article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
relativistic flow problems are important in many astrophysical phenomena including gamma - ray burst ( grb ) , active galactic nuclei ( agn ) , as well as microquasar and pulsar wind nebulae , among others .
apparent superluminal motion is observed in many jets of extragalactic radio sources associated with agn . according to the currently accepted standard model ,
this implies the jet flow velocities as large as @xmath0 of the speed of light @xcite .
similar phenomena are also seen in microquasars such as grs 1915 + 105 @xcite and gro j1655 - 40 @xcite thus by similar arguments relativistic flows are thought to play a role . in the case of grb ,
the observed non - thermal spectrum implies that the source must be optically thin , which can be used to put a limit on the minimum lorentz factor within those bursts @xcite .
this argument shows that the source of grb must be in highly relativistic motion .
this conclusion is further confirmed by the rapidly increasing number of grb afterglow observations by the swift satellite @xcite . to understand physical processes in those phenomena quantitatively , high resolution multi - dimensional simulations are crucial .
jim wilson and collaborators pioneered the numerical solution of relativistic hydrodynamics equations @xcite .
starting with these earliest papers this has typically been done in the context of general relativistic problems such as accretion onto black holes and supernovae explosions .
the problem was recognized to be difficult to solve when the lorentz factor becomes large @xcite and a solution with an implicit adaptive scheme was demonstrated in one dimension .
unfortunately , this approach is not generalizable to multi - dimensions .
however , in the past two decades accurate solvers based on godunov s scheme have been designed that have adopted shock - capturing schemes for newtonian fluid to the relativistic fluid equations in conservation form ( for a review see * ? ? ?
such schemes , called high - resolution shock - capturing ( hrsc ) methods , have been proven to be very useful in capturing strong discontinuities with a few numerical zones without serious numerical oscillations .
we will discuss a number of them in section 3 .
studies involving astrophysical fluid dynamics in general are benefiting tremendously from using spatial and temporal adaptive techniques
. smoothed particle hydrodynamics @xcite is a classic example by being a lagrangian method .
increasingly also variants of @xcite s structured adaptive mesh technique are being implemented .
this is also true in relativistic hydrodynamics @xcite where certainly the work of @xcite showed that a serial amr code could solve problems even highly efficient parallel fixed grid codes would have difficulty with . in this paper
we discuss our implementation of different hydrodynamics solvers with various reconstruction schemes as well as different time integrators on top of the _ enzo _ framework previously developed for cosmology @xcite .
this new code we call _ r_enzo _ _ is adaptive in time and space and is a dynamically load balanced parallel using the standard message passing interface . in the following we
briefly summarize the equations being solved before we give details on the different solvers we have implemented .
section 4 discusses the adaptive mesh refinement strategy and implementation .
we then move on to describe various test problems for relevant combinations of solvers , reconstruction schemes in one , two and three dimensions , with and without amr .
section 6 presents an application of our code to three - dimensional relativistic and supersonic jet propagation problem .
section 7 discusses 3d grb jet simulation .
we summarize our main conclusions in section 8 .
the basic equations of special relativistic hydrodynamics ( srhd ) are conservation of rest mass and energy - momentum : @xmath1 and @xmath2 where @xmath3 is the rest mass density measured in the fluid frame , @xmath4 is the fluid four - velocity ( assuming the speed of light @xmath5 ) , @xmath6 is the lorentz factor , @xmath7 is the coordinate three - velocity , @xmath8 is the energy - momentum tensor of the fluid and semicolon denotes covariant derivative . for a perfect fluid
the energy - momentum tensor is @xmath9 where @xmath10 is the relativistic specific enthalpy , @xmath11 is the specific internal energy , @xmath12 is the pressure and @xmath13 is the spacetime metric .
srhd equations can be written in the form of conservation laws @xmath14 where the conserved variable @xmath15 is given by @xmath16 and the fluxes are given by @xmath17 @xcite has shown that system ( [ 2.4 ] ) is hyperbolic for causal eos , i.e. , those satisfying @xmath18 where the local sound speed @xmath19 is defined as @xmath20.\label{cs}\ ] ] the eigenvalues and left and right eigenvectors of the characteristic matrix @xmath21 , which are used in some of our numerical schemes , are given by @xcite .
the conserved variables @xmath15 are related to the primitive variables by @xmath22 where @xmath23 .
the system ( [ 2.4 ] ) are closed by an equation of state ( eos ) given by @xmath24 . for an ideal gas ,
the eos is , @xmath25 where @xmath26 is the adiabatic index .
we use method of lines to discretize the system ( [ 2.4 ] ) spatially , @xmath27 where @xmath28 refers to the discrete cell index in @xmath29 directions , respectively .
