolvent by using the notation @xmath225^{-1}$ ] .
similarly , we define @xmath226^{-1}$ ] .
we begin with some notation .
we consider an interval @xmath227 of size @xmath228 containing @xmath229 such that @xmath230 for concreteness , let @xmath231 define , in addition , @xmath232 we consider a smooth function @xmath233,$ ] such that @xmath234 on @xmath235 .
it is known that there exists an almost analytic extension @xmath236 of @xmath237 such that @xmath238 and @xmath239 for any @xmath240 we shall use these properties of @xmath236 in the sequel .
we divide the proof of proposition [ pr : main ] into several steps , deferring the proof of some technical ingredients to the following section .
we extend a method due to hunziker to prove proposition [ pr : main ] , see @xcite or @xcite .
let @xmath247 be a punctured neighbourhood of @xmath88 such that @xmath248 and @xmath249 let @xmath250 be a contour that encloses i and @xmath144 for @xmath21 inside @xmath251 we have that @xmath252 where @xmath212 denotes the spectral projection onto the eigenspace associated to the eigenvalue @xmath99 of @xmath253 , that is @xmath254 where @xmath255 denotes a circle centered at @xmath99 with radius chosen so that @xmath256 and the _ regular part _ , @xmath257 is given by @xmath258 which can be written as @xmath259 where @xmath21 is inside @xmath260 note that @xmath261 and @xmath262 . by stone s theorem ,
@xmath268 \psi_j \right ) dz.\ ] ] since @xmath0 and @xmath109 are dilatation analytic , this implies for @xmath269 @xmath270 where @xmath271 it follows from lemma [ lemma_neumann ] , below , that we can expand @xmath272 into a neumann series , which is convergent under our assumptions on @xmath1 and @xmath263 if fermi s golden rule holds .
we obtain @xmath273 for @xmath266 and hence the claim of the lemma is proven . .
it follows from the spectral theorem that @xmath279 where @xmath280 are the spectral projections of @xmath281 ; see for example @xcite .
furthermore , @xmath282 , where @xmath283 ( respectively @xmath284 ) denotes the vacuum in @xmath285 ( in @xmath175 ) . inserting this into @xmath286 and using ( [ use_of_lambda ] ) , we get @xmath287 > from proposition [ theorem_hgeqsigma ] , we know that the spectrum of @xmath182 is of the form pictured in figure [ spectrum_hgeqsigma ] .
in particular , a gap of order @xmath263 opens between the non - degenerate eigenvalue @xmath200 and the essential spectrum of @xmath182 . by proposition [ lambda_diff_estimate ] ,
the same holds for @xmath253 instead of @xmath182 , with @xmath99 replacing @xmath200 , since @xmath288 and we assumed that @xmath289 .
let us begin to estimate @xmath290 by considering the contribution of the regular part , @xmath291 in @xmath292 . by applying green s theorem
, we find that @xmath293 where @xmath294 and @xmath295 and @xmath296 denote respectively the curve and the domain pictured in figure [ curve_gamma ] , such that the interval @xmath297 strictly contains @xmath227 . by proposition [ theorem_hgeqsigma ] and ( [ eq : regular ] ) ,
the regular part @xmath298 in @xmath299 is analytic in @xmath300 and satisfies @xmath301 where @xmath207 is a positive constant .
we also have from @xmath302 that @xmath303 and from @xmath304 that @xmath305 and from proposition [ corollary_diff_projections ] , below , that @xmath306 for some positive constant c. thus , by @xmath307 our assumptions on @xmath236 and the fact that @xmath308 we get @xmath309 where @xmath310 and any @xmath240 similarly , for the contribution of the regular part @xmath311 in @xmath312 we have the following estimate @xmath313 where @xmath314 and any @xmath240 next , we estimate the singular part of @xmath315 it is given by @xmath316 where we use the notation @xmath317 by proposition [ corollary_diff_projections ] , we know that @xmath318 we deform the path of integration as we did above , adding a circle @xmath319 of radius @xmath320 around @xmath99 .
this yields @xmath321 dz \\ & + \frac{1}{2 i \pi } \int _ { \mathcal{c}_\rho } e^{-itz } \widetilde{f } ( z ) \left [ \frac { c_g^\sigma ( \overline{\theta } ) } { z - \overline { \lambda_{j , g } } } - \frac { c_g^\sigma ( \theta ) } { z - \lambda_{j , g } } \right ] dz \\ & + \frac{1}{2 i \pi } \iint _ { \mathrm{d } ( \gamma_3 ) \setminus \mathrm{d}_\rho } e^{-itz } ( \partial _ { \bar{z } } \widetilde{f } ) ( z ) \left [ \frac { c_g^\sigma ( \overline{\theta } ) } { z - \overline { \lambda_{j , g } } } - \frac { c_g^\sigma ( \theta ) } { z - \lambda_{j , g } } \right ] dz d\overline{z } , \end{split}\ ] ] for all @xmath322 sufficiently small , where @xmath323 denotes the disc of radius @xmath320 centered at @xmath92 and @xmath324 the first integral can be estimated by using arguments similar to those used to estimate the regular part , ( [ agsigma=1+o ] ) , and the fact that @xmath325 we then obtain that for @xmath326 @xmath327 dz \right | = o ( \delta g^2 \sigma^{-2}e^ { -t \gamma_3 } ) .\ ] ] similarly , since @xmath328 on @xmath329 , we see that the third integral in the r.h.s .
of @xmath330 is independent of @xmath320 , for @xmath320 sufficiently small , and that for any @xmath331 @xmath332 dz \right | = o ( |\gamma_3/\delta|^n ) .\ ] ] it remains to estimate the second integral on the right hand side of @xmath330 .
taking the limit as @xmath333 leads to the `` residue '' @xmath334 .
since , by construction , @xmath335 on @xmath336 , we get @xmath337 dz = c_g^\sigma ( \theta ) e^ { -it \lambda_{j , g } } .\ ] ] the claim of the lemma follows from @xmath338 . recall that @xmath343 where @xmath344 it follows from and lemma [ lemma_neumann ] that can be improved if one uses instead of @xmath109 a state that is a better approximation to the resonance eigenstate . ]
latexmath:[\[\label{eq : estimate_bn } uniformly in @xmath246 together with @xmath346 and the assumption on @xmath263 and @xmath347 it follows that @xmath348 uniformly in @xmath349 one can similarly show that @xmath350 and hence the claim of the lemma follows .
* of proposition [ pr : main]*. it follows from lemmata [ lm : dilatation ] , [ lm : a_estimate ] and [ theorem_btheta ] that for @xmath351 @xmath352 let @xmath353 for some @xmath354 we optimize the estimate on the error term by choosing @xmath355 and hence the claim of the proposition is proven . * of theorem [ main_theorem]*. proposition [ pr : main ] implies that , for @xmath356 @xmath357 which , together with the boundedness of the unitary operator @xmath358 and proposition [ pr : main ] , for arbitrary @xmath359 yields @xmath360 let @xmath594 be the projection on the vacuum @xmath152 in @xmath595 .
we prove theorem [ thm : resonpoles ] for the set @xmath149 chosen explicitly as @xmath596 for the nelson model and as @xmath597 for the qed one .
since @xmath598 , the set @xmath149 is dense in @xmath118 .
we conduct the proof for the nelson model only .
to extend it to the qed one uses the methodology of section [ sec : qedproof ] . as in sections
[ sec : hamiltonians ] - [ sec : ham ] , the symbol @xmath7 stands for the dilatation transformation , of the nelson hamiltonian @xmath599 .
the rg analysis @xcite shows that given @xmath560 , there exist @xmath600 and @xmath601 s.t .
for @xmath602 and @xmath603 , the spectrum of the operator @xmath7 in the half - plane @xmath604 lies in the union of wedges @xmath605 where @xmath606 , @xmath607 and @xmath608 is a positive number .
moreover , the apices , @xmath99 , of these wedges are the eigenvalues of @xmath7 .
if , in addition , condition ( c ) holds for @xmath88 then @xmath609 we take @xmath610 with @xmath611 and @xmath612 .
we want to estimate @xmath613 using an infrared cut - off as in section [ sec : hamiltonians ] , we decompose @xmath614 see .
the infrared cut - off hamiltonian @xmath615 has an eigenvalue at @xmath144 we use the second resolvent equation @xmath616 let @xmath617 and let @xmath618 be the projection onto the vacuum state of @xmath619 and @xmath620 then @xmath621 where , as above , @xmath622 by our condition on @xmath21 we can pick @xmath5 so that @xmath623 i.e. @xmath624
. then @xmath625 ( more generally , the l.h.s . is bounded by @xmath626 for @xmath627 . )
furthermore , an elementary analysis of the @xmath628photon sectors shows that @xmath629 hence , by the definition of @xmath630 we have that , for all @xmath631 @xmath632 next , to estimate @xmath633 , see eq .
, we use the representation . applying to @xmath182 a renormalization group analysis as in @xcite
, one can show that the spectrum of @xmath253 is of the form pictured in figure [ spectrum_hgesigma_2 ] , and that for @xmath634 and @xmath635 @xmath636 which , together with , implies , for @xmath634 and @xmath637 the estimate @xmath638 for some constant c. finally we estimate the last term on the r.h.s .
recall that @xmath641 where @xmath642 below , we let @xmath643 as @xmath644 .
hence we have to estimate @xmath645 for any @xmath646 .
we claim that @xmath647 this estimate is proven in the proposition at the end of this section . iterating the last term on the r.h.s . of eq .
we see that the worst term is @xmath648 .
we use the decomposition .
since the operator @xmath649 is in normal form , we see that the term coming from sandwiching it between the first term on the r.h.s . of vanishes .
thus , it remains to consider the terms @xmath650 @xmath651\ ] ] @xmath652\ ] ] and the term obtained by switching the right and left factors in ( 6.17 ) . to estimate , we first observe that , due to , we have that , for @xmath434 @xmath655 assume @xmath656 .
using estimates , and ( or ) , the fact that @xmath657 and standard estimates on the creation and annihilation operators , and remembering the condition that @xmath658 we obtain the bound @xmath659 this together with yields @xmath660 it follows from , and that , for @xmath1 small enough , @xmath663 where @xmath664 and @xmath665 , for some constant c , provided @xmath656 . we now pick @xmath666 where @xmath667 . by our assumption ,
@xmath668 and therefore @xmath669 then , for this choice of @xmath670 @xmath671 let @xmath672 .
the last estimate , together with the relation @xmath673 implies ( i ) and ( ii ) in theorem [ thm : resonpoles ] , with @xmath674 .
@xmath675 the expression for @xmath676 can be improved if one uses to conclude that , for @xmath677 @xmath678 which is better than .
this estimate leads to the inequality @xmath679 which has a better r.h.s than . .
to prove we use the rg approach .
here we only point out particularities of the present problem and outline the general strategy ; technical details can be found in @xcite ( see also @xcite for the qed case ) .
since we do not go into details , we use the feshbach - schur map , rather than the smooth feshbach - schur map , to underpin our construction .
the former ( @xcite ) is simpler to formulate but the latter ( @xcite ) is easier to handle technically .
our strategy follows ( @xcite ) .
first we apply the feshbach - schur map @xmath681 associated to the projection @xmath682 where @xmath683 . for @xmath684 and @xmath685 ,
the operator @xmath686 is in the domain of @xmath687 .
indeed , an easy estimate shows that the operator @xmath688 is invertible on @xmath689 and @xmath690\overline{p}_{\rho_0 } \left [ \overline{p}_{\rho_0 } h_{g,\theta}^ { \sigma } \overline{p}_{\rho_0 } - z \right ] ^{-1}\overline{p}_{\rho_0 } \| \le \mathrm{c}$ ] . since @xmath691^{-1/2 }
w_{g,\theta}^ { \leqslant \sigma } [ h_f^ { \leqslant \sigma } + \sigma ] ^{-1/2 } \| \leq \mathrm{c } g \sigma^{\mu}$ ] , we see by neumann series expansion that the operator @xmath692 is invertible on @xmath689 and @xmath693^{-1}\overline{p}_{\rho_0 } \| \le \mathrm{c}/\sigma$ ] .
hence the operator @xmath686 is in the domain of @xmath694 , as claimed .
next , we note that @xmath695 where the operator @xmath696 acts on @xmath697 and is given by @xmath698 where @xmath699^{-1 } \overline{p}_{\rho_0 } w_{g,\theta}^ { \leqslant \sigma } $ ] . by the isospectrality of the feshbach - schur map ( see @xcite )
, we have that @xmath700 is an eigenvalue of @xmath7 iff @xmath701 is an eigenvalue of @xmath696 . to investigate the spectral properties of @xmath702 we make use of the renormalization group method .
as a first step , we rewrite the operator @xmath696 in a generalized normal form . to this end
we expand the resolvent on the r.h.s . in a neumann series in @xmath526 and normal order the creation and annihilation operators not entering the expression for @xmath703 .
this brings the operator @xmath696 to the form ( see @xcite ) @xmath704 where @xmath705 is a number ( more precisely , a complex function of @xmath21 and other parameters ) , @xmath706 is a differentiable function of @xmath703 and @xmath707 is an operator in the generalized normal form that is a sum of terms with at least one creation or annihilation operator .
a standard computation gives that @xmath708 , with @xmath709^{-1 } \overline{p}^{\geqslant\sigma}_{g,\theta } g_{x,\theta}^ { \leqslant \sigma } ( k)\psi^{\geqslant\sigma}_{g,\theta } ) dk + h.o.t.,\ ] ] @xmath710 @xmath711 f(h_f^++ ) a(k ) g_x,^ ( k ) dk dk)^_g , ) + h.o.t .
, @xmath712 where @xmath713^{-1 } \overline{p}_{\rho_0 } $ ] .
clearly , @xmath714 let @xmath715 stand for either @xmath61 or @xmath60 , @xmath716 .
we define the _ scaling transformation _ @xmath717 \to \mathcal{b}[\mathcal{f}_s^ { \leqslant \sigma/\rho } ] $ ] , by @xmath718 and the dilatation transform , by @xmath719
. now we rescale the operator @xmath696 as @xmath720 . the new operator acts on @xmath721 .
the last estimate in and an estimate on the derivative of @xmath706 as a function of @xmath703 , which we do not display here , show that the operator @xmath722 is in the domain of the feshbach - schur map @xmath723 , provided @xmath724 and @xmath725 ( the latter inequality is also considered as a restriction on @xmath21 ) .
if we neglect the term @xmath707 in @xmath726 ( see ) then the remaining operator has the vacuum @xmath152 as an eigenvector corresponding to the eigenvalue @xmath701 , provided @xmath21 solves the equation @xmath727 .
one can show ( @xcite ) that this equation has a unique solution @xmath728 . by the isospectrality mentioned above ,
this is our first approximation to @xmath99 .
now we introduce the decimation map @xmath729 on the domain of the decimation map @xmath730 we define the renormalization map @xmath731 as @xmath732 by the above , the operator @xmath722 is in the domain of the decimation map @xmath733 and therefore in the domain the renormalization map @xmath731 , provided @xmath734 and @xmath735 .
iterating this map as in @xcite we obtain a sequence of operators @xmath736 ( hamiltonians on scales @xmath737 ) acting on the space @xmath721 .
again , one argues that @xmath701 is an approximate eigenvalue of the operators @xmath738 provided @xmath21 satisfies the equations @xmath739 .
namely , one proves that the equations @xmath740 in @xmath21 have have unique solutions @xmath741 satisfying @xmath742 and @xmath743 ( see @xcite , proposition v.3 ) .
consequently , @xmath741 converge , @xmath744 as @xmath745 . by the isospectrality of @xmath746
we conclude that the operator @xmath722 has a simple eigenvalue @xmath747 provided @xmath748 ( see @xcite , theorem v.2 ) .
hence , by the isospectrality of the feshbach - schur map , the operator @xmath7 has a unique eigenvalue @xmath99 in the disc @xmath749 and this eigenvalue satisfies . since on the other hand @xmath750 is the unique eigenvalue of the operator @xmath751 bifurcating from the eigenvalue @xmath128 of @xmath201 , we conclude that @xmath99 is the unique eigenvalue of the operator @xmath7 emerging from the eigenvalue @xmath128 of @xmath201 .
v. bach , j. frhlich and i. m. sigal : quantum electrodynamics of confined non - relativistic particles . adv . in math . * 137 * , 299 - 395 , ( 1998 ) . v. bach , j. frhlich and i. m. sigal : renormalization group analysis of spectral problems in quantum field theory . adv . in math . *
137 * , 205 - 298 , ( 1998 ) .
v. bach , j. frhlich and i. m. sigal : spectral analysis for systems of atoms and molecules coupled to the quantized radiation fields .
phys . * 207*(2 ) , 249 - 290 , ( 1999 ) .
@xmath753 laboratoire de mathmatiques edppm , umr - cnrs 6056 , universit de reims , moulin de la housse - bp 1039 , 51687 reims cedex 2 , france .
_ current address _
: institut for matematiske fag , aarhus universitet , ny munkegade , 8000 aarhus c , denmark .
e - mail : [email protected] .
in this section , we study the operator @xmath361 used in the previous section as an approximation of @xmath7 .
we use the feshbach - schur map , @xcite , defined for a projection @xmath362 and a closed operator @xmath363 whose domain is contained in @xmath364 , by @xmath365^{-1 } \overline{p } h p,\ ] ] where @xmath366 . note that the domain of @xmath367 consists of operators @xmath363 such that @xmath368^{-1}|_{\mathrm{ran } ( \overline{p } ) } , \quad p h \overline{p } \left [ \overline{p } h \overline{p } \right ] ^{-1}|_{\mathrm{ran } ( \overline{p } ) } ,
\quad \left [ \overline{p } h \overline{p } \right ] ^{-1 } \overline{p } h p\ ] ] extend to bounded operators .
we begin with the following proposition .
[ theorem_hgeqsigma ] suppose @xmath369 then , for @xmath269 such that @xmath370 and @xmath371 , the spectrum of @xmath372 in the disc @xmath373 consists of a single eigenvalue @xmath374 furthermore , there exists @xmath375 such that , for all @xmath21 in @xmath376 such that @xmath377 , @xmath378 for some positive constant @xmath207 . .
let @xmath379 and @xmath380 .
for @xmath371 and @xmath381 , one can show that , for any @xmath382 , @xmath383^{-1 } \left ( - w_{g,\theta}^ { \geqslant \sigma } \left [ \overline{p}_\theta h_{0,\theta}^ { \geqslant \sigma } - z \right ] ^{-1 } \right ) ^n \right \|
\leq \mathrm{c}_1 \sigma^{-1 } \left ( \mathrm{c}_2 g \sigma^{-1/2 } \right ) ^n,\ ] ] where @xmath384 , @xmath385 are positive constant . hence for @xmath386 and any @xmath21 in @xmath387 , the operator @xmath388 is invertible and its inverse is given by the convergent neumann series as @xmath389^{-1}= \left [ \overline{p}_\theta h_{0,\theta}^ { \geqslant \sigma } - z \right ] ^{-1 } \sum _ { n \geq 0 } \left ( - w_{g,\theta}^ { \geqslant \sigma } \left [ \overline{p}_\theta h_{0,\theta}^ { \geqslant \sigma } - z \right ] ^{-1 } \right ) ^n .\ ] ] this implies that the operator @xmath390 is in the domain of @xmath391 .
moreover , and lead to the estimates @xmath392^{-1 } \right \| \le \mathrm{c}\sigma^{-1 } , \ \ n=0,1,\ ] ] for some positive constant @xmath207 .
note that our choice of @xmath393 yields @xmath394 .
therefore @xmath395^{-1 } \overline{p}_\theta w_{g,\theta}^ { \geqslant \sigma } p_\theta .\ ] ] the non - degeneracy of @xmath88 implies that @xmath396 can be written as @xmath397 p_\theta $ ] , where @xmath398 is a function from @xmath399 . following @xcite ( see also proposition [ theorem_diff_resonances ] below ) , we have @xmath400 for some @xmath375 , where @xmath401 with @xmath402^{-1 } \overline{p}_{p , j } g_x ( k ) p_{p , j } \mathcal{u}_\theta^{-1 } dk , \label{z_j^od } \\ & z_{j,\theta}^ { \mathrm{d } } = \int _ { \mathbb{r}^3 } \mathcal{u}_\theta p_{p , j } \overline{g_x } ( k ) p_{p , j } g_x ( k ) p_{p , j } \mathcal{u}_\theta^{-1 } \frac { dk } { \omega(k ) } . \label{z_j^d}\end{aligned}\ ] ] using the leibniz rule and the fact that @xmath403^{-1 } = \left [ \overline{p}_\theta h_{0,\theta}^ { \geqslant \sigma } - z \right ] ^{-2},\ ] ] one can prove , by differentiating with respect to @xmath21 , that @xmath404 is an analytic function on @xmath373 , and that @xmath405 , provided that @xmath406 is sufficiently small .
this implies that @xmath407 is a bijection on @xmath408 .
the isospectrality of the feshbach map ( see @xcite ) tells us that @xmath409 on the other hand , it follows from the usual perturbation theory , applied to the isolated non - degenerate eigenvalue @xmath88 of @xmath410 , that the spectrum of @xmath182 is not empty in @xmath411 , for @xmath1 sufficiently small .
hence there exists a unique @xmath200 in @xmath412 such that @xmath413 , that is @xmath414 to prove , we use the following identity ( see @xcite ) : @xmath415^{-1 } = & \left [ p_\theta - \left [ \overline { p } _ \theta \left ( h_{g,\theta}^ { \geqslant \sigma } - z \right ) \overline{p}_\theta \right ] ^{-1 } \overline { p } _ \theta w_{g,\theta}^ { \geqslant \sigma } p_\theta \right ] \left [ \mathcal{f } _
{ p_\theta } ( h_{g,\theta}^ { \geqslant \sigma } - z ) \right ] ^{-1}\\ & \times \left [ p_\theta - p_\theta w_{g,\theta}^ { \geqslant \sigma } \overline{p}_\theta \left [ \overline { p } _ \theta \left ( h_{g,\theta}^ { \geqslant \sigma } - z \right ) \overline{p}_\theta \right ] ^{-1 } \right ] \\ & + \left [ \overline { p } _ \theta \left ( h_{g,\theta}^ { \geqslant \sigma } - z \right ) \overline{p}_\theta \right ] ^{-1 } \overline { p } _
\theta , \end{split}\ ] ] which holds for @xmath21 in @xmath416 .
the simple form of @xmath417 ( [ eq : resest ] ) and the fact that @xmath418 by lead to @xmath419^{-1 } \right \| \leq \mathrm{c}_1 \left ( \frac { 1 + g^2 \sigma^{-1 } } { | z - \lambda_{j , g}^ { \geqslant \sigma } | - \mathrm{c}_2 g^{2+\varepsilon } } + \sigma^{-1 } \right ) , \ ] ] for some positive constants @xmath420
. hence the proposition is proven for @xmath21 in @xmath421 such that @xmath422 .
[ corollary_diff_projections ] let @xmath424 as in proposition [ theorem_hgeqsigma ] and choose @xmath269 such that @xmath370 . then , for @xmath1 small enough , @xmath425 where @xmath426 is a positive constant . .
let @xmath427 denote a circle centered at @xmath88 , with radius @xmath428 , so that @xmath429 . since we have assumed @xmath430 , for @xmath1 sufficiently small , @xmath427 contains both @xmath200 and @xmath88 .
thus , @xmath431 dz.\ ] ] we expand @xmath432 into a neumann series @xmath433 one can show by following the method of @xcite that , for @xmath434 , @xmath435 hence , using that @xmath436 and @xmath437 , we obtain that for all @xmath438 , @xmath439 where @xmath384 and @xmath385 denote positive constants . inserting this in @xmath440 and using the fact that the radius of @xmath427 is equal to @xmath428 , we obtain @xmath441 provided that @xmath442 is sufficiently small . hence the proposition is proven . using a renormalization group analysis , we will prove in proposition [ lambda_diff_estimate ] below the following estimate of the difference between the eigenvalues @xmath99 and @xmath200 of @xmath7 and @xmath361 : @xmath443 for any @xmath444 .
here we prove a weaker estimate , which holds only for @xmath445 , but which does not require the use of a renormalization group analysis .
besides , it is sufficient to obtain the statement of theorem [ main_theorem ] , with the slightly weaker error term @xmath446 for the nelson model , and @xmath447 for the qed one . . for @xmath1 and @xmath263 small enough , we choose @xmath450 , @xmath451 , such that @xmath452 for @xmath320 such that @xmath453 , let @xmath454
. following @xcite , @xmath99 satisfies @xmath455 where @xmath456 , with @xmath457 and @xmath458 given by @xmath459-@xmath460 , and @xmath461^{-1 } \overline{p}_\theta w_{g,\theta } p_\theta - g^2 q_\theta \label{eq : rem2 } \\ & \mathrm{rem}_3 = g^2 \left [ q_\theta - z_{j,\theta } \right ] , \label{eq : rem3 } \\ & \mathrm{rem}_4 = p_\theta w_{g,\theta } \left ( \overline{p}_\theta \left [ \overline{p}_\theta h_{0,\theta } - \lambda_{j , g } \right ] ^{-1 } \overline{p}_\theta w_{g,\theta } \right ) ^2 p_\theta , \\ & \mathrm{rem}_5 = p_\theta w_{g,\theta } \sum _ { n \geq 3 } \left ( \overline{p}_\theta \left [ \overline{p}_\theta h_{0,\theta } - \lambda_{j , g } \right ] ^{-1 } \overline{p}_\theta w_{g,\theta } \right ) ^n p_\theta .\end{aligned}\ ] ]
here we have set @xmath462 g_{x,\theta } ( k ) p_\theta dk,\ ] ] where @xmath463 , and @xmath464 .
using the expression of @xmath465 and estimates similar to ( * ? ? ?
* lemmas iv.6-iv.12 ) or ( * ? ? ?
* lemma 3.16 ) , we claim that @xmath466 the first bound in easily follows from @xmath467 the second one follows from normal - ordering and using again . finally ,
the last bound in follows from computing the difference in and using the estimate @xmath468 for some positive constant @xmath384 .
now it is proved in @xcite that @xmath469 .
let us estimate these terms more precisely : we claim that @xmath470 to prove the first bound in , we decompose @xmath471 into @xmath472 and estimate each term separately by normal ordering .
for instance , let us compute @xmath473 it follows from a pull - through formula and the canonical commutation rules that the `` worst '' term we have to estimate from the rhs of is @xmath474 one can see that @xmath475 for some positive constant @xmath384 . inserting this together with into
, we get @xmath476 where @xmath385 is a positive constant .
since the other terms could be estimated in the same way , the first bound in follows ; the second bound in can be obtained by using similar computations ( see also ( * ? ? ? * lemma 3.16 ) ) .
for @xmath477 the eigenvalue @xmath478 of @xmath182 is given by the formulas @xmath479 , except that @xmath471 and @xmath480 are replaced respectively by @xmath481 and @xmath482 .
for the terms analogous to @xmath457 and @xmath458 we have by a straightforward computations that @xmath483^{-1 } \overline{p}_{p , j } g_x ( k ) p_{p , j } \mathcal{u}_\theta^{-1 } dk = 0 , \label{diff_zod } \\
& \int _ { |k| \leq \sigma } \mathcal{u}_\theta p_{p , j } \overline{g_x } ( k ) p_{p , j } g_x ( k ) p_{p , j } \mathcal{u}_\theta^{-1 } \frac { dk } { \omega(k ) } = o \left ( \sigma^{1 + 2\mu } \right ) , \label{diff_zd}\end{aligned}\ ] ] where in we used the fact that @xmath484 .
hence , with the obvious notation , @xmath485 furthermore , eqns .
@xmath479 still hold for @xmath200 . hence remembering the assumptions @xmath486 , @xmath487 we obtain @xmath488 optimizing with respect to @xmath320 leads to the claim of the proposition .
[ lemma_neumann ] let @xmath5 in @xmath6 , @xmath489 and let @xmath424 be such that @xmath490 .
then for all @xmath491 and @xmath438 , we have the estimate : @xmath492 where @xmath420 are positive constants . .
recall that @xmath493 from the spectral representation and the decomposition @xmath299 , we can write @xmath494^{-1},\ ] ] where @xmath495 . with the definition , we have @xmath496 it follows from proposition [ theorem_hgeqsigma ] that for all @xmath21 in @xmath227 and for @xmath497 @xmath498 besides , for @xmath497 @xmath499^{-1 } \right \|
\leq \left | \mathrm{im } ( \lambda_{j , g}^ { \geqslant \sigma } ) \right |^{-1+n } = o(g^{-2(1-n)}),\ ] ] provided that fermi s golden rule holds . from @xmath500 and @xmath501 ( where @xmath502 denotes the projection onto the vacuum in @xmath175 ) , we get @xmath503^{-1 } \right \| = o ( g^{-1 } \sigma^{1/2+\mu } ) .\ ] ] similarly leads to @xmath504 the claim of the lemma then follows from @xmath505 , the assumption @xmath506 , and proposition [ lambda_diff_estimate ] .
now we extend the analysis above to the standard hamiltonian of non - relativistic qed introduced in ( [ eq : nrqed ] ) , section [ sec : introduction ] .
let now @xmath7 be the dilatation deformation of the hamiltonian @xmath507 defined in .
we keep the notation of sections [ sec : hamiltonians ] - [ sec : ham ] .
the results and proofs of sections [ sec : mainresult ] - [ sec : ham ] go through without a change except for the proof of lemma [ theorem_btheta ] . in the non - relativistic qed case , @xmath471 is given by @xmath508 where we used the notation @xmath509 , and similarly for @xmath510 .
the quantized vector potential @xmath511 is given by ( [ eq : vectorpot ] ) , and the constant @xmath512 is given by @xmath513 .
here we have @xmath514 where @xmath515 .
hence the qed hamiltonian satisfies the condition similar to with @xmath516 .
we show now how to overcome this difficulty ( a different way to proceed is to use the pauli - fierz transform @xcite ) . in our sketch of the proof of lemma [ theorem_btheta ] , we begin with the most singular term , @xmath517 , of the expansion in section [ sec : mainresult ] .
thus we have to bound the term @xmath518 .
the part of @xmath519 involving the difference of the eigenvalues is estimated in the same way as before .
namely , using that @xmath520 and that , by proposition [ lambda_diff_estimate ] , @xmath521 , we obtain that @xmath522 now we estimate the remaining part @xmath523 of @xmath524 . using the relation @xmath525-g a_\theta(x_j)$ ] the term @xmath526 can be written as @xmath527 \cdot a^{\leqslant \sigma}_\theta ( x ) + i , \ ] ] where @xmath528 \cdot a^{\leqslant \sigma}_\theta ( x ) : = \sum_{j } [ h^\sigma_{g,\theta } , x_{j } ] \cdot a^{\leqslant \sigma}_\theta ( x_{j})$ ] and @xmath529 $ ] .
furthermore , using that @xmath528 \cdot a^{\leqslant \sigma}_\theta ( x)=[h^\sigma_{g,\theta } , x \cdot a^{\leqslant \sigma}_\theta ( x)]- x \cdot [ h^\sigma_{g,\theta } , a^{\leqslant \sigma}_\theta ( x)]$ ] , we obtain @xmath530 + i + ii,\ ] ] where @xmath531 $ ] .
we can now rewrite the operator @xmath532 as @xmath533 let @xmath237 be a ( vector-)function of @xmath58 . to estimate the expression above we will use the following estimates @xmath534 the first two estimates are standard ( see e.g. @xcite ) . to prove the last inequality one uses eqns - .
in addition we need the following estimate for any @xmath535 @xmath536 where , recall , @xmath537 , and @xmath538^{1/2}$ ] .
eqn follows in the same way as , provided we prove that still holds if one replaces @xmath182 and @xmath539 respectively by @xmath540 and @xmath541 .
to prove the latter property , we note that @xmath542 where @xmath543 and similarly for @xmath544 , @xmath545 , @xmath546 and @xmath547 . using that @xmath548 and the fact that @xmath549 has the same eigenvalues as @xmath550 we write @xmath551^{-1}\overline{p}^\delta_\theta= \left [ e^{-\theta } h_f^ { \geqslant \sigma } + \lambda_j - z \right ] ^{-1 } ( p_{p , j , \theta } ^\delta \otimes \overline{p}_\omega^ { \geqslant \sigma } ) \ ] ] @xmath552^{-1}(\overline{p}_{p , j , \theta } ^\delta \otimes i).\ ] ] using this decomposition we conclude , similarly to , that for @xmath553 and @xmath21 in @xmath554 , we have the for some positive constants @xmath384 , @xmath385 @xmath555^{-1 } \left ( - w_{g,\theta}^ { \delta , \geqslant \sigma } \left [ \overline{p}^\delta_\theta h_{0,\theta}^ { \delta , \geqslant \sigma } - z \right ] ^{-1 } \right ) ^n \right \|
\leq \mathrm{c}_1 \sigma^{-1 } \left ( \mathrm{c}_2 g \sigma^{-1/2 } \right ) ^n.\ ] ] now , the first two terms in eqn have only one resolvent each . using estimates eqns and , with @xmath556 and @xmath557 , and eqns - with @xmath558
, we obtain for these terms , times @xmath559 , with @xmath560 , the estimate @xmath561 .
the operator @xmath562\ ] ] has better infrared behaviour than the original operator @xmath563 by an extra factor @xmath564 or @xmath565 , which , due to , with @xmath556 and @xmath566 , and - , gives the estimate @xmath567 .
finally , the term @xmath568 is quadratic in @xmath569 . putting it to the normal form and using the estimates and leads to the estimate @xmath570 . collecting the above estimates and using that @xmath571 , we arrive at @xmath572 next , we pull @xmath559 , with @xmath560 sufficiently small , from @xmath109 and use the above estimate to obtain @xmath573 therefore , the largest contribution to @xmath517 comes from the term that involves the difference of the eigenvalues . taking into account the factor @xmath263 obtained from the @xmath21 integration yields that @xmath574 one can estimate the operators @xmath575 similarly . in particular , we claim that @xmath576 for @xmath577 .
consider for example the term @xmath578 since @xmath579 and @xmath580 , we have that @xmath581 by pulling @xmath582 from @xmath98 we have using ( [ eq : estb1factor ] ) that @xmath583 using ( [ wqed5 ] ) , we have @xmath584 it follows from ( [ standardest1 ] ) and ( [ rtildeest])-([eq : expdelta_tilder ] ) and from that @xmath585 since the operator @xmath568 is quadratic in @xmath586 , we obtain by putting it to the normal form and using again ( [ standardest1 ] ) and ( [ rtildeest])-([eq : expdelta_tilder ] ) @xmath587 finally , as above the fact than @xmath588 has better infrared behavior than @xmath586 by the factor @xmath565 leads to @xmath589 by pulling @xmath582 from @xmath109 and using ( [ eq : b2factor])-([eq : b2est6 ] ) we find that @xmath590 together with a factor of @xmath263 obtained from the @xmath21 integration , ( [ eq : b2est1 ] ) , ( [ eq : b2est2 ] ) and ( [ eq : b2est5 ] ) yield the estimate @xmath591 | we study the mathematical theory of quantum resonances in the standard model of non - relativistic qed and in nelson s model .
in particular , we estimate the survival probability of metastable states corresponding to quantum resonances and relate the resonances to poles of an analytic continuation of matrix elements of the resolvent of the quantum hamiltonian . |
both observational and theoretical evidence indicates that classical be stars are near main sequence b - type stars with gaseous circumstellar disks ( e.g. see review by @xcite ) ; however , the mechanism(s ) responsible for creating these disks are not well understood .
optical spectroscopic studies have uncovered a myriad of different line profile morphologies commonly observed in classical be stars @xcite , including double - peaked asymmetrical emission lines @xcite .
the ratio of the violet ( v ) to red ( r ) flux within individual lines , hereafter referred to as a v / r ratio , is observed to vary on quasi - periodic timescales of several to ten years , as documented for a large set of be stars by @xcite .
while numerous theories have been proposed to explain this observed phenomenon , as summarized within @xcite , these features are most commonly interpreted as evidence of one - armed density waves within these circumstellar disks @xcite .
note however that the presence of phase lags between polarimetric and h@xmath1 observations has led @xcite to suggest these waves might be more spiral - like than previously thought .
interferometric studies @xcite have also provided evidence that be disks deviate from axisymmetry , namely they have provided evidence of prograde one - armed oscillations in several disks .
it is not uncommon to observe multiple emission lines in a spectrum of a classical be star that exhibit a v / r asymmetry , and with few exceptions the same orientation of this asymmetry is observed in all lines .
@xcite summarizes several of these exceptions , in which a phase lag appears to be present between optical hydrogen and fe ii lines @xcite and between different lines in the hydrogen balmer series @xcite .
@xcite also suggested that a phase lag exists between the h@xmath1 and hydrogen pa 15 - 18 lines of @xmath4 cas presented by @xcite ; however , we believe this suggestion is rather dubious given the poor resolution of these data and clear evidence of contamination from nearby lines .
few moderate resolution infrared ( ir ) spectroscopic studies of classical be stars have been performed to date . @xcite and @xcite have investigated a large set of be stars in the h and k band at sufficient resolution to begin to search for detailed line profile effects .
the first tentative evidence of ir v / r reversals , analogous to the optical v / r phase lags summarized by @xcite , was presented in @xcite , who noted that the brackett @xmath4 and he i 2.058@xmath5 m lines had opposite orientations in one of their targets . in this letter , we present the initial results from a near - contemporaneous , optical and ir spectroscopic observing campaign designed to provide insight into the detailed density structure of classical be circumstellar disks , hence diagnose the mechanisms responsible for creating these disks . detailed modeling of these data , as well as follow - up observations , will be presented in subsequent papers . in section 2
we outline the observational data to be discussed in this work .
an analysis and discussion of these data are presented in section 3 .
finally , a summary of our main results and suggested follow - up work is presented in section 4 .
the infrared ( ir ) spectroscopic data presented in this paper were obtained at the 3 m nasa infrared telescope facility ( irtf ) , using the spex medium resolution spectrograph @xcite .
we used a 0.3 x 15 arc - second slit to provide wavelength coverage from 0.8 - 2.4@xmath5 m at r @xmath62000 .
observations of a0v stars were obtained at @xmath615 minute intervals throughout our run , at a similar hour angle and airmass to our science targets , to serve as telluric standards .
further details regarding the telluric correction method commonly applied to spex data are presented in @xcite .
all observations were reduced using spextool @xcite .
these data were taken during a period of extremely poor weather conditions on mauna kea , thus we have not attempted to perform a rigorous flux calibration . while these observing conditions did introduce small artifacts in wavelength regimes outside of those presented here ( see @xcite ) , we are fully confident that the line profile morphologies and line - to - continuum ratios of the data presented in this paper are not influenced by the presence of any residual telluric features .
the optical spectroscopic data presented here were obtained at the 1 m telescope of the ritter observatory , using its fiber - fed echelle spectrograph .
nine non - overlapping orders , each of width 70 , were obtained over the wavelength range 5300 - 6600 at r @xmath626000 .
standard iraf techniques were used to reduce these data , including procedures to apply bias , flat field , and wavelength calibration corrections to all data .
further details regarding the reduction of ritter data can be found in @xcite .
a summary of the optical and ir data to be discussed in this paper are presented in table [ photsummary ] .
note that we have not corrected any of our optical or ir spectral line data for underlying photospheric absorption components .
as demonstrated by our near - contemporaneous optical and ir observations of 48 librae ( see figure [ 48lib ] ) , double - peaked asymmetrical profiles in be spectra often exhibit the same v / r orientation for all observed transitions .
@xcite reported phase lags between optical fe ii and hydrogen lines in several of his observations of be stars ; we found evidence of a similar phenomenon in our ir observation of the suggested classical be star ngc 2439:wbbe1 ( see @xcite ) .
we obtained two h@xmath1 spectra of the well studied classical be star @xmath3 tau ( see figure [ ztauoptical ] ) one month before and one month after our ir observations , and note that both h@xmath1 profiles exhibit triple - peaked profiles characterized by an overall v @xmath2 r asymmetry .
the v / r period is often several years , so the line profiles are generally stable on timescales of several months @xcite .
we are confident that @xmath3 tau exhibited a h@xmath1 profile similar to that shown in figure [ ztauoptical ] at the time of our ir observation , and was not in the midst of a cycle of irregular fluctuations of the type described by @xcite , based on our long baseline of previous observations of the star . at the resolution of spex ,
@xmath3 tau s ir spectrum shows many emission - lines characterized by a range of morphologies , including single - peaked profiles , asymmetrical double - peaked profiles , and asymmetrical double - peaked shell profiles . as shown in figure [ ztauoptical ] , hydrogen brackett transitions in these data exhibit shell - like profiles , which are asymmetrical with v @xmath0 r. other ir emission - lines which have asymmetrical profiles with a similar v / r orientation to the h i brackett series include o i 0.8446 , 1.129 , 1.317 @xmath5 m ; fe ii 1.086 , 1.113 @xmath5 m ; he i 1.083 @xmath5 m ; h i paschen @xmath4 and @xmath7 ; and the h i pfund lines at 2.382 , 2.392 , 2.403 @xmath5 m .
these data exhibit a fundamentally different v / r morphology than our near - contemporaneous h@xmath1 profile . as @xmath3 tau is known to experience cyclical v / r variations , likely due to the presence of a density wave in its disk @xcite
, we suggest that the observed opposite v / r orientation of the ir lines and h@xmath1 can be explained by the presence of a density wave in the disk which has a significantly different average azimuthal morphology ( e.g. a spiral wave ) , combined with different loci for line formation .
this idea is illustrated in figure [ diagram ] , where we show schematically the disk of @xmath3 tau with a one - armed spiral density perturbation : note that lighter shading denotes regions of density enhancement while darker shading denotes density decrements .
the h i ir lines are much more optically thin than the optical lines , hence they form within the inner disk , while the optical h i lines originate from a much larger area . as a result of the different azimuthal properties of the disk with respect to the distance from the star , in certain phases of the v / r cycle the net density enhancement of
the inner disk may occur in an opposite side ( with respect to the line of sight ) to the net density enhancement of the entire disk , thus producing line profiles with opposite v / r ratios .
for instance , in figure [ diagram ] the v @xmath0 r morphology of br-12 is produced by a net density enhancement in the inner disk region which is moving towards the observer , and the v @xmath2 r morphology of h@xmath1 is produced by a net density enhancement in the outer disk region which is moving away from the observer .
@xcite , @xcite , and @xcite outline a new monte carlo code for solving the radiative transfer in the circumstellar disks of be stars .
this fully three - dimensional non - lte monte carlo code produces a self - consistent determination of the hydrogen level populations , electron temperature , and gas density of a circumstellar disk , given a prescription of the central star and assuming a static , nearly - keplerian disk with a given mass loss rate , truncation radius , and a prescription for the gas viscosity . in @xcite
, the authors used this code to study the circumstellar disk of @xmath3 tau .
following @xcite , they adopted a luminosity class of iv and a spectral type of @xmath8 , which corresponds to an effective temperature of @xmath9 and a stellar radius of @xmath10 . assuming an outer disk radius of @xmath11 and a mass loss rate of @xmath12 the authors were able to fit the observed sed from visible to mid ir wavelengths , and optical ( @xmath13 320010500 ) spectropolarimetry . using this model , we calculated intensity maps of h@xmath1 and br-12 for @xmath3 tau which , as seen in figure [ allmodel ] , illustrate that br-12 is formed in the inner disk ( @xmath14 ) whereas h@xmath1 is produced over a much more extended region , thereby confirming our previous assertion that our optical versus ir h i data probe distinctly different disk regions .
future iterations of the model will incorporate complex density profiles , such as one - armed waves , which will enable us to test our exploratory discussion in a more quantitatve manner .
we note that v / r reversals have been predicted by at least one other theory in the refereed literature .
@xcite predicted instances in which a ratio of r@xmath0v in a h@xmath1 line would reverse to a ratio of r@xmath2v in ir h i lines , due to line self - absorption in the inner region of edge - on disks .
however , the presence of this reversal depends critically on the existence of a large expansion component in the disk velocity at the location where the lines form , an unlikely scenario given the very small radial velocities now believed to characterize inner disk regions @xcite .
we have presented the first results of near - contemporaneous optical and ir spectroscopic observations of classical be stars .
we report the discovery of asymmetrical infrared emission lines ( he i , o i , fe ii , h i paschen , brackett , and pfund series ) in @xmath3 tau which have oppositely oriented v / r ratios to that seen at h@xmath1 .
we suggest these data indicate that the density wave which characterizes @xmath3 tau s disk has a significantly different average azimuthal morphology in the inner disk region as compared to its outer disk region .
this would also indicate that the fe ii lines are formed in the inner disk , in the same spatial location as the ir lines are formed .
these features should travel through the disk on the cyclical timescale generally reported for density waves in @xmath3 tau @xcite ; hence , we would predict continued anti - correlation amongst these line profiles over the course of several observing seasons .
we have begun a long - term optical and ir spectroscopic monitoring program to trace this evolution .
as demonstrated by the initial results presented in this letter , such an observational approach has the potential to provide powerful constraints to theoretical efforts to map the detailed density distribution of be circumstellar disks , hence better understand the mechanism(s ) which create these disks .
we thank our referee , myron smith , for comments which helped to improve this manuscript .
we also thank the staff of ritter observatory for their assistance in providing data presented in this paper .
this research was supported by a nasa ltsa grant to ksb ( nag5 - 8054 ) , a research corporation cottrell scholar award ( ksb ) , a nasa gsrp fellowship to jpw ( ngt5 - 50469 ) , a sigma xi grant to jpw , and the nsf reu program at the university of toledo ( afk ) .
jeb and acc acknowledge support from nsf grant ast-0307686 ; acc is also supported by fapesp grants 01/12589 - 1 and 04/07707 - 3 .
observations at the ritter observatory are supported by the nsf under the prest program , grant number ast-0440784 .
baade , d. 1985 , a&a , 148 , 59 berio , p. 1999 , a&a , 345 , 203 bjorkman , j.e . &
carciofi , a.c .
2005 , in astronomical polarimetry : current status and future directions , ed .
a. adamson , c. aspin , c. j. davis , & t. fujiyoshi , asp conf . ser .
343 , 417 carciofi , a.c . &
bjorkman , j.e .
2006 , apj , 639 , 1081 carciofi , a.c .
2006 , apj , 652 , 1617 carciofi , a.c .
, bjorkman , j.e . , miroshnichenko , a.s . ,
magalhes , a.m. , & bjorkman , k.s .
2007 , in active ob stars : laboratories for stellar and circumstellar physics , eds .
s. owocki , s. stefl & a. okazaki ( asp conf .
series 361 ; san francisco : asp ) , in press chalabaev , a. & maillard , j.p .
1983 , a&a , 127 , 279 clark , j.s . & steele , i.a .
2000 , a&as , 141 , 65 cushing , m.c . ,
vacca , w.d .
, & rayner , j.t .
2004 , pasp , 116 , 362 dachs , j. , hummel , w. , & hanuschik , r.w .
1992 , a&as , 95 , 437 guo , y. , huang , l. , hao , j. , cao , h. , guo , z. , & guo , x. 1995 , a&as , 112 , 201 hanuschik , r.w . , hummel , w. , dietle , o. , & sutorius , e. 1995 , a&a , 300 , 163 hanuschik , r.w .
, hummel , w. , sutorius , e. , dietle , o. , & thimm , g. 1996 , a&as , 116 , 309 hanuschik , r.w .
2000 , in iau colloq .
175 , the be phenomenon in early - type stars , ed . m.a .
smith , h.f .
henrichs , & j. fabregat ( asp conf .
214 ; san francisco : asp ) , 518 kogure , t. & suzuki , m. 1984 , pasj , 36 , 191 mcdavid , d. , bjorkman , k.s . ,
bjorkman , j.e . , & okazaki , a.t .
2000 , in iau colloq .
175 , the be phenomenon in early - type stars , ed . m.a .
smith , h.f .
henrichs , & j. fabregat ( asp conf .
214 ; san francisco : asp ) , 460 mclaughlin , d.b .
1961 , jrasc , 55 , 73 morrison , n.d .
, knauth , d.c . , mulliss , c.l . ,
& lee , w. 1997 , pasp , 109 , 676 okazaki , a.t .
1991 , pasj , 43 , 75 okazaki , a.t .
1997 , a&a , 318 , 548 porter , j.m . & rivinius , t. 2003 , pasp , 115 , 1153 quirrenbach , a. et al .
1997 , apj , 479 , 477 rayner , j.t .
2003 , pasp , 115 , 362 reig , p. , negueruela , i. , coe , m.j .
, fabregat , j. , tarasov , a.e . ,
& zamanov , r.k .
2000 , mnras , 317 , 205 rivinius , th .
, stefl , s. , & baade , d. 1999 , a&a , 348 , 831 slettebak , a. 1982 , apjs , 50 , 55 steele , i.a . & clark , j.s .
2001 , a&a , 371 , 643 telting , j.h . ,
waters , l.b.f.m . ,
persi , p. , & dunlop , s.r .
1993 , a&a , 270 , 355 vacca , w.d . ,
cushing , m.c . , & rayner , j.t .
2003 , pasp , 115 , 389 vakili , f. et al .
a&a , 335 , 261 waters , l.b.f.m . & marlborough , j.m .
1992 , a&a , 253 , l25 wisniewski , j.p .
2005 , ph.d .
thesis , university of toledo wisniewski , j.p .
& bjorkman , k.s .
2006 , apj , 652 , 458 wood , k. , bjorkman , k.s . , & bjorkman , j.e .
1997 , , 477 , 926 spectrum of 48 librae observed on 2004 march 11 exhibits an asymmetrical profile characterized by v @xmath0 r. _ right : _ a near contemporaneous infrared spectrum obtained on 2004 march 6 also exhibits hydrogen brackett lines having asymmetrical profiles with v @xmath0 r. the agreement between the orientation of these line profile morphologies is the standard scenario expected in observations of classical be stars .
[ 48lib],title="fig:",width=302 ] spectrum of 48 librae observed on 2004 march 11 exhibits an asymmetrical profile characterized by v @xmath0 r. _ right : _ a near contemporaneous infrared spectrum obtained on 2004 march 6 also exhibits hydrogen brackett lines having asymmetrical profiles with v @xmath0 r. the agreement between the orientation of these line profile morphologies is the standard scenario expected in observations of classical be stars .
[ 48lib],title="fig:",width=302 ] spectrum of @xmath3 tau obtained on 2004 february 28 ( left panel ) clearly exhibits an asymmetrical triple - peaked profile characterized by v @xmath2 r. although not shown here , an optical spectrum obtained on 2004 april 12 exhibits the same morphology .
right : _ this ir spectrum of @xmath3 tau obtained on 2004 march 5 is characterized by asymmetrical hydrogen brackett lines having v @xmath0 r. although not picture here , additional he i , o i , fe ii , and h i ( paschen and pfund series ) emission lines in this spectrum exhibit this same v @xmath0 r morphology .
[ ztauoptical],title="fig:",width=302 ] spectrum of @xmath3 tau obtained on 2004 february 28 ( left panel ) clearly exhibits an asymmetrical triple - peaked profile characterized by v @xmath2 r. although not shown here , an optical spectrum obtained on 2004 april 12 exhibits the same morphology .
_ right : _ this ir spectrum of @xmath3 tau obtained on 2004 march 5 is characterized by asymmetrical hydrogen brackett lines having v @xmath0 r. although not picture here , additional he i , o i , fe ii , and h i ( paschen and pfund series ) emission lines in this spectrum exhibit this same v @xmath0 r morphology .
[ ztauoptical],title="fig:",width=283 ] tau s disk based on observed v / r line profile variations .
our ir ( left panel ) and optical ( right panel ) spectroscopic data probe different physical locations within the @xmath3 tau disk , hence probe different regions of the density perturbation located within the disk .
[ diagram],width=302 ] tau disk at h@xmath1 ( upper panel ) and h i br-12 ( lower panel ) .
the scale of these images is 40 stellar radii ( @xmath6224 solar radii ) .
these images clearly demonstrate that h@xmath1 is produced over a much larger disk area , whereas h i br-12 is produced from a more inner disk region .
[ allmodel],width=302 ] lccc object & observatory & date & exposure times + @xmath3 tau & i & 2004 march 5 & 60 s. + @xmath3 tau & r & 2004 february 28 & 1200 s. + @xmath3 tau & r & 2004 april 12 & 1200 s. + 48 lib & i & 2004 march 6 & 150 s. + 48 lib & r & 2004 march 11 & 3600 s. + | the first results from a near - contemporaneous optical and infrared spectroscopic observing program designed to probe the detailed density structure of classical be circumstellar disks are presented .
we report the discovery of asymmetrical infrared emission lines of he i , o i , fe ii , and the brackett , paschen , and pfund series lines of h i which exhibit an opposite v / r orientation ( v @xmath0 r ) to that observed for the optical balmer h@xmath1 line ( v @xmath2 r ) in the classical be star @xmath3 tau .
we interpret these data as evidence that the density wave which characterizes @xmath3 tau s disk has a significantly different average azimuthal morphology in the inner disk region as compared to the outer disk region . a follow - up multi - wavelength observational campaign to trace the temporal evolution of these line profile morphologies , along with detailed theoretical modeling ,
is suggested to test this hypothesis . |
the existence of a permanent electric dipole moment ( edm ) in leptons , neutrons or neutral atoms is direct evidence for time reversal ( t ) violation .
because of the cpt theorem , the search for edms can provide us valuable information about sources of cp violation . though a phase in the cabibbo kobayashi maskawa matrix is enough to account the level of cp violation in kaon and b meson decays , it can not explain the observed matter / anti matter asymmetry in the universe .
physics that can ( and new physics at the weak scale more generically ) should produce edms not far from current upper limits .
so far no edms have been observed , but experiments are continually improving . here
we are interested in the conclusions that can be drawn from the measured upper limit @xcite in @xmath0 , a diamagnetic atom .
the largest part of its edm most likely comes from t violation in the nucleus , caused by a t violating ( and parity violating ) component of the nucleon
nucleon interaction .
the atomic edm is generated by the subsequent interaction of the nucleus with the electrons .
that interaction is more subtle than one might think . if the nucleus and electrons were non
relativistic point
particles interacting solely via electrostatic forces , the electrons would rearrange in response to a nuclear edm to cancel it essentially exactly . fortunately , as was shown by schiff @xcite , the finite size of the nucleus leads to a residual atomic edm .
it turns out , however , that the relevant nuclear quantity is not the nuclear edm but rather the nuclear `` schiff moment '' @xmath3 which is the nuclear ground
state expectation value , in the substate @xmath4 with angular momentum projection @xmath5 equal to the angular momentum @xmath6 , of the @xmath7component of the `` schiff operator '' @xmath8 here @xmath9 is the charge of the proton , @xmath10 is the mean squared radius of the nuclear charge distribution , and the sum is restricted to protons .
for the schiff moment to exist , p and t must be violated by the nuclear hamiltonian .
we assume that whatever its ultimate source , the t violation works its way into a meson mediated p and t violating nn interaction generated from a feynman graph containing a meson propagator , the usual strong meson nn strong vertex and a ( much weaker ) p and t violating meson nn vertex .
the second vertex can take three different forms in isospin space .
references @xcite showed that short range nuclear correlations and a fortuitous sign make the contribution of @xmath11 and @xmath12exchange to the interaction small compared to that of pion
exchange if the t violating coupling constants of the different mesons are all about the same , and so we neglect everything but pion exchange . the most general p and t
odd nn potential then has the form ( @xmath13 ) @xmath14\left(\bm{\sigma}_a-\bm{\sigma}_b\right)-\right.\nonumber\\ & & \mbox{}\left.-\frac{{\rm \bar{g}}_1}{2}\left(\tau_a^z-\tau_b^z\right ) \left(\bm{\sigma}_a+\bm{\sigma}_b\right)\right\}\cdot \left(\bm{r}_a-\bm{r}_b \right)\frac{{\rm exp}\left(-m_\pi|\bm{r}_a-\bm{r}_b|\right ) } { m_\pi|\bm{r}_a-\bm{r}_b|^2}\left[1+\frac{1}{m_\pi|\bm{r}_a-\bm{r}_b| } \right]~,\end{aligned}\ ] ] where @xmath15 is the mass of the pion , @xmath16 that of the nucleon , @xmath17 , @xmath18 is the strong @xmath1nn coupling constant , and the @xmath19 are the isoscalar ( @xmath20 ) , isovector ( @xmath21 ) , and isotensor ( @xmath22 ) pt odd @xmath1nn coupling constants . a word of caution here : more than one sign convention for the @xmath23 s is in use .
our @xmath24 and @xmath25 are defined with a sign opposite to those used by flambaum et .
al @xcite and by dmitriev et .
al @xcite .
the goal of this paper is to calculate the dependence of the schiff moment of @xmath0 on the t violating @xmath1nn couplings ( we leave the dependence of these couplings on fundamental sources of cp violation to others ) so that models of new physics can be quantitatively constrained .
an accurate calculation is not easy because the schiff moment depends on the interplay of the schiff operator with complicated spin and space dependent correlations induced by the the two body interaction @xmath26 . in the early calculation by flambaum , khriplovich and
sushkov almost two decades ago @xcite , the correlations were taken to be admixtures of simple 1particle 1hole excitations into a slater determinant produced by a one body wood saxon potential .
more recent work @xcite made significant improvements by treating the correlations in the rpa after generating an approximately self consistent one body potential .
however , that work used only the relatively schematic landau migdal interaction ( in addition to @xmath26 ) in the rpa and mean field equations , and did not treat pairing self consistently .
the reliance on a single strong interaction makes it difficult to analyze uncertainty .
such an analysis seems to be particularly important in @xmath0 , the system with the best experimental limit on its edm .
the calculated schiff moment of refs .
@xcite in that nucleus depends extremely weakly on the isoscalar coefficient @xmath27 , a result of coincidentally precise cancellations among single particle and collective excitations
. they might be less precise when other interactions are used .
here we make several further improvements .
our mean field , which we calculate in @xmath28hg before treating core polarization by the valence nucleon , includes pairing and is exactly self consistent .
pairing changes the rpa to the quasiparticle
rpa ( qrpa ) , a continuum version of which we use to obtain ground state correlations .
most importantly , we carry out the calculation with several sophisticated ( though still phenomenological ) skyrme nn interactions , the appropriateness of which we explore by examining their ability to reproduce measured isoscalar
e1 strength ( generated by the isoscalar component of the schiff operator ) in @xmath29pb .
the use and calibration of more than one such force allows us to get a handle on the uncertainty in our final results .
the rest of this paper is organized as follows : section ii describes our approach and the skyrme interactions we use , and includes their predictions for strength distributions that bear on the schiff moment .
section iii presents our results and an analysis of their uncertainty , including a calculation in the simpler nucleus @xmath30pb that allows us to check the size of effects we omit in @xmath0 .
section iv is a conclusion .
our schiff moment is a close approximation to the expectation value of the schiff operator in the completely self consistent one quasiparticle ground state of @xmath0 , constructed from a two body interaction that includes both a skyrme potential and the p and t violating potential @xmath26 .
it is an approximation because we do not treat @xmath26 in a completely self consistent way , causing an error that we estimate to be small in the section [ sec : results ] .
in addition , we do not actually carry out the mean field calculation in @xmath0 itself .
instead , we start from the hf+bcs ground state of the even even nucleus @xmath31 and add a neutron in the @xmath32 level .
we then treat the core polarizing effects of this neutron in the qrpa .
a self consistent core with qrpa core polarization is completely equivalent to a fully self consistent odd a calculation @xcite .
we omit one part of the qrpa core polarization , again with an estimate showing its contribution to be insignificant . a good way to keep track of the two interactions and their effects is to formulate the calculation ( and corrections to it ) as a sum of goldstone like diagrams , following the shell model effective operator formalism presented , e.g. , in ref.@xcite . the one difference between our diagrams and he usual `` brandow '' kind is that our fermion lines will represent bcs quasiparticles rather than pure particles or holes .
our diagrams reduce to the familiar kind in the absence of pairing .
we begin , following a spherical hf+bcs calculation in @xmath31 ( in a 20fm box with mixed volume and surface pairing fixed as in ref .
@xcite ) , by dividing the hamiltonian into unperturbed and residual parts .
the unperturbed part , expressed in the quasiparticle basis , is @xmath33 where @xmath34 is the kinetic energy and @xmath35 the skyrme interaction , with subscripts that refer to the numbers of quasiparticles the operator creates and destroys . the residual piece and @xmath36 zero .
] is @xmath37 the interaction @xmath26 can also be expanded in terms of quasiparticle creation and annihilation operators ; all the terms are included in @xmath38 , though @xmath39 vanishes because @xmath26 is a pseudoscalar operator .
the valence `` model space '' of effective operator theory is one dimensional : a quasiparticle with @xmath40 and @xmath41 ( i.e. a particle , since it is not part of a pair ) in the the @xmath42 level .
the unperturbed ground state @xmath43 is simply this one quasiparticle state .
then the expectation value of @xmath44 , eq .
( [ eqn : sm ] ) , in the full correlated ground state @xmath45 is given by @xmath46s^z \left[1+\left ( \frac{q}{\epsilon_{{a}}-h_0}\right)h_{\rm res}+\cdots\right]|\phi_{{a}}\rangle~. \label{eqn : per4}\ ] ] here @xmath47 is the single quasiparticle energy of the valence nucleon , the operator @xmath48 projects onto all other single quasiparticle states , @xmath49 is a normalization factor that , we will argue later , is very close to one , and the dots represent higher order terms in @xmath38 . to evaluate the expression ,
we write @xmath44 in the quasiparticle basis as @xmath50 ( @xmath51 vanishes for the same reason as @xmath39 ) . .
diagram ( vi ) is the complex conjugate of diagram ( iii ) . ]
the zeroth order contribution to the schiff moment in eq.([eqn : per4 ] ) vanishes because the schiff operator can not connect two states with the same parity , and also because the schiff operator acts only on protons while the valence particle in @xmath0 is a neutron .
( there are no center of mass corrections to the effective charges @xcite . )
the terms that are first order in @xmath38 do not include the strong interaction @xmath35 because it has a different parity from the schiff operator .
thus the lowest order contribution to the schiff moment is @xmath52 q^{\dag}_{{a}}|-\rangle+{\rm c.c.}~ , \label{eqn : order1}\ ] ] where @xmath53 is the creation operator for a quasiparticle in the valence level @xmath54 and @xmath55 is the no quasiparticle bcs vacuum describing the even even core , so that @xmath43 is just @xmath56 . the contribution of eq .
( [ eqn : order1 ] ) in an arbitrary nucleus can be represented as the sum of the diagrams in fig .
[ fig : order1 ] , the rules for which we give in the appendix . in @xmath0 , because the valence particle is a neutron , only diagrams ( iii ) and ( vi ) are nonzero
. we can interpret diagram ( iii ) as the schiff operator exciting the core to create a virtual three
quasiparticle state , which is then de excited back to the ground state when the valence neutron interacts with the core through @xmath26 .
this diagram and its partner ( vi ) are what was evaluated by flambaum et al .
@xcite , though their mean field was a simple wood saxon potential , their @xmath26 was a zero range approximation that did nt include exchange terms , and they neglected pairing .
the broken line represents the action of the schiff operator ( as in fig.[fig : order1 ] ) , the zig - zag line represents the p and t violating interaction ( also as in fig .
[ fig : order1 ] ) , and the looped line represents a generic skyrme interaction . ] core polarization , implemented through a version of the canonical basis qrpa code reported in ref .
@xcite ( with residual spurious center of mass motion removed following ref.@xcite ) , can be represented by a subset of the higher order diagrams . because @xmath26 is so weak ,
we need only include it in first order .
the higher order terms in @xmath35 that we include have the effect of replacing the two quasiparticle bubble in ( iii ) and ( vi ) of fig .
[ fig : order1 ] with chains of such bubbles ( see fig . [
fig : qrpa ] ) , as well adding diagrams in which the qrpa bubble chains are excited through a strong interaction of the core with the valence neutron .
we therefore end up evaluating diagrams labeled a , b1 , and b2 , in fig .
[ fig : orderrpa ] ( plus two more of type b in which the interaction @xmath26 is below the schiff operator ) .
the explicit expression for diagram a , the first on the left in the figure , is @xmath57 acts below @xmath44 . ] here , @xmath58 represents the qrpa amplitudes ( the sum appears because the matrix elements of all our operators are real ) and @xmath59 is the energy of the collective state @xmath60 in @xmath31 .
the quasiparticle matrix elements @xmath61 and @xmath62 are related to the usual particle matrix elements @xmath63 and @xmath64 through the transformations discussed in the appendix . in the absence of qrpa correlations ,
the @xmath65 and @xmath66 amplitudes are 1 or 0 , and eq .
( [ eqn : diaga ] ) reduces to that associated with diagrams ( iii ) and ( vi ) of fig . [
fig : order1 ] . diagrams b1 and b2 of fig .
[ fig : orderrpa ] have the explicit expressions @xmath67 the factor 2 accounts for diagrams not shown in fig.[fig : orderrpa ] in which @xmath26 acts _ below _ the qrpa bubble , and the @xmath68 s are quasiparticle energies . the difference between eq .
( [ eqn : diagb1 ] ) and eq .
( [ eqn : diagb2 ] ) is mainly in the three - quasiparticle intermediate states .
a complete qrpa calculation that is first order in @xmath26 would also include versions of diagram a in which @xmath26 trades places with one of the @xmath35 s in the bubble sum .
we do nt evaluate such diagrams but estimate their size ( which we find to be small ) rom calculations in the simpler nucleus @xmath69 in the next section .
we also use that nucleus to examine other low order diagrams not included in the bubble sum of diagram a. why do we expect the qrpa subset of diagrams to be sufficient ? the reason is that they generally account for the collectivity of virtual excitations in a reliable way when calculated with skyrme interactions . we illustrate this statement below with some calculations of isoscalar
e1 strength in @xmath2 .
we carry out the calculation with 5 different skyrme interactions : our preferred interaction sko@xmath70 @xcite ( preferred for reasons discussed in ref .
@xcite ) , and the older , commonly used interactions siii @xcite , skm@xmath71 @xcite , sly4 @xcite , and skp @xcite .
to get some idea of how well they will work , we calculate the strength distribution of the isoscalar
e1 operator @xmath72 this operator is interesting because it is the isoscalar version of the schiff operator ( the isoscalar version of the second term in the schiff operator acts only on the center of mass and so does nt appear in @xmath73 ) .
the isoscalar
e1 strength , measured , e.g. , in @xmath2 @xcite , seems to fall mainly into two peaks .
the high energy peak , related to the compressibility coefficient @xmath74 , @xcite , is observed to lie between 19 and 23 mev , depending on the experimental method used @xcite .
recent interest has focused on a smaller but still substantial low
energy peak around @xmath75 , which has been studied theoretically in the rpa @xcite as well as experimentally @xcite .
figure [ fig : isgdrpb ] shows the predictions of several skyrme interactions in the rpa , with widths of 1 mev introduced by hand following ref .
@xcite , for the isoscalar
e1 strength distribution in @xmath2 .
the figure also shows the locations of the measured low and high energy peaks of davis et al . @xcite and of clark
nearly all self consistent rpa calculations , including ours ( except with skp ) over predict the energy of the larger peak by a few mev @xcite .
siii does a particularly poor job .
the predicted low
energy strength is closer to experiment , though usually a little too low .
table [ tb : isgdrpb ] summarizes the situation .
unfortunately , the data are not precise enough to extract much more than the centroids of the two peaks .
since the schiff strength distribution in @xmath0helps determine the schiff moment , it would clearly be useful to have better data , either in that nucleus or a nearby one such as @xmath2 . the isovector
e1 strength distribution also bears on the schiff moment through the second term of the schiff operator , but it is well understood experimentally and generally reproduced fairly well by skyrme interactions .
.comparison between experimental and theoretical results for the centroids of the low- and high - energy peaks in the distribution of isoscalar e1 strength in @xmath69 .
the experimental results are from refs .
@xcite and @xcite .
@xcite identifies only the high - energy peak . )
the theoretical distributions are from self - consistent hf+rpa calculations with five different skyrme interactions . [
cols="<,^,^,^",options="header " , ] the several skyrme interactions we use all give different results , but the spread in numbers is about a factor of four in the isoscalar channel , two in the isotensor channel , and much less for the large isovector coefficient @xmath76 .
it is possible that all the interactions are systematically deficient , but we have no evidence for that . in any event ,
the effective interaction is not the only source of uncertainty .
we have evaluated only a subset of all diagrams , and although it is not obvious whether all the rest should be evaluated with effective interactions that are determined through hartree
fock or rpa based fits , there are some that should certainly be included and we would like to estimate their size
. the diagrams labeled c and d in fig . [ fig : others ] are the leading terms in bubble chains that would result from including @xmath26 in the hartree fock calculation ( c ) and in the qrpa calculation ( d ) in @xmath31 .
we evaluated both sets in the simple nucleus @xmath69 ( which has no pairing at the mean field level ) , and found that diagram c can be nearly as large as the type
b diagrams in the isovector and isotensor channels .
the same is true of diagram d in the isotensor channel . in that nucleus
, however , diagram a is much larger than all the others and essentially determines the @xmath77 . in @xmath0
, we only evaluated diagram c , but found that even though diagrams a and b can cancel there , they do so the most in the isoscalar channel , where the diagrams c and d are smallest . in the end diagram c never amounts to more than 10% of the sum of the a and b diagrams , and usually amounts to much less .
including the higher order ( qrpa ) terms in the bubble chain will only reduce the diagram - c contribution , so we conclude that it can be neglected .
we are not positive that the same statement is true of diagram d , but unless it is much larger in @xmath0 than in @xmath69 ( none of the other diagrams are ) , it can be neglected too .
the diagram labeled e represents a correction from outside our framework that is of the same order as the terms we include .
we evaluated it in @xmath69 ; it is uniformly smaller than those of type c and d. unless the situation is very different in @xmath0 , it can be neglected as well .
the fact that these extra diagrams are all small is not terribly surprising ; they all bring in extra energy denominators and/or interrupt the collective bubble .
but the value of which we estimated through calculations in @xmath69(and @xmath0 in the case of diagram c ) .
we have omitted the labels on the the interactions ; they are the same as in the earlier figures . ]
we have also not included the normalization factor @xmath49 in equation ( [ eqn : per4 ] ) .
when calculated to second order in @xmath69 , it is about 1.05 , independent of the skyrme interaction used .
though this factor could be larger in rpa order because of low lying phonons , most strength is pushed up by the rpa and we do not anticipate a large increase . it is reasonable to assume these statements are true in @xmath0 as well . at short distances
the nn potential is strongly repulsive and the associated short
range correlations should be taken into account .
reference @xcite , however , reports that the correlations reduce matrix elements of the effective one body p and t violating pion
exchange potential only by about 5% , and in ref .
@xcite , which calculates the schiff moment of @xmath78 , their effects are smaller than 10% .
we are not missing much by neglecting them , though we would be if we included a @xmath11meson exchange potential .
when all is said and done , the uncertainty is dominated by our uncertainty in the effective interaction .
our preferred interaction is sko@xmath70 , which ( to repeat ) gives the result @xmath79~. \label{eqn : myresults2}\ ] ] if instead we average the results from the five interactions , we get @xmath80~ , \label{eqn : averesults}\ ] ] the range of results in tab . [ tb : skyrme ] is a measure of the uncertainty .
our goal has been a good calculation of the dependence of the schiff moment of @xmath0 , the quantity that determines the electric dipole moment of the corresponding atom , on three p and t violating @xmath1nn coupling constants .
the current experimental limit on the dipole moment of the @xmath0 atom , @xmath83 , together with the theoretical results of ref.@xcite , @xmath84 , yields the constraint @xmath85 .
the @xmath77 calculated in this paper then give a constraint on the three @xmath86 . in obtaining the @xmath77 we have included what we believe to be most of the important physics , including a pion
exchange p and t violating interaction , collective effects that are known to renormalize strength distributions of schiff like operators , pairing at the mean field level , self consistency , and finally , several different skyrme interactions .
the last of these , together with an examination of effects we omitted , allows us to give the first real discussion of uncertainty for a calculation in this experimentally important nucleus .
we conclude that while the isovector coefficient @xmath76 is not very different from the initial estimate of ref .