@xmath30 , @xmath31 and @xmath32 are the fluxes at the cell interface . as discussed by @xcite ,
if using a high order scheme to reconstruct flux spatially , one must also use the appropriate multi - level total variation diminishing ( tvd ) runge - kutta schemes to integrate the ode system ( [ ode ] ) .
thus we implemented the second and third order tvd runge - kutta schemes coupled with amr .
the second order tvd runge - kutta scheme reads , @xmath33 and the third order tvd runge - kutta scheme reads , @xmath34 where @xmath35 is the final value after advancing one time step from @xmath36 . for an explicit time
integration scheme , the time step is constrained by the courant - friedrichs - lewy ( cfl ) condition .
the time step is determined as @xmath37 where @xmath38 is a parameter called cfl number and @xmath39 is the local largest speed of propagation of characteristics in the direction @xmath40 whose explicit expression can be found in @xcite . generally speaking , there are two classes of spatially reconstruction schemes ( see e.g. leveque 2002 ) .
one is reconstructing the unknown variables at the cell interfaces and then use exact or approximate riemann solver to compute the fluxes .
another is direct flux reconstruction , in which we reconstruct the flux directly from fluxes at the cell centers . to explore the coupling of different schemes with amr as well as exploring
which method is most suitable for a specific astrophysical problem , we implement several different schemes in both classes . to reconstruct unknown variables , we have implemented piecewise linear method ( plm , van leer 1979 ) , piecewise parabolic method ( ppm , colella & woodward 1984 , mart & mller 1996 ) , the third - order convex essentially non - oscillatory scheme ( ceno , liu & osher 1998 ) .
these are used to reconstruct the primitive variables since reconstructing the conserved variables can produce unphysical values in srhd . furthermore , unphysical values of three - velocities may arise during the reconstruction especially for ultrarelativistic flows .
so we either use @xmath41 to do the reconstruction or we also reconstruct the lorentz factor and use it to renormalize the reconstructed three - velocity when they are unphysical . for direct flux reconstruction ,
we have implemented plm and the third and fifth order weno scheme of @xcite .
direct flux reconstruction using weno was first used to solve srhd problems by @xcite .
they showed that the fifth order weno scheme works well with the third order runge - kutta time integration . in our implementation
, we followed their description closely . for the plm and ceno schemes
, we used a generalized minmod slope limiter @xcite . for given @xmath42 , @xmath43 , where @xmath44 . for @xmath45
it reduces to the monotonized central - difference limiter of @xcite .
we found that this generalized minmod slope limiter behaves much better than a traditional minmod limiter @xcite especially for strong shear flows . in our calculation , @xmath46 is used by default .
for the ppm scheme , we used the parameters proposed by @xcite for all the test problems .
for weno , we used the parameters suggested in the original paper of @xcite . in the first class of reconstruction methods ,
given the reconstructed left and right primitive variables at interfaces , the flux across each interface is calculated by solving the riemann problem defined by those two states .
an exact riemann solver is quite expensive in srhd @xcite .
thus we have implemented several approximate riemann solvers including hll @xcite , hllc @xcite , local lax - friedrichs ( llf , kurganov & tadmor 2000 ) and the modified marquina flux @xcite .
the hllc scheme is an extension of the hll solver developed by @xcite for newtonian flow which was extended to two - dimensional relativistic flows by @xcite .
the improvement of hllc over hll is restoring the full wave structure by constructing the two approximate states inside the riemann fan .
the two states can be found by the rankine - hugoniot conditions between those two states and the reconstructed states . with this modification , hllc indeed behaves better than other riemann solvers in some 1d ( 5.1.7 ) and 2d ( 5.2 ) test problems .
but when we apply hllc to three dimensional jet simulation , we found that hllc suffers from the so called carbuncle " artifact well - known in the computational fluid dynamics literature @xcite .
we have used hllc to run many other two - dimensional test problems designed to detect the carbuncle artifact and confirmed this shortcoming .
we found that the hllc solver is unsuitable for many multi - dimensional problems .
the discussion of these problems will be presented elsewhere . in this work
, we will only apply hllc to two test problems showing that the hllc solver has less smearing at contact discontinuities than other schemes .
has compared the hll scheme , the llf scheme and the modified marquina flux formula using 1d and 2d test problems .
they found those three schemes give similar results for all their test problems . in our tests , we found similar results .
however , the modified marquina flux formula is not as stable as hll in problems with strong transverse flows and llf is more diffusive than hll .
so in the following discussion we will only show the results using hll in most of the tests if there is no difference among those three schemes . in the following discussion
, we will denote a specific hydro algorithm by x - y where x is the flux formula and y is the reconstruction scheme .
for example , f - weno5 denotes direct flux reconstruction using fifth order weno .
we used the third order runge - kutta method for all the tests in this work .
since primitive variables are needed in the reconstruction process , after every rk time step , we need to convert conserved variables to primitive variables . while conserved variables can be computed directly from primitive variables using eqs .