@xcite , the isoscalar coefficient , which determines the limit one can set on the qcd parameter @xmath87 , is smaller by between about 9 and 40 ( with the former our preferred value ) and the isotensor parameter @xmath88 by a factor between about 7 and 16 ( with our preferred value about 10 ) .
the uncertainty in these numbers comes primarily from our lack of knowledge about the effective interaction .
there is good reason to make better measurements of low lying dipole strength , particularly in the isoscalar channel .
they would help to unravel the details of nuclear structure that determine the schiff moment .
we thank j. dobaczewski and j. terasaki for helpful discussions .
this work was supported in part by the u.s .
department of energy under grant de - fg02 - 97er41019 and by the fundao para a cincia e a tecnologia ( portugal ) .
j. h. de jesus thanks the institute for nuclear theory at the university of washington for its hospitality and the department of energy for partial support during the completion of this work .
there are some differences between our rules for quasiparticle diagrams and the usual rules for particle hole diagrams .
the main one is that one and two body operators are written in a quasiparticle basis and do not conserve quasiparticle number , leading to different expressions for matrix elements .
an example is the generic quasiparticle operator @xmath89 , which contains two quasiparticle creation and no destruction operators .
its matrix elements will be written @xmath90 , which means that it creates two quasiparticle states @xmath91 out of the quasiparticle vacuum @xmath55 . in
what follows , in " refers to lines with arrows pointing toward the vertex and out " to lines pointing away from it .
a diagram should be read from top to bottom , and from left to right .
the rules are then : 1 .
each operator @xmath92 contributes @xmath93 ; 2 .
each operator @xmath89 contributes @xmath94 ; because the diagram is read from the left , the label out " is on the line that is further to the left ; 3 .
each operator @xmath95 contributes @xmath96 ; 4 .
each operator @xmath97 contributes @xmath98 ; 5 .
each operator @xmath99 contributes @xmath100 ; 6 .
each operator @xmath101 contributes @xmath102 ; 7 .
each operator @xmath103 contributes @xmath104 ; 8 .
each operator @xmath105 contributes @xmath106 ; 9 .
the diagram should be summed over all intermediate states ; 10 .
energy denominators are evaluated by operating with @xmath107 between the action of every two operators in the diagram , giving @xmath108^{-1}$ ] , where @xmath109 are quasiparticle energies ; 11 .
the phase for each diagram is @xmath110 , where @xmath111 is the number of closed loops .
a factor of @xmath112 is included for each pair of lines that start at the same vertex and end at the same vertex .
we first summarize some important quantities involving the quasiparticle creation and annihilation operators @xmath113 and @xmath114 , which are defined in terms of the usual particle operators @xmath115 and @xmath116 by @xmath117 here @xmath118 from the anti
commutation rules for @xmath115 and @xmath116 , we derive the following anti commutation rules for the quasiparticle operators in eq .
( [ eqn : quasiparticles2 ] ) @xmath119 using definition ( [ eqn : quasiparticles2 ] ) and properties ( [ eqn : qpproperties1 ] ) and ( [ eqn : qpproperties2 ] ) , one can write a one body operator in second quantization as @xmath120 where @xmath121 here , the subscripts indicate the number of quasiparticle creation and annihilation operators involved . in the same way , a two body operator takes the form @xmath122 where @xmath123 with @xmath124 , @xmath125 and @xmath126 .
the matrix elements of quasiparticle operators are related to those of the usual one and two body operators .
we show how this works for the operator @xmath127 ; the generalization to other operators follows automatically . @xmath127 adds two quasiparticles to any state . from @xmath128 and eq .
( [ eqn : t20 ] ) , we write @xmath129 since @xmath130~ , \label{eqn : t20b}\ ] ] eq .
( [ eqn : t20a ] ) becomes @xmath131 | we calculate the schiff moment of the nucleus @xmath0 , created by @xmath1nn vertices that are odd under parity ( p ) and time reversal ( t ) .
our approach , formulated in diagrammatic perturbation theory with important core
polarization diagrams summed to all orders , gives a close approximation to the expectation value of the schiff operator in the odd a hartree
fock
bogoliubov ground state generated by a skyrme interaction and a weak p and t odd pion exchange potential . to assess the uncertainty in the results , we carry out the calculation with several skyrme interactions , the quality of which we test by checking predictions for the isoscalar e1 strength distribution in @xmath2 , and estimate most of the important diagrams we omit . |
the energy behaviour of the scattering amplitude may be consistently described by two different scenarios for the asymptotic regime @xcite .
one is called the _ weak coupling _ of the pomerons . in this case , at very high energy , @xmath7 , the cross sections tend to the universal constant value _ tot as s . in order not to violate unitarity ,
the triple - pomeron coupling must vanish with vanishing transverse momentum , @xmath8 , transferred through the pomeron @xcite g_3p q^2_t as q_t 0 .
[ eq : weak ] another possibility is called the _ strong coupling _
scenario @xcite . here , at a very high energies , the cross sections grows as _
tot^ with 0<2 , and the bare vertex g_3p|_q_t 0 constant .
[ eq : strong ] the present data are usually described within the froissart - like limit of the second scenario ( with @xmath9 ) .
however to reach asymptotics we need * very * high energy the energy at which the slope of the elastic amplitude , @xmath10 is dominated by the second term , that is when @xmath11 .
this is far beyond the energies available at present .
another possibility , to distinguish between the _ weak _ and _ strong _ approaches , is to study the @xmath8 dependence of the bare triple - pomeron vertex @xcite .
thus , it is important to extract the _ bare _ vertex before its behaviour is affected by absorptive corrections . in ref .
@xcite we analysed the data in the triple - pomeron region accounting for absorptive effects in the framework of a two - channel eikonal , see also @xcite .
that is , we performed a triple - regge analysis of the available @xmath12 data for @xmath13 and @xmath14 ( where @xmath15 and @xmath16 is the mass of system @xmath17 ) , allowing for screening ( absorptive ) effects . to be precise , we fitted the cern - isr and @xmath18 dependence . ]
@xcite , fnal fixed - target @xcite and tevatron @xcite data for @xmath13 and @xmath14 .
the differential distributions for the fnal fixed target and tevatron experiments can be found in ref .
both the ` strong ' and ` weak ' coupling scenarios were considered .
we found that the data favoured the ` strong ' coupling scenario @xcite .
however , there is another possibility which should be studied . in the ` weak ' coupling scenario considered in @xcite a ` scalar ' form of the vanishing of the coupling as @xmath19 , ( [ eq : weak ] ) ,
was assumed . as was pointed out by v.n .
gribov @xcite , it is natural to have a vector form of the triple - pomeron vertex which vanishes as @xmath19 . indeed in feynman diagrams for the @xmath20 amplitude we never deal with @xmath21 , but rather with vector @xmath22 multiplied by some vector @xmath23 that characterizes the final state @xmath17 .
thus we may have a weak coupling in which the vertex of @xmath24 dissociation has the vectorial form , was considered in @xcite . ]
t ( that is _ t ) .
[ eq : pol ] an example is photon - exchange . in this case
the vector form of the @xmath24 vertex comes from current conservation ( gauge invariance ) , and is clearly seen in the weizscker - williams approach @xcite , where the polarisation vector of a coulomb - like photon , @xmath25 , is replaced by @xmath26 . here
, the ( coulomb ) photon plays the role of pomeron exchange and ` @xmath27 ' is the momentum fraction transferred through the photon ( or pomeron ) .
the ` polarisation ' structure , ( [ eq : pol ] ) , will change the predicted cross sections , since it leads to different screening corrections .
however , can this ` vectorial weak ' scenario describe the data in the triple - pomeron domain ?
triple - regge diagram .
the conjugate momenta to @xmath28 are @xmath29 . if @xmath30 , then @xmath31.,height=151 ] if , for the moment , we neglect the screening correction , then the @xmath32 triple - regge diagram of fig . [ fig:3rb ] gives the contribution = _ j(0)_i^2(t)g_iij(t)()^2_i(t)-2()^_j(0)-1 , where @xmath33 is the coupling of reggeon @xmath34 to the proton , @xmath35 is the trajectory of reggeon @xmath34 , and @xmath16 is the mass of the system @xmath17 produced by proton dissociation ( in fig .
[ fig:3rb ] this system is described by reggeon @xmath36 ) .
let us first recall the ` strong ' coupling case , ( [ eq : strong ] ) , that we studied in @xcite .
for this we used a simple exponential parametrisation of the triple - regge vertices g_iij(t ) = g_iij(0 ) exp(b_iij(q^2+|q^2-k^2_t ) ) , where the momenta are defined in fig .
[ fig:3rb ] , and where @xmath37 .
screening effects were then included by working in impact parameter , @xmath38 , space and using suppression factors of the form exp(@xmath39 ) . recall that in the eikonal approach the elastic amplitude has the form t_el = 1-e^-/2 , where the opacity = _ i = p , r _ i with _ p=_p^2 ( 0)()^_i(0)-1 , where @xmath40 is the @xmath41-slope of the pomeron exchange amplitude .
a similar expression holds for the reggeon opacity @xmath42 . to determine the @xmath8 or @xmath41 dependence we took the fourier transforms with respect to the impact parameters specified in fig .
[ fig:3rb ] .
we then obtained as @xmath30 . ]
= ae^i_t _ 2 f_i(b_2)e^i_t
_ 3 f_i(b_3 ) f_j(b_1 ) , [ eq:3rb ] where f_i(b_2 ) = d^2q_t _
i(q_t)()^-^_i q^2_t e^b^_iijq^2e^i_t _ 2 , [ eq : fi ] f_j(b_1 ) = d^2k_t
_ j(k_t)()^-^_j k^2_t e^-b^_iijk^2_t , [ eq : fj ] and where the @xmath8-independent factors are collected in @xmath43 a = _ j(0)_i^2(0)g_iij(0)()^2_i(t_min)-2()^_j(0)-1 .
[ eq : a ] these equations are relevant for the strong triple - pomeron scenario , see ( [ eq : strong ] ) . in ref .
@xcite we also studied the ` scalar ' weak triple - pomeron coupling ansatz , ( [ eq : weak ] ) . to do this
we included a factor @xmath8 in the integrand of the expression ( [ eq : fi ] ) for @xmath44 when @xmath45 , and also @xmath46 in the analogous formula for @xmath47 . to obtain the screening corrections , for a single - channel eikonal
, we then included in the integrands on the right - hand side of ( [ eq:3rb ] ) the factors ( -(_2-_1)/2 ) exp(-(_3-_1)/2 ) s(_2-_1 ) s(_3-_1 ) .
that is , we computed .
|_iij = a f_j(b_1 ) |i(b_1)|^2 , [ eq:3rbscreen ] where @xmath48 is given by i(b_1 ) e^i_t _ 2 f_i(b_2 ) s_i(_2-_1 ) .
[ eq : sj ] here @xmath8 is the transverse momentum of the outgoing proton , which is now the sum of the transverse momentum of the pomeron coupling to the triple - pomeron vertex and that of the screening pomeron . these were the ` strong ' and ` weak ' scenarios studied in detail in @xcite .
as mentioned above , we found that the data favoured the ` strong ' coupling scenario .
our concern here is the possibility that we may have a ` weak ' triple - pomeron coupling with a vectorial structure . that is the coupling in fig .
[ fig:3rb ] is of the form g_3p(t ) = g_3p(0 ) exp(b_iij(q^2+|q^2-k^2_t))_,q_t|q_t .
the final product means that an additional factor @xmath49 will occur in the integrand of ( [ eq : fi ] ) .
the presence of a vectorial @xmath49 means that after the angular integration in ( [ eq : fi ] ) , we now obtain the bessel function @xmath50 , and not @xmath51 as was found in @xcite .
also , in the impact parameter , @xmath38 , representation the amplitude corresponding to ( [ eq : fi ] ) takes a vector form _ i ( ) = f(b ) . due to the factor @xmath50
, the amplitude vanishes at @xmath52 , where the screening effect , exp@xmath53 , is at its maximum .
thus , we now anticipate a weaker screening for the triple - pomeron term .
moreover , as the amplitude now has vector form ( analogous to that occurring in the @xmath54 contribution discussed in section 4.3 of @xcite ) we have to consider the components @xmath55 and @xmath56 separately . that is @xmath48 of ( [ eq : sj ] )
is given by |i|^2=|i_x|^2+|i_y|^2 , [ eq : ixy ] where it is convenient to direct @xmath27 along @xmath49 .
the generalisation of the formalism to a two - channel eikonal is straightforward .
it has been presented in section 4.2 of @xcite .
cross section data that are fitted using the weak - vector ( continuous curves ) , weak - scalar ( dashed curves ) and strong ( dot - dashed ) triple - pomeron coupling ansatzes .
( @xmath57 ) . here ,
the curves corresponding to the ( strong , weak ) coupling fits of the fnal data have been normalised ( down , up ) by 15@xmath58 at @xmath59 546 gev and by 10@xmath58 at @xmath59 1800 gev , to allow for the normalisations found for these data in the respective fits.,height=566 ] -dependence of @xmath12 at @xmath60 and @xmath61 obtained using the weak - vector triple - pomeron coupling , with that corresponding to the weak - scalar triple - pomeron coupling used in @xcite , together with the data available at these kinematic values.,height=642 ] -dependence of @xmath12 at @xmath62 and @xmath63 obtained using the weak - vector triple - pomeron coupling , with that corresponding to the weak - scalar triple - pomeron coupling used in @xcite , together with the data available at these kinematic values.,height=642 ] , measured in the collider experiments at fnal @xcite , obtained in the weak - vector triple - pomeron coupling fit .
the individual triple - regge contributions are also shown.,height=529 ] as can be seen from figs .
[ fig : weak]@xmath64[fig : tev ] , the presently available data in the triple - regge region are well described assuming that the triple - pomeron vertex has a vectorial ` weak ' form .
[ fig : weak ] also shows the results of two previous fits , taken from ref .
@xcite , which were obtained with the assumption that the triple - pomeron vertex had , first , a ` strong ' coupling and , then , a scalar ` weak ' coupling . all three fits used the same data sets .
recall that these three scenarios differ in how the triple - pomeron coupling vanishes as @xmath19 : @xmath65 @xmath66 @xmath67 table 1 compares the values of parameters obtained for the vectorial weak scenario with those obtained in ref .
@xcite for the strong and scalar weak scenarios .
note that for the case of vectorial ` weak ' coupling we get practically the same @xmath68 as in the more popular ` strong ' coupling scenario .
the secondary reggeon contributions , that is the @xmath69 , @xmath70 and @xmath71 terms , coincide , within the error bars , in all three scenarios .
.the values of the _ `` bare '' _ triple - regge couplings @xmath72 of ( [ eq : a ] ) , and slopes @xmath73 of ( [ eq : fi ] , [ eq : fj ] ) , obtained in the three optimum fits to the @xmath12 data .
gev units are used ; so , for example , the couplings @xmath74 have units of @xmath75 .
the parameters for the strong and weak - scalar fits are taken from ref .
recall that all the slopes @xmath73 are set to zero , except for those of the @xmath76 and @xmath71 vertices in the weak coupling fit . [ cols="^,^,^,^",options="header " , ] unlike that for the weak - scalar , the weak - vector triple - pomeron contribution has no dip at very low @xmath77 gev@xmath78 ( see figs .
[ fig : tdep2 ] , [ fig : tdep21800 ] and [ fig : tdeplhc ] ) .
the _ vector _ contribution vanishes at @xmath79 since we have no direction .
more rapid decrease of the @xmath76 contribution at large @xmath80 , seen in fig .
[ fig : tev ] , is due to the longitudinal part of the momentum transfer -t = q^2_t+q^2_=(q^2_t+^2 m^2_p)/(1- ) . in
the vectorial ` weak ' scenario the triple - pomeron contribution vanishes as @xmath1 , while for fixed @xmath41 , the value of @xmath8 decreases as @xmath18 increases .
measurements of @xmath12 at the lhc , especially at small @xmath81 , should be able to distinguish between the three scenarios , see fig .
[ fig : tdeplhc ] .
[ t ] -dependence of the @xmath12 at @xmath62 and @xmath82 tev obtained using the weak - vector ( continuous curves ) , weak - scalar ( dashed ) and strong ( dot - dashed ) triple - pomeron coupling scenarios.,height=453 ]
recall that in perturbative qcd , the leading - order bfkl triple - pomeron vertex does not vanish , but takes a non - zero constant value at @xmath79 .
this result corresponds to interactions at very small distances .
however , at larger distances , relevant for @xmath19 , the absorptive effects caused by enhanced multi - pomeron diagrams could modify the perturbative qcd result leading to a vanishing triple - pomeron coupling as @xmath19 .
this vanishing behaviour also looks natural if we bear in mind confinement , which does not allow colour - induced interactions at large distances . from this point of view
it looks encouraging that the present triple - regge data may be well described within the vectorial ` weak ' approach , after accounting for absorptive corrections .
the only minor problem is the inelastic @xmath83 diffractive photoproduction observed at hera @xcite . due to the small cross section of the @xmath83-proton interaction ,
the absorptive effects in this case are very small .
the ratio , @xmath84 , of the cross section with proton dissociation , @xmath85 , integrated over the mass region @xmath86 gev , to that of ` elastic ' photoproduction , @xmath87 , was measured as a function of momentum transfer @xmath41 . at the smallest value of @xmath88 gev@xmath78 measured by zeus@xcite , this ratio @xmath89 @xcite . on the other hand , using the parameters of our ` weak - vector ' fit from table 1 , we predict a smaller value @xmath90 here @xmath91 and @xmath92 denote the contributions of the @xmath76 and @xmath71 terms respectively .
this should be compared to the predictions @xmath93 and @xmath94 obtained using the strong and ` weak - scalar ' fits of ref .
indeed , in @xcite we concluded that the weak - scalar triple - pomeron coupling was disfavoured both by the fit to the triple - regge data and by the @xmath83 data . on the other hand ,
the triple - regge data can be reasonably accommodated by the weak - vector coupling , although this triple - pomeron coupling is still a bit disfavoured , in comparison with the strong coupling regime , by the @xmath83 data .
unfortunately , ( a ) we have no data at smaller @xmath95 and ( b ) the cross section of inelastic @xmath96 diffractive photoproduction has never been published in a journal .
thus , these data can not be considered as a strong argument against the ` weak - vector ' coupling scenario .
it would be very interesting to measure the @xmath41 dependence of high - mass diffractive dissociation , @xmath12 , at the lhc in order to choose between the possible vectorial ` weak ' and ` strong ' coupling asymptotic regimes .
mgr and egsl thank the ippp at the university of durham for hospitality .
the work was supported by the russian state grant rsgss-3628.2008.2 .
xx for recent reviews see + m.g .
ryskin , a.d .
martin and v.a .
khoze , in proc . of gribov memorial workshop ( budapest ,
may 2005 ) , p. 115
, arxiv : hep - ph/0506272 ; + m. g. ryskin , a. d. martin , v. a. khoze and a. g. shuvaev , j. phys .
g * 36 * , 093001 ( 2009 ) [ arxiv:0907.1374 [ hep - ph ] ] . .
y. i. azimov , v. a. khoze , e. m. levin and m. g. ryskin , nucl .
b * 89 * , 508 ( 1975 ) ; + y. i. azimov , v. a. khoze , e. m. levin and m. g. ryskin , sov . j. nucl .
* 23 * ( 1976 ) 449 [ yad .
* 23 * ( 1976 ) 853 ] .
luna , v.a .
khoze , a.d .
martin and m.g .
ryskin , eur .
j. * c59 * ( 2009 ) 1 .
azimov , e.m .
levin , m.g .
ryskin and v.a .
khoze , proc .
9th winter school of the leningrad nuclear physics institute , feb .
1974 ( in russian ) , cern translation 74 - 8 ; + y. i. azimov , v. a. khoze , e. m. levin and m. g. ryskin , sov .
j. nucl .
* 21 * , 215 ( 1975 ) [ yad .
* 21 * , 413 ( 1975 ) ] . c. f. von weizscker , z. phys .
* 88 * ( 1934 ) 612 ; + e. j. williams , phys .
rev . * 45 * , 729 ( 1934 ) .
zeus collaboration : abstract 549 , int .
europhysics conf . on hep , aachen ,
july 2003 .
zeus collaboration : s. chekanov et al .
, nucl . phys . *
b695 * ( 2004 ) 3 ; + h1 collaboration : c. adloff et al . , eur .
j. * c10 * ( 1999 ) 373 ; + zeus collaboration : s. chekanov et al .
j. * c24 * ( 2002 ) 345 ; + h1 collaboration : a. aktas et .
al . , eur .
j. * c46 * ( 2006 ) 585 . | we study the case when the triple - pomeron vertex is assumed to have a vectorial form , that is , the amplitude of high - mass diffractive dissociation vanishes as @xmath0 as @xmath1 .
we find that the available data in the triple - reggeon region may be well described in such a ` weak ' coupling scenario , providing that absorptive effects are taken into account .
we compare this weak ( vector ) coupling scenario with the strong and weak ( scalar ) coupling scenarios .
corresponding predictions are presented for an lhc energy of 14 tev .
ippp/10/40 + dcpt/10/80 + + * the possibility that the triple - pomeron coupling + * e.g.s .
luna@xmath2 , v.a .
khoze@xmath3 , a.d . martin@xmath4 and m.g .
ryskin@xmath3 + @xmath5 instituto de fsica e matemtica , universidade federal de pelotas , caixa postal 354 , cep 96010 - 900 , pelotas , rs , brazil + @xmath4 institute for particle physics phenomenology , university of durham , durham , dh1 3le + @xmath6 petersburg nuclear physics institute , gatchina , st .
petersburg , 188300 , russia |
the large conceptual interest for the properties of graphene was motivated first by the relativistic - like effects in the honeycomb structure and the opportunity for the high temperature relativistic integer quantum hall effect @xcite .
next , the interest was also stimulated by the topological insulating properties , based on the helical edge states of graphene , that support the quantum spin hall effect ( qshe ) .
the topological phase of the graphene , predicted by kane and mele @xcite , is induced by the _
intrinsic _ spin - orbit ( iso ) coupling , which opens a _ topological gap _ between the dirac cones located at the points @xmath0 and @xmath1 in the brillouin zone .
the gap is filled with helical states stretching along the edges , which appear in pairs and carry opposite spins in opposite directions .
one has to remind that the helical states are protected against disorder by the time - reversal symmetry of the hamiltonian , but they are not protected against the spin - flip processes involved by the rashba - type coupling or against a staggered sublattice potential ; a phase diagram can be theoretically obtained in the space of the coupling parameters corresponding to the different interactions @xcite . because of the very small spin - orbit coupling , the qshe could not be proved experimentally in graphene .
the experimental endevor moved toward other 2d systems which show topological properties , like cdte / hgte / cdte @xcite or alsb / inas / gasb / alsb @xcite quantum wells , and toward the 3d topological insulators @xcite .
the honeycomb lattice remains however under investigation , as optical and synthetic such lattices ( where the magnetic flux and the spin - orbit coupling strength can be artificially tuned ) were obtained @xcite .
another line of investigation consists in finding techniques for the enhancement of the spin - orbit coupling by introducing adatoms in graphene @xcite or using other 2d materials like silicene @xcite .
when discussing different topological systems , attention should be paid to the existence and behavior of the different types of edge states ( helical or chiral ) , which depend on various factors as the lattice structure , geometry of the sample , spin - orbit interaction , presence of the magnetic field .
we remind , for instance , that even for vanishing spin - orbit interaction , edge states are supported by the zig - zag graphene ribbon , but not by the arm - chair ribbon .
this was proved by solving the dirac equation with proper boundary conditions @xcite or by calculating the zak invariant @xcite . however , when the intrinsic spin - orbit coupling is considered , the edge states , which become now spin - polarized and helical , are present both in the zig - zag @xcite and arm - chair @xcite ribbons .
the relevance of the sample geometry can be noticed also by unfolding the ribbon and imposing everywhere vanishing boundary conditions .
for the finite - size plaquette , we find that the helical edge states extend all around the perimeter , looking however different along the zig - zag and arm - chair margin , respectively ( see fig.2 ) .
when the system is subject to a magnetic field , which breaks the time - reversal symmetry , we expect interesting peculiarities of the edge states under the mixed effect of the magnetic field and spin - orbit ( so ) interaction .
even for the torus geometry ( i.e. , with periodic boundary conditions along the both directions ) , when the edge states are missing , the hofstadter energy spectrum exhibits relevant aspects in the simultaneous presence of the intrinsic so coupling , rashba - type so interaction and perpendicular magnetic field @xcite .
it turns out that the topological gap opened around @xmath2 @xcite closes with increasing magnetic flux , and it is _ weak _ in the sense that it is annihilated by the rashba coupling .
it was also found in the graphene ribbon subjected to a magnetic field that an additional staggered potential @xcite induces spin imbalanced regions in the spectrum , where the number of spin - up and spin - down states are different @xcite .
for the confined graphene system , a spin imbalance will be detected in this paper as due to the splitting induced by the iso coupling , and interesting consequences for the charge and spin transport will be put forward . in this paper
we study the confined honeycomb lattice obtained by imposing vanishing boundary conditions all along the perimeter .
this approach simulates the mesoscopic case , provides some specific new properties , and allows for the calculation of the transport properties and disorder effects .
the plaquette exhibits both zig - zag and arm - chair boundaries as in fig.1 , and the first question concerns the fate of the helical states familiar from the cylinder ( ribbon ) geometry .
the combined effect of the intrinsic so coupling and perpendicular magnetic field on the spectral properties of the graphene plaquette are discussed in the next section .
some specific spectral properties anticipate new aspects of the charge and spin transport , which are presented in section iii .
the robustness of the spectral properties against the anderson disorder is analyzed in a subsection .
the disordered spectrum corresponding to helical states exhibits a tulip - like picture due to existence in the graphene spectrum of some highly degenerated energies ( corresponding to the saddle points in the infinite model ) . in section
iii we show how the symmetry properties of the spin - dependent electron transmittance give rise to particular features of the charge and spin currents , which are calculated in the landauer - bettiker formalism for a four - lead device . both the spin and charge hall conductance exhibit supplementary plateaus corresponding to gaps characterized by the imbalance between the edge states with opposite spins .
the conclusions are summarized in the last section .
in this section we reveal new spectral properties of topological graphene in perpendicular magnetic field , insisting on the features of the different types of edge states that result by imposing vanishing boundary conditions all around the perimeter of the plaquette .
the localization of the wave function and the robustness against disorder are discussed .
we remind that the hofstadter spectrum of the graphene sheet in the absence of the spin - orbit coupling looks like a double butterfly @xcite , and exhibits both relativistic dirac - landau bands in the middle and conventional bloch - landau bands at the extremities of the spectrum @xcite , separated by well - defined gaps . in the case of
the _ finite _ plaquette , the vanishing boundary conditions and the perpendicular magnetic field generate _
chiral _ edge states that fill the gaps .
the sign of chirality is determined by the direction of the magnetic field , and one has to mention that the relativistic and conventional edge states show opposite chirality .
a second class of edge states in the system are the _ helical _ ones , which appear in the presence of the iso coupling , and are located in the topological gap opened by this interaction .
our aim in this section is : i ) to note the evolution of the helical states with the magnetic field , ii ) to evaluate the degree of localization along the edges of the helical and chiral states , iii ) to identify domains of imbalance between the densities of spin - up and spin - down edge states ( where the charge and spin currents should become anomalous ) , iv ) to see the effect of the anderson disorder on the energy spectrum and on the edge localization of helical states . adopting the tight - binding representation , as the 2d honeycomb lattice contains two atoms @xmath3 and @xmath4 per unit cell
, we define corresponding creation and annihilation operators @xmath5 , where @xmath6 is the spin index and @xmath7 are the cell indices ( see fig.1 ) .
the hamiltonian defined on the honeycomb lattice can be written as : @xmath8 where the first term describes the tunneling between the nearest neighbors , while the second one represents the intrinsic spin - orbit interaction . in the presence of a perpendicular magnetic field , described by the vector potential @xmath9 ,
the first term reads : @xmath10 @xmath11 are the atomic energies , @xmath12 is the hopping integral between the sites @xmath3 and @xmath4 , and the peierls phase due to the magnetic field equals @xmath13 , where the magnetic flux through the unit cell @xmath14 is expressed in quantum flux units @xmath15 .
the intrinsic spin - orbit hamiltonian @xcite conserves the electron spin @xmath16 , and invokes the hopping to the six next - nearest - neighbors , keeping also in mind the chirality of the trajectory between the two sites . in the presence of the magnetic field ,
the hopping terms acquire a supplementary phase , and the hamiltonian can be written in a compact form as @xcite : @xmath17 where @xmath18 is the spin - orbit coupling constant , @xmath19 expresses the clock- or anticlockwise chirality of the trajectory between the next - nearest - neighbors , and the phases @xmath20 should be calculated by the integration of the vector potential along each trajectory .
the hamiltonian ( 3 ) contains many terms and , for the reader s sake , we write it in detail , and show also the illustrative fig.1 : @xmath21 the phases in the above equation are the following : @xmath22 it is worth to note some symmetry properties of the energy spectrum . since the hamiltonian ( 1 ) commutes with @xmath16 , its spectrum is the union of the spin - up and spin - down eigenvalues @xmath23 , where @xmath24 ( @xmath25 being the total number of sites on the finite lattice ) .
let @xmath26 be the index of the lowest eigenvalue for both spin - up and spin - down subsets . with this notation ,
the symmetry of the hamiltonian ( 1 ) generates the property @xmath27 . in words
, this means that if the energy @xmath28 belongs to the spin - up subset of the spectrum , the energy @xmath29 exists also in the spectrum , but belongs to the spin - down subset .
one has also to note that the usual periodicity with the magnetic flux @xmath30 , which is valid at @xmath31 , is replaced by @xmath32 in the case of non - vanishing spin - orbit coupling @xcite .
an eigenfunction of the hamiltonian ( 1 ) with @xmath33 corresponding to a helical edge state is shown in fig.2 .
it is to observe that the state stretches along the whole perimeter of the plaquette , but the aspect along the zig - zag edges differs from that one along the arm - chair edges . for vanishing so coupling , the low flux range of the hofstadter butterfly of the finite - size graphene plaquette shows a thin , quasi - degenerate band at @xmath2 , as it can be noticed in fig.3(left ) .
these states correspond to the landau band indexed by @xmath34 in the periodic geometry , and their number depends on the dimension of the plaquette .
the significant changes that appear when the iso coupling is introduced are only partially studied in the presence of the magnetic field .
one knows that the topological gap existing at @xmath33 persists at low flux , but closes with increasing @xmath14 .
this gap is called weak in @xcite , and we keep the terminology .
however , the origin and properties of the edge states filling the weak topological gap of the mesoscopic grahene plaquette have not been studied yet .
they result from the simultaneous presence of the magnetic field and iso interaction , and have to justify the survival of the qshe at non - vanishing magnetic field ( see fig.12 ) .
-1.5 cm the analysis of the edge states located in the topological gap will be done by inspection of fig.3(right ) and fig.4 .
one may identify a first class of states resulting from the splitting in magnetic field of the doubly - degenerated helical states existing at @xmath33 .
these states are drawn in fig.3(right ) with thicker lines .
one notices that at low flux the splitting that separates the spin - up and the spin - down levels increases linearly with @xmath14 , however , at some higher magnetic flux , all these states merge into bands that border the weak topological gap ( colored in red for spin - up and blue for spin - down ) .
since @xmath35 and @xmath36 show opposite signs , the states of opposite spins continue to carry opposite spin currents .
besides , the weak topological gap accommodates also a second category of edge states , which are chiral states stemming from the adjacent relativistic gaps .
we can see , for instance , that spin - down states ( in blue ) coming from the relativistic gap below cross the red band ( composed of spin - up states ) , enter the weak topological gap , and eventually merge the blue band that border the topological gap from above ( and similarly for the spin - up red lines entering the topological gap from above ) .
the both types of edge states filling the weak topological gap at @xmath37 , although of different provenance ( helical or chiral ) , show opposite currents for opposite spins , so that the qshe survives at any magnetic field as far as the gap remains open . in what follows
we examine the degree of localization of different edge states .
the information regarding the edge localization can be obtained from the quantity : @xmath38 where the sum is taken over all sites @xmath39 that belong to the plaquette boundary @xcite .
the data in fig.4 , calculated at @xmath40 , indicate that the states that are close to @xmath2 are strongly localized along the edges .
this is expected , but it is less expected that the helical states that converge toward the bands confining the weak topological gap at @xmath41 are pushed away from edges , such that @xmath42 eventually vanishes at the respective energies .
this denotes that , while evolving into the two bands , the helical states lose their localized character and become more similar to bulk states .
the same fig.4 shows that outside the topological gap , in the relativistic gaps where all states are of chiral - type , the edge localization depends significantly on the spin orientation . at the same time , from fig.3(right ) one can see that the derivative @xmath43 is also spin - dependent . at @xmath44 , for instance , both the edge localization and the magnetic moment of the spin - up states are higher then for the opposite spin .
a small spin - orbit coupling ( meaning @xmath45 ) affects visible only the center of the spectrum occupied by the topological gap and relativistic bands and gaps .
the extremities of the spectrum , corresponding to the conventional landau bands / gaps , being less sensitive to the spin - orbit coupling .
this statement is proved by fig.5 , which shows eigenvalues @xmath46 and their corresponding index @xmath47 , at a given flux .
the ( quasi ) horizontal lines correspond to the energy gaps ( where the edge states are rare ) and the steps correspond to the bands ( where the bulk states are dense ) .
the difference between the two lines corresponding to opposite spins is well visible in the energy range [ -1,1 ] , while for energies outside this range the lines overlap , meaning an indistinguishable spin - orbit splitting .
we have to make two observations concerning the behavior of the spin - dependent edge states in the relativistic gaps : i ) in contradistinction to the case of the topological gap ( described in the previous subsection ) the chirality @xmath43 of edge states shows now the same sign , independently of the spin orientation .
a difference appears however in what concerns the magnitude of the derivative , which again is more pronounced for the internal gaps and less evident at higher energies .
ii ) the hofstadter butterfly exhibits the splitting of each relativistic band in two spin - dependent subbands .
the small spin - orbit gap created inbetween is filled with edge states of both spin , however , essentially , the number of spin - up states differs from the number of states with spin - down .
this denotes the existence in the energy spectrum of spin - imbalanced gaps induced by the iso coupling @xcite .
this finding should not be overlooked as it is associated obviously with an imbalance of the spin currents , which may account for a non - zero qshe in the corresponding energy range .
the explicit calculation in the next section of the spin - dependent electron transmittance confirms this prediction .
the sketch in fig.6 describes the manner in which the imbalanced gap arises . on the left
, we show the first two relativistic gaps , separated by the relativistic band , in the case of vanishing iso coupling , when all states are spin - degenerate .
it is known that the number of edge states crossing the fermi level at a given flux in the first relativistic gap is @xmath48 @xcite , while in the second gap the number is 6 . in the right panel ,
the degeneracy of the band and of the edge states is lifted in the the presence of the iso coupling , and a small spin - orbit gap arises between the spin - down ( blue ) and spin - up ( red ) subbands .
now , we are interested in the number and the spin of the edge states occurring in this gap . to this aim , considering , for instance , the upper half of the spectrum , let us notice that the spin - up ( red ) edge state crosses the spin - down subband ( blue ) and enter the spin - orbit gap .
( at the same time , the spin - down edge state is absorbed in the subband of the same spin . )
next , we notice that three spin - down edge states originating from the blue subband emerges in the spin - orbit gap , then cross the subband of opposite spin ( red ) , and eventually enter the second relativistic gap .
altogether , it turns out that there are four edge states in the gap we look at , namely @xmath49 and @xmath50 , fact that justifies the term of spin - imbalanced gap ( see also the note @xcite ) .
both the helical and chiral edge states are topological states which are robust with respect to disorder , being protected however by different symmetries .
the helical states are protected by the time - reversal symmetry ( trs ) which is preserved by the iso interaction , while the chiral edge states , despite the trs breaking , are robust against disorder due to the strong magnetic field that imposes the chiral motion and impedes the backscattering .
so , it is pertinent to ask whether the two types of states are equally robust .
up till now , we could not give a definite answer to this question , and here we restrict ourselves to follow the evolution with the disorder of the spectral properties only at @xmath33 .
we use the anderson disorder model characterized by the parameter @xmath51 defining the width of the diagonal disorder .
-1 cm the general aspect of the disordered spectrum of the confined graphene lattice shown in fig.7a is determined by the existence of regions that respond differently to the increase of the disorder strength .
one knows that , at low disorder , the topological gap is not affected , but , on the other hand , the energy ranges with very high density of states about @xmath52 ( which correspond in the periodic model to the saddle points @xmath53 in the brillouin zone ) are very sensitive to any disorder .
the consequence is a specific _ tulip_-like shape of the spectrum in the topological range @xmath54 $ ] , depicted in blue in fig.7a .
the qualitative explanation of this shape is the following : the disordered potential broadens the very dense spectrum close to @xmath55 , where the level spacing increases with the disorder strength @xmath51 and produces a compression on the topological levels located in the middle of the spectrum .
since , according to the von neumann - wigner theorem @xcite , the energy levels can not cross each other , the result is the tulip shape of the levels in the topological range . -1.0
cm the level spacing analysis helps also to understand the disorder effects on the energy spectrum .
let us define the level spacing as @xmath56 , where @xmath57 and @xmath58 means the average over all disorder configurations .
the level spacing distributions @xmath59 calculated numerically at low disorder for several @xmath47-s corresponding to states in the topological gap are shown in fig.7b .
the distributions can be well fitted with gaussian functions .
since the same distribution is exhibited also by the edge states in the integer quantum hall phase @xcite , one may conclude that the gaussian distribution of the level spacing is the feature of the states that are robust against disorder .
it is to note in fig.7b that the curves show different widths , namely the states in the center of the topological gap exhibit a narrower width ( being more robust ) than those located near the gap margin .
additional information can be obtained by calculating the edge localization eq.(6 ) as function of @xmath51 . in fig.8a
we find that the disordered helical states remain localized near edges as long as @xmath51 is small .
however , @xmath60 falls down with increasing disorder , meaning that the states extend gradually inside the plaquette , and eventually become disordered metallic - like states spread over the whole plaquette area if @xmath61 .
an example of such a uniformly distributed state originating from a helical state is shown in fig.8b .
of course , the level spacing distribution should change also from the gaussian to a wigner- dyson distribution , however this topic will be discussed elsewhere .
in this section , we simulate a four - lead device by attaching leads to a graphene plaquette , and calculate the longitudinal and transverse resistances corresponding to both spin and charge currents .
we emphasize specific properties of the transmittance matrix @xmath62 in the presence of the iso coupling that generate a uncommon behavior of the spin and charge quantum hall effect .
we find that , besides the usual plateaus of the hall conductance at @xmath63 , the iqhe gets new intermediate plateaus at @xmath64 with @xmath65 .
next , we find that these plateaus are associated with a non - vanishing quantum spin conductance @xmath66 , the sign being opposite to the usual spin hall conductance that occurs in the topological gap
. the changes of both iqhe and qshe can be observed in fig.12 , which represents the main result of the section .
these transport effects have not been explored up till now , and it turns out that they shows up in energy gaps with spin imbalance , where @xmath67 .
the transport calculations are based on the landauer - bttiker formalism .
the landauer - bttiker approach requires that the hamiltonian ( 1 ) be completed with terms describing the leads ( @xmath68 ) and the coupling between leads and the graphene plaquette ( @xmath69 ) .
then the total hamiltonian reads : @xmath70 where the last two terms are considered to be spin independent .
the quantities which enter the expression of the spin - dependent electron transmittance between the leads @xmath71 and @xmath72 are the lead - plaquette coupling @xmath73 , the matrix element of the retarded green function corresponding to the total hamiltonian eq.(7 ) and the lead density of states : @xmath74 where @xmath75 is the green function of leads .
the symmetries of the energy spectrum , discussed in the previous section , determine the properties of the green function @xmath76 , and reflect eventually , via eq.(8 ) , in the symmetries of the spin - resolved transmittance matrix : @xmath77 an inspection of the hofstadter spectrum shows that the gaps located in the central part of the spectrum ( i.e. , topological and relativistic gaps ) open at lower magnetic fluxes in comparison with the conventional landau gaps located at higher energies .
that is , it is hard to find a magnetic flux that allows to evidentiate simultaneously all the types of edge states occuring in different energy gaps . intending to compare all the three regimes ( topological , relativistic and conventional landau ) , we need to perform the calculations at a relatively small flux , and we select @xmath78 . on the other hand , at such a small flux , the transmittance @xmath79 ( that hops over a lead ) gets rather large values in some energy ranges affecting the quantum hall effect ; the implications will be discussed below .
the transmittances calculated according to eq.(8 ) are illustrated in fig.9 . in fig.9a , the plot of @xmath80 and @xmath81 prove the symmetry @xmath82 expressed by eq.(9 ) .
the two transmittances allow an easy identification of the topological gap in the middle of the energy - axis ( approximately in the range @xmath83 $ ] ) , where @xmath84 , denoting the presence of two channels of opposite spin running in opposite directions , i.e. , the well - known condition for qshe . in fig.9b
we compare the transmittance @xmath80 with @xmath85 , which carry the opposite spin but runs in the same direction .
the two curves coincides at high energy , however they show a significant shift in the central region of the spectrum .
this shift of the transmittances is an obvious consequence of the shifted spin - dependent energies in the spectrum , which has been already noticed in fig.5 .
the inset depicts the energy range that comprises the topological gap and the spin - imbalanced gap about @xmath86 , where we have to realize that @xmath87 and @xmath88 .
as the other transmittances vanish , we may conclude that in this gap there are four active channels , one of spin - up and three of spin - down , all of them running in the same direction . in the next subsection ,
when calculating the transverse resistance , we shall see that this spin - imbalance yields unusual plateaus of both integer and spin hall effect .
the relatively small value of the magnetic flux , at which we are compelled to perform the transport calculations , makes it interesting to discuss the behavior of @xmath89 .
we remind that at strong magnetic fields , the gaps corresponding to the quantized plateaus are characterized by @xmath90=integer , while all the other transmittances vanish , including @xmath89 . at such strong fields , the edge states are localized very close to the perimeter of the plaquette , and the negligible value of @xmath79 may be considered as a measure of the high degree of edge localization . for @xmath78
the transmittance @xmath91 is shown in fig.9a .
one notices the vanishing of @xmath92 in the topological gap , fact that attests the strong localization of the helical edge states . on the other hand ,
the large values shown in a rather wide range about @xmath52 can be associated with the bulky character of the quantum states , accompanied by the absence of quantum plateaus in the corresponding energy range . without spin - flip processes ,
the system behaves as a two independent spin fluids .
then , the particle current in a multi - lead device can be written in the linear approach as @xmath93 , where @xmath94 stand for the lead indices , @xmath95 is the current through the lead @xmath71 and @xmath96 is the potential at the contact site @xmath72 .
summing up the contributions of the two spin , the total charge- and spin - currents flowing through the lead @xmath71 read : @xmath97 since the transmitance matrix @xmath98 is already known , the spin - resolved longitudinal and hall resistances can be calculated according to the landauer - bttiker formalism as : @xmath99 where @xmath100 is any 3x3 subdeterminant of the transmittance matrix .
the comparative study of the hall resistance , with and without so coupling , is instructive .
this is done in fig.10a , b where @xmath101 , calculated at the flux @xmath40 , is superimposed over the corresponding hofstadter spectrum . when the spectrum and the resistance are spin independent ( the case @xmath102 )
, the two quantities looks like in fig.10a , where the antisymmetry @xmath103 is obvious . if @xmath104 , the previous antisymmetry at the reflection @xmath105 is lost since one has now to inverse also the spin .
indeed , from eq.(9 ) and eq.(11 ) , it follows immediately that @xmath106 . it is to emphasize that in fig.10b the spin - dependent hall resistances shows the plateaus @xmath107 and @xmath108 , respectively , which extend over both the topological gap and the the first relativistic gap , each one crossing the band of opposite spin .
this behavior of the resistance is the immediate result of the fact that the chiral edge states located in the first relativistic gap extend also in the topological gap , as underlined in the previous section and observable in fig.3(right ) .
since the two spins act as two parallel channels , the total charge and spin resistances ( @xmath109 and @xmath110 , respectively ) are given by : @xmath111 where the indices @xmath112 and @xmath113 stand for longitudinal and transversal ( hall ) , respectively .
it is straightforward to prove the relationships : @xmath114 that show the different symmetries of the charge and spin resistance . in the previous section
, we mentioned the presence in the relativistic range of the energy spectrum of small gaps , induced by the spin - orbit splitting , where a spin imbalance exists .
obviously , the imbalance gives rise to @xmath115 , and , in line with eq.(12 ) , this fact indicates the presence of a net spin current . in other words , the mesoscopic graphene plaquette exhibits qshe not only in the weak topological gap , located symmetrically about @xmath2 , but also in some other energy stripes , where the number of spin - up edge states differs from the spin - down edge states .
another peculiar aspect is that , contrary to the situation in the topological gap , the two spins flow in the same direction , as the chirality of the edge states is the same no matter the spin . -2
cm the result of the numerical calculation for @xmath116 and @xmath117 is presented in fig.11 together with the spin - dependent densities of states @xcite .
the shift between the spin - up and spin - down resistances , visible in the energy stripes about @xmath118 and @xmath119 , is confirmed by a similar shift between the two spin - dependent densities of states .
a last comment concerns the sign of the qshe : in the topological gap one has @xmath107 and @xmath108 ; then , according to eq.(12 ) , the sign of the total hall resistance @xmath120 is positive .
on the other hand , in the spin - imbalanced gap at @xmath86 , one has @xmath121 , and @xmath108 , meaning that @xmath120 is negative
. this opposite sign of @xmath120 follows from the fact that , in the spin - imbalanced gap , the spin - up and spin - down edge sates show the same chirality ( which is not the case in the topological gap ) .
the longitudinal and hall resistances being known from eq.(12 ) , in the spirit of the experimental work @xcite , we shall plot in fig.12 the corresponding hall conductances calculated as : @xmath122 in what concerns the charge conductance @xmath123 , one has to observe not only the vanishing value in the topological range and the known plateaus at 2 , 6 and 10 in the relativistic one , but also some unexpected plateaus at 4 and 8 ( in units @xmath124 ) .
a similar behavior is proved by the spin hall conductance @xmath125 , that shows the expected value 2@xmath126 in the topological range , and then vanishes everywhere except the same energy stripes where the unusual values of the charge hall conductance occur . in the respective stripes
the spin hall conductance equals @xmath127 .
according to the previous discussions , it is obvious that they appear in the spin - imbalanced gaps generated by the intrinsic spin - orbit interaction in the presence of the magnetic field .
the main result consists in finding anomalous plateaus @xmath128 of the qshe outside the topological gap , namely in the range of the _ spin - imbalanced _ gaps . in the same places , the iqhe exhibits also uncommon intermediate plateaus at @xmath129 .
the spin - imbalanced gaps are characterized by a non - equal number of spin - up and spin - down edge states .
they are due to the splitting generated by iso coupling , as sketched in fig.6 .
since we consider a small spin - orbit coupling ( @xmath130 ) , this type of gaps appear in the relativistic range of the energy spectrum . in the weak topological gap ,
the degeneracy of the helical states is lifted by the magnetic field @xmath4 .
the states evolve with @xmath4 and merge into the bands that confine the gap . during this process
the states lose the edge character , as shown in fig.4 . at some higher magnetic fields ,
the topological gap gets filled with edge states of chiral origin , i.e. , coming from the neighboring dirac - landau gaps .
these states come also in pairs of opposite spin and chirality ( see fig.3(right ) ) , so that the qshe survives , and continues to equal @xmath131 , even at higher magnetic fields , as long as the gap remains open .
we have noticed that , outside the topological gap , in the relativistic gaps of the energy spectrum , the degree of localization of the edge states depends on the spin .
moreover , the derivatives @xmath35 and @xmath36 are different , meaning that also the diamagnetic moments carried by states of opposite spins differ in magnitude . we have looked for disorder effects on the topological gap and helical edge states at vanishing magnetic field . although being robust at low disorder , the localization along the edges of the helical states is lost at higher disorder strength ( see fig.8a ) , when the states become of metallic - type , being distributed uniformly on the whole plaquette . we find also that the level spacing of the disordered helical states follow a gaussian distribution at low disorder ( see fig.7b ) . for @xmath132 ,
the topological gap at @xmath33 becomes progressively narrower with increasing disorder under the compression of the levels that stem from the highly quasi - degenerated regions about @xmath52 , resulting in a _
tulip_-like spectrum ( as in fig.7a ) .
the transmittance matrix is spin dependent and its symmetries are shown in eq.(9 ) and fig.9a . the symmetry of @xmath133 results in the properties of the spin and charge hall resistance of the four - lead graphene device , which we are interested in .
the numerical calculations are performed at relatively small magnetic flux ( @xmath78 ) in order to catch the effect of all gaps specific to graphene ( i.e. , of dirac - landau and conventional landau type ) .
this allows also to show the energy dependence of @xmath134 , the transmittance that kills the quantum plateaus when it takes large values ( noticeable in fig.9a ) .
we acknowledge support from pnii - id - pce research programme ( grant no . 0091/2011 ) and core programme ( contract no .
45n/2009 ) .
k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , m. i. katsnelson , i. v. grigorieva , s. v. dubonos , and a. a. firsov , nature * 438 * , 197 ( 2005 ) . c. l. kane and e. j. mele , phys .
lett . * 95 * , 226801 ( 2005 ) . c. l. kane and e. j. mele , phys .
. lett . * 95 * , 146802 ( 2005 ) .
b. a. volkov and a. o. pankratov , pisma zh .
. fiz . * 42 * , 145 ( 1985 ) .
b. a. bernevig , t. l. hughes , s. c. zhang , science * 314 * , 1757 ( 2006 ) .
m. knig , s. wiedmann , c. brne , a. roth , h. buhmann , l. w. molenkamp , x. l. qi , s. c. zhang , science * 318 * , 766 ( 2007 ) . c. liu ,
t. l. hughes , x.l .
qi , k. wang , and s.c .
zhang , phys .
lett . * 100 * , 236601 ( 2008 ) .
l. du , i. knez , g. sullivan , r. r. du , arxiv:1306.1925 [ cond-mat.mes-hall ] . m. z. hasan and m. z. kane , rev . mod . phys . * 82 * , 3045 ( 2010 ) .
y. ando , j. of phys .
japan * 82 * , 102001 ( 2013 ) .
m. gibertini , a. singha , v. pellegrini , m. polini , g. vignale , a. pinczuk , l. n. pfeiffer , and k. w. west , phys . rev .
b 79 , 241406(r ) ( 2009 ) .
l. tarruell , d. greif , t. uehlinger , g. jotzu , and t. esslinger , nature * 483 * , 302 ( 2012 ) . c. weeks ,
j. hu , j. alicea , m. franz , and r. wu , phys .
x * 1 * , 021001 ( 2011 ) .
j. balakrishnan , g. k. w. koon , m. jaiswal , a. h. castro neto , and b. zyilmaz , nature physics * 9 * , 284 ( 2013 ) .
m. ezawa , j. phys .
soc . of jpn . * 81 * , 064705 , ( 2012 ) .
c. c. liu , w. feng , and y. yao , phys .
lett . * 107 * , 076802 ( 2011 ) .
l. brey , h. a. fertig , phys .
b * 73 * , 235411 ( 2006 ) .
p. delplace , d. ullmo , g. montambaux , phys .
b * 84 * , 195452 ( 2011 ) .
o. shevtsov , p. carmier , c. petitjean , c. groth , d. carpentier , and x. waintal , phys .
x * 2 * , 031004 ( 2012 ) .
w. beugeling , n. goldman , and c. m. smith , phys . rev .
b * 86 * , 075118 ( 2012 ) . in the continuous model for the graphene ribbon
@xcite the width of the gap at @xmath135 is @xmath136 , where @xmath18 is the spin - orbit coupling constant . in the continuous model ,
the relativistic bands depend on the magnetic field as @xmath138 , while the conventional landau bands are linear . for the finite plaquette described by the tight - binding model ,
this remains true at low magnetic flux , as it can be seen in fig.3(left ) .
the differences are due to the discrete lattice structure and to the confinement .
n. m. r. peres , f. guinea , and a. h. castro neto , phys .
b * 73 * , 125411 ( 2006 ) . another way to consider
the spin - imbalance is by noting that the spin - orbit gap is an intersection of the first spin - up relativistic gap with the second spin - down relativistic gap .
then one has obviously @xmath49 and @xmath50 , i.e. , the number of spin - up and spin - down edge states are different in the spectral range corresponding to the intersection . | we study the electronic properties of the confined honeycomb lattice in the presence of the intrinsic spin - orbit ( iso ) interaction and perpendicular magnetic field , and report on uncommon aspects of the quantum spin hall conductance corroborated by peculiar properties of the edge states .
the iso interaction induces two specific gaps in the hofstadter spectrum , namely the weak topological gap defined by beugeling _
et al _ [ phys .
rev .
b * 86 * , 075118 ( 2012 ) ] , and spin - imbalanced gaps in the relativistic range of the energy spectrum . we analyze the evolution of the helical states with the magnetic field and with increasing anderson disorder .
the edge localization of the spin - dependent states and its dependence on the disorder strength is shown .
the quantum transport , treated in the landauer - bttiker formalism , reveals interesting new plateaus of the quantum spin hall effect ( qshe ) , and also of the integer quantum hall effect ( iqhe ) , in the energy ranges corresponding to the spin - imbalanced gaps .
the properties of the spin - dependent transmittance matrix that determine the symmetries with respect to the spin , energy and magnetic field of the longitudinal and transverse resistance are shown .
= 1 |
when a drop of liquid solution dries the suspended particles deposit at the periphery leaving a thicker boundary .
this phenomena is named as ` coffee - ring ' effect from the familiar stains of dried coffee drops . ` coffee - ring '
stains are the outcome of self - assembly process of the solute molecules .
the ring is formed at droplet periphery as evaporation being highest there solution carrying solutes move there to supply more solvent .
thus solute molecules accumulate and get deposited there to form ` coffee - ring ' @xcite .
coffee - ring stains are very generalized phenomena @xcite and are observed for versatile materials , such as colloids @xcite , polymers @xcite , biomolecules @xcite , liquid crystals @xcite , nanoparticles @xcite , carbon nanotubes @xcite .
although the mechanism of ring formation is driven by convection controlled by evaporation and surface tension of solvent @xcite , there are other parameters such as nature of solute molecules @xcite , the surface over which the droplet lies @xcite , droplet size @xcite , surrounding environment @xcite which also control this phenomena . depending on these factors the droplets dry leaving the solute particles over the surface in various patterns . also by adjusting marangoni flow induced by surface tension gradient ,
coffee - ring formation can be completely suppressed and uniform deposition of solutes is obtained then @xcite
. patterns formed by polymer molecules are our matter of interest for its scientific @xcite as well as technological relevance .
polymer droplets differ a lot from the usual coffee - ring stains which are formed from dispersions of colloidal particles .
the drying mechanism and consequent self - assemble behaviour of polymer molecules are employed in different types of surface treatment and coating , films , spray drying for drug delivery @xcite , inkjet printing @xcite , data storage , surface - adhered proteins assays , dna mapping technique @xcite and in soft patterning @xcite .
polymers are long chain macromolecules whose dynamical behavior is governed by their entangled chains .
entanglement which is an unique property of polymers can be challenged by confining geometry . within thin films of one
dimensional confinement polymer molecules loss their inter - molecular entanglement and get layered @xcite .
droplets are another example of confining space @xcite where due to surface effects behaviour of polymer molecules deviates from their bulk phase characters @xcite . confinement
may even lead to the crystallization of polymers @xcite .
our aim was to investigate how molecular entanglement affects the self - organization behaviour of polymers within a macroscopic droplet .
here we report our results on molecular weight and concentration dependence of the coffee - ring stains left by the dried droplets of polymer .
we have worked with the droplets of atactic polystyrene ( ps ) of molecular weights 3700 ( ps3 ) , 18100 ( ps18 ) , 44000 ( ps44 ) , 114200 ( ps1c ) , 560900 ( ps5c ) and 994000 ( ps9c ) gm / mole , all bought from sigma - aldrich , usa . ps solutions of varied concentrations were prepared by dissolving it in toluene ( merck , germany ) which is a ` good ' solvent @xcite for polystyrene @xcite .
concentrations were maintained well below the critical ` overlap concentration ' @xcite of each molecular weight .
millimeter sized droplets were deposited on si ( 111 ) substrates ( ted pella , inc . ,
usa ) using micro liter syringe . before depositing ,
si ( 111 ) substrates were sonicated in toluene as it helps in the dewetting of polymer solutions @xcite .
all experiments were performed at ambient condition with temperature lying in the range 23@xmath0c - 25@xmath0c and relative humidity varied within a range of 35% - 45% .
optical microscope images were recorded using olympus bx51 of nanonics multiview1000 microscope .
digital image analysis software ( dias ) was used for recording both still and video images .
imagej software ( national institute of health , usa ) was used for image analysis .
1 shows the optical microscope images of dried polystyrene droplets .
we observe the patterns have changed with molecular weight and concentrations .
these versatile patterns are the consequences of the contact line ( cl ) motion which is determined by the properties of solution and substrate . for droplets of all concentrations and molecular weight , at least within the range observed by us ,
the contact line moved towards the droplet center , although the details of motion varied significantly .
this variation produced varieties of patterns . here contact line motion exhibited a two step process @xcite .
initially the contact line remained pinned and after a certain time it started to recede . for low molecular weights ( image a1 and a2 of fig .
1 ) we observe a relatively faint boundary line and an almost uniform patch covering the whole area of droplet .
however , for a higher range of molecular weights , multiple rings are observed ( image b4 , c4 , d4 and e4 of fig .
these multiple rings structure is absent at higher concentrations of higher molecular weights .
1-d6 and 1-e6 show the pattern from a dried drop of ps9c of concentrations 2mg / ml and 3mg / ml , respectively .
these patterns have a very deep boundary line and some broad dark lines directed towards the droplet center . in the drying process
, rings are formed when contact line remains pinned . during the pinning period
the solvent continues its evaporation leaving the solutes at the pinning point and deposited solutes form a ring . for low molecular weight ( panel a of fig .
1 ) the contact line starts its motion almost instantaneously after deposition of droplet and a faint boundary line is produced .
also the motion of contact line continues uninterruptedly without being pinned anywhere which leaves the solute polymer molecules over the whole area of the droplet .
it produces a patch after complete evaporation which appears as a dewetting pattern at low concentration for diluteness of the initial concentration @xcite
. a closer look towards the droplet edge shows contact line leaves the solutes in the shape of small spheroidal structures ( fig .
polymer molecules of very low molecular weights remain in a coil shape within a dilute solution of good solvent .
their intra and inter molecular entanglements are also low .
the molecules remain in a ` granular ' state . when droplet is formed from a solution containing such molecules within a highly evaporating solvent , the solvent starts rapid evaporation .
the evaporation rate remains maximum near the droplet periphery due to the large area for curvature .
as the solvent gets evaporated , the contact line moves inward rapidly leaving the polymer molecules on its way .
at the same time , as the solvent evaporates , the concentration of the solution within droplet increases , the polymer chains go towards their ` overlapping region ' , for which the central part of the droplet contains a continuous film of polymers even for a dilute solution of ps3 ( fig .
thus in the drying process , polymer molecules change their behaviour from granular to collective . in case of droplets of higher molecular weights as entanglement is enhanced , the collective behaviour of polymer molecules dominates .
then contact line takes longer time to initiate its motion towards the center resulting a thick boundary line .
the pattern shows ring - like structures ( image b4 , c4 , d4 and e4 of fig .
1 ) . these rings appear when the cl stops during its ` stick - slip ' motion .
this ` stick - slip ' motion is due to the sequential pinning - depinning of the cl and polymers get deposited whenever the cl remains pinned @xcite .
we have also observed multiple branched rings ( fig .
1-a6 ) when whole part of the contact line does not move simultaneously @xcite .
the tendency of getting delayed to start motion of the contact line continues for more concentrated droplets of higher molecular weights ( ps9c ) which we have reported elsewhere @xcite .
we have observed for concentrations of 2mg / ml onwards the whole part of the contact line moved simultaneously towards the center without being pinned anywhere and leaves a continuous film of polymer molecules over the whole area . fig .
2 reveals the difference with the lower molecular weight droplets .
2 illustrates formation of continuous polymer film when molecular weight has enhanced maintaining same concentration i.e. monomer number remaining fixed .
hence this phenomena can be attributed towards the entangling behaviour of polymer chains which introduces a ` collective ' behaviour within the polymer molecules .
the broad dark lines directed inward the droplet carries the signature of ` collective ' motion of strongly entangled polymers . fig .
4 illustrates the delay to start the contact line motion for two representative concentrations ( 2mg / ml and 3mg / ml ) along with its exponential dependence @xmath1 which leads to a saturation with increasing molecular weight ( @xmath2 ) . in fig .
4 @xmath3 denotes time taken to initiate contact line motion and @xmath4 denotes saturation value for a given concentration . with increasing concentration
the values of the parameters @xmath4 , @xmath5 and @xmath6 increase . as coffee - ring effect is a droplet - size dependent phenomena @xcite , measurements were done for almost equal sized droplets .
we have observed self - assembling behaviour of polymer molecules within drying droplets .
depending on the molecular weight , hence entanglement , polymer molecules organize themselves .
the patterns left after evaporation of solvent represent that organization .
the pattern and drying mechanism reveals ` granular ' behaviour of polymer molecules at low molecular weight .
their collective behaviour emerges at high molecular weight .
thus patterns are governed by the entangling nature and confinement does not take over at this length scale .
these results show , besides the surface curvature , entanglement is the most influential factor in the evaporation process of polymer solutions within a droplet .
it indicates molecular weight can be used a tuning parameter in drying process although more quantitative understanding of the physical process behind it is very much necessary to determine the shape of final pattern .
deegan , r. d. ; bakajin , o. ; dupont , t. f. ; huber , g. ; nagel , s. r. ; witten , t. a. _ nature _ , * 389 * , 827 ( 1999 ) .
han , w. ; lin , z. _ angew .
_ , * 51 * , 1534 ( 2012 ) .
deegan , r. d. _ phys .
e _ , * 61 * , 475 ( 2000 ) .
kaneda , m. ; hyakuta , k. ; takao , y. ; ishizuka , h. ; fukai , j. _ langmuir _ , * 24 * , 9102 ( 2008 ) .
gorr , h. m. ; zueger , j. m. ; barnard , j. a. _ langmuir _ , * 28 * , 4039 ( 2012 ) .
smalyukh , i. i. ; zribi , o. v. ; butler , j. c. ; lavrentovich , o. d. ; wong , g. c. l. _ phys .
_ , * 96 * , 177801 ( 2006 ) .
tsige , m. ; mahajan , m. p. ; rosenblatt , c. ; taylor , p. l. _ phys .
e _ , * 60 * , 638 ( 1999 ) .
majumder , m. ; rendall , c. s. ; eukel , j. a.;wang , j. y. l. ; behabtu , n. ; pint , c. l. ; liu , t .- y . ; orbaek , a. w. ; mirri , f. ; nam , j. ; barron , a. r. ; hauge , r. h. ; schmidt , h. k. ; pasquali , m. _ j. phys .
b _ , * 116 * , 6536 ( 2012 ) . zgheib , n. ; putaux , j .- l . ; thill , a. ; dagosto , f. ; lansalot , m. ; bourgeat - lami , e. _ langmuir _ , * 28 * , 6163 ( 2012 ) .
zhang , s. ; li , q. ; kinloch , i. a. ; windle , a. h. _ langmuir _ , * 26 * , 2107 ( 2010 ) .
yunker , p. j. ; still , t. ; lohr , m. a. ; yodh , a. g. _ nature _ , * 476 * , 308 ( 2011 ) .
kajiya , t. ; monteux , c. ; narita , t. ; lequeux , f. ; doi , m. _ langmuir _ , * 25 * , 6934 ( 2009 ) .
cormier , s. l. ; mcgraw , j. d. ; salez , t. ; raphal , e. ; dalnoki - veress , k. _ phys .
_ , * 109 * , 154501 ( 2012 ) .
dou , r. ; derby , b. _ langmuir _ , * 28 * , 5331 ( 2012 ) .
suzuki , s. ; nakajima , a. ; yoshida , n. ; sakai , m. ; hashimoto , a. ; kameshima , y. ; okada , k. _ chem .
_ , * 445 * , 37 ( 2007 ) .
gonuguntla , m. ; sharma , a. _ langmuir _ , * 20 * , 3456 ( 2004 ) .
shen , x. ; ho , c .- m . ; wong , t .- s .
_ j. phys .
b _ , * 114 * , 5269 ( 2010 ) . massa , m. v. ; dalnoki - veress , k. _ phys .
_ , * 92 * , 255509 ( 2004 ) .
lin , z. ; granick , s. _ j. am .
_ , * 127 * , 2816 ( 2005 ) .
hu , h. ; larson , r. g. _ j. phys .
b _ , * 110 * , 7090 ( 2006 ) .
kaya , d. ; belyi , v. a. ; muthukumar , m. _ j. chem .
_ , * 133 * , 114905 ( 2010 ) .
tsapis , n. ; bennett , d. ; jackson , b.;weitz , d. a. ; edwards , d. a. _ proc .
acad . sc .
_ , * 99 * , 12001 ( 2002 ) .
wang , j. z. ; zheng , z. h. ; li , h. w. ; huck , w. t. s. ; sirringhaus , h. _ nature materials _ , * 3 * , 171 ( 2004 ) .
sirringhaus , h. ; kawase , t. ; friend , r. h. ; shimoda , t. ; inbasekaran , m. ; wu , w. ; woo , e. p.
_ science _ , * 290 * , 2123 ( 2000 ) .
hu , h. ; larson , r. g. _ langmuir _ , * 21 * , 3963 ( 2005 ) .
harris , d. j. ; conrad , j. c. ; lewis , j. a. _ phil .
a _ , * 367 * , 5157 ( 2009 ) .
chattopadhyay , s. ; datta , a. ; giglia , a. ; mahne , n. ; das , a. ; nannarone , s. _ macromolecules _ , * 40 * , 9190 ( 2007 ) .
miller , d. s. ; bertics , p. j. ; murphy , c. j. ; de pablo , j. j. ; abbott , n. l. _ science _ , * 332 * , 1297 ( 2011 ) .
biswas , n. ; ichikawa , m. ; datta , a. ; sato , y. t. ; yanagisawa , m. ; yoshikawa , k. _ chem .
_ , * 539 * , 157 ( 2012 ) .
yakhno , t. _ j. coll .
_ , * 318 * , 225 ( 2008 ) .
kawakatsu , t. statistical physics of polymers ; springer , ( 2004 ) .
sherrington , d. c. _ chem .
_ , 2275 ( 1998 ) .
strobl , g. the physics of polymers ; springer - verlag , ( 2007 ) .
mller - buschbaum , p. _ euro .
j. e _ , * 12 * , 443 ( 2003 ) .
chen , y .- j .
; suzuki , k. ; mahara , h. ; yamaguchi , t. _ chem . phys .
_ , * 529 * , 74 ( 2012 ) .
maheshwari , s. ; zhang , l. ; zhu , y. ; chang , h .- c .
lett . _ , * 100 * , 044503 ( 2008 ) .
shmuylovich , l. ; shen , a. q. ; stone , h. a. _ langmuir _ , * 18 * , 3441 ( 2002 ) .
biswas , n. ; datta , a. _ aip conf .
_ , * 1512 * , 142 ( 2013 ) . + figure 1 : optical microscope images of ps droplets of different molecular weights and concentrations . sets a , b , c , d , e correspond concentrations 0.5mg / ml , 1mg / ml , 1.5mg / ml , 2mg / ml , 3mg / ml , respectively .
sets 1 , 2 , 3 , 4 , 5 correspond ps3 , ps18 , ps44 , ps1c , ps5c , ps9c , respectively .
all scale bars are 2 mm .
+ figure 2 : magnified view of peripheral part of droplets : ( a ) ps3 of 1mg / ml , ( b ) ps3 of 3mg / ml , ( c ) ps9c of 1mg / ml and ( d ) ps9c of 3mg / ml . + figure 3 : magnified view of central part of droplets : ( a ) ps3 of 1mg / ml , ( b ) ps3 of 3mg / ml , ( c ) ps9c of 1mg / ml and ( d ) ps9c of 3mg / ml . + figure 4 : time taken to initiate contact line motion with molecular weight .
scatter points are data , solid lines are fit by eq . | dried droplets of polymer solutions of different molecular weights and concentrations leave various types of ` coffee - ring ' patterns .
these patterns are consequence of contact line motion .
we have observed for very low molecular weight the droplet periphery part contains spheroidal structures whereas central part contains continuous layer . for
higher molecular weight contact line exhibits ` stick - slip ' motion .
whereas for very high molecular weight contact line delays enough to start motion and finally moves uninterruptedly leaving a continuous layer of polymer .
we have explained this phenomena in terms of chain entanglement which is resultant of molecular weight and solution concentration .
depending on the entanglement chains can exhibit ` granular ' and ` collective ' behavior even when monomer numbers remain same within the droplets . |
supersymmetric extension of the standard model is one of interesting and promising candidates for physics above the weak scale .
even if supersymmetry ( susy ) is realized at high - energy regime , it must be broken below the weak scale , because the superpartners of the standard model particles have not been observed .
it is well known that susy breaking inside our visible sector often leads to phenomenological problems , for which one usually needs to introduce the hidden sector for susy breaking and its mediation .
recently , several important properties of susy breaking were clarified in the framework of global susy with the canonical khler potential ; the susy - breaking vacuum of f - term scalar potential has pseudo - moduli or tachyonic direction at tree level . in addition
, a superpartner of a massless fermion remains massless or becomes tachyonic ( even ) when susy is broken in renormalizable wess - zumino models .
the latter fact is called the komargodski - shih ( ks ) lemma in this paper . the general proof of these properties
has recently been given in @xcite .
the above result has important phenomenological implications , particularly to the gauge mediation of susy breaking .
the gauge mediation is one of interesting mechanisms to mediate susy breaking in the hidden sector to the visible sector at quantum level ( for a review , see @xcite ) .
the presence of pseudo - moduli and the ks lemma imply vanishing gaugino masses at the one - loop order or the existence of tachyonic direction @xcite .
that is phenomenologically unfavorable since gaugino masses are relatively light compared with susy - breaking scalar masses .
such a problem was found empirically in a direct gauge mediation model @xcite and a solution has been proposed with higher - energy local minimum @xcite .
it is natural to expect that the result would not hold if one changes any of the model assumptions , that is , the tree - level f - term scalar potential of global susy wess - zumino model with the canonical khler form .
a possible remedy is to consider the effect of non - canonical ( higher - dimensional ) khler potential @xcite such as in supergravity .
this change however might spoil the predictability / calculability of gauge mediation . as for the loop - level effect
, it would be interesting to combine the gauge mediation with the anomaly mediation @xcite without losing their merits ( see e.g. @xcite ) .
another interesting way out is to include the non - vanishing d - term scalar potential of additional gauge group . in this paper
, we systematically study the susy breaking with f and d - term scalar potentials of global susy models in which the khler form is canonical .
the models are classified into two categories under the criterion that the f - flatness conditions have a solution or not .
we then call models are in the first class when there exists no solution , i.e. f - term susy breaking .
on the other hand , we call models are in the second class when there exists a solution satisfying all the f - flatness conditions but it is destabilized by the d - term contribution . a non - vanishing ( charged ) f term is generally required to have a non - vanishing d term , and therefore susy - breaking models with f and d terms belong to each of these two classes . in the first class ,
the pseudo - moduli appear in the f - term scalar potential . as we will show , once the d - term contribution
is included , such a pseudo - moduli direction has a runaway behavior , where the expectation value runs away to the infinity and the vacuum is undetermined . in the second class , susy breaking can be realized when and only when a non - vanishing fayet - iliopoulos ( fi ) term is introduced .
we thus find that the second class of models is useful to avoid the discussion of komargodski and shih and to generate one - loop order gaugino masses in the visible sector through gauge mediation .
this paper is organized as follows . in section [ sec : fterm ] , we give a brief review on some important properties of susy - breaking models only with the f - term scalar potential . in section
[ sec : dterm ] , we systematically study generic aspects of f and d - term susy breaking
. the above two classes of models are investigated in details . in section [ sec : gaugino_mass ] , we present a simple model of gauge mediation and show that the gaugino mass is generated on the stable vacuum . section [ sec : summary ] is devoted to summarizing our results . in appendix
, we summarize the stationary conditions for the first class of models discussed in section [ sec : dterm ] .
in this section , we briefly review some general features of f - term susy - breaking models , explaining the existence of a pseudo - moduli direction and the ks lemma as well as their implications to the gauge mediation .
we consider a model with the following general renormalizable superpotential @xmath0 where @xmath1 are chiral superfields and the subscript @xmath2 denotes the species of them .
the khler potential has the canonical form ; @xmath3 .
then the scalar potential @xmath4 receives the f - term contribution and is written as @xmath5 where @xmath6 represents the derivative of @xmath7 with respect to @xmath1 . by @xmath8 and @xmath9 , respectively . a similar notation for the field derivative
is used for the potential @xmath4 and the d term @xmath10 . ] here and hereafter we follow the conventional notation that chiral superfields and their lowest scalar components are denoted by the same letters .
let us suppose that the model has a susy - breaking minimum at @xmath11 where at least one of the f components is non - vanishing , @xmath12 for some @xmath2 . throughout this paper
, we use the superscript @xmath13 to clarify that it is evaluated at the minimum of the f - term scalar potential , @xmath14 .
the minimum satisfies the stationary condition @xmath15 since @xmath16 corresponds to the fermion mass matrix , the condition ( [ eq : wij ] ) implies that there is a massless fermion along the direction @xmath17 ( @xmath18 ) .
this is nothing but the massless nambu - goldstone fermion corresponding to susy breaking .
the boson mass - squared matrix at the minimum is given by @xmath19 \sum_k w^{(0)}_{ijk}\overline{w}^{(0)}_{\bar{k } } & \sum_k w_{ik}^{(0)}\overline{w}_{\bar{k}\bar{j}}^{(0 ) } \end{pmatrix } , \ ] ] in the basis @xmath20 .
consider a massless fermion denoted by the eigenvector @xmath21 , that is , @xmath22 .
along its superpartner direction @xmath23 , the boson mass term in the scalar potential is evaluated as @xmath24 which can be made negative by a phase rotation of the eigenvector @xmath25 .
hence this must vanish , @xmath26 for a positive semi - definite hermite @xmath27 , that is , as long as the vacuum is stable .
in other words , a superpartner of a massless fermion must be massless or tachyonic ( even ) when susy is broken @xcite . in particular , eq . ( [ eq : wwv ] )
is explicitly written down for the massless nambu - goldstone fermion for susy breaking ; @xmath28 the f - term potential has a tree - level flat direction ( pseudo - moduli ) parametrized as @xmath29 that is easily confirmed by showing that @xmath6 is invariant along this direction : @xmath30 with the help of eqs .