( [ d ] ) , ( [ sj ] ) and ( [ tau ] ) , the inverse operation is not straightforward .
one needs to solve a quartic equation for the ideal gas eos and a nonlinear equation for more complicated eos .
iteration methods are used even for ideal gas eos , because computing the solution of a quartic is expensive .
following @xcite , we have used a newton - raphson ( nr ) iteration to solve a nonlinear equation for pressure to recover primitive variables from conserved variables .
typically , the nr iteration needs only @xmath47 to @xmath48 steps to converge .
we have also implemented cylindrical and spherical coordinates following the description of zhang & macfadyen ( 2006 ) .
this affects three parts of the code .
firstly , the geometric factors are incorporated into the flux when updating the conserved variables .
secondly , there will be geometric source terms .
thirdly , the flux correction in amr ( 4.4 ) is modified by geometric factors .
structured adaptive mesh refinement ( amr ) was developed by @xcite and @xcite to achieve high spatial and temporal resolution in regions where fixed grid resolution is insufficient . in structured amr , a subgrid will be created in regions of its parent grid needing higher resolution .
the hierarchy of grids is a tree structure .
each grid is evolved as a separate initial boundary value problem , while the whole grid hierarchy is evolved recursively . _
r_enzo _ _ is built on top of the amr framework of _ enzo _ @xcite .
_ enzo _ s implementation of amr follows closely the berger & colella paper and has been shown to be very efficient for very high dynamic range cosmological simulations ( see e.g. abel et al .
the pseudocode of the main loop for the second order runge - kutta method reads , @xmath49 the rebuildhierarchy function called at the end of every time step is at the heart of amr .
its pseudocode as implemented originally in _ enzo _ reads , @xmath50 when a new subgrid is created , the initial values on that grid are obtained by interpolating spatially from its parent grid . in this case , we apply the conservative second order order interpolation routine provided by _ enzo _ to conserved variables . but in this process sometimes the interpolated values can violate the constraint @xmath51 . if this happens , we will then use first order method for that subgrid . before the first runge - kutta step for a grid at level @xmath52 ( in the following discussion
, we use the convention that top grid has level @xmath53 ) , we will need the boundary condition at time @xmath54 , which is derived by interpolating from its parent grid .
then at the later steps of runge - kutta scheme , one needs the boundary condition at time @xmath55 . since the variables of its parent grid has already been evolved to time @xmath56 , which is greater than time @xmath55
, we can obtain the boundary conditions at time @xmath55 for a grid at level @xmath52 by interpolating both temporally and spatially from its parent grid .
there are two exceptions to this procedure .
first , if a cell of fine grid abuts the box boundary , then we just use the specified boundary condition for that cell .
second , if a cell abuts another grid at the same level , we copy the value from that grid . because of the above mentioned problem for interpolating conserved variables , when interpolating boundary values , we apply the second order interpolation to primitive variables .
since for ultrarelativistic flows spatially interpolating three - velocity can lead to unphysical values , we also interpolate the lorentz factor and then use it to renormalize the interpolated three - velocity . in the test problems discussed in section [ sec : test ] , we mainly used two general purpose refinement criteria that have been widely used in amr code @xcite .
in the first one , we compute the slope @xmath57 where @xmath58 is typically density , pressure and velocities , @xmath11 is a small number typically taken to be @xmath59 .
when @xmath60 is larger than a minimum slope , typically @xmath61 , a cell will be flagged for refinement . in the second one ,
for every cell we compute @xmath62 which is the ratio of the second and first derivatives with a safety factor in the denominator .
unless otherwise stated , we use @xmath63 .
when @xmath64 is larger than a critical value @xmath65 , a cell will be flagged for refinement .
typically we use @xmath66 . to fully exploit amr ,
it is desirable to design more specific refinement criterions that are most efficient for a specific astrophysical problems .