( [ eq : wij ] ) and ( [ eq : wijk ] ) .
we have discussed the ks lemma and the presence of pseudo - moduli for the susy - breaking wess - zumino models .
these two properties hold even when a model contains higher - order superpotential terms such as in the generalized oraifeartaigh model discussed in @xcite .
the most generic superpotential form of generalized oraifeartaigh models studied in @xcite is @xmath31 the numbers of @xmath32 and @xmath33 fields , @xmath34 and @xmath35 respectively , satisfy @xmath36 , and @xmath37 are generic functions of @xmath38 . in this case , some f components in @xmath32 become nonzero in the susy - breaking vacuum , and the pseudo - moduli direction is therefore a linear combination of @xmath32
. many concrete examples of susy breaking is classified into this form and share the vacuum property discussed in this section .
it was shown @xcite that f - term susy - breaking models lead to a phenomenological problem when they are used for the hidden sector of gauge mediation ; vanishing gaugino masses or unstable hidden sector somewhere in the pseudo - moduli space . to see this ,
let us start with the susy - breaking wess - zumino model in the canonical form @xmath39 the susy - breaking minimum satisfies @xmath40 and the scalar potential has the corresponding pseudo - moduli direction .
we assume for simplicity that @xmath41 is non - vanishing for generic @xmath42 , that is , massless eigenstates are irrelevant for gauge mediation and removed from the following discussion . by definition , the determinant is a polynomial of @xmath42 and is expanded as @xmath43 unless @xmath44 and @xmath45 for @xmath46 , one always finds a solution for @xmath47 at some point @xmath48 on the pseudo - moduli direction .
this fact means that the mass matrix @xmath49 has a zero eigenvalue and the corresponding eigenvector @xmath25 such that @xmath50 .
then the ks lemma , eq .
( [ eq : wwv ] ) , is found to imply @xmath51 if there is no tachyonic direction anywhere in the pseudo - moduli space .
this contradicts with our assumption that @xmath52 is not identically zero . as a result
, @xmath41 should not have any zero , and must be independent of @xmath42 ( @xmath53 for @xmath54 ) ; @xmath55 in gauge mediation , some fields @xmath56 play as the messenger of susy breaking which , by definition , have the gauge interaction to the standard model vector multiplets .
the gauge invariance implies the mass matrix form @xmath57 the above statement , eq .
( [ eq : const ] ) , holds for each of sub - matrices , i.e. , @xmath58 this fact leads to a problem that the standard - model gaugino masses are not generated at the leading order of gauge interactions ; @xmath59 thus the vacuum structure of f - term susy breaking has the phenomenological problem ( or unstable vacuum ) and is not suitable for gauge mediation . in the next section
, we will examine how this situation is changed when the d - term scalar potential is included .
to circumvent the phenomenologically unfavorable result of f - term susy breaking , we examine in this section how the situation can change with the d - term potential included .
for this purpose , we first present some useful relations between the f and d components , and then investigate the vacuum of susy - breaking models with f and d terms .
some discussions of the d - term potential in gauge mediation are given in @xcite . before proceeding the model classification
, we give several useful relations between f and d components , satisfied identically and only on the vacuum . in this paper
we discuss the models with superpotential @xmath7 and u(1 ) gauge symmetry .
the full scalar potential consists of the f and d terms @xmath60 where the d - term contribution is explicitly given by @xmath61 here @xmath62 denotes the gauge coupling constant and @xmath63 the u(1 ) charge of @xmath1 .
the u(1 ) gauge - invariant lagrangian generally contains the linear term of vector superfield ( the fi term @xcite ) with the coefficient @xmath64 , and the d component is then modified as @xmath65 the formulas given below are valid in the presence of the fi term .
the u(1 ) transformation acts on @xmath1 as @xmath66 where @xmath67 is an infinitesimal parameter , and the u(1 ) gauge invariance of the ( super)potential leads to @xmath68 the field derivatives of this identity are also satisfied ; @xmath69 where @xmath70 and @xmath71 are not summed over .
they are rewritten by using the derivatives of the d component , @xmath72 and @xmath73 , as @xmath74 other field identities can be obtained with various contractions , e.g. , @xmath75 on the vacuum , the stationary conditions impose additional constraints between f and d terms : @xmath76 by multiplying @xmath77 , @xmath33 , @xmath78 and taking the summation over @xmath70 , we obtain @xmath79 where we have used the field identities given above .
the last relation means an important fact that the d term gives a non - vanishing contribution to the scalar potential only if some f terms of u(1)-charged multiplets are non - vanishing . as seen below
, this result is useful to classify and analyze the models with f and d terms .
the susy vacuum has the vanishing scalar potential , that is , all of f and d - flatness conditions are simultaneously satisfied .
@xmath80 thus susy - breaking models with f and d terms are classified into two categories .
we call a model is in the first class when there is no solution satisfying all the f - flatness conditions at the same time
. the second class of models has a solution of the f - flatness conditions but that is destabilized by the d - term potential , which causes susy breaking .
as mentioned above , the d - term effect appears only if @xmath81 is non - vanishing at the vacuum .
thus the above two classes are all possible situations leading to susy breaking . for susy
breaking , only f terms are effective in the first class of models , and the combination of f and d terms is important in the second class .
we will study the first and second classes of models in the subsections [ subsec:1class ] and [ subsec:2class ] , respectively .
the results are summarized in table [ table : class ] .
the third and fourth columns are our main results in this paper .
the scalar potential in the first class of models turns out to have a runaway direction no matter whether the fi term is added or not , and thus this class can not be applied to phenomenology as it stands . for the second class , susy is broken if a proper fi term is introduced , and it would be useful for model building such as gauge mediation .
.classification of susy - breaking models with f and d terms .
[ cols="<,^,^,^ " , ] as mentioned before , the non - vanishing d term requires charged f terms and then implies two pairs of vector - like chiral multiplets as the minimal set of messengers .
we have the superpotential for the messenger fields , @xmath82 when the messenger mass scales are such that @xmath83 , the vacuum ( [ eq : model - min ] ) remains stable and the standard - model gauge group is unbroken . as a result
, the susy - breaking f terms are still given by ( [ eq : model - fd ] ) and the bosonic mass - squared matrix for the messengers becomes @xmath84 for a small susy breaking @xmath85 , the gaugino mass @xmath86 in the visible sector is generated through the one - loop diagrams ( figure [ fig : gaugino ] ) and evaluated as @xmath87 where @xmath88 is the gauge coupling of the standard model group and @xmath89 denotes the dynkin index for the messenger representation @xmath90 . while the diagrams in figure 2 are expressed in terms of the mass insertion method , the result of gaugino mass ( [ eq : gaugino ] ) is easily shown to be independent of it and exact in the leading order .
the result explicitly shows that , in susy - breaking models with both f and d terms , one can indeed circumvent the argument in @xcite and generate gaugino masses at the one - loop order of gauge mediation .
denotes the corresponding superpartner of the scalar @xmath91.,title="fig:",width=245 ] + denotes the corresponding superpartner of the scalar @xmath91.,title="fig:",width=245 ] denotes the corresponding superpartner of the scalar @xmath91.,title="fig:",width=245 ]
in this paper we have systematically studied the susy - breaking models with f and abelian d terms .
the analysis has been performed with the canonical khler form and the tree - level scalar potential .
the models are classified under the condition whether there exist a f - flatness solution or not .
the first class of models has no solution for the f - flatness conditions , i.e. f - term susy breaking , and the second one has a solution which is destabilized by the d term in the full potential . in the first class of models
, there exists a pseudo - moduli direction in the f - term scalar potential as studied in @xcite .
this direction could be lifted in the full scalar potential by a non - vanishing d term with finite values of scalar fields .
we have however shown that there is a runaway direction in the full potential and the d term goes to zero . in the second class of models ,
susy breaking requires a fi term added with a proper coefficient . applying a simple model in the latter class to the gauge mediation ,
the visible - sector gaugino mass is found to be generated on the stable vacuum at the one - loop order .
the authors thank t. okada for useful discussions .
is supported by japan society for the promotion of science ( jsps ) and special postdoctoral researchers program at riken .
t.k . and k.y .
are supported in part by the grant - in - aid for scientific research nos .
20540266 and 23740187 , and for the global coe program `` the next generation of physics , spun from universality and emergence '' from the ministry of education , culture , sports , science and technology of japan .
is also supported by keio gijuku academic development funds .
we here examine the stationary conditions of the potential along the direction ( [ eq : runaway ] ) .
the first derivatives of the f and d - term potentials with respect to @xmath92 are evaluated as @xmath93 + { \cal o}\big(\frac{1}{z_\infty}\big ) \ , \end{aligned}\ ] ] where we have used the field identities ( [ eq : wqphi ] ) and ( [ eq : qww ] ) , and also the minimum conditions of @xmath94 , i.e. eqs .
( [ eq : wij ] ) and ( [ eq : wijk ] ) .
the coefficients @xmath95 satisfy the equations ( [ eq : wwc ] ) and ( [ eq : qwc ] ) so that ( [ eq : runaway ] ) is a runaway direction . as for the generic form ( [ eq : most - or ] ) , the superpotential is linear in @xmath32 and the f terms of @xmath33 are assumed to be zero , and we find @xmath96 ( v_f)_{\bar\phi_i}\ , & = \sum_{\phi_j } \big(\sum_{x_\ell } w_{\phi_jx_\ell}^{(0)}c_{x_\ell}^{(1)}+ \omega^2\!\!\sum_{x_\ell,\phi_m } w_{\phi_j\phi_mx_\ell}^{(0)}\ !
\overline{w}_{\bar{x}_\ell}^{(0)}c_{\phi_m}^{(2)}\big ) \sum_{x_k}\overline{w}_{\bar\phi_i\bar\phi_j\bar{x}_k}^{(0 ) } w_{x_k}^{(0 ) } \ , \label{eq : vfphij } \\ ( v_d)_{\bar{x}_i } & = g^2\omega q_{x_i}^{}\overline{w}_{\bar{x}_i}^{(0 ) } \big[\sum_{x_j}\omega q_{x_j}^{}\big(\bar{x}_j^{(0)}c_{x_j}^{(1 ) } + w_{x_j}^{(0)}c_{x_j}^{(2 ) } \big ) + \text{c.c.}\big ] \ , \label{eq : dxj } \\ ( v_d)_{\bar\phi_i}\ , & = 0 \ , \end{aligned}\ ] ] up to @xmath97 .
notice that some combination of @xmath98 s are determined by the runaway conditions of the potential and expressed with @xmath99 , and generally @xmath100 are free parameters .
as a result , in the direction ( [ eq : runaway ] ) , the stationary conditions for the full potential can be satisfied ; @xmath101 with properly fixed @xmath100 ( and also @xmath102 consistent with the runaway behavior ) .
the number of parameters is clearly large enough to have a solution for the conditions @xmath103 . as an example
, we discuss the stationary conditions for the model in section [ subsubsec : runaway_example ] . along the direction ( [ eq : example_runaway ] ) with a real @xmath104
, we find the stationary conditions of the full potential @xmath105 v_{\bar{\phi}_1 } & = m_1m_2 ^ 2\big(c_{x_2}^{(1)}+ m_1 c_{\phi_1}^{(2)}\big ) = 0 \ , \\[1 mm ] v_{\bar{\phi}_2 } & = m_1 ^ 2m_2\big(c_{x_1}^{(1)}+ m_2 c_{\phi_2}^{(2)}\big ) = 0 \ , \end{aligned}\ ] ] by explicitly writing down eqs .
( [ eq : vfphij ] ) and ( [ eq : dxj ] ) .
the condition for the charge - neutral field @xmath106 is trivial .
together with the runaway condition ( [ eq : example_condition ] ) , they give 6 conditions among 12 real parameters , @xmath107 , @xmath108 , and @xmath109 .
we thus find runaway solutions , an example of which is described by @xmath110 x_0 & = z_\infty f_0
\ , \\[.5 mm ] x_1 & = -z_\infty f -\frac{1}{z_\infty}\,\frac{m_2 ^ 2-m_1 ^ 2}{2m_1 ^ 2m_2 ^ 2}\,f \ , \\[.5 mm ] x_2 & = + z_\infty f \ .\end{aligned}\ ] ] r. kitano , h. ooguri and y. ookouchi , phys .
d * 75 * , 045022 ( 2007 ) [ arxiv : hep - ph/0612139 ] ; c. csaki , y. shirman and j. terning , jhep * 0705 * , 099 ( 2007 ) [ arxiv : hep - ph/0612241 ] ; s. abel , c. durnford , j. jaeckel and v. v. khoze , phys . lett .
b * 661 * , 201 ( 2008 ) [ arxiv:0707.2958 [ hep - ph ] ] ; n. haba and n. maru , phys .
d * 76 * , 115019 ( 2007 ) [ arxiv:0709.2945 [ hep - ph ] ] ; r. essig , j. f. fortin , k. sinha , g. torroba and m. j. strassler , jhep * 0903 * , 043 ( 2009 ) [ arxiv:0812.3213 [ hep - th ] ] .
a. pomarol and r. rattazzi , jhep * 9905 * , 013 ( 1999 ) [ arxiv : hep - ph/9903448 ] ; r. rattazzi , a. strumia and j. d. wells , nucl .
b * 576 * , 3 ( 2000 ) [ arxiv : hep - ph/9912390 ] ; r. sundrum , phys .
d * 71 * , 085003 ( 2005 ) [ arxiv : hep - th/0406012 ] ; k. hsieh and m. a. luty , jhep * 0706 * , 062 ( 2007 ) [ arxiv : hep - ph/0604256 ] ; m. endo and k. yoshioka , phys .
d * 78 * , 025012 ( 2008 ) [ arxiv:0804.4192 [ hep - ph ] ] ; t. kobayashi , y. nakai and m. sakai , jhep * 1106 * , 039 ( 2011 ) [ arxiv:1103.4912 [ hep - ph ] ] .
for example , m. luo and s. zheng , jhep * 0901 * , 004 ( 2009 ) [ arxiv:0812.4600 [ hep - ph ] ] ; l. f. matos , arxiv:0910.0451 [ hep - ph ] ; k. intriligator and m. sudano , jhep * 1006 * , 047 ( 2010 ) [ arxiv:1001.5443 [ hep - ph ] ] ; t. t. dumitrescu , z. komargodski and m. sudano , jhep * 1011 * , 052 ( 2010 ) [ arxiv:1007.5352 [ hep - th ] ] . | we systematically study supersymmetry breaking with non - vanishing f and d terms . we classify the models into two categories and find that a certain class of models necessarily has runaway behavior of scalar potential , while the other needs the fayet - iliopoulos term to break supersymmetry .
the latter class is useful to have a simple model of gauge mediation where the vacuum is stable everywhere and the gaugino mass is generated at the one - loop order . |
the finite - dimensional edwards - anderson spin glass @xcite is a model for disordered systems which has attracted much attention over the last decades .
the opinion on its nature , especially for three dimensional systems , is still controversial @xcite . beside trying to address the problem with the help of analytic calculations and simulations at finite temperature , it is possible to investigate the behavior of the model by means of ground - state calculations @xcite . since obtaining spin - glass ground states
is computationally hard @xcite , the study is restricted to relatively small systems . recently
a new algorithm , the cluster - exact approximation ( cea ) @xcite was presented , which allows in connection with a special genetic algorithm @xcite the calculation of true @xcite ground states for moderate system sizes , in three dimensions up to size @xmath1 . by applying this method
it is possible to study the ground - state landscape of systems exhibiting a @xmath0 degeneracy @xcite . for a thermodynamical correct evaluation
it is necessary that each ground state contributes to the results with the same weight , since all ground states have exactly the same energy .
recently it was shown @xcite , that the genetic cea causes a bias on the quantities describing the @xmath0 landscape .
the aim of this paper is to analyze the algorithm with respect to its ground - state statistics .
the reasons for the deviation from the correct behavior are given and an extension of the method is outlined , which guarantees thermodynamical correct results . in this work , three - dimensional edwards - anderson
( ea ) @xmath2 spin glasses are investigated .
they consist of @xmath3 spins @xmath4 , described by the hamiltonian @xmath5 the sum runs over all pairs of nearest neighbors .
the spins are placed on a three - dimensional ( d=3 ) cubic lattice of linear size @xmath6 with periodic boundary conditions in all directions .
systems with quenched disorder of the interactions ( bonds ) are considered .
their possible values are @xmath7 with equal probability .
to reduce the fluctuations , a constraint is imposed , so that @xmath8 .
the article is organized as follows : next a description of the algorithms is presented .
then it is shown for small systems , that the method does not result in a thermodynamical correct distribution of the ground states . in section four
, the algorithm and its different variants are analyzed with respect to the ground - state statistics . in the last section a summary
is given and an extension of the method is outlined , which should guarantee thermodynamical correct results .
the algorithm for the calculation bases on a special genetic algorithm @xcite and on cluster - exact approximation @xcite .
cea is an optimization method designed specially for spin glasses .
its basic idea is to transform the spin glass in a way that graph - theoretical methods can be applied , which work only for systems exhibiting no bond - frustrations .
now a short sketch of these algorithms is given , because later the influence of different variants on the results is discussed .
genetic algorithms are biologically motivated .
an optimal solution is found by treating many instances of the problem in parallel , keeping only better instances and replacing bad ones by new ones ( survival of the fittest ) .
the genetic algorithm starts with an initial population of @xmath9 randomly initialized spin configurations (= _ individuals _ ) , which are linearly arranged using an array .
the last one is also neighbor of the first one .
then @xmath10 times two neighbors from the population are taken ( called _ parents _ ) and two new configurations called _ offspring _ are created . for that purpose
the _ triadic crossover _ is used which turned out to be very efficient for spin glasses : a mask is used which is a third randomly chosen ( usually distant ) member of the population with a fraction of @xmath11 of its spins reversed . in a first step the offspring are created as copies of the parents .
then those spins are selected , where the orientations of the first parent and the mask agree @xcite .
the values of these spins are swapped between the two offspring
. then a _ mutation _ with a rate of @xmath12 is applied to each offspring , i.e. a randomly chosen fraction @xmath12 of the spins is reversed .
next for both offspring the energy is reduced by applying cea : the method constructs iteratively and randomly a non - frustrated cluster of spins . during the construction of the cluster a local gauge - transformation of the spin variables
is applied so that all interactions between cluster spins become ferromagnetic .
[ figceaexample ] shows an example of how the construction of the cluster works for a small spin - glass system . to increase the performance , spins adjacent to many unsatisfied bonds are more likely to be added to the cluster
this may introduce a bias on the resulting distribution of the ground states . later this scheme ( `` bias '' )
is compared to a variant ( `` same '' ) , where all spins may contribute to the cluster with the same probability . for 3d @xmath2
spin glasses each cluster contains typically 55 percent of all spins .
the non - cluster spins remain fixed during the following calculation , they act like local magnetic fields on the cluster spins . consequently , the ground state of the gauge - transformed cluster is not trivial , although all interactions inside the cluster are ferromagnetic .
since the cluster exhibits no bond - frustration , an energetic minimum state for its spins can be calculated in polynomial time by using graph - theoretical methods @xcite : an equivalent network is constructed @xcite , the maximum flow is calculated @xcite and the spins of the cluster are set to orientations leading to a minimum in energy .
please note , that the ground state of the cluster is degenerate itself , i.e. the spin orientations can be chosen in different ways leading all to the same energy .
it is possible to calculate within one single run a special graph , which represents all ground states of the cluster @xcite , and select one ground state randomly .
this procedure is called `` broad '' here . on the other hand
, one can always choose a certain ground state of the cluster directly .
usually this variant , which is called `` quick '' here , is applied , because it avoids the construction of the special graph .
but this again introduces a certain bias on the resulting distribution of the ground states .
later the influence of the different methods of choosing ground states is discussed .
this cea minimization step is performed @xmath13 times for each offspring . afterwards each offspring is compared with one of its parents . the offspring
/ parent pairs are chosen in the way that the sum of the phenotypic differences between them is minimal .
the phenotypic difference is defined here as the number of spins where the two configurations differ .
each parent is replaced if its energy is not lower ( i.e. not better ) than the corresponding offspring .
after this whole step is conducted @xmath10 times , the population is halved : from each pair of neighbors the configuration which has the higher energy is eliminated . if more than 4 individuals remain the process is continued otherwise it is stopped and the best individual is taken as result of the calculation .
the following representation summarizes the algorithm .
= = = = = = genetic cea(@xmath14 , @xmath9 , @xmath15 , @xmath12 , @xmath13 ) + * begin * + create @xmath9 configurations randomly + ( @xmath16 ) * do * + + @xmath17 * to * @xmath10 * do * + + select two neighbors + create two offspring using triadic crossover + do mutations with rate @xmath12 + both offspring * do * + + @xmath18 * to * @xmath13 * do * + + construct unfrustrated cluster of spins + construct equivalent network + calculate maximum flow + construct minimum cut + set new orientations of cluster spins + + offspring is not worse than related parent + + replace parent with offspring + + + half population ; @xmath19 + + one configuration with lowest energy + * end * the whole algorithm is performed @xmath20 times and all configurations which exhibit the lowest energy are stored , resulting in @xmath21 statistically independent ground - state configurations ( _ replicas _ ) . a priori nothing about the distribution of ground states raised by the algorithm is known .
thus , it may be possible that for one given realization of the disorder some ground states are more likely to be returned by the procedure than others .
consequently , any quantities which are calculated by averaging over many independent ground states , like the distribution of overlaps , may depend on a bias introduced by the algorithm . for a thermodynamical correct evaluation all ground states have to contribute with the same weight , since they all have exactly the same energy . for the preceding work ,
the distribution of the ground states determined by the algorithm was taken .
the method was utilized to examine the ground state landscape of two - dimensional @xcite and three - dimensional @xcite @xmath2 spin glasses by calculating a small number of ground states per realization .
some of these results depend on the statistics of the ground states , as it will be shown in the next section for the @xmath22 case . on the other hand ,
the main findings of the following investigations are not affected by the bias introduced by genetic cea : the existence of a spin - glass phase for nonzero temperature was confirmed for the three - dimensional spin glass @xcite .
the method was applied also to the @xmath2 random - bond model to investigate its @xmath0 ferromagnetic to spin - glass transition@xcite .
finally , for small sizes up to @xmath23 all ground - state valleys were obtained by calculating a huge number of ground states per realization and applying a new method called _ ballistic search _ @xcite .
in this section results describing the ground - state landscape of small three - dimensional @xmath2 spin glasses are evaluated .
it is shown that the data emerging from the use of raw genetic cea and from a thermodynamically correct treatment differ substantially .
several ground states for small systems of size @xmath24 were calculated .
1000 realizations of the disorder for @xmath25 and 100 realizations for @xmath26 were considered .
the parameters ( @xmath9 , @xmath15 , @xmath12 , @xmath13 ) , for which true ground states are obtained , are shown in @xcite .
for all calculations the variants bias and quick were used to obtain maximum performance .
the effect of different variants on the results is discussed in the next section .
two schemes of calculation were applied : * for each realization @xmath27 runs of genetic cea were performed and all states exhibiting the ground - state energy stored .
consequently , this scheme reflects the ground - state statistics which is determined solely by the genetic cea method .
configurations which have a higher probability of occurrence contribute with a larger weight to the results .
* for each realization the algorithm was run up to @xmath28 times .
each particular state was stored only once . for later analysis
the number of times each state occurred was recorded .
additionally , a systematic local search was applied to add possibly missing ground states which are related by flips of free spins to states already found .
finally , a @xmath29 realization exhibits 25 different ground states on average . for a @xmath30 realization on average 240 states
were found and 6900 states for @xmath26 .
+ for the evaluation of physical quantities every ground state is taken with the same probability in this scheme .
thus , the statistics obtained in this way reflect the true @xmath0 thermodynamic behavior . to analyze the ground - state landscape ,
the distribution of overlaps is evaluated . for a fixed realization @xmath31 of the exchange interactions and two replicas @xmath32 , the overlap @xcite is defined as @xmath33 the ground state of a given realization is characterized by the probability density @xmath34 .
averaging over the realizations @xmath35 , denoted by @xmath36_{j}$ ] , results in ( @xmath37 = number of realizations ) @xmath38_{j } = \frac{1}{z } \sum_{j } p_j(q ) \label{def_p_q}\ ] ] because no external field is present the densities are symmetric : @xmath39 and @xmath40 .
so only @xmath41 is relevant .
the result of @xmath41 for @xmath26 is shown in fig .
[ figpqlfive ] . for the true thermodynamic result
small overlaps occur less frequent than for the data obtained by the application of pure genetic cea .
large overlap values occur more often .
this deviation has an influence on the way the spin glass behavior is interpreted .
the main controversy about finite - dimensional spin glasses mentioned at the beginning is about the question whether for the infinite system @xmath41 shows a long tail down to @xmath42 or not @xcite . to investigate the finite size behavior of @xmath41
the fraction @xmath43 of the distribution below @xmath44 is integrated : @xmath45 the development of @xmath43 as a function of system size @xmath6 is shown in fig .
[ figxl ] .
the datapoints for the larger sizes @xmath46 , obtained using pure genetic cea , are taken from former calculations @xcite .
these values are more or less independent of the system size , while the correct thermodynamic behavior shows a systematic decrease . whether for @xmath47 the long tail of @xmath41 persists can not be concluded from the data , because the systems are too small .
nevertheless , the true @xmath0 behavior differs significantly from the former results .
to understand , why genetic cea fails in producing the thermodynamical correct results , in this section the statistics of the ground states , which is determined by the algorithm , is analyzed directly . for the case where all ground states were calculated using a huge number of runs , the frequencies each ground state occurred were recorded . in fig .
[ fighistogramm ] the result for one sample realization of @xmath48 is shown .
the system has 56 different ground states . for each state
the number of times it was returned by the algorithm in @xmath28 runs is displayed .
obviously the large deviations from state to state can not be explained by the presence of statistical fluctuations .
thus , genetic cea samples different ground states from the same realization with different weights . to make this statement more precise ,
the following analysis was performed : two ground states are called _ neighbors _ , if they differ only by the orientation of one spin .
all ground states which are accessible from each other through this neighbor - relation are defined to be in the same ground - state _
valley_. that means , two ground states belong to the same valley , if it is possible to move from one state to the other by flipping only free spins , i.e. without changing the energy . for all realizations
the valleys were determined using a method presented in @xcite , which allows to treat systems efficiently exhibiting a huge number of ground states .
then the frequencies @xmath49 for each valley @xmath50 were computed as the sum of all frequencies of the states belonging to @xmath50 . in fig .
[ fighistosample ] the result is shown for a sample @xmath48 realization , which has 15 different ground state valleys .
large valleys are returned by the algorithm more frequently , but @xmath49 seems to grow slower than linearly .
a strict linear behavior should hold for an algorithm which guarantees the correct @xmath0 behavior .
for averaging @xmath49 has to be normalized , because the absolute values of the frequency differ strongly from realization to realization , even if the size @xmath51 of a valley , i.e. the number of ground states belonging to it , is the same . for each realization
, the normalized frequency @xmath52 is measured relatively to the average frequency @xmath53 of all valleys of size 1 : @xmath54 if a realization does not exhibit a valley consisting only of one ground state , the frequency @xmath55 of the smallest valley @xmath56 is taken .
it is assumed , that the normalized frequency exhibits a @xmath57 dependence , which is justified by the results shown later .
consequently , for the case the size @xmath58 of the smallest valley is larger than one , @xmath59 is chosen .
the value of @xmath60 is determined self - consistently .
the result for @xmath29 of @xmath52 as a function of the valley - size @xmath51 is presented in fig .
[ fighistolthree ] . a value of @xmath61 was determined .
please note , that the fluctuations for larger valleys are higher , because quite often only one valley was available for a given valley - size .
the algebraic form is clearly visible , proving that genetic cea overestimates systematically the importance of small ground - state valleys .
for @xmath30 a value of @xmath62 was obtained , while the @xmath26 case resulted in @xmath63 .
consequently , with increasing system size , the algorithm fails more and more to sample configurations from different ground - state valleys according to the size of the valleys .
this explains , why the difference of @xmath64 between the correct result and the values obtained in @xcite increases with growing system size .
similar results were obtained for two - dimensional systems . for @xmath26
a self - consistent value of @xmath65 was found , while the treatment of @xmath66 systems resulted in @xmath67 .
here only a slight finite - size dependence occurs .
this may explain the fact , that the width of the distribution of overlaps , even calculated only by the application of pure genetic cea , seems to scale to zero @xcite . in the second section of this paper
two variants of the algorithm were presented , which may be able to calculate ground states more equally distributed . to investigate this issue ,
similar ground - state calculations were conducted for @xmath30 and again @xmath52 was calculated . for the case , were same was used instead of bias , a value @xmath68 was determined self - consistently . using broad instead of quick resulted in @xmath69 .
finally , by applying same and broad together , @xmath70 was obtained .
consequently , applying different variants of the method decreases the tendency of overestimating small valleys , but the correct thermodynamic behavior is not obtained as well .
even worse , broad and same are considerably slower than the combination of quick and bias .
so far it was shown , that genetic cea fails in sampling ground states from different valleys according the size of the valleys .
now we turn to the question , whether at least states belonging to the same valley are calculated with the correct thermodynamic distribution . by investigating the frequencies of different ground states belonging to the same valley it was found again , that these configurations are not equally distributed .
but it is possible to study this issue in a more physical way . for that purpose ground states of 100 @xmath71 realizations
were calculated .
then the valley structure was analyzed .
the average distribution of overlaps was evaluated , but only contributions of pairs of states belonging to the same valley were considered .
for comparison , for the same realizations a long @xmath0 monte - carlo ( mc ) simulation was performed , i.e. randomly spins were selected and flipped if they were free .
the ground states were used as starting configurations . since
a mc simulation ensures the correct thermodynamic distribution of the states , all ground states of a valley appear with the same frequency , if the simulation is only long enough .
a length of 40 monte - carlo steps per spin were found to be sufficient for @xmath71 .
the result for the distribution of overlaps @xmath72 restricted to the valleys is displayed in fig .
[ figpqvalley ] .
significant differences between the datapoints from the pure genetic cea and the correct @xmath0 behavior are visible .
consequently , the algorithm does not sample configurations belonging to the same ground - state valley with the same weight as well .
in this work the genetic cluster - exact approximation method is analyzed .
the algorithm can be used to calculate many independent true ground states of ea spin glasses .
the results from the raw application of the method and from calculations of _ all _ ground states for small system sizes were compared . by evaluating the distribution of overlaps
is was shown , that genetic cea imposes some bias on the ground - state statistics .
consequently , the results from the application of the raw method do not represent the true @xmath0 thermodynamics . to elucidate the behavior of the algorithm
the statistics of the ground states were evaluated directly .
it was shown , that different ground states have dissimilar probabilities of occurrence . to understand this effect better ,
the ground - state valleys were determined .
the genetic cea method finds configurations from small ground - state valleys relative to the size of the valley more often than configurations from large valleys .
additionally , within a valley the states are not sampled with the same weight as well .
it was shown that two variants of the algorithm , which decrease its efficiency , weaken the effect , but it still persists . summarizing , two effects are responsible for the biased ground - state sampling of genetic cea : small valleys are sampled too frequently and the distribution within the valleys is not flat
. for small system sizes it is possible to calculate all ground states , so one can obtain the true thermodynamic average directly .
but already for @xmath26 there are realizations exhibiting more than @xmath28 different ground states .
since the ground - state degeneracy grows exponentially with system size @xcite larger systems can not be treated in this way .
the following receipt should overcome these problems and should allow to obtain the true thermodynamic @xmath0 behavior for larger systems : * calculate several ground states of a realization using genetic cea . *
identify the ground states which belong to the same valleys .
* estimate the size of each valley .
this can be done using a variant of ballistic search @xcite , which works by flipping free spins sequentially , each spin at most once .
the number of spins flipped is a quite accurate measure for the size of a valley . *
sample from each valley a number of ground states , which is proportional to the size of the valley .
this guarantees , that each valley contributes with its proper weight .
each state is obtained by performing a @xmath0 mc simulation of sufficient length , starting with true ground - state configurations .
since mc simulations achieve a thermodynamical correct distribution , it is guaranteed that the states within each valley are equally distributed .
please note , that it is not necessary to calculate all ground states to obtain the true thermodynamic behavior , because it is possible to estimate the size of a valley by analyzing only some sample ground states belonging to it .
furthermore , it is even only necessary to have configurations from the largest valleys available , since they dominate the ground - state behavior .
this condition is fulfilled by genetic cea , because large valleys are sampled more often than small valleys , even if small valleys appear too often relatively . from the results presented here
it is not possible to deduce the correct @xmath0 behavior of the infinite system , because the system sizes are too small . using the scheme outlined above , it is possible to treat system sizes up to @xmath73 @xcite .
the author thanks k. battacharya and a.w .
sandvik for interesting discussions .
the work was supported by the graduiertenkolleg `` modellierung und wissenschaftliches rechnen in mathematik und naturwissenschaften '' at the _ interdisziplinres zentrum fr wissenschaftliches rechnen _ in heidelberg and the _ paderborn center for parallel computing _ by the allocation of computer time .
the author announces financial support from the dfg ( _ deutsche forschungsgemeinschaft _ ) . for reviews on spin glasses
see : k. binder and a.p .
young , rev .
phys . * 58 * , 801 ( 1986 ) ; m. mezard , g. parisi , m.a .
virasoro , spin glass theory and beyond , world scientific , singapur 1987 ; k.h .
fisher and j.a .
hertz , spin glasses , cambridge university press , 1991 [ 0.45 ] for @xmath26 .
the dashed line shows the old result obtained by computing about @xmath74 independent ground states per realization using genetic cea .
the solid line shows the same quantity for the case , where all existing ground states were used for the evaluation , i.e. where the correct @xmath0 thermodynamic behavior is ensured .
, title="fig : " ] [ 0.45 ] of the distribution of overlaps for @xmath75 as a function of system size @xmath6 .
the upper points ( circles ) where obtained by calculating about @xmath74 independent ground states per realization using genetic cea .
the lower points ( triangles ) show the result ( @xmath76 ) for the case , where all existing ground states were used for the evaluation , i.e. where the correct @xmath0 thermodynamic behavior is ensured .
, title="fig : " ] [ 0.45 ] is found by genetic cea for @xmath29 .
the frequency is normalized so that @xmath77 ( see text ) . the probability that a cluster is found increases with the size of the cluster , but slower than linearly .
the line shows a fit @xmath57 with @xmath61 , title="fig : " ] [ 0.45 ] of overlaps for @xmath71 restricted to pairs of ground states belonging to the same valley .
the full line shows the result for the case , where the statistics of the ground state is determined by the genetic cea algorithm .
the data represented by the dashed line was obtained using states which are equally distributed within each valley , which was guaranteed by performing a @xmath0 mc simulation .