when a cell is overlayed by a finer level grid , then the coarse grid value is just the conservative average of the fine grid values .
on the other hand , when a cell abuts a fine grid interface but is not itself covered by any fine grid , we will do flux correction for that cell , i.e. we will use the fine grid flux to replace the coarser grid flux in the interface abutting the fine grid.(see berger & colella 1989 for more detailed description of flux correction ) . for this purpose , note that the second order runge - kutta method can be rewritten as @xmath67 and the third order runge - kutta method can be rewritten as @xmath68 thus for example , when we do flux correction in the @xmath69-direction for interface @xmath70 , we will use @xmath71 and @xmath72 to correct the coarser grid conserved variables for the second and third order runge - kutta method , respectively . _
r_enzo _ _ uses the _ enzo _ parallel framework which uses dynamically load balancing using the message passing interface ( mpi ) library . at run time
, the code will move grids among processors according to the current load of every processor to achieve a balanced distribution of computational load among processors .
the computational load of a processor is defined as the total number of active cells on that processor and level .
relativistic riemann problems have analytical solutions @xcite , thus they are ideal for tesing srhd codes . in the following discussion , subscript @xmath73 and @xmath74 refer to the left and right initial states , respectively .
the initial discontinuity is always at @xmath75 . we will report the error between numerical solutions and analytical solutions using @xmath76 norm defined as @xmath77 where @xmath58 is the numerical solution
, @xmath78 is the analytical solution and @xmath79 is the cell width . this test and the following one are fairly standard and all modern srhd codes can match the analytical solution quite well ( see martt & mller 2003 for a summary of different codes performance on those two tests ) .
the initial left and right states for this problem are @xmath80 and @xmath81 , @xmath82 , @xmath83 .
the gas is assumed to be ideal with an adiabatic index @xmath84 .
the initial discontinuity gives rise to a transonic rarefaction wave propagating left , a shock wave propagating right and a contact discontinuity in between .
this problem is only mildly relativistic with a post - shock velocity @xmath85 and shock velocity 0.83 .
the results using four hydro solvers are shown in fig .
[ fig : riemann1 ]
. the cfl number used is @xmath86 .
the @xmath76 errors are shown in fig .
[ fig : l1_riemann1 ] , from which we can see the four schemes behave essentially identical for this problem .
the order of global convergence rate is about 1 for all four schemes , which is consistent with the fact that there are discontinuities in the problem .
the initial left and right states for this problem are @xmath87 and @xmath88 , @xmath89 , @xmath83 .
the gas is assumed to be ideal with an adiabatic index @xmath84 .
this test is more relativistic than the previous one . while the wave structure is the same , the thermodynamically relativistic initial left state gives rise to a relativistic shock propagating at a lorentz factor @xmath90 and a very thin dense shell behind the shock with width @xmath91 at @xmath92 .
the cfl number used is @xmath86 .
the results using four hydro algorithms are shown in fig .
[ fig : riemann2 ] .
the @xmath76 errors are shown in fig .
[ fig : l1_riemann2 ] .
we can see that for this problem ppm and weno have smaller @xmath76 errors than plm and ceno .
this is due to their ability to better resolve the thin shell .
the initial left and right states for this test are @xmath93 and @xmath94 , @xmath95 , @xmath96 .
the gas is assumed to be ideal with an adiabatic index @xmath97 .
this test mimics the interaction of a planar jet head with the ambient medium .
the decay of the initial discontinuity gives rise to a strong reverse shock propagating to the left , a forward shock propagating to the right and a contact discontinuity in between .
the results are shown in fig .
[ fig : riemann3 ] .
the cfl number is @xmath86 .
the @xmath76 errors are shown in fig .
[ fig : l1_riemann3 ] . as can be seen in figs .
[ fig : riemann3 ] & [ fig : l1_riemann3 ] , for this problem ppm and weno behave better than plm and ceno : there is almost no oscillation behind the reverse shock and they capture the contact discontinuity with fewer cells .
especially ppm captures the contact discontinuity with only @xmath98 cells and has the smallest @xmath76 error . for this and the following two problems
, we will consider non - zero transverse velocities in the initial states .
the initial state is identical to blast wave problem ii except the presence of transverse velocities .
those problems were first discussed analytically by @xcite . since
then various groups have shown that when transverse velocities are non - zero , in some cases those problems become very difficult to solve numerically unless very high spatial resolution is used @xcite .
in realistic astrophysical phenomena transverse velocities are usually very important ( see e.g. aloy & rezzolla 2006 ) , thus solving those problems accurately is of great importance . as an easy first case
, we will consider non - zero transverse velocity only in the low pressure region .
the initial left and right states are @xmath99 and @xmath100 .
the gas is assumed to be ideal with an adiabatic index @xmath101 .
the results are shown in fig . [
fig : transverse_c ] .
the cfl number is @xmath86 .
the @xmath76 errors are shown in fig .