, title="fig : " ] | the genetic cluster - exact approximation algorithm is an efficient method to calculate ground states of ea spin glasses .
the method can be used to study ground - state landscapes by calculating many independent ground states for each realization of the disorder .
the algorithm is analyzed with respect to the statistics of the ground states and the valleys of the energy landscape .
furthermore , the distribution inside each valley is evaluated .
it is shown that the algorithm does not lead to a true @xmath0 thermodynamic distribution , i.e. each ground state has not the same frequency of occurrence when performing many runs .
an extension of the technique is outlined , which guarantees that each ground states occurs with the same probability . * keywords ( pacs - codes ) * : spin glasses and other random models ( 75.10.nr ) , numerical simulation studies ( 75.40.mg ) , general mathematical systems ( 02.10.jf ) . |
in the past decade the interest in the physical properties of correlated electronic systems has greatly increased .
one reason is the strong local correlations on transition metal ions in cuprate superconductors and manganites , and the corresponding unusual properties of these compounds .
the parent undoped compounds are mott - hubbard or charge transfer insulators , while doping leads to correlated metals in which the kinetic energy of charge carriers competes with magnetic order.@xcite one of the most spectacular consequences is the onset of high - temperature superconductivity in the cuprates .
it is believed that a satisfactory description of the normal phase properties is a prerequisite for the understanding of the microscopic mechanism of pairing in high - temperature superconductors .
the electronic states in cuo@xmath0 planes of cuprate superconductors are usually described in terms of the emery model which includes the hybridization between cu(@xmath1 ) and o(@xmath2 ) states.@xcite however , hole doping leads to the formation of local zhang - rice singlets , @xcite and the essential excitations in the cuprates within a window of a few ev around the chemical potential are well reproduced using the effective two - dimensional ( 2d ) hubbard model with extended hopping @xcite and a large local coulomb interaction @xmath3 , as shown by various numerical studies of the @xmath4-@xmath5 and hubbard model.@xcite recently it was shown @xcite that the effective hubbard model has to include hopping beyond the nearest - neighbors .
the second nearest - neighbor hopping changes the dispersion of the quasiparticle ( qp ) states , and is therefore crucial for understanding angular resolved photoemission ( arpes ) data of the antiferromagnetic ( af ) insulator sr@xmath0cuo@xmath0cl@xmath0.@xcite both second and third neighbor hopping parameters follow from the down - folding procedure in electronic structure calculations,@xcite and influence the shape of the fermi surface .
they have a particular relation to the value of the superconducting transition temperature at optimal doping.@xcite the superconductivity occurs in the cuprates under doping @xmath6 of a half - filled ( @xmath7 ) af insulator , and is accompanied by a gradual modification of the magnetic order .
the nature of magnetic correlations in doped materials is therefore a central issue in the theory of the cuprate superconductors .
undoped la@xmath0cuo@xmath8 is a _ commensurate _
af insulator , while doping by sr into la@xmath9sr@xmath10cuo@xmath8 results in short - range af order within _ incommensurate _
magnetic structures.@xcite such an incommensurate magnetic order was indeed found analytically,@xcite in hartree - fock ( hf),@xcite and in a slave - boson approximation.@xcite however , in order to understand the transport properties one has to go beyond an effective single - particle description and include the dynamics due to local electron correlations .
a sufficiently accurate treatment of local electron correlations remains one of the challenging problems in modern solid state theory .
although an important progress in the present understanding of strongly correlated fermion systems occurred recently due to numerical methods , such as quantum monte carlo ( qmc ) and exact diagonalization ( ed ) , an analytic treatment that maintains local correlations is needed to investigate the consequences of strong correlations in the thermodynamic limit .
an attractive possibility is the limit of large spatial dimension ( @xmath11 ) , where the diagrams in the perturbative expansion collapse to a single site and the fermion dynamics is described by a _
local self - energy_.@xcite this allows a mapping of lattice models onto quantum impurity models , which can then be solved self - consistently using the dynamical mean - field theory ( dmft).@xcite the dmft was quite successful for the hubbard model with nearest - neighbor hopping @xmath4 at half - filling , where it predicts the mott transition to the insulating state ( @xmath7).@xcite this was also found by logan , eastwood , and tusch @xcite for the @xmath11 case using an analytic method .
attempts to use dmft at arbitrary filling , however , made it clear that the local self - energy becomes particularly important in systems with magnetic long - range order ( lro ) , which are easily destabilized when the correlation effects are overestimated .
the self - energy therefore plays a decisive role and has to be described _ beyond _ second order perturbation theory ( sopt).@xcite this has made the application of dmft to magnetically ordered systems notoriously difficult .
recently we have shown that the screening of local coulomb interaction by the particle - particle diagrams plays a crucial role in stabilizing the incommensurate magnetic lro in doped systems.@xcite the advantage of using dmft becomes clear by looking at the single hole problem , which can be solved exactly in the @xmath12 limit.@xcite the method becomes exact because the quantum fluctuations are of higher order in the @xmath13-expansion than the leading potential term which originates from the ising part of the superexchange interaction @xmath14 .
therefore , applying dmft to the @xmath15 case might still capture the essential features that result from the coupling of a moving hole to local spin fluctuations .
we will show below that in fact such quantities as the spectral function , the qp band , and the size of the qp spectral weight are well reproduced within the dmft , which for a single hole includes only those processes which are present in the @xmath16 model .
although this approach becomes exact only in the @xmath12 limit,@xcite it gives a sufficient accuracy of the one - particle spectral function even in finite dimension @xmath15.@xcite the dmft allows us to calculate the optical conductivity in the @xmath11 limit of metzner and vollhardt @xcite from the knowledge of the local self - energy without further approximations.@xcite the studies performed in this limit for the nonmagnetic systems already allowed a qualitative reproduction of such experimental observations in the cuprates as the increase of the drude peak with doping , and a temperature and doping dependent mid - infrared peak.@xcite the paper is organized as follows .
the self - consistent procedure to determine a local self - energy within the dmft is introduced in sec .
it consists of the hf potential and the dynamical part due to spin fluctuations which uses a coulomb interaction renormalized by particle - particle scattering .
the formalism to calculate the one - particle and optical excitation spectra in the spin spiral ( ss ) states is developed in sec .
next we analyze the numerical results for the one - particle spectral properties in sec .
iv , where we show how they change with doping and with increasing temperature .
the optical properties are presented in sec .
v ; there we discuss the new effects in the optical conductivity , scattering rate and the effective mass which arise due to extended hopping and by increasing the value of coulomb interaction @xmath3 .
vi presents a short summary and conclusions .
we consider the spectral and optical properties of the minimal model for strongly correlated electrons in high - temperature superconductors , the hubbard model with extended hopping,@xcite @xmath17 where @xmath18 is a creation operator of an electron with spin @xmath19 at site @xmath20 , and @xmath21 .
the hopping elements @xmath22 , @xmath23 and @xmath24 stand for the nearest neighbor , second - nearest neighbor , and third - neighbor hopping on a 2d square lattice , and serve to model the electronic states of the charge - transfer type in the cuprates . for convenience
we choose @xmath4 as the energy unit .
it is interesting to note that hopping beyond nearest - neighbors contributes to the energy and other properties not only in a 2d model , but also in the limit of @xmath12 .
the energy contributions due to more distant neighbors are finite due to the scaling of the hopping parameters on a hypercubic lattice .
it is given by @xmath25 ( see refs . ) , where @xmath26 is the distance between @xmath20 and @xmath27 defined by the bond metric , and gives the scaling factors @xmath28 for first- , and @xmath29 for second- and third - nearest neighbor hopping , respectively , as the latter sites are two bonds apart . as mentioned above , we adopt the limit of infinite dimensions to determine the spectral properties of the hubbard hamiltonian on a square lattice in the thermodynamic limit . in order to simplify the numerical evaluation of the self - energy
we introduce an ansatz for the modified magnetic order in the doped systems , and assume incommensurate ss structures with a large but finite periodicity .
this approach captures the essence of the competition between the weakened short - range af order , and the kinetic energy induced by hole doping,@xcite and allows to treat the systems in the thermodynamic limit at low temperature .
the spiral states are characterized by the amplitude of the local magnetization , @xmath30 which is independent of the site index @xmath20 .
the direction of the magnetic moment at each site @xmath20 is specified in the _ global reference frame _ by two spherical angles , @xmath31 , and therefore the original fermion operators , @xmath32 , are transformed to the new fermions quantized with respect to _ the local quantization axis _ at each site,@xcite @xmath33_{\lambda\sigma } , \label{rotation}\ ] ] where @xmath34 is the rotation matrix , @xmath35 and @xmath36 are polar and azimuthal angle , respectively , and @xmath37 and @xmath38 are pauli spin matrices .
this transforms the hubbard hamiltonian ( [ generic ] ) to the following form , @xmath39_{\sigma\sigma^{\prime}}c^{}_{j\sigma^{\prime}}\ ! + u\ ! \sum_i n_{i\uparrow}n_{i\downarrow}. \label{hubbard}\ ] ] in the ss states we take the polar angle site - independent , @xmath40 , and the azimuthal angle is given by the wave - vector @xmath41 of the spiral , @xmath42 .
using the periodicity of the @xmath43 matrix in eq .
( [ hubbard ] ) , one finds after a fourier transformation that the kinetic energy takes a simple @xmath44 matrix form , @xmath45 where @xmath46 is the electron dispersion in a noninteracting system in the global reference frame . here
we limit ourselves to plane spirals , and choose @xmath47 .
therefore , the order parameter rotates in the @xmath48 plane , @xmath49 $ ] .
double spirals were shown to be unstable in the 2d @xmath4-@xmath5 model,@xcite and we have no reason to believe that they might be stabilized by further neighbor hopping . in order to construct the leading local part of the self - energy , we use the dmft and consider the impurity model coupled to the lattice by the effective field ( for more details see ref . ) .
the anderson model of a magnetic impurity coupled to a conduction band with ss order consists of a `` non - degenerate impurity orbital '' at site @xmath50 , with the fermion operators @xmath51 , and the conduction electron bath as an `` effective ss conduction band '' described by the operators @xmath52 , @xmath53_{\sigma \sigma^\prime } c_{{\bf k}\sigma^\prime}^ { } \nonumber \\ & + & \sum_{{\bf k } \sigma \sigma^\prime } \left [ f_{o\sigma}^{\dagger}\ , [ v_{\bf q}({\bf k})]_{\sigma\sigma^\prime}^{}c_{{\bf k}\sigma^\prime}^ { } + h.c . \right ] + u n^f_{o\uparrow } n^f_{o\downarrow } , \label{anderson}\end{aligned}\ ] ] where @xmath54 is an impurity energy level , and @xmath55 is an effective one - particle energy of the same functional form as @xmath56 ( [ ekin ] ) . the hybridization @xmath44 matrix in the local reference frame , @xmath57 \nonumber \\ &
+ & \frac{1}{2}v_{{\bf k}+\frac{\bf q}{2 } } \left [ \cos\frac{\theta}{2 } ( \hat{1}-\hat{\sigma}_z ) + \sin\frac{\theta}{2}\hat{\sigma}_-\right ] , \label{hma}\end{aligned}\ ] ] where @xmath58 , is given by the individual hybridization elements in the global reference frame , @xmath59 .
the hamiltonian ( [ anderson ] ) is quadratic in @xmath60 , and the bath of conduction electrons can be integrated out giving raise to an effective action of the impurity electrons which is of the usual form,@xcite @xmath61 where @xmath62 are grassmann variables for the @xmath63-electrons .
weiss effective field _
@xmath64 is a @xmath65 matrix in spin space , @xmath66^{-1}{\hat v}_{\bf q}^{\dagger}({\bf k } ) .
\label{impb}\end{aligned}\ ] ] for a plane spin spiral with @xmath47 the weiss effective field becomes a diagonal matrix in spin space , @xmath67 note that this result only depends on the functional form of eqs .
( [ ekin])([hma ] ) , and not on the parameters , except that it holds for a plane spiral .
it implies that the local spin fluctuations are decoupled from the local charge fluctuations , and simplifies the present self - consistent calculation for ss states within the dmft approach , as all local quantities including the self - energy @xmath68 , are diagonal .
in the spirit of the dmft approach , we approximate the green function using a _ local self - energy_,@xcite @xmath69 where @xmath70 is the chemical potential . the lattice ( finite ) dimensionality enters via the one - particle energies @xmath56 and gives rise to the @xmath71-dependence of the spectral function .
the lattice one - particle green function ( [ localg ] ) is described by a @xmath44 matrix @xmath72 in spin space , where @xmath73 are fermionic matsubara frequencies .
the corresponding local lattice green function , @xmath74 , is diagonal in spin space due to the parity of the kinetic energy @xmath56 .
the self - energy consists of the hf part , @xmath75 with @xmath76 , and the spin - fluctuation ( sf ) part , @xmath77 , which is determined by the many - body effects . using the _ cavity method _
@xcite for a hypercubic lattice at @xmath11 , we verified that the dynamical weiss field , @xmath78 , can be computed from the dyson equation of the anderson impurity model ( [ anderson ] ) with broken spin - symmetry , @xmath79 eqs .
( [ localg ] ) and ( [ cavity ] ) are fundamental in the dmft,@xcite and can be solved self - consistently , provided an expression for the self - energy is known .
the calculations at finite temperature @xmath80 require the knowledge of the free energy , @xmath81 , being a thermodynamic potential for a system of @xmath82 electrons .
it has to be minimized to find a stable ss state which determines the system properties .
the free energy may be found from the grand canonical potential , @xmath83 , using the standard approach for quantum many - body systems , @xmath84.@xcite for a translationally invariant lattice model with _ local self - energy _
one finds the functional form of the grand canonical potential , @xmath85 \nonumber \\ & - & \beta^{-1 } \sum_{{\bf k}\nu } \ln\det \left [ 1 - \hat{g}^{0}_{\bf q}({\bf k},i\omega_{\nu } ) { \hat \sigma}^{\rm sf}_{\bf q}(i\omega_{\nu})\right ] \nonumber \\ & - & \beta^{-1}n \sum_{\nu}tr\left[{\hat\sigma}^{\rm sf}_{\bf q}(i\omega_{\nu } ) { \hat g}_{\bf q}(i\omega_{\nu})\right ] , \label{kbpo}\end{aligned}\ ] ] with @xmath86 , and @xmath87 is the self - energy discussed below .
the functional ( [ kbpo ] ) is stationary , i.e. , @xmath88 ensures that the minimum of the grand canonical potential has been found , and determines the self - energy from the luttinger - ward functional , @xmath89=\frac{1}{n}\ ; \frac{\delta \phi^{\rm sf}[{\hat g}_{\bf q } ] } { \delta { \hat g}_{{\bf q},\sigma\sigma}}. \label{luwa}\ ] ] our perturbative expansion is constructed around the symmetry - broken hf state , hence the grand canonical potential of the `` non - interacting '' reference system includes a correction term to avoid double - counting and reads , @xmath90 \nonumber \\ & - & un \langle n_{o \uparrow } \rangle \langle n_{o \downarrow } \rangle .
\label{kbpohf}\end{aligned}\ ] ] the spectrum which defines @xmath91 is given by the green s function in the hf approximation , @xmath92 the luttinger - ward functional @xmath93 $ ] in eq . ( [ kbpo ] )
is defined via the diagrammatic expansion of @xmath94 in terms of the full green s function @xmath95 .
the self - energy of the infinite - dimensional hubbard model is a local dynamical quantity and involves only the local component of the green s function ( [ localg ] ) .
this implies that @xmath96=n\phi^{\rm sf}_{\rm imp}[g]$],@xcite meaning that the functional @xmath96 $ ] can be approximated by some infinite subset of the one - particle irreducible closed feynman diagrams of the anderson impurity model ( [ anderson ] ) .
we take for @xmath96 $ ] the sum of all particle - hole diagrams,@xcite and the effective particle - hole interaction @xmath97,@xcite @xmath98 + \beta^{-1}\bar{u } \sum_{\mu } \chi^{(0)}_{\uparrow\downarrow}(i\omega_{\mu } ) \nonumber \\&+ & \frac{1}{2}\beta^{-1 } \bar{u}^{2 } \sum_{\mu } \chi^{(0)}_{\uparrow\downarrow}(i\omega_{\mu } ) \chi^{(0)}_{\downarrow\uparrow}(i\omega_{\mu } ) , \\
% \phi_{\| } & = & \frac{1}{2}\beta^{-1 } \sum_{\mu } \ln \left[1-\bar{u}^{2 } \chi^{(0)}_{\uparrow\uparrow}(i\omega_{\mu } ) \chi^{(0)}_{\downarrow\downarrow}(i\omega_{\mu } ) \right ] \nonumber\\ & + & \frac{1}{2}\beta^{-1 } \bar{u}^{2 } \sum_{\mu } \chi^{(0)}_{\uparrow\uparrow}(i\omega_{\mu } ) \chi^{(0)}_{\downarrow\downarrow}(i\omega_{\mu } ) , \label{kbdia}\end{aligned}\ ] ] where @xmath99 is the noninteracting particle - hole susceptibility .
self - consistency would require that @xmath100 $ ] ; here instead we apply the non - self - consistent procedure introduced by bulut , scalapino , and white,@xcite and approximate @xmath93\to\phi^{\rm sf}[{\cal g}^{0}_{\bf q}]$ ] .
it has been shown that this procedure may be regarded to be a reasonable approximation as the thermodynamic potential ( [ kbpo ] ) is stationary and one expects not to move too far away from its minimum .
it is known to be notoriously difficult to obtain an analytic expression for the self - energy , and so far mostly an ansatz within the iterative perturbation scheme ( ips ) based on sopt has been used.@xcite the ansatz introduces an approximate form of self - energy which starts from the sopt and allows to reproduce the exact results in certain limits .
although this approach reproduces the correct large @xmath3 limit,@xcite it overestimates the correlation effects in the nonmagnetic states , and thus becomes uncontrollable in the intermediate @xmath3 regime .
therefore , it can not be applied to investigate the phase stability and dynamics in the magnetic states of the hubbard model .
we have verified that the af lro disappears in the 2d hubbard model ( @xmath101 ) at half - filling for @xmath102 for @xmath101 , if the formula introduced by kajueter and kotliar@xcite is used ( see sec .
[ sec : mott ] ) .
the sf part of the self - energy , @xmath77 follows from the kadanoff - baym potential ( [ kbpo ] ) containing a class of diagrams up to infinite order , @xmath103 here we approximated @xmath104 $ ] by @xmath105 $ ] and avoid self - consistency .
the transverse part in eq . ( [ sigma ] ) resembles the self - energy derived by the coupling of the moving hole to transverse spin - fluctuations , as derived using the spin - wave theory.@xcite however , the longitudinal part is not included in the latter approach , and we find that it can not be neglected in the relevant regime of parameters for high temperature superconductors .
the self - energy in a magnetic system is calculated using the weiss effective field ( [ cavity ] ) in the symmetry - broken magnetic state . the transverse , @xmath106 ^ 2 } { 1-{\bar u}\chi^{(0)}_{\bar{\sigma}\sigma}(i\omega_{\mu } ) } , \label{trans}\ ] ] and longitudinal , @xmath107 susceptibility in eq .
( [ sigma ] ) are found in random phase approximation ( rpa ) with renormalized interaction @xmath108 . here
the non - interacting susceptibilities , @xmath109 , are calculated from the dynamical weiss field green function ( [ cavity ] ) .
we would like to emphasize that the renormalized interaction @xmath97 is not a fitting parameter,@xcite but results from the static screening by particle - particle diagrams which leads to@xcite @xmath110,\ ] ] where the particle - particle vertex is again determined by the weiss field , @xmath111 this screening effect gives the magnetic structure factor @xcite and the self - energy @xcite calculated from eq .
( [ sigma ] ) in good agreement with the qmc results , and depends on the underlying magnetic order .
it is largest in the paramagnetic states and vanishes in the nel state at @xmath7 for @xmath112 , and is thus very important when the magnetic phase diagrams are considered.@xcite the proposed self - energy ( [ sigma ] ) expresses the spin - fluctuation exchange interaction @xcite with an effective potential due to particle - particle scattering.@xcite eqs .
( [ localg ] ) , ( [ cavity ] ) and ( [ sigma ] ) represent a solution for the one - particle green function within the dmft .
they have been solved self - consistently and the energetically stable spiral configuration was found .
this procedure is further justified by the sum rule,@xcite @xmath113 which is well fulfilled in the present approach with @xmath114.@xcite we also show below ( sec .
[ sec : mott ] ) that the local self - energy ( [ sigma ] ) leads to an overall satisfactory agreement with the qmc and ed data .
a complete theory of photoemission ( pes ) would require an analysis not only of the one - particle green s function but also of the three - particle green s functions .
we would like to point out that quantitative calculations of the three - particle green s function for strongly correlated systems has not yet proven feasible .
however , some aspects of the problem can be discussed in terms of the one - electron spectrum , provided that `` final - state '' or `` particle - hole '' interactions can be neglected . under this assumption
the problem simplifies and the pes spectrum may be determined using the one - particle green function alone .
such an approach which neglects final - state and particle - hole interactions has been applied with success to interpret @xcite the dispersion found in the arpes data of the copper oxides.@xcite here we shall derive the relation of the pes spectra to the one - electron spectral function within the `` sudden '' approximation , where final state interactions are neglected.@xcite . to be specific ,
let us consider a transition from a state @xmath115 with energy @xmath116 into a state of the form @xmath117 , in which we treat the photoelectron in state @xmath118 as dynamically decoupled , but retain the full many - body interactions of the electrons in the bulk described by the hubbard model hamiltonian .
the pes intensity for the magnetic system with an incommensurate magnetic order is nontrivial within the dmft approach , as one can not use a bogoliubov transformation to establish the relation between the measured electrons and their local states in the ss state .
the outgoing photoelectron is observed in the _ global reference system _ , whereas the quantum states of the bulk @xmath115 have to be considered within the _ local reference system _ for the spin degrees of freedom . for clarity
we write in the following the operators for the scattered states in capital letters and the operators for the electronic states of the solid described by the present model hamiltonian in lower - case letters . at finite temperature @xmath80
we consider the probability density of the absorption of a photon with frequency @xmath119 in a grand canonical ensemble and obtain , @xmath120 with @xmath121 and the partition function @xmath122 . the amplitude of the transition , @xmath123_{\sigma\sigma^\prime } c_{{\bf p}^\prime-\frac{\bf q}{2 } \sigma^\prime}^ { } | n \rangle \label{matphoto}\ ] ] is determined by the optical matrix element , @xmath124 \nonumber \\ & + & \frac{1}{2}\delta_{{\bf p},{\bf p}^\prime-{\bf q } } [ \cos\frac{\theta}{2}(\hat{1}+\hat{\sigma}_z ) -\sin\frac{\theta}{2 } \hat{\sigma}_+ ] , \label{oma}\end{aligned}\ ] ] and can be calculated using the bloch wave - functions in the _ global reference system _ , @xmath125 where is the polarization vector . the operator @xmath126 conserves total momentum in the scattering plane , so that @xmath127 , where @xmath128 is a 2d lattice vector , and @xmath129 is the photoelectron momentum component in the 2d plane . for solids
the outgoing wave solution is the `` time - inverted low - energy electron diffraction ( leed ) state''.@xcite the leed state consists of an incoming plane wave , reflected plane waves , and a combination of bloch waves inside the solid which fulfill the matching boundary conditions . in lowest order
we have one ( damped ) bloch wave travelling away from the surface .
in the time inverted ( complex conjugated ) state the bloch wave travels towards the surface , and goes over in a plane wave outside .
the leed scattered waves become incoming waves on time inversion , and give no contribution to the photocurrent .
the photoelectron is usually detected at energies which are much higher than the typical energy regime described by the hubbard model , and therefore the bloch waves occupy high - energy quantum states which are initially unoccupied , @xmath130 hence , we obtain for the plane ss state ( @xmath47 ) , @xmath131 where @xmath132 is the fermi function , and the following identity , valid only for plane spirals ( @xmath47 ) , has been used ( @xmath133 for @xmath134 ) , @xmath135 within the `` sudden '' approximation the measured pes spectra near the fermi energy can therefore be related to the one - electron spectral function eq .
( [ prophotof ] ) of the system with local spin - quantization axes , defined by @xmath136 therefore , the total one - particle excitation spectra is described by the spectral function , @xmath137 where @xmath138 is given by eq .
( [ localg ] ) , and a numerical broadening @xmath139 .
the electron occupation number @xmath140 normalized per one spin , equal to the one - electron removal sum , can be obtained without analytic continuation of the matsubara green s function ( [ localg ] ) by performing a direct summation over the matsubara frequencies , @xmath141 finally , we calculate also the total densities of states in the af and ss states using the derived spectral functions ( [ apes ] ) , @xmath142 we derived the complex optical conductivity @xmath143 for the spiral magnetic order following the formalism introduced by shastry and sutherland,@xcite and by scalapino , white and zhang.@xcite their derivation has to be generalized to the case of extended hopping
. moreover , as the symmetry is locally broken in a magnetic system with local quantization axes , the calculation of the optical conductivity is not straightforward .
the hubbard hamiltonian ( [ hubbard ] ) within the local reference system for the spin quantization axis and first- , second- , and third - nearest neighbor hopping elements , @xmath144 , respectively , has an electron kinetic energy , @xmath145_{\sigma\sigma^{\prime } } c^{}_{l\sigma^{\prime}}\right .
\nonumber \\ & + & \left .
c^{\dagger}_{l\sigma^{\prime}}\left[{\cal r}^{\dagger}(\omega_l ) { \cal r}(\omega_i)\right]_{\sigma^{\prime}\sigma}c^{}_{i\sigma}\right ] , \label{kinpart}\end{aligned}\ ] ] where @xmath146 indicates a restricted sum , with @xmath147 around a given lattice site @xmath20 , and @xmath148 , @xmath149 are unit lattice vectors .
we introduce a directed `` bond metric '' , @xmath150 , which is a distance between two sites , @xmath20 and @xmath151 , _ on the lattice _ , and counts the number of @xmath152oriented bonds that connect site @xmath20 with site @xmath151 , respectively , e.g. , @xmath153 and @xmath154 if the electron hops to a third - nearest neighbor with amplitude @xmath24 along an @xmath155oriented link . here
@xmath156 is the unitary matrix which transforms the original fermions , @xmath157 , into the fermions quantized with respect to local quantization axes at each site , @xmath158 , introduced in eq .
( [ rotation ] ) . in what follows
we are interested in the current response to a vector potential along the @xmath159-direction of the 2d square lattice @xmath160 . in the presence of a vector potential ,
the hopping term is modified by the peierls phase factor,@xcite either @xmath161 or @xmath162 , for @xmath163 or @xmath164 , respectively . expanding these phase factors in the usual manner up to second order
@xmath165 one finds , @xmath166 .
\label{kinpotpart}\ ] ] here @xmath167 is the @xmath159-component of the paramagnetic current density , @xmath168_{\sigma\sigma^{\prime } } c^{}_{l\sigma^{\prime}}\right .
\nonumber \\ & - & \left .
c^{\dagger}_{l\sigma^{\prime}}\left[{\cal r}^{\dagger}(\omega_l ) { \cal r}(\omega_i)\right]_{\sigma^{\prime}\sigma}c^{}_{i\sigma}\right ] , \label{parcurdens}\end{aligned}\ ] ] and @xmath169 is the kinetic - energy contribution due to the @xmath155oriented links , weighted by the metric factor connecting site @xmath20 with site @xmath151 , @xmath170_{\sigma\sigma^{\prime } } c^{}_{l\sigma^{\prime } } \right .
\nonumber \\ & + & \left .
c^{\dagger}_{l\sigma^{\prime } } \left[{\cal r}^*(\omega_l){\cal r}(\omega_i)\right]_{\sigma^{\prime}\sigma } c^{}_{i\sigma}\right ] , \label{kinx}\end{aligned}\ ] ] after performing a fourier transformation one finds the average contribution of kinetic energy ( [ kinx ] ) per one site , @xmath171^{}_{\sigma\sigma'}c^{}_{{\bf k}\sigma'}\rangle , \label{xkin}\ ] ] with the coupling between the transformed elements at momenta @xmath172 and @xmath173 due to the magnetic order , @xmath174 and @xmath175 . as
usually , the optical conductivity in long - wavelength limit @xmath176 , @xmath177 , is determined by the current response to a vector potential along the @xmath159-direction,@xcite and one finds using the kubo linear response theory , @xmath178 where @xmath179 is the current - current correlation function , @xmath180 the latter correlation function is given exactly by the particle - hole bubble diagram,@xcite where for @xmath176 , @xmath181_{\sigma \sigma ' } c^{}_{{\bf k}\sigma ' } , \label{curr}\ ] ] and for the present ss state , @xmath182 with @xmath183 .
the advantage of using the dmft with the local self - energy is that the vertex corrections to the current - current correlation function ( [ currcorr ] ) disappear , and the optical conductivity can be calculated without further approximations.@xcite we have verified for large variety of doping levels and temperatures that the following optical sum rule , @xmath184 is always fulfilled within the numerical accuracy , in contrast to the approaches which can not be derived in a diagrammatic way .
( [ sumo ] ) is also used to define the plasma - frequency @xmath185 , @xmath186 for the discussion of the complex conductivity function , it is convenient to introduce the following parametrization by the scattering rate , @xmath187 , and the effective mass @xmath188 ( @xmath189 is the electron mass),@xcite @xmath190 > from the real part of the optical conductivity ( [ optcon ] ) we find in the limit @xmath191 the static conductivity , @xmath192 with the drude weight @xmath193 which may be obtained from the zero - temperature extrapolation of the current - current correlation function in the upper complex plane,@xcite @xmath194 . \label{druden}\ ] ] the optical conductivity allows to determine the in - plane static resistivity , @xmath195 where the static conductivity @xmath196 is obtained as in eq .
( [ dcc ] ) .
we present the results obtained for the optical conductivity and static resistivity in sec .
[ sec : optic ] , and show that the magnetic order in the doped compounds has directly measurable consequences for these quantities .
the ground state of the hubbard model with nearest neighbor hopping ( @xmath101 ) on a square lattice is an af insulator . the insulating behavior and the gap develop gradually at half - filling with increasing @xmath3 starting from @xmath197 due to the perfect nesting instability , leading to a slater gap .
this gap changes into a mott - hubbard gap under increasing @xmath3 , and the system approaches the limit of a heisenberg antiferromagnet.@xcite this regime of large @xmath3 was found to be difficult for a quantitative description within the dmft approches,@xcite as an accurate determination of the energy gains due to af long - range order is there of crucial importance .
therefore , the attempts to describe the af order based on the sopt within the ips failed and the magnetic order disappeared at larger @xmath3.@xcite in contrast , the qmc calculation in the @xmath12 limit gave a stable af state for large @xmath198.@xcite we treat here the range of large @xmath199 ( @xmath200 ) as a test case for our analytic method .
the calculations were performed at a low temperature @xmath201 which allows to describe the magnetic excitations ( @xmath202 ) .
they gave an af ground state at @xmath7 which reproduces correctly the localization of electrons in the limit of @xmath112 .
the magnetization @xmath203 ( [ moment ] ) is only slightly reduced by the dynamical effects with respect to its hf value , and approaches the hf limit at @xmath112 .
the ground state is the nel af state , as found in the @xmath12 limit.@xcite thus , we reduced the self - consistently obtained values of the mean - field magnetization @xmath203 ( [ moment ] ) by a factor 0.606 in order to simulate the known reduction of @xmath203 by intersite quantum fluctuations in a 2d lattice,@xcite @xmath204 after this reduction the calculated values of @xmath205 approach the value of 0.606 in the limit @xmath112 [ fig .
[ fig : magnet](a ) ] .
one finds also a very good agreement with the qmc data @xcite at @xmath206 and 4 , and a reasonable agreement at @xmath207 .
in contrast , the af gap @xmath208 is significantly reduced from its hf value [ fig . [
fig : magnet](b ) ] .
this reduction follows from a drastic change of the one - particle spectra by dynamical effects which lead to qp states at the edge of the mott - hubbard gap which are accompanied by a large incoherent part at higher energies .
also the reduction of the gap found in ed,@xcite comes out correctly as shown in ref . .
for example , we have found a gap of @xmath209 at @xmath207 , while the corresponding gap in the hf calculation is @xmath210 .
this gap reduction can also be captured by the leading dynamical correlations described within the sopt , but only in the regime of @xmath211 .
the discrepancy between the sopt and the dmft results increases with increasing @xmath3 , with the gap and the magnetization @xmath205 being too small , and finally the af order disappears and the gap closes at @xmath212 .
this shows a very limited applicability of the approaches using the self - energy based on the sopt,@xcite which are known to underestimate the region of stability of magnetic states and fail at large @xmath3 due to the uncontrolled increase of the correlation energy in nonmagnetic states .
the spectral functions found within the dmft ( [ apes ] ) are dominated by the lower hubbard band ( lhb ) , i.e. , pes part at @xmath213 , and the upper hubbard band ( uhb ) , i.e. , inverse photoemission ( ipes ) part at @xmath214 , separated by a large gap ( fig .
[ fig : spectraaf ] ) .
both the pes and ipes spectrum show two distinct energy regimes : ( i ) narrow qp peaks at low energies , i.e. , at the edge of the mott - hubbard gap , and ( ii ) incoherent and more extended features at higher energies @xmath215 .
the overall shape of the density of states @xmath216 agrees very well with the ed data for a @xmath217 cluster.@xcite the spectra have a characteristic @xmath71-dependence with the overall weight moving from the pes to ipes part along the @xmath218 and @xmath219 directions , where @xmath220 , @xmath221 and @xmath222 , in qualitative agreement with qmc data.@xcite the spectra obey the particle - hole symmetry of the model , with spectra symmetric with respect to @xmath223 at the @xmath224 and @xmath225 points .
the spectrum at the @xmath226 point is a mirror image of the one at the @xmath227 point .
the qp maxima near the mott - hubbard gap resemble those found in the @xmath4-@xmath5 model in ed or within the self - consistent born approximation,@xcite in spite of using a local self - energy in the present scheme .
this shows that the local many - body problem solved within the dmft suffices to capture the low - energy scale relevant for the qp propagation .
moreover , unlike in the @xmath16 model which results in the ladder spectrum for a single hole,@xcite the qp _ can propagate _ , as they couple to the spin flips of the mean - field bath around site @xmath228 at which the many - body problem is being solved .
the qp dispersion is @xmath229 [ fig .
[ fig : band0](a ) ] , with the maxima along the af brillouin zone ( bz ) , and remains very close to that found in the @xmath4-@xmath5 model.@xcite in the hf approximation , the electron occupation factors @xmath230 are larger for the states which belong to the af bz than for the remaining states outside the af zone . on comparing the weights of the electronic states with momenta @xmath71 and @xmath231
, one finds that also in the dmft the electron weights are much larger within than outside of the af bz [ fig .
[ fig : band0](b ) ] .
the overall pes weight is smoothly distributed in the 2d bz , with the maximum ( minimum ) at the @xmath227 ( @xmath226 ) point , respectively .
this result agrees well with a qmc simulations , and the present data show the same step - like behavior of the electron occupation factor @xmath140 when crossing the @xmath224 point along the @xmath218 direction as the qmc data at @xmath232 and @xmath233.@xcite .5 cm .values of the model parameters used for the presented calculations ; the parameter sets chosen for la@xmath9sr@xmath10cuo@xmath8 and yba@xmath0cu@xmath234o@xmath235 follow from the down - folding procedure of ref . . [ cols="^,^,^,^,^,^,^",options="header " , ] [ table2 ] the @xmath71-resolved spectral functions ( figs .
[ fig : spectra8 ] and [ fig : spectra4 ] ) allow us to identify the generic features of the doped antiferromagnets described by the hubbard model , in the regime of large @xmath3 .
first of all , the spectra are still dominated by the large _ mott - hubbard gap _ which separates the lhb from the uhb .
the mott - hubbard gap develops from the respective gap at half - filling and is considerably reduced from @xmath3 by the qp subbands which occur next to the large gap both in the lhb and in the uhb .
this large gap is accompanied by a smaller _ pseudogap _ @xmath236 between the occupied ( pes ) and unoccupied ( ipes ) part of the lhb at low temperature @xmath201 ( taking @xmath237 ev it corresponds to @xmath238 k ) .
this pseudogap results from the ss order , and separates the majority and minority spin states ( with respect to the local coordinates at each site ) .
it is best visible along the @xmath239 and @xmath240 direction at @xmath241 , and becomes somewhat wider and less distinct in the ss(1,0 ) spiral at higher doping @xmath242 .
we emphasize that the two features below and above the chemical potential @xmath243 originate from the same qp peak at half - filling .
this shows that the qp found in the spectral function of one hole in the @xmath4-@xmath5 ( or hubbard ) model can not describe the regime with finite doping as the _ rigid band picture breaks down_. the pseudogap is visible along the @xmath239 direction starting from @xmath244 , and the maximum above the chemical potential @xmath243 grows gradually towards the @xmath224 point .
consider first the case of lower doping @xmath241 ( underdoped case ) .
one finds that most of the spectral weight at the @xmath224 point is still at @xmath213 , with a sharp qp peak at @xmath245 .
increasing @xmath71 along the @xmath240 direction gives a transfer of the overall weight to higher energies , and the qp peak below @xmath243 gradually looses the intensity , while the peak above @xmath243 takes over around @xmath246 .
however , the feature at @xmath213 is still well visible as a shadow qp band ( fig . [
fig : spectra8 ] ) , with a width @xmath229 .
thus , the qp band of the @xmath4-@xmath5 model is drastically modified at finite doping , and a new energy scale @xmath247 due to the pseudogap accompanies the dispersive qp feature below the chemical potential .
a similar situation is found also at higher doping @xmath242 ( overdoped regime ) , and the pseudogap is quite pronounced along the @xmath239 and @xmath240 directions ( fig .
[ fig : spectra4 ] ) .
however , except for the neighborhood of the @xmath248 point , more spectral weight is found at high energies . already at the @xmath224-point
one finds that the peak at @xmath249 has a higher intensity than the one below @xmath243 .
it becomes gradually weaker when the @xmath226 point is approched , and disperses in the energy range @xmath250 , while the feature below @xmath243 still has a similar dispersion @xmath229 as in the @xmath241 case .
we note that the pseudogap increases to @xmath251 .
moreover , one finds that the qp dispersion is broader at @xmath242 , indicating the gradual weakening of the local magnetic order with increasing doping .
the spectra are drastically changed , in particular in the low - energy range of @xmath252 when the temperature is increased . at @xmath253
the ss order is unstable against the af order which we interpret as a crossover to the small regions of the short - range order with the preferably af ordering of nearest - neighbor spins .
the spectra found for doping @xmath241 at @xmath254 consist of broad maxima which correspond to the lhb and uhb , respectively , and only a single maximum is found in @xmath255 next to the @xmath224-point . these data , and also the spectral functions for @xmath256 reported earlier,@xcite
agree remarkably well with the results of qmc calculations.@xcite the spectra at @xmath242 and @xmath254 are quite similar to those at lower doping @xmath241 , with more weight in the ipes part of the lhb .
we do not intend to present detailed analysis of the spectra obtained using the extended hopping parameters which correspond to the electronic structure of la@xmath9sr@xmath10cuo@xmath8 and yba@xmath0cu@xmath234o@xmath235 , respectively .
instead , we point out the important similarities and differences to the hubbard model as far as the ss states are concerned .
consider first the effective parameters of la@xmath9sr@xmath10cuo@xmath8 .
first of all , a narrow qp band is also found below the fermi energy ( fig .
[ fig : lsco ] ) , but the measured dispersion between the @xmath227 and @xmath224 point is @xmath257 ( @xmath258 ) at @xmath241 ( @xmath242 ) , while it amounts only to @xmath259 at @xmath260 . note that the energies of the qp peak are much closer to each other for the present parameters than in the hubbard model with @xmath101 , where one finds instead the dispersion of @xmath261 ( @xmath262 ) at @xmath241 ( @xmath242 ) at @xmath263 , while it amounts to @xmath264 in the undoped case .
this gradual widening of the qp dispersion with increasing doping may be understood as a consequence of the admixture of ferromagnetic components with increasing doping in the ss(1,1 ) states .
the same trend is also observed for the parameters of yba@xmath0cu@xmath234o@xmath235 , where one finds the qp states in the pes part separated by @xmath265 ( @xmath266 ) between the @xmath227 and @xmath224 point at @xmath267 ( 0.25 ) ( fig .
[ fig : bisco ] ) , while this splitting is only @xmath268 at half - filling . finally , the finite hopping elements to the second and third neighbors stabilize the ss(1,1 ) state with respect to the ss(1,0 ) state also at higher doping ( see fig . [
fig : mu ] ) , and therefore the intensity at the @xmath224 point does not cross the fermi level even at @xmath242 for both parameter sets .
in fact , taking @xmath269 mev ( @xmath270 ) , the qp state at the @xmath224 point is found at @xmath271 ev , and does not change significantly as a function of doping ( fig .
[ fig : bisco ] ) . in contrast , in the arpes experiments for bi@xmath0sr@xmath0cacu@xmath0 the qp state at @xmath224-point is found at energy @xmath272 ev ( @xmath273 ev ) in the underdoped ( optimally doped ) compound.@xcite this indicates that either an improved solution of the many - body problem is still required , or the actual magnetic order in these compounds might be different from ss states .
however , the observed increase of the onset of incommensurability with increasing @xmath3 and @xmath23 is consistent with the observations made by igarashi and fulde @xcite and with qmc calculations of duffy and moreo.@xcite we already pointed out@xcite a very good agreement between the calculated density of states ( [ dos ] ) and the results of ed by dagotto _
et al_.@xcite here we present instead a comparison between the density of states obtained within the dmft method and that found in the hf approximation ( fig .
[ fig : dos ] ) .
first of all , one notices a narrower gap which separates the qp subbands in the dmft of width @xmath229 , instead of the hf one - particle states , on the scale of @xmath274 .
this part of the spectral density might also be reproduced in effective single - particle approaches , as for instance in the slave - boson mean - field theory .
however , the incoherent parts which extend on the energy scale down to @xmath275 result from many - body scattering and can only be reproduced if the dynamical part of the self - energy is included .
the overall width of the subbands at @xmath213 and @xmath214 is @xmath276 , respectively , as known from the analysis of the @xmath4-@xmath5 model in ed and in qmc calculations.@xcite it is evident that due to the changes of @xmath216 in the range of @xmath277 with respect to qp band in the undoped system , the low - energy part of the spectrum can not be reproduced in a renormalized one - particle theory .
the pseudogap in the doped systems is not visible in the hf density of states , and it remains a challenge whether an effective one - particle theory which captures this essential new energy scale could be constructed .
as expected , the agreement between the hf and dmft density of states improves somewhat at higher doping @xmath242 , where the mott - hubbard gap is gradually lost , and the system approaches the single - particle limit .
we note , however , that the gap between the lhb and the uhb relies in our approach on the magnetic order , and a more accurate approach in the strongly doped regime at large @xmath3 would instead have to include the scattering on local moments . in spite of a very good agreement between the present dmft approach and the ed
data,@xcite it is interesting to investigate to what extent the analytic formula for the self - energy ( [ sigma ] ) describes the hole dynamics in a doped system .
therefore , we performed also a dmft - qmc calculation of local @xmath278 for ss states , and the corresponding densities of states , shown in fig .
[ fig : qmc ] .
the qp peaks are very close to each other at half - filling , while the incoherent states at higher energies in the lhb and uhb have almost the same weights , but are moved to somewhat higher energies in the qmc calculation .
the increase of the spectral weight close to the fermi level is well pronounced in the latter calculation at @xmath241 , but one finds instead _ a pseudogap smaller by a factor close to five_. however , one should realize that the present calculation performed at low temperature @xmath201 corresponds in practice to the ground state , while the same temperature in qmc includes already thermal fluctuations which considerably reduce the size of the pseudogap . indeed
, we find using the ed method to solve the self - energy within dmft the pseudogap in the ss state of @xmath279 .
it might be expected that this reduction of the energy scale would result in a better quantitative description of the spectral functions and the related excitations across the pseudogap , leading to a reduced energy scale for the low - energy features of the optical conductivity ( sec .
[ sec : optic ] ) .
we also found more extended energy range of the incoherent states which belong to the uhb in the qmc calculation .
altogether , the comparison with the dmft - qmc calculation demonstrates that the analytic method developed in this paper is very useful for a qualitative insight into the possible changes of magnetic states under doping and their consequences for the properties measured in experiment .
the evolution of the spectral functions @xmath255 and the density of states @xmath216 with doping , reported in secs . [
sec : mott ] and [ sec : pes ] , motivates an investigation of the optical properties . here
we make use of the theory introduced in sec . [
sec : transport ] , where we have shown how the optical conductivity can be derived from the local self - energy in the present dmft treatment . as an illustrative example , we concentrate on the optical conductivity found for the hubbard model with the nearest - neighbor hopping ( @xmath280 ) and @xmath263 .
we present the optical data in figs .
[ fig : opticlo ] and [ fig : optichi ] at two temperatures : @xmath201 and @xmath281 . while the magnetic order is af at half - filling , the ss states characterized by the wave - vector @xmath282 $ ] change with doping and temperature . at lower doping @xmath241
we find a ss(1,1 ) with @xmath283 ( 0.09 ) at @xmath201 ( @xmath284 ) , respectively , while at higher doping @xmath242 , a ss(1,0 ) state ( @xmath285 ) with @xmath286 ( 0.23 ) , or an equivalent ss(0,1 ) state , is found instead . at half - filling one
finds a large gap below @xmath287 at @xmath207 and no drude peak which shows that the system is in the insulating phase.@xcite the conductivity at @xmath288 is incoherent and originates from the excitations across the mott - hubbard gap .
this changes drastically when the system is doped and two new features occur at lower energy : the drude peak , and the mid - gap state at @xmath289 , both with the increasing intensity between @xmath241 and @xmath242 at low temperature ( fig .
[ fig : opticlo ] ) .
these features are accompanied by an incoherent background of the excitations within the lhb .
the peak at @xmath289 corresponds to excitations across the pseudogap ; as such it is more influenced by the increasing temperature in the underdoped regime , where the ss(1,1 ) state is less robust than the ss(1,0 ) state in the overdoped regime .
below @xmath290 the frequency dependent scattering rate @xmath291 and the effective mass @xmath188 can also be divided into two regions : ( i ) above @xmath289 the scattering rate increases monotonically with increasing frequency , ( ii ) below @xmath289 it has a maximum at energy @xmath292 ( @xmath293 ) for @xmath241 ( @xmath294 ) , and drops to zero for @xmath191 at finite doping .
this behavior for @xmath191 and @xmath295 is consistent with the fermi liquid behavior which follows from the local approximation to the self - energy ( [ localg ] ) .
a finite value at @xmath223 is a numerical effect due to finite broadening of the spectra ( @xmath296 ) .
the frequency region in which the scattering is suppressed has a direct relation to the existence of a pseudogap region in the single - particle spectral function @xmath255 , reported in sec .
[ sec : spectra ] , and indicates that ss lro reduces the scattering of the charged carriers in the energy range @xmath297 . at the same time , the effective mass @xmath298 rises to a maximum value of @xmath299 within the pseudogap region , and is found to be rather independent of hole doping .
as the temperature increases to @xmath284 , the pseudogap disappears and the region of suppressed scattering is filled up in the underdoped regime with @xmath241 , while the low scattering persists for @xmath300 at @xmath242 ( see fig .
[ fig : optichi ] ) . at the same time , the mid - gap state in the real part of the optical conductivity changes into a smooth feature which extends down to the drude peak for @xmath241 , contrary to the case with @xmath242 where the spectral weights of the above two features remain well separated .
this is clearly related to the behavior observed in @xmath255 with increasing temperature , where the pseudogap along the @xmath240 direction filled up with spectral weight as @xmath80 increased for @xmath241 .
these changes of the mid - gap state with temperature are due to the changes of the magnetic correlations in the doped systems included in our calculations which do not distinguish between long - range and short - range magnetic order , but treat local dynamical correlations . however , there are indications that the mid - gap feature results from an interplay between short - range magnetic order and electron correlations.@xcite therefore , the rather strong evolution of the low - energy weight with increasing temperature shown in figs .
[ fig : opticlo ] and [ fig : optichi ] may be overestimated in the present treatment of the self - energy which does not allow to get a metal - insulator transition _ without _ an accompanying magnetic lro .
we also note that the mid - gap states are likely a bare consequence of the strongly correlated nature of optical and one - particle excitations in the hubbard model,@xcite and it is still a challenge to describe them better in a theory which would treat the af and paramagnetic states with local moments on equal footing .
the frequency - dependent scattering rate allows us to find a _ crossover temperature _
@xmath301 at which the pseudogap closes .
we estimated that @xmath302 for @xmath241 , and observed a monotonic increase of @xmath303 up to @xmath304 . at @xmath284
the effective mass increases up to @xmath305 within the pseudogap at @xmath241 . at half - filling and
@xmath284 one finds that the charge - transfer gap is only slightly reduced from its value at @xmath201 , and the insulating behavior is accompanied by af lro .
we estimated the nel temperature for @xmath263 to be @xmath306 .
further evidence for a characteristic crossover temperature @xmath301 may be found in the behavior of the in - plane dc resistivity ( [ dcr ] ) .
the resistivity received a lot of attention in connection with the observed normal state pseudogap in the electronic excitation spectrum,@xcite and from theoretical point of view.@xcite in fact , the physical origin of linear @xmath80-dependence of @xmath307 for samples of high-@xmath308 compounds close to the optimal doping level remains puzzling .
the results for @xmath309 obtained for the hubbard model at two doping levels , @xmath241 and @xmath242 , are shown in fig .
[ fig : zstar ] . at low temperatures , @xmath310
, the resistivity shows fermi - liquid behavior for both hole densities , i.e. , @xmath311 .
as usually in the dmft calculations,@xcite the @xmath312-dependence of @xmath309 originates from the low - frequency behavior of the imaginary part of the local self - energy .
in the regime of high temperatures ( @xmath313 ) , the resistivity increases linearly with temperature which is due to temperature independence of the spectral functions @xmath314 at @xmath315 , and the high temperature limit of the derivative , @xmath316 , thus leading to @xmath317 , i.e. , @xmath318 .
as the temperature is lowered , the magnetic moments gradually build up , and a kink in the resistivity appears . therefore ,
the increase in the resistivity as the temperature is lowered can be attributed to the enhancement in the scattering of electrons by local spin fluctuations . on the contrary , for large hole doping
the system is a better metal , hole spin correlations are gradually lost , and the increase of resistivity is less pronounced in the temperature region @xmath319 at @xmath242 .
the maximum of @xmath309 for @xmath241 is located almost exactly at @xmath320 ( @xmath321 taking the experimental value of the superexchange @xmath322 mev ) , where the pseudogap in the single - particle excitation spectra opens leading to a suppression of the effective scattering rate @xmath323 , as discussed previously .
this defines the crossover temperature @xmath301 .
remarkably , the change from a linear to a nearly - linear @xmath80-dependence , @xmath324 ( @xmath139 ) , of the in - plane dc resistivity of la@xmath9sr@xmath10cuo@xmath8 was found to be at @xmath325 k for @xmath326 and was attributed to the opening of a pseudogap in the electronic excitation spectrum.@xcite however , the saturation of the resistivity @xmath327 can not be observed in a real system as the carriers also couple to other bosonic excitations , e.g. to phonons , which are neglected here . upon lowering the temperature below @xmath301 , at @xmath328 for both @xmath241 and for @xmath242
, we observed a nearly linear @xmath80-dependence of @xmath309 .
in this temperature range the ss wave vector , @xmath282 $ ] , becomes strongly temperature dependent , and maintains the directional deviation from the af wave vector , @xmath329 , with @xmath330 for @xmath241 and @xmath331 ( @xmath285 ) for @xmath242 . in both cases @xmath332 increases from @xmath333 with decreasing temperature and saturates at its ground - state value @xmath334 below @xmath335 . in the linear regime ( @xmath336 )
the resistivity can be fitted quite well by a linear @xmath80-dependence , as expected for the ss states,@xcite @xmath337 , with @xmath338 ( -0.25 ) for @xmath241 ( 0.25 ) , respectively , where the increase of the negative temperature coefficient @xmath339 is a further manifestation of the gradual loss of local magnetic moments as doping is increased . on the contrary , in the paramagnetic phase of the hubbard model at @xmath11 one finds @xmath340.@xcite furthermore , the slope of @xmath309 in the low temperature regime is given by @xmath341 independent of hole density .
this value is larger by about a factor of 2.5 than the respective slope found in the retraceable - path approximation , @xcite and in the ed studies at finite temperature,@xcite being @xmath342 and 0.60 , respectively , and demonstrates that the changes in the magnetic order with increasing temperature influence significantly the system resistivity .
unfortunately , such effects can not be studied in the ed method due to the small size of considered clusters . in order to further support our observation that the crossover temperature @xmath301 is related to the pseudogap in the single - particle excitation spectrum we plot in the inset of fig .
[ fig : zstar ] an average of the single - particle spectral weight within an energy window @xmath343 around the fermi energy @xmath223 , defined by,@xcite @xmath344 similarly , a measure for the temperature dependence of the density of states at the fermi energy @xmath345 is obtained from the local green s function ( [ localg ] ) , @xmath346 . in the low temperature limit @xmath345 can be obtained from the relation @xmath347,@xcite which gives @xmath348 ( @xmath349 ) for @xmath241 ( @xmath242 ) , respectively .
however , one finds that the one - particle density of states at the fermi energy does not evolve smoothly to the low temperature values , but instead states are depleted from the region @xmath350 as @xmath80 is reduced below @xmath302 ( @xmath351 ) for @xmath241 ( @xmath242 ) , respectively . in particular , we observed a faster loss of the qp weight with momentum @xmath352 for @xmath241 ( fig .
[ fig : zstar ] ) .
this shows that the opening of the pseudogap in the one - particle excitation spectrum at @xmath353 coincides with the suppression of the effective scattering rate @xmath323 .
experimentally , the resistivity changes from a linear to a nearly - linear @xmath80-dependence at @xmath301 of the order of 500 k. although our calculations do not allow to interpret the linear part of @xmath309 at high temperature @xmath354 as only the electronic degrees of freedom are included , we note that the enhanced slope of @xmath309 at low temperature @xmath355 k and the negative temperature coefficient agree qualitatively with the experimental results for yba@xmath0cu@xmath234o@xmath356 in the underdoped regime.@xcite our calculations confirm the conjecture of shraiman and siggia of a nearly - linear @xmath80-dependence of the resistivity for a system with ss magnetic order.@xcite these features can be seen as generic fingerprints of incommensurate magnetic correlations .
similar changes in the optical excitation spectra as a function of hole doping were also found using the effective single - band models with the parameters representative for la@xmath9sr@xmath10cuo@xmath8 ( fig .
[ fig : opticla ] ) and yba@xmath0cu@xmath234o@xmath235 ( fig .
[ fig : opticbi ] ) , respectively . due to somewhat larger values of the effective @xmath3 , the gap in the optical spectra increases to @xmath357 and @xmath358 in these two compounds .
one finds again that the drude weight and the mid - gap state appear in the conductivity of doped systems . for the parameters of la@xmath9sr@xmath10cuo@xmath8 ( yba@xmath0cu@xmath234o@xmath235 ) the region of
suppressed scattering extends up to @xmath359 ( @xmath360 ) at @xmath241 .
this regime of low @xmath119 gives an enhanced effective mass @xmath361 for both sets of model parameters .
at larger doping @xmath242 the coherence of the charge carriers is enhanced by @xmath23 and @xmath24 hopping , and one finds a significantly reduced effective scattering between charged carriers , extending with roughly no structure over a rather broad energy range .
simultaneously , the effective mass @xmath362 is only little enhanced at low energies . the overall shape of @xmath363 ( fig .
[ fig : opticla ] ) shows a qualitatively similar behavior to the optical conductivity of la@xmath9sr@xmath10cuo@xmath8 reported by uchida _
et al._.@xcite at low doping the mid - gap band centered at @xmath364 ( corresponding to @xmath365 ev for @xmath322 mev and the present parameters with @xmath366 ) is clearly distinguishable from the drude contribution .
it moves to higher energy @xmath367 ( 0.7 ev ) at @xmath242 .
it is quite remarkable that our dmft calculations reproduce qualitatively the structures observed in the frequency dependent effective scattering rate @xmath291 and in the effective mass @xmath188 of la@xmath9sr@xmath10cuo@xmath8.@xcite in particular , the strong doping dependence of @xmath291 and @xmath188 show the same trends , namely a pronounced reduction of scattering and effective carrier mass for the heavily doped systems , and further justifies the importance of extended hopping parameters in the cuprates .
this behavior originates from an increase of qp weight in the single - particle excitation spectrum induced by doping .
puchkov _ et al._@xcite reported extensive studies of the infrared properties of yba@xmath0cu@xmath234o@xmath235 , bi@xmath0sr@xmath0cacu@xmath0o@xmath368 , and other high-@xmath308 compounds .
they found that the far - infrared effective scattering rate @xmath291 and the effective mass @xmath188 differ significantly between underdoped and optimally doped samples above @xmath308 .
the optimally doped samples show a structureless and lower effective scattering rate @xmath291 and a nearly constant and unrenormalized mass @xmath298 . on the contrary , in the underdoped samples the scattering between the charged carriers below @xmath369 ev
is strongly suppressed and @xmath188 is enhanced in the low - energy region .
these observations are in remarkably good agreement with our findings and supports our conclusion that the observed doping dependence of @xmath291 and @xmath188 originate from an increased coherence of the one - particle excitation spectra reported in [ sec : spectra ] , and experimentally observed in arpes spectra of bi@xmath0sr@xmath0cacu@xmath0o@xmath368 by kim
_ et al._.@xcite the suppression of @xmath291 below @xmath370 ev originates from the opening of a pseudogap in the one - particle excitation spectrum . using @xmath322 mev and the value of @xmath371 adequate for yba@xmath0cu@xmath234o@xmath235 , we find the energy threshold below which qp scattering is strongly suppressed in the weakly doped system at @xmath372 ev . unfortunately
, this is about a factor of five larger than the experimental value for underdoped yba@xmath0cu@xmath234o@xmath235 being @xmath373 .
similar discrepancies in the energy of the qp state with momentum @xmath353 were reported in sec .
[ sec : spectra ] . finally , we compare the drude weight @xmath193 ( [ druden ] ) and the kinetic energy density associated with @xmath159-oriented links @xmath374 ( [ xkin ] ) for the three model parameter sets in fig . [ fig : drude ] .
the dmft gives @xmath375 for the hubbard model at half - filling ( @xmath260 ) with @xmath263 which is by a factor @xmath376 smaller than the hf result and is in excellent agreement with the value of @xmath377 obtained in qmc calculations.@xcite we also found an overall satisfactory agreement of @xmath374 as a function of doping with ed data of dagotto _
et al._.@xcite the kinetic energy @xmath378 increases with doping not only because the actual carrier density changes , but also as a consequence of changing wave vector in the ss state @xmath379 $ ] with a gradually increasing pitch @xmath380 allowing for coherent electronic transport through the system . in agreement with qmc data,@xcite
we observe that increased extended hopping amplitudes accelerate the ss formation and result in a stronger increase of @xmath374 with @xmath381 for the yba@xmath0cu@xmath234o@xmath235 than for la@xmath9sr@xmath10cuo@xmath8 model parameters .
the @xmath159-directed kinetic energy shows a linear doping dependence in the regime of low hole doping , and @xmath374 changes by a factor of @xmath382 ( @xmath383 ) with respect to half - filling in the case of the yba@xmath0cu@xmath234o@xmath235 ( la@xmath9sr@xmath10cuo@xmath8 ) model parameters , one finds an faster increase of total spectral weight in the case of stronger hopping to second and third neighbors , as realized in yba@xmath0cu@xmath234o@xmath235 .
the calculated total optical spectral weights are @xmath384 following the optical sum rule ( [ sumo ] ) , and we made a quantitative comparison with the experimental data .
the doping dependence of the total integrated spectral weight below the charge - transfer band edge at 1.5 ev reported by cooper _
@xcite for la@xmath9sr@xmath10cuo@xmath8 is strikingly similar to the numerical data of fig .
[ fig : drude ] .
the model reproduces a rapid increase of spectral weight up to @xmath385 sr - doping and a rather doping independent spectral weight in the range of @xmath386 .
the increase of @xmath384 with increasing doping is faster for the parameters of yba@xmath0cu@xmath234o@xmath235 , with the integrated spectral weight increased by @xmath387 at @xmath388 with respect to its value at @xmath260 .
this value compares again very well , taking the simplicity of the effective single - band hubbard model , with a factor of @xmath389 found by orenstein _
@xcite in the compound with highest @xmath308 . at @xmath260
we find a vanishing drude weight for all three sets of model parameters , and the system is an insulator .
this is of course an expected result at half - filling , but in the present context it serves as a test of the internal consistency of theory , as the kinetic energy term @xmath390 in eq .
( [ druden ] ) , and has to be compensated by the current - current correlation function @xmath391 in the limit of low temperature . at small hole doping
we observed an almost perfect linear increase of the drude weight with @xmath381 for all three sets of model parameters which is an indication of strong electron correlations near the mott insulator at half - filling.@xcite .
such a behavior is compatible with a picture of a dilute hole gas in a background with ss lro which contributes to the optical response .
however , the crossover to a metal due to increasing doping has been analyzed recently using scaling theory,@xcite and ed technique combined with scaling theory,@xcite which give @xmath392 for small doping concentration @xmath381 in a 2d @xmath4-@xmath5 model .
this last result is in sharp contrast to the present picture of a dilute hole gas in an af or ss background , and might indicate that other correlations are realized in the spin background when the system is doped , namely that the dilute hole gas is unstable towards microscopic phase separation , such as realized in polaronic solutions or stripe phases .
the present results demonstrate a substantial transfer of spectral weight to low energy in the doped systems .
we already pointed out earlier @xcite that the spectral weight transferred into the lhb in the one - particle spectra agrees with the predictions of perturbation theory in the strongly correlated regime.@xcite in the optical spectra for the hubbard model at @xmath207 one finds that the weight transferred into the region below the mott - hubbard gap is increased by a factor @xmath393 with respect to @xmath241 when the system is doped to @xmath242 .
this change is significant as the total weight obtained from ( [ sumo ] ) via @xmath374 remained roughly constant [ see fig . [
fig : drude](a ) ] , indicating a spectral weight transfer from the high- to the low - energy region in the single - particle excitation spectrum.@xcite in particular , the weight transfer is in favor of the drude weight [ fig .
[ fig : drude](b ) ] , which increased in the same doping range by a factor @xmath394 although the hole density increased only by a factor @xmath395 .
these changes in the coherent optical weight are consistent with the observation made in sec .
[ sec : spectra ] that the single - particle excitation spectra become _ more coherent _ as the hole density increases .
we reported a generalization of the dmft to the magnetically ordered states and showed that this method allows for a very transparent study of spectral properties of the hubbard model at and close to half - filling .
the crucial step is the derived formula for the self - energy using the berk - schrieffer @xcite spin - fluctuation exchange interaction with an effective potential due to particle - particle scattering.@xcite we have demonstrated that this treatment of the many - body effects reproduces the leading dependence on doping and temperature and gives a very favorable comparison with the available numerical data obtained in the qmc and ed calculations for a 2d hubbard model .
although the @xmath71-dependence of the self - energy was not included , the spectral functions for a single hole in a mott - hubbard insulator agrees well with the known structure for the @xmath4-@xmath5 model,@xcite and gives the pes spectrum consisting of a _ coherent _ qp peak with a dispersion @xmath229 , and an _ incoherent _ part of width @xmath276 at lower energies .
we have verified that the qp weight agrees well with the ed data in the range of @xmath396,@xcite and supports the string picture.@xcite furthermore , the calculation reveals a nontrivial relation between the electron occupation factors , @xmath140 , and the qp weights , @xmath397 , and shows that the maximum of @xmath397 is shifted away from the @xmath398 point , in agreement with the ed results of eskes and eder.@xcite our study has shown that doping of a mott - hubbard insulator leads to an incommensurate magnetic order at low temperatures , which depends on the actual values of the hopping parameters and the coulomb interaction @xmath3 .
this kind of magnetic order induces a _ pseudogap in the one - particle spectra _ which is one of the generic features of the doped mott - hubbard insulators .
the dependence of the pseudogap on the incommensurate magnetic order explains why it could not be observed in ed data on small clusters at finite temperature,@xcite or in the infinite - dimensional hubbard model in the paramagnetic state.@xcite this new energy scale due to a _ pseudogap of magnetic origin _ demonstrates a combination of physics arising from the slater picture and the mott - hubbard description of strongly correlated electron systems .
the coherent qp states survive in the doped systems , in agreement with the qmc and ed results . however , the numerical studies suggest that a strong @xmath71-dependence of self - energy might be necessary to describe the spectra , as the qp dispersions change .
this failure of the rigid qp band picture finds here quite a different explanation : _ the changes of the qp dispersion follow from the incommensurate magnetic order _ which develops with doping , and the leading effects in the hole dynamics are still _ captured by a local self - energy_. the one - particle and optical spectra are interrelated , and the opening of a pseudogap at low temperatures leads to a mid - gap state next to the drude peak in the optical conductivity , both with growing intensity under increasing doping .
such features , observed in the ss states at low temperatures , as the suppressed scattering rate and large effective mass in the underdoped regime , and almost no enhancement of the effective mass in a broad energy range in overdoped systems are in remarkably good _ qualitative _ agreement with the experimental findings in the cuprates.@xcite this is consistent with the reduced density of states @xmath399 at the fermi energy at low temperature . with increasing temperature the value of @xmath399 increases , which could not be explained in paramagnetic calculations performed within the dmft approach .
it should be realized that such a strong temperature dependence of @xmath399 should have important consequences for several measurable quantities in the normal phase , as for example knight shift .
here we limited ourselves to the qualitative consequences of the extended hopping @xmath23 and @xmath24 for the one - particle and optical spectra .
first of all , the qp dispersion is strongly influenced by these parameters , and at half filling reproduces the experimental width and dispersion of the qp band in sr@xmath0cuo@xmath0cl@xmath0.@xcite second , the deviation of the characteristic @xmath41-vector from the af vector @xmath400 increases as a function of doping in ss states , and this process is accelerated by a finite value of second- ( @xmath401)and third - neighbor ( @xmath402 ) hopping .
this explains why the systems with extended hopping are more metallic which is indicated by the low effective mass and larger drude weight .
the dependence of the magnetic order on temperature has also rather drastic consequences for the measurable quantities .
the onset of magnetic order below a characteristic temperature results in quite different one - particle and optical spectra at low temperatures from those obtained in a paramagnetic phase .
the changes of the spiral @xmath41-vector with decreasing temperature allow to introduce a crossover temperature @xmath301 , below which the low - lying excitations are gradually modified along with the changes in local magnetic order .
such a modification gives a quasi - linear resistivity , and verifies the conjecture of shraiman and siggia.@xcite in spite of very good agreement for the undoped systems , however , we identified several important features which do not agree with the experiments in the doped cuprates even on a qualitative level , that might indicate that either a more accurate treatment of the many - body problem is necessary , or more complex magnetic structures are stabilized in these compounds : ( i ) the value of the pseudogap in the one - particle spectra and the accompanying energy scale for the suppressed scattering rate in the optical conductivity are overestimated by a factor close to five with respect to the experimental observations .
( ii ) the ss(1,1 ) state obtained for the la@xmath9sr@xmath10cuo@xmath8 model parameters leads to a different splitting of the magnetic scattering peak in neutron experiments than the experimentally observed ( 1,0 ) and ( 0,1 ) splittings.@xcite ( iii ) the incommensurate order deviates too fast from the af state for the model parameters of yba@xmath0cu@xmath234o@xmath235 which results in different spin - spin correlations than those observed in experiment , and a qp peak at the @xmath224-point moving to too low energies.@xcite ( iv ) the doping behavior of the pseudogap and of the related crossover temperature @xmath301 is opposite to the one observed in the cuprates . in these materials
the pseudogap and @xmath301 decrease upon doping , whereas here the corresponding quantities increase from the @xmath241 to the @xmath242 case . with increasing doping charge fluctuations
become more and more important which gives rise to the suppression of magnetic order , and consequently the pseudogap closes .
however , such correlations are _ underestimated _ in the present treatment , and one finds instead a persisting pseudogap .
( v ) finally , the spiral spin ordering in the ( 1,1 ) direction contrasts with experimental evidences from neutron scattering in the cuprates , suggesting that stripe ordering might play a prominent role in these systems at very low temperatures.@xcite we have found a phase separation at low doping levels and therefore the presently studied dilute hole gas in ss states is unstable towards magnetic polarons or stripe phases at doping levels lower than @xmath403 .
this motivates a further search for more complex magnetic ground states with incommensurate order , and more accurate methods to describe them in theory .
summarizing , we presented a successful formulation of the dmft for strongly correlated magnetic systems , which opens a possibility of further applications in transition metal oxides .
in contrast to the earlier formulations based on the modified second - order formula for the self - energy,@xcite the present self - energy which describes the dynamical effects in the propagation of a hole coupled to spin fluctuations allows to obtain stable magnetic solutions : af ordering at half - filling and ss in doped systems .
although it is likely that better variational states , possibly with stripe ordering,@xcite could be found it is expected that the presented spectral and optical properties are generic for strongly - correlated systems with incommensurate order parameter . a better understanding of the cuprates , however , requires a further development of theory which should be able to capture the gradual changes of local magnetic correlations in doped mott - hubbard systems under increasing temperature .
it is our pleasure to thank e. arrigoni , p. horsch and d. munzar for valuable discussions .
a.m.o . acknowledges the support by the committee of scientific research ( kbn ) of poland , project no .
2 p03b 175 14 .
j. baa , a. m. ole , and j. zaanen , phys .
b * 52 * , 4597 ( 1995 ) ; p. w. leung , b. o. wells , and r. j. gooding , phys . rev .
b * 56 * , 6320 ( 1997 ) ; r. eder , y. ohta , and g. a. sawatzky , phys . rev .
b * 55 * , r3414 ( 1997 ) ; o. p. sushkov , g. a. sawatzky , r. eder , and h. eskes , phys . rev .
b * 56 * , 11 769 ( 1997 ) ; j. van den brink and o. p. shushkov , phys . rev .
b * 57 * , 3518 ( 1998 ) .
j. m. tranquada , j. d. axe , n. ichikawa , a. r. moodenbaugh , n. nakamura , and s. uchida , phys .
* 78 * , 338 ( 1997 ) ; p. dai , h. a. mook , and f. doan , phys . rev .
lett . * 80 * , 1738 ( 1998 ) ; h. a. mook _
_ , nature * 395 * , 580 ( 1998 ) .
t. pruschke , d. l. cox , and m. jarrell , phys .
b * 47 * , 3553 ( 1993 ) ; m. jarrell , j. k. freericks , and t. pruschke , adv . phys . *
44 * , 187 ( 1995 ) ; m. j. rozenberg _
et al . _ ,
lett . * 77 * , 105 ( 1995 ) . | we present spectral and optical properties of the hubbard model on a two - dimensional square lattice using a generalization of dynamical mean - field theory to magnetic states in finite dimension .
the self - energy includes the effect of spin fluctuations and screening of the coulomb interaction due to particle - particle scattering . at half - filling
the quasiparticles reduce the width of the mott - hubbard ` gap ' and have dispersions and spectral weights that agree remarkably well with quantum monte carlo and exact diagonalization calculations . away from
half - filling we consider incommensurate magnetic order with a varying local spin direction , and derive the photoemission and optical spectra .
the incommensurate magnetic order leads to a pseudogap which opens at the fermi energy and coexists with a large mott - hubbard gap .
the quasiparticle states survive in the doped systems , but their dispersion is modified with the doping and a rigid band picture does not apply .
spectral weight in the optical conductivity is transferred to lower energies and the drude weight increases linearly with increasing doping .
we show that incommensurate magnetic order leads also to mid - gap states in the optical spectra and to decreased scattering rates in the transport processes , in qualitative agreement with the experimental observations in doped systems .
the gradual disappearence of the spiral magnetic order and the vanishing pseudogap with increasing temperature is found to be responsible for the linear resistivity .
we discuss the possible reasons why these results may only partially explain the features observed in the optical spectra of high temperature superconductors . 2 |
high - harmonic generation ( hhg ) is the key light source in the extreme ultraviolet region .
one of the most important applications of the hhg is the generation of attosecond pulses@xcite , which permits subfemtosecond - resolution studies of fundamental electronic processes in atoms and molecules and provides a valuable tool for attosecond molecular physics , nanolithography , et al . @xcite .
the main mechanism of hhg is described by the three - step model , in which an electron undergoes ionization , acceleration in the strong electric field , and recombination accompanied by the emission of high - energy photons @xcite .
recently the generation of high harmonics in gases in the vicinity of metal nanostructures ( nanostructrure - assited high - harmonic generation , na - hhg ) @xcite and micrometer - sized hollow metallic fibers@xcite was experimentally studied .
such setup allows to use significantly reduced pump intensities due to plasmonic field enhancement in the vicinity of nanostructures .
a theoretical investigation of hhg using field enhancement by metallic nanostructures was provided by ref .