[ fig : l1_bt1 ] .
we can see that all four hydro algorithms behaves similarly well , except that plm and ceno shows some small oscillations around the contact discontinuity .
next , we consider non - zero transverse velocity in the high pressure region . in this case , the problem becomes more difficult to solve numerically @xcite .
the initial left and right states for this problem are @xmath102 and @xmath103 .
the gas is assumed to be ideal with an adiabatic index @xmath84 .
the high pressure region is connected to the intermediate state by a rarefaction wave .
since the initial normal velocity in the high pressure region is zero , the slope of the adiabat increases rapidly with transverse velocity , thus a large initial transverse velocity will lead to a small intermediate pressure and a small mass flux .
the results using a uniform grid and two amr runs are shown in fig .
[ fig : transverse_d ] .
the hydro solver used for this figure is hll - plm .
the cfl number is @xmath104 .
the @xmath76 error are shown in fig .
[ fig : l1_transverse_d ] .
it can be seen that for the run with 400 uniform grid cells , the numerical solution is inadequate , as previously found by @xcite .
this is mainly due to the poor capture of the contact discontinuity .
we have tried to run this problem with various algorithms but only obtained accurate solutions by dramatically increasing the resolution .
cccc 400 & 400 & 400 + 800 & 448 & 421 + 1600 & 455 & 442 + 3200 & 470 & 468 + 6400 & 501 & 473 + 12800 & 518 & 476 + 25600 & 562 & 520 now we introduce transverse velocity in both region .
the initial left and right states for this problem are @xmath105 and @xmath106 .
the gas is assumed to be ideal with an adiabatic index @xmath101 .
this problem is more difficult than the previous one due to the formation of an extremely thin shell between the rarefaction wave tail and the contact discontinuity @xcite .
the results with a uniform grid and two amr runs are shown in fig .
[ fig : transverse_e ] .
the hydro solver used for this run is hll - plm .
the cfl number used is @xmath86 .
the @xmath76 errors are shown in fig .
[ fig : l1_transverse_e ] . table 1 shows the equivalent resolution and the actual number of cells used for this and the previous tests .
it can be seen for the highest resolution calculation our code uses about four hundred times less grid cells than the corresponding uniform grid calculation .
thus amr allows us to achieve very high resolution while significantly reducing the computational cost . for this test
we set up a one - dimensional riemann problem that mimics the interaction of jet with an overpressured cocoon .
the initial left and right states for this problem are @xmath107 and @xmath108 .
the gas is assumed to be ideal with @xmath109 .
those values mimic the conditions of the jet - cocoon boundary in model c2 of @xcite .
the result using four different riemann solver with plm on uniform grid are shown in fig .
[ fig : jetcocoon400 ] .
it can be seen that solutions use hll , llf and direct flux reconstruction have large positive fluctuations in the normal velocity at the rarefaction wave .
it is interesting to note that only hllc does not suffer from this shortcoming which is probably due to the ability of hllc to resolve the contact discontinuity compared to the other riemann solvers in the code .
if those fluctuations also happen in higher dimensional jet simulation , then one would expect that the normal velocity fluctuation seen in this test would lead to an artificially extended cocoon . in fig .
[ fig : jet1d_amr ] the result of using hll - plm and hllc - plm with amr is shown .
it can been seen that the fluctuation in the hll scheme becomes smaller with higher resolution .
[ fig : l1_jet1d ] shows the @xmath76 error for those two schemes with different levels of refinement . to test the code in higher dimension ,
we first study the two - dimensional shock tube problem suggested by @xcite and latter also used by various groups ( see e.g. zhang & macfadyen 2006 , mignone & bodo 2005a , lucas - serrano et al .
this test is done in a two - dimensional cartesian box divided into four equal - area - constant states : @xmath110 where ne means northeast corner and so on .
the grid is uniform @xmath111 .
the gas is assumed to be ideal with an adiabatic index @xmath109 .
we use outflow boundary conditions in all four directions and the cfl number is @xmath86 . the results are shown in fig .