@xcite where the lewenstein model was extended to incorporate the field inhomogeneity in the hot spots and electron absorption from the metal surfaces .
after that , in a number of publications the quantum description of both the field inhomogeneity and electron absorption at surface was provided by several groups @xcite , along with the studies of the high - energy photon emission in plasmonic field @xcite , plasmon - assisted as - pulse generation without ceo - stabilization @xcite , hhg in periodic bowtie arrays @xcite , and theoretical studies of na - hhg employing nanoplasmonic field enhancement of coupled ellipsoids @xcite . besides , na - hhg in the vicinity of a rough metallic surface @xcite , in an array of coupled metal nanospheres @xcite , and the generation of isolated attosecond pulses using polarization gating by crossed bowtie nanostructures @xcite have been investigated .
quite notably , experimental investigation of na - hhg in tapered hollow metallic waveguides @xcite demonstrated the possibility to achieve high harmonic numbers using low - intensity pump .
a typical disadvantage of na - hhg is the low efficiency ( @xmath010@xmath1 ) of the spectral transformation due to the short propagation length , compared to typical efficiencies in noble gases of 10@xmath2 .
in contrast to other systems , the na - hhg efficiency is difficult to increase by increasing the propagation length due to significant phase mismatch . besides , the approach of ref .
@xcite was criticized in ref .
@xcite , which interprets the experimental peaks as fluorescence peaks from the noble - gas atoms rather than as high harmonics .
this controversy apply only to set - ups with rather low hhg efficiencies like in @xcite .
therefore it is desirable to investigate the possibilities to increase the efficiency of the na - hhg .
the key factor which limits the nanostructure - assisted hhg efficiency is a large phase mismatch .
it can be overcome by periodic modulation of the medium properties , so - called quasi - phase - matching .
previously , in the context of standard hhg in gases , several approaches @xcite were proposed and implemented for this aim . in the context of na - hhg , quasi - phase - matching in a metal nanoparticle - gas composite @xcite was proposed . however , in view of the development of the hhg generation in very thin waveguides @xcite , it is important to investigate the setups to increase the hhg efficiency in the framework of the waveguide design . in the context of the na - hhg ,
the most suitable of these approaches is the corrugation of the waveguide : the periodic change of the waveguide radius leads to the modulations of the pump beam intensity with a predesigned period .
since the hhg strongly nonlinearly depends on the pump intensity , this results in the quasi - phase - matching .
this approach was used in several experimental and theoretical publications for the standard hhg @xcite . here
we investigate quasi - phase - matched na - hhg waveguides with radii comparable or below the wavelength .
the corrugation ( localized reduction of the inner radius ) of such waveguides will result in pump field enhancement , leading to lowered pump - intensity requirements . by placing the corrugations at the specific period
, one can expect to achieve the quasi - phase - matching between the pump wave and a high harmonic of a certain order . in this paper , we investigate such a design using the first - principle analysis of the modal field and its enhancement , accompanied by the ( 1 + 1)d model for the pump and harmonic field propagation .
we show that one can indeed achieve the quasi - phase matching in such a system , leading to the increased efficiency for a chosen harmonics .
the optimization of the waveguide geometry is performed .
the paper is organized as follows . in section 2
we present the general setup , the distribution of the modal field and its enhancement as the function of the waveguide geometrical parameters . in section 3 ,
the numerical approach is presented in detail . in section 4
, we present and discuss the output harmonics spectra as well as the predicted dependence of the efficiency on the propagation length for quasi - phase - matched and non - quasi - phase - matched situations .
in fig . [ scheme ] , a scheme of the proposed setup is presented .
we consider a plane waveguide ( invariant in @xmath3 direction ) which consists of two metal slabs with thickness of 50 nm each , separated by the distance @xmath4 .
such plane geometry of the waveguide can be more easily manufactured experimentally ( e.g. by lithography ) than a waveguide with a circular hole .
the thickness of 50 nm is larger than the skin layer depth and is sufficient to localize the pump light in the inner core of the waveguide .
the corrugations consist of arc - shaped metal areas with radius @xmath5 separated by distance @xmath6 , which protrude inside the waveguide , as depicted in fig .
[ scheme ] .
the @xmath3-polarized near - ir pump enters the waveguide from the left .
excitation of localized plasmons and associated field enhancement occur in the vicinity of the corrugations .
the high harmonics which are emitted in gas propagate to the right through the waveguide core , adding up constructively in the case of the quasi - phase - matching . before presenting the nonlinear numerical model for the propagation of the pump beam and the harmonics , let us analyze the linear properties of such a waveguide .
the linear propagation of pump light was analyzed using the jcmwave commercial software , which provides rigorous solution of the maxwell equations , with silver properties being properly described by both real and imaginary parts of the dielectric function . in fig .
[ maps](b ) one can see the typical distribution of the light intensity inside the waveguide for the cw pump wave .
note that the periodic oscillations of the field amplitude in the @xmath7 direction are due to the resolved optical wavelength . in the vicinity of the corrugations ,
the intensity is enhanced by a factor up to 15 .
also , the intensity of the pump reduces with the propagation due to the modal loss of the waveguide . in fig .
[ maps](a ) , the spatial map of the phase of the field is presented .
one can see the oscillations of the phase with period equal to the wavelength in the waveguide , which is different from the free - space wavelength due to the waveguide contribution to dispersion , as well as change of the phase at the metal interface . from the data presented in fig .
[ maps ] both real and imaginary parts of the effective refractive index of the waveguide were extracted .
note that here and hereafter we neglect the energy transfer between the transverse modes , since considered propagation distances are small compared to the distance over which such transfer takes place . in fig .
[ deps ] the dependence of the effective refractive index of the waveguide on its geometrical parameters is presented .
one can see that both reduction of the core size and increase of the corrugation lead to the increase of the real part of the @xmath8 , which is caused by the tighter light localization in the @xmath9 direction .
the numerical dependence of the imaginary part of the refractive index , which corresponds to the modal losses , is more complicated .
it exhibits minima at @xmath10 nm and 650 nm , and is almost independent on the corrugation size for @xmath5 up to 85 nm .
based on the presented data , we have chosen the case @xmath10 nm and @xmath11 nm for further study .
it is characterized by relatively low loss as well as relatively low real part of @xmath12 , which means that harmonics with higher order can be quasi - phase - matched .
let us now present the numerical model which was used to simulate the propagation of pump pulse propagation and the harmonic pulse .
the numerical treatment of the problem under consideration includes three components : microscopic description of the hhg process , propagation of the pump pulse , and the propagation of the generated harmonics .
the harmonic generation was modeled using the strong - field formalism developed by lewenstein _
@xcite , using the following expression for the dipole moment : @xmath13 where @xmath14 is the pump frequency , @xmath15 is the dipole moment of the outer - shell orbital , @xmath16 is the saddle - point canonical momentum , @xmath17 is the vector - potential of the enhanced electric field @xmath18 .
the enhancement @xmath19 is equal to the maximum along @xmath9 direction of the plasmonic enhancement inside the waveguide at the position @xmath7 .
however , one should also take into account that only a part of the transverse cross - section is characterized by this high enhancement .
this is accounted for by mulitplying the harmonic dipole moment by the pre - factor @xmath20 , which is @xmath9-direction full - width half - maximum of the enhancement divided by the waveguide width @xmath4 .
the classical action is defined by @xmath21 ^ 2/m_e)dt'$ ] with @xmath22 being the ionization potential .
no approximation of zero velocity of the ionized electron was used , which allowed for the accurate description of the spectra also well below the harmonic threshold .
analytic expressions for dipole transition moments of the hydrogen - like atoms @xcite were utilized . for the propagation of the pump pulse
we have numerically solved a ( 1 + 1)d unidirectional propagation equation : @xmath23 where @xmath24 is the wavenumber and @xmath25 is the nonlinear polarization , determined by @xmath26 where @xmath27 is the third - order susceptibility .
we note that the third - order susceptibility has a very minor effect on the propagation .
the plasma contribution to the refractive index was neglected , since the corresponding change of the refractive index is significantly lower than the waveguide contribution to the effective refractive index .
finally , the propagation of the high - order harmonics is governed by the ( 1 + 1)d propagation equation , which incorporates a high - harmonic polarization using harmonic source dipole moment @xmath28 calculated as described above , as well as harmonic reabsorption by the argon gas . for further details of the numerical treatment
see e.g. @xcite .
in fig . [ spectrum ] , the output harmonic spectrum is shown for the input pulse and waveguide parameters listed in the caption .
the period of the modulation of @xmath29 @xmath30 corresponds to the first - order quasi - phase matching length of the 15th harmonic , given by @xmath31 $ ] .
therefore one can observe a pronounced maximum in the spectrum for the 15th harmonic , with contrast of roughly two orders of magnitude .
note that the predicted cutoff at 23rd harmonic corresponds to the enhanced intensity of 40 tw/@xmath32 .
this intensity is in agreement with the pump intensity of 3.5 tw/@xmath32 and the maximum enhancement factor of around 10 . to confirm the quasi - phase - matching in the considered case , in fig .
[ evol ] we present the evolution of the harmonic energy with propagation for different harmonic numbers .
one can see that for the quasi - phase - matched 15th harmonic , all the contributions from the different corrugations add up constructively .
this leads to a steady growth of the energy proportional to the square of the propagation distance , which overgoes into a decay after the waveguide loss reduces the pump intensity and harmonic reabsorption comes into play . on the other hand ,
the evolution of the energy of the 11th and the 19th harmonics shows incoherent addition of contributions from different corrugations , since the corrugation period does not correspond to the quasi - phase - matching period for these harmonics .
we have predicted that quasi - phase - matching in nanometer - scale corrugated hollow waveguides can be used to enhance the efficiency of high harmonic generation .
the rigorous analysis of the modal properties of such waveguides was utilized together with the ( 1 + 1)d model for the pump beam and harmonics propagation .
the optimization of the corrugated waveguide geometry is performed .
german research council ( hu 1593/2 - 1 ) .
a. paul , r. a. bartels , r. tobey , h. green , s. weiman , i. p. christov , m. m. murnane , h. c. kapteyn , and s. backus , `` quasi - phase - matched generation of coherent extreme - ultraviolet light '' , nature * 421 * , 51 - 54 ( 2003 ) .
g. sansone , f. kelkensberg , j. f. prez - torres , f. morales , m. f. kling , w. siu , o. ghafur , p. johnsson , m. swoboda , e. benedetti , f. ferrari , f. lpine , j. l. sanz - vicario , s. zherebtsov , i. znakovskaya , a. lhuillier , m. yu .
ivanov , m. nisoli , f. martn , and m. j. j. vrakking , `` electron localization following attosecond molecular photoionization , '' nature * 465 * , 763 - 766 ( 2010 ) .
i .- y . park , s. kim , j. choi , d .- h .
lee , y .- jin kim , m. f. kling , m. i. stockman , and seung - woo kim , `` plasmonic generation of ultrashort extreme - ultraviolet light pulses , '' nature photonics * 5 * , 677 - 681 ( 2011 ) .
m. f. ciappina , ss s. acimovic , t. shaaran , j. biegert , r. quidant , and m. lewenstein , `` enhancement of high harmonic generation by confining electron motion in plasmonic nanostrutures , '' opt .
express * 20 * , 26261 - 26267 ( 2012 ) .
t. shaaran , m. f. ciappina , and m. lewenstein , t. shaaran , m. f. ciappina , and m. lewenstein , `` quantum - orbit analysis of high - order harmonic enhancement using high - order harmonic generation : the role of the field inhomogeneity , '' j. of mod . opt . * 59 * , 1634 ( 2012 ) .
t. shaaran , m. f. ciappina , r. guichard , j. a. prez - hernndez , l. roso , m. arnold , t. siegel , a. zar , and m. lewenstein , `` high - order - harmonic generation by enhanced plasmonic near - fields in metal nanoparticles , '' phys .
a * 87 * , 041402(r ) ( 2013 ) .
m. f. ciappina , t. shaaran , r. guichard , j. a. prez - hernndez , l roso , m arnold , t siegel , a zair , and m. lewenstein , `` high energy photoelectron emission from gases using plasmonic enhanced near - fields , '' laser phys . lett . * 10 * , 105302 ( 2013 ) .
j. a. prez - hernndez , m. f. ciappina , m. lewenstein , l. roso , and a. zair , `` beyond carbon k - edge harmonic emission using a spatial and temporal synthesized laser field , '' phys .
lett . * 110 * , 053001 ( 2013 ) .
n. pfullmann , c. waltermann , m. noack , s. rausch , t. nagy , c. reinhardt , m. kovacev , v. knittel , r. bratschitsch , d. akemaier , a. htten , a. leitenstorfer , and u. morgner , `` bow - tie nano - antenna assisted generation of extreme ultraviolet radiation , '' new j. of phys . * 15 * , 093027 ( 2013 ) .
stebbings , f. smann , y.y .
yang , a. scrinzi , m. durach , a. rusina , m.i .
stockman , and m.f . kling , `` generation of isolated attosecond extreme ultraviolet pulses employing nanoplasmonic field enhancement : optimization of coupled ellipsoids , '' new journal of physics * 13 * , 073010 ( 2011 ) . y .- y .
yang , a. scrinzi , a. husakou , q .- g .
li , s. l. stebbings , f. suessmann , h .- j .
yu , s. kim , e. rhl , j. herrmann , x .- c .
lin , and m. f. kling , `` high - harmonic and single attosecond pulse generation using plasmonic field enhancement in ordered arrays of gold nanoparticles with chirped laser pulses , '' opt .
express * 21 * , 2195 ( 2013 ) .
a. husakou , f. kelkensberg , j. herrmann , and m. j. j. vrakking , `` polarization gating and circularly - polarized high harmonic generation using plasmonic enhancement in metal nanostructures , '' opt .
express * 19 * , 25346 ( 2011 ) .
m. sivis , m. duwe , b. abel , and c. ropers , `` plasmonic enhancement of high harmonic generation revisited : predominance of atomic line emission , '' xviiith international conference on ultrafast phenomena , epj web of conferences * 41 * , artice 01014 ( 2013 ) .
x. zhang , a. l. lytle , t. popmintchev , x. zhou , h. c. kapteyn , m. m. murnane , and oren cohen , `` quasi - phase - matching and quantum - path control of high - harmonic generation using counterpropagating light , '' nat .
phys . * 3 * , 270 - 275 ( 2007 ) .
a. willner , f. tavella , m. yeung , t. dzelzainis , c. kamperidis , m. bakarezos , d. adams , m. schulz , r. riedel , m. c. hoffmann , w. hu , j. rossbach , m. drescher , n. a. papadogiannis , m. tatarakis , b. dromey , and m. zepf , `` coherent control of high harmonic generation via dual - gas multijet arrays , '' phys .
lett . * 107 * , 175002 ( 2011 ) .
x. zhang , z. sun , y. wang , g. chen , z. wang , r. li , z. zeng , and z. xu , `` high - order harmonic and attosecond pulse generation for a few - cycle laser pulse in modulated hollow fibres , '' j. phys .
b : at . mol .
. phys . * 40 * , 2917 - 2925 ( 2007 ) . | the high harmonic generation in periodically corrugated submicrometer waveguides is studied numerically .
plasmonic field enhancement in the vicinity of the corrugations allows to use low pump intensities . simultaneously ,
periodic placement of the corrugations leads to quasi - phase - matching and corresponding increase of the high harmonic efficiency .
the optimization of waveguide geometry is performed , and the resulting spectra are analyzed by the means of ( 1 + 1)d numerical model . |
the discovery of accelerating expansion of the universe by the observations of type ia supernovae in 1998 @xcite motivated the search for dark energy and modified gravity .
the gravitational force decays at a scale larger than @xmath2 if graviton has a mass @xmath3 , so massive gravity may be used to explain the cosmic acceleration . naively , the mass of graviton should be very small so that gravity is still approximately a long range force , therefore , it is expected that the mass of graviton is about hubble scale @xmath4 .
dvali , gabadadze and porrati proposed that general relativity is modified at the cosmological scale @xcite . in this model , there are a continuous tower of massive gravitons .
the first attempt of a theory of gravity with massive graviton was made by fierz and pauli @xcite .
however , the linear theory with the fierz - pauli mass is in contradiction with solar system tests @xcite .
recently , de rham , gabadadze and tolley introduced a nonlinear theory of massive gravity @xcite that is free from bouldware - deser ghost @xcite .
the cosmological solutions for massive gravity were sought in @xcite . the first homogenous and isotropic solution was found for spatially open universe in @xcite and the massive graviton term is equivalent to a cosmological constant .
the same solutions were then found for spatially open and closed universe in @xcite .
in addition to the equivalent cosmological constant solution , more general cosmological solutions were also found in @xcite by taking the de sitter metric as the reference metric .
we follow the approach in @xcite and proposed a new approach to find more general cosmological solutions .
the theory of massive gravity is base on the following action @xcite @xmath5 where @xmath6 is the mass of the graviton , the mass term @xmath7
@xmath8 ^ 2-[\mathcal{k}^2],\\ \label{massv3 } \mathcal{u}_3=[\mathcal{k}]^3 - 3[\mathcal{k}][\mathcal{k}^2]+2[\mathcal{k}^3],\\ \label{massv4 } \mathcal{u}_4=[\mathcal{k}]^4 - 6[\mathcal{k}]^2[\mathcal{k}^2]+8[\mathcal{k}^3][\mathcal{k}]-6[\mathcal{k}^4],\end{gathered}\ ] ] and @xmath9 the tensor @xmath0 is defined by four stckelberg fields @xmath10 as @xmath11 the reference metric @xmath12 is usually taken as the minkowski one . the cosmological solution was first found in @xcite for an open universe and the mass term behaves like an effective cosmological constant with @xmath13 the same solution was then found in @xcite for a flat universe by considering an arbitrary spatially isotropic metric and a spherically symmetric ansatz for the stckelberg fields . in @xcite
, the authors obtained the solution by assuming isotropic forms for both the physical and reference metrics . for a general case with positive , negative and zero curvature , kobayashi _ et al .
_ found the same solution with @xmath14 for the particular choices of parameters @xmath15 and @xmath16 @xcite , @xmath17 in @xcite , the authors assumed that the spacetime metric takes the form @xmath18 with the spatial metric @xmath19 and generalized the reference metric from minkowski metric to de sitter metric , @xmath20 where the stckelberg fields are assumed to be @xmath21 , @xmath22 , so that the tensor @xmath0 takes the homogeneous and isotropic form , @xmath23\gamma_{ij}\},\ ] ] and the functions @xmath24 are @xmath25 they then found three branches of cosmological solutions , two of them correspond to the effective cosmological constant ( [ efflambda ] ) and exist for spatially flat , open and closed cases .
they also found a new solution @xcite @xmath26}{df}=\frac{\dot a}{n}.\ ] ] for the flat case , @xmath27 , substituting the de sitter function @xmath28=e^{h_c f(t)}$ ] into equation ( [ masssol1 ] ) , we obtain the effective energy density and pressure for the massive graviton , @xmath29,\\ % \begin{split } % \rho_g=&-m_g^2m_{pl}^2\left(1-\frac{h}{h_c}\right)\left[3(\alpha_3 + 4\alpha_4)\frac{h^2}{h_c^2}\right.\\ % & \left.-3(1 + 5\alpha_3 + 8\alpha_4)\frac{h}{h_c}+6 + 12\alpha_3 + 12\alpha_4\right ] , % \end{split}\\ \label{effp1 } \begin{split } p_g=&m_g^2m_{pl}^2\left[-3(\alpha_3 + 4\alpha_4)\frac{h^3}{h^3_c}\left(1+\frac{\dot h}{h^2}\right ) + 6 + 12\alpha_3 + 12\alpha_4\right.\\ & \left.-(3 + 9\alpha_3 +
12\alpha_4)\frac{h}{h_c}\left(3+\frac{\dot h}{h^2}\right ) + ( 1 + 6\alpha_3 + 12\alpha_4)\frac{h^2}{h^2_c}\left(3 + 2\frac{\dot h}{h^2}\right)\right ] .
\end{split}\end{gathered}\ ] ] so when @xmath30 , @xmath31 .
the friedmann equations are @xmath32 for the flat case , substituting equations ( [ effrho1 ] ) and ( [ effp1 ] ) into friedmann equations ( [ frweq1 ] ) and ( [ frweq2 ] ) , we get @xcite @xmath33\\ \frac{m_g^2}{h_0 ^ 2}\frac{h(z)}{h_c}\left[-(\alpha_3 + 4\alpha_4)\frac{h^2(z)}{h_c^2 } + ( 1 + 6\alpha_3 + 12\alpha_4)\frac{h(z)}{h_c}-3(1 + 3\alpha_3 + 4\alpha_4)\right]\\ = -e^2(z)+\omega_m(1+z)^{3(1+w_m)}-2\frac{m_g^2}{h_0 ^ 2}(1 + 2\alpha_3 + 2\alpha_4 ) . \end{split}\ ] ] @xmath34\right\}=3\omega_{m}(1+w_m)(1+z)^{3(1+w_m ) } , \end{split}\ ] ] where @xmath35 .
the effective equation of state @xmath36 for the massive graviton is @xmath37}.\ ] ] since @xmath31 when @xmath38 , so if @xmath39 , we find that @xmath40 which is inconsistent with current observations , therefore @xmath41 . if @xmath42 , then we can not recover the standard cosmology @xmath43 in the past unless we fine tune the value of @xmath44 to be very small . from equation ( [ frweq4 ] ) , we see that the standard cosmology is recovered when @xmath45 and @xmath46 . at very early times , @xmath47 ,
the universe evolves according to @xmath48 .
if it was radiation dominated in the very early times , then the universe evolves faster as @xmath49 instead of @xmath50 . for the special case @xmath51 , friedmann equation
is simplified to @xmath52 at @xmath53 , @xmath54 , we get @xmath55 as discussed above , @xmath56 , so @xmath57 must be negative for this special case .
the sign of @xmath57 is not important because we can always redefine the potential term so that the graviton mass is positive . without loss of generality ,
we assume that @xmath58 , and @xmath59 with @xmath60 . for the special case @xmath51 , equation ( [ conseq1 ] ) gives @xmath61 in this case , we have only two free parameters @xmath62 and @xmath63 , and equation ( [ acceq4 ] ) gives @xmath64 if @xmath65 , then @xmath66 , and the model becomes the @xmath1cdm model .
this is shown in fig .
[ fig1 ] for @xmath67 and @xmath68 . fitting this model to the three year supernova legacy survey ( snls3 ) sample of 472 sne ia data with systematic errors
@xcite , and the baryon acoustic oscillation ( bao ) measurements from the 6dfgs @xcite , the distribution of galaxies @xcite and the wigglez dark energy survey @xcite , we find the best fit values are @xmath69 , @xmath70 , @xmath71 with @xmath72 . as discussed above , when @xmath65 , the model becomes the @xmath1cdm model which is independent of the value of @xmath63 , so @xmath63 can not be constrained from above by the observational data , and we get @xmath73 at @xmath74 confidence level .
the more detailed observational constraints are done in @xcite . and the effective equation of state @xmath75 of massive graviton .
the blue lines are for the model with de sitter metric as the reference metric and the black lines are for the model taking @xmath76 as power law form .
, width=384 ]
in summary , starting with the homogeneous and isotropic metric ( [ frwmetric ] ) and tensor ( [ stuckphi1 ] ) , we obtain the friedmann equations ( [ frweq1 ] ) and ( [ frweq2 ] ) with the effective energy density and pressure for the massive graviton , @xmath77-a)\{6(1 + 2\alpha_3 + 2\alpha_4)a^2 -(3 + 15\alpha_3 + 24\alpha_4)a b_k[f]+3(\alpha_3 + 4\alpha_4)b_k[f]^2\},\\ \label{effp2 } \begin{split } p_g=&\frac{m_g^2m^2_{pl}}{a^2}\{[6 + 12\alpha_3 + 12\alpha_4-(3 + 9\alpha_3 + 12\alpha_4){\dot f}]a^2 -2[3 + 9\alpha_3 + 12\alpha_4\\ & -(1 + 6\alpha_3 + 12\alpha_4){\dot f}]a b_k[f ] + [ 1 + 6\alpha_3 + 12\alpha_4 - 3(\alpha_3 + 4\alpha_4){\dot f}]b_k^2[f]\ } , \end{split}\end{gathered}\ ] ] and the equation of motion for the function @xmath78 which leads to the three branches of solutions @xmath79=\frac{(1 + 6\alpha_3 + 12\alpha_4\pm\sqrt{1 + 3\alpha_3 + 9\alpha_3 ^ 2 - 12\alpha_4})}{3(\alpha_3 + 4\alpha_4)}a(t),\\ \label{masssol5 } \frac{db_k[f]}{df}=\frac{\dot a}{n}.\end{gathered}\ ] ] when we take the solution ( [ masssol4 ] ) , we get the effective cosmological constant solution ( [ efflambda ] ) independent of the choice of spatial curvature @xmath80 . for @xmath81 ,
the solution was obtained in @xcite by taking the reference metric @xmath12 as minkowski and the same homogeneous and isotropic tensor @xmath0 ( [ stuckphi1 ] ) with @xmath82=f(t)$ ] . the same solution ( [ efflambda ] ) was obtained in @xcite for all values of @xmath80 for the particular parameters ( [ 1parm ] ) , but the tensor @xmath0 is not homogeneous and isotropic . for the flat case @xmath27 , gratia , hu and wyman obtained the same cosmological constant solution ( [ efflambda ] ) by using another inhomogeneous and anisotropic tensor @xmath0 @xcite .
motohashi and suyama obtained the cosmological constant solution for the @xmath27 case with isotropic forms for both the physical and reference metrics @xcite .
however , the cosmological constant solution ( [ efflambda ] ) is just the consequence of the equation of motion of the function @xmath78 once we assumed the homogeneous and isotropic form for the metric ( [ frwmetric ] ) and the tensor @xmath0 ( [ stuckphi1 ] ) .
since the cosmological constant solution ( [ efflambda ] ) was obtained by different methods for different sepcial cases , this suggests that this solution exists for the general case . the method proposed in @xcite not only gives the solution for the general case , but also gives additional new dynamic solutions .
this suggests that the solution ( [ masssol5 ] ) should be quite general even without the assumption of the reference metric @xmath12 as de sitter .
therefore , we propose that more solutions can be found with equations ( [ frwmetric ] ) and ( [ stuckphi1 ] ) by assuming more general form of @xmath82 $ ] .
note that the specific form of @xmath0 in equation ( [ stuckphi1 ] ) may be obtained from minkowski , de sitter , or isotropic reference metrics .
follow the above argument , we assume a power law form @xmath82=(h_c f(t))^{\gamma/(\gamma-1)}$ ] with @xmath83 , then the solution to equation ( [ masssol5 ] ) is @xmath84=\left(\frac{a(\gamma-1)h}{\gamma h_c}\right)^\gamma,\ ] ] and the effective energy density becomes @xmath85 .
\end{split}\ ] ] the effective pressure of massive graviton is @xmath86\right.\\ & + ( 1 + 6\alpha_3 + 12\alpha_4)\left(\frac{\gamma-1}{\gamma}\right)^{2\gamma } a^{2\gamma-2}\left(\frac{h}{h_c}\right)^{2\gamma } \left[1 + 2\gamma\left(1+\frac{\dot h}{h^2}\right)\right]\\ & \left.-3\gamma(\alpha_3 + 4\alpha_4)\left(\frac{\gamma-1}{\gamma}\right)^{3\gamma } a^{3\gamma-3}\left(\frac{h}{h_c}\right)^{3\gamma } \left(1+\frac{\dot h}{h^2}\right)\right\}. \end{split}\ ] ] again when the hubble parameter is in the range @xmath87 , the standard cosmology is recovered . to have a long history of matter and radiation domination , we require @xmath88 . for the special case @xmath51 ,
friedmann equations are @xmath89 @xmath90 with @xmath91\left[\left(\frac{\gamma-1}{\gamma}\right)^\gamma -2\beta_2^\gamma\right]}.\ ] ] in this case , we have three free parameters @xmath62 , @xmath92 and @xmath63 . again if @xmath65 , @xmath66 , the model is equivalent to the @xmath1cdm model and is independent of the values of @xmath63 and @xmath92 .
this is shown in fig .
[ fig1 ] for @xmath67 and @xmath68 .
in conclusion , more general cosmological solutions which are consistent with the observational data can be found by taking homogeneous and isotropic form for both the metric @xmath93 and the tensor @xmath0 without specifying the form of reference metric , even though the tensor @xmath0 may be obtained with minkowski , de sitter or isotropic reference metrics .
in addition to the cosmological constant solution , more richer dynamics can be found in these solutions .
the mass of graviton is in the order of @xmath94 .
this work was partially supported by the national basic science program ( project 973 ) of china under grant no .
2010cb833004 , the national nature science foundation of china under grant nos .
10935013 and 11175270 , and the fundamental research funds for the central universities . 30 natexlab#1#1[2]#2 , et al . ,
, et al . , ( ) . , , , ( ) . , , ( ) . , , ( ) .
, , , ( ) . , , ( ) .
, et al . , ( ) . , , , ( ) . , , , ( ) .
, , , ( ) . , , , ( ) .
, , , , ( ) . , ( ) . , , , ( ) . ,
( ) , arxiv : 1206.3852 . , , ( ) . , , , , , ( ) .
, , , , , ( ) . , , , , arxiv : 1209.0457 . , , , ( ) .
, , , , arxiv : 1210.3627 .
, et al . , ( ) .
, et al . , ( ) . , et al . , ( ) . , et al . , ( ) . , arxiv : 1210.5396 | we argue that more cosmological solutions in massive gravity can be obtained if the metric tensor and the tensor @xmath0 defined by stckelberg fields take the homogeneous and isotropic form .
the standard cosmology with matter and radiation dominations in the past can be recovered and @xmath1cdm model is easily obtained .
the dynamical evolution of the universe is modified at very early times . |
clusters of galaxies are powerful laboratories in investigating the formation and evolution of galaxies .
rich clusters are dominated by early - type ( e+s0 ) galaxies .
a detailed study of the properties of these galaxies provides beneficial insights in their formation and evolution .
in particular , the paradigm of hierarchical galaxy formation can be critically tested . in the local universe
numerous investigations revealed that e+s0 form a very homogeneous galaxy population ( e.g. , @xcite ; @xcite ) .
their structural parameters ( effective surface brightness @xmath3 and the size as described by the effective radius @xmath4 ) and kinematics ( velocity dispersion @xmath5 ) represent a tight correlation in three dimensional parameter space , the fundamental plane ( @xcite ) .
furthermore , they show a small scatter in their relations of colours ( e.g. , mg(@xmath6 ) ) , @xmath7 ratios and absorption line indices with velocity dispersion ( e.g. , mg@xmath5 ) .
however , the question arises whether e+s0s are truly one single family or rather a diverse group with different formation and evolutionary processes .
therefore , one aim of this work is to explore if there are differences between the properties of elliptical and s0 galaxies .
previous spectroscopic studies were limited to a small number of the more luminous galaxies . to overcome bias and selection problems of small samples , we focus in this study of the clusters abell 2218 ( @xcite ) and abell 2390 on a large number of objects ( @xmath8 ) , spanning a wide range in luminosity , in case of a2390 @xmath9 , and a wide field - of - view ( fov ) of @xmath10 ( @xmath11 mpc@xmath12 ) .
we adopt a cosmology for a flat universe with @xmath13 , @xmath14 and @xmath15kms@xmath16mpc@xmath16 .
the objects were selected on the basis of ground - based gunn @xmath17-band images and a combination of defined colour regions using a similar selection procedure as described in ziegler et al .
( 2001 ) . in particular , in the selection procedure for the mask design special care was taken that galaxies cover the whole fov .
additional imaging data obtained with the 200 inch hale telescope is available in @xmath18 and @xmath19 and _ hst _ wfpc2 observations in the f555w and f814w filter .
during two observing runs ( 09/1999 and 07/2000 ) we observed three masks using the mosca spectrograph at the calar alto 3.5 m telescope with total exposure times between @xmath208 and @xmath2012 hours each . in total , we obtained 63 high @xmath21 spectra of 50 different early - type galaxies , out of which 15 are situated within the _ hst _ field .
only one object was revealed to be a background galaxy , which shows that our sample selection was highly efficient .
the instrumental resolution in the spectral regime of @xmath22 and @xmath23 was 5.5 fwhm ( @xmath24 km s@xmath16 ) .
the average @xmath21 is @xmath25 .
the reduction of the spectra was carried out using standard reduction techniques and velocity dispersions were calculated with the fcq method ( @xcite ) .
structural parameters were determined by fitting the surface brightness profile with either an @xmath26 or a combination of an @xmath26 plus exponential law profile ( @xcite ) .
further details concerning the reduction and analysis are outlined in fritz et al .
in fig.[fig : fj ] the faber - jackson ( fj ) relation for the complete sample of 94 e+s0 galaxies is shown . as a local reference we use the coma sample of @xcite ( jfk95 ) .
overall , the luminosity evolution is small ( @xmath28 mag ) . the fundamental plane ( fp ) in rest frame gunn @xmath29 is shown in fig.[fig : fp ] .
the sub - sample with _ hst _ structural parameters comprises 34 e+s0 galaxies , splitted into 17 ellipticals , 2 e / s0 , 8 s0 , 3 sb0/a , 3 sa and 1 sab galaxy that enter the fp . with this large sample ,
possible differences of the galaxies properties can be explored for various sub - populations .
both clusters have a similar behaviour within and along the fp , with no hint for a slope change .
for the 34 cluster e+s0 galaxies we find a combined evolution of @xmath1 mag .
lenticular galaxies ( e / s0 , s0 , sb0/a , sa , sab ) have a zeropoint offset of @xmath30 mag with respect to coma and seem to inhabit a certain band within the fp ( fig.[fig : fp ] , right ) .
ellipticals exhibit a similar mild evolution of @xmath31 .
both sub - samples have a small scatter ( @xmath200.10 mag ) .
we construct a spectroscopic sample of 48 e+s0 in the rich cluster abell 2218 ( @xmath32 ) and 50 e+s0 in abell 2390 ( @xmath33 ) .
since both cluster samples are very similar , we combine them and analyse a total number of 98 e+s0 at @xmath0 . for a subsample of 34 galaxies
we explore the evolution of the fp and study their @xmath7 ratios . on average , all 34 cluster
e+s0 galaxies show a mild evolution of @xmath34 .
s0 galaxies show lower @xmath7 ratios of @xmath35 with respect to the local reference . for the ellipticals
we again deduce a moderate evolution of @xmath36 .
elliptical and s0 galaxies seem to comprise similar sub - populations with no significant differences .
an analysis based on the @xmath7 ratios revealed a mean formation redshift of @xmath37 , which is consistent with stellar population models of passive evolution .
this project is a collaboration with r. g. bower and i. smail ( durham / uk ) and r. l. davies ( oxford / uk ) who have contributed to these results .
we thank the calar alto staff for efficient observational support .
af and blz acknowledge financial support by the volkswagen foundation ( i/76520 ) and the dfg ( zi663/1 - 1 , zi663/2 - 1 ) . | we present a spectroscopic sample of 48 early - type galaxies in the rich cluster abell 2218 and 50 galaxies in abell 2390 .
since both samples are very similar , we combine them and investigate a total number of 98 early - type galaxies at @xmath0 .
a subsample of 34 galaxies with _ hst _ structural properties is used to construct the fundamental plane .
elliptical and s0 galaxies show a zeropoint offset of @xmath1 mag with respect to the local coma fp . both sub - samples , ellipticals and lenticulars , exhibit a similar , mild evolution and small scatter .
the moderate amount of luminosity evolution is consistent with stellar population models of passive evolution , if @xmath2 is assumed . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.