[ fig : shocktube ] for four schemes .
this problem does not have analytical solutions to compare with , but comparing our result with other groups result shows good agreement .
the cross in the lower left corners of ( a ) hll - plm and ( d ) f - weno is a numerical artifact due to the inability to maintain a contact discontinuity perfectly , which are absent in the results using the hllc solver ( c ) .
this agreed with the result of @xcite that the hllc solver behaves better in this problem than other riemann solvers because of its ability of resolve contact discontinuities .
we first study a test problem with uniform grid in cartesian coordinate that has been used by several groups to test the symmetric properties of a three - dimensional srhd code @xcite .
the initial setup of this test consists of a cold spherical inflow with initially constant density @xmath112 and constant velocity @xmath113 colliding at the box center .
this problem is run using three - dimensional cartesian coordinates so it allows one to evaluate the symmetry properties of the code @xcite . when gas collides at the center , a reflection shock forms . behind the shock
, the kinetic energy will be converted completely into internal energy .
thus the downstream velocity @xmath114 is zero and the specific internal energy is given by the upstream specific kinetic energy @xmath115 using the shock jump condition , the compression ratio @xmath116 and the shock velocity can be found to be @xmath117 in the unshocked region ( @xmath118 ) , the gas flow will develop a self - similar density distribution , @xmath119 the initial state are @xmath120 .
we chose a small value for pressure because a grid - based code can not handle zero pressure .
a cfl number @xmath121 is used for this problem , as other groups @xcite .
we chose to use llf - plm for this problem because this turns out to be the most stable solver for this problem .
[ fig : rssr_1d ] shows the one - dimensional cut though axis and diagonal direction and fig . [
fig : rssr_2d ] shows a contour through @xmath122 plane , both at @xmath92 .
it can be seen from those plots that our code keeps the original spherical symmetry quite well . since in a cartesian box the simple outflow boundary condition is inconsistent with the initial spherical inflow setup , we evolve this problem only to @xmath92 , at which point all the mass in the original box has just entered the shocked region ( see fig . [
fig : rssr_1d ] ) . after that time
, the evolution would be affected by the unphysical boundary condition . in this test
, we study a spherical blast wave in three - dimensional cartesian coordinates .
there is no analytical solution for this problem .
thus for the sake of comparison , we set up the same problem as other groups @xcite .
the center of the blast wave source is located at the corner @xmath123 of the box .
the initial conditions are @xmath124 where @xmath125 is the distance to the center @xmath126 .
an ideal gas with an adiabatic index of @xmath109 is assumed .
the left boundaries at x , y , z directions are reflecting while others are outflow . we use a top grid of @xmath127 zones with two more levels of refinement and a refinement factor of 2 for this calculation ( equivalent resolution @xmath128 ) .
f - plm is used for the result shown and the cfl number is @xmath121 . the results are given in fig .
[ fig : blast3dcorner_1 ] which shows the cut along x - axis and diagonal direction . for comparison
, we run a high resolution one - dimensional simulation using spherical coordinates .
the three - dimensional run in cartesian coordinates agrees with the one - dimensional high resolution run .
furthermore , it can be seen that the spherical symmetry of the initial condition is preserved rather well in the three - dimensional cartesian case .
finally we study another blast wave problem for which the center of the blast wave source is located at the box center .
this problem also does not have analytical solution but it has been studied by @xcite so our result can be compared to theirs .
the initial conditions are @xmath129 an ideal gas eos with an adiabatic index of @xmath130 is used .
we stop the run at @xmath131 , roughly the same ending time of @xcite .
a top grid of @xmath132 zones with four levels of refinement and a refinement factor of two is used ( equivalent resolution @xmath133 ) .
thus our resolution is roughly @xmath134 times of @xcite .
we used hll - plm and a cfl number of @xmath86 for this calculation .
[ fig : blast3d_1 ] plots the numerical solution for all cells centered on the highest level in the two dimensional slice at @xmath135 at @xmath136 as a function of radius from the center ( 0.5,0.5,0.5 ) .
the position and amplitude of the high density shell agrees with the calculation of @xcite . and
it can be seen that the spherical symmetry is preserved rather well in our code .
having validated our code using various test problems , we can apply it to astrophysical problems .
we will study two typical astrophysical relativistic flow problems in this work , agn jets and grb jets .
both topics have been studied extensively with 2d simulations before , but very few 3d calculations have been done in both cases . consequently we will focus on 3d simulations here . in this section ,
we study a relativistic supersonic jet in three dimensions .
we set up the problem using the same parameters as model c2 of @xcite .
this model has also been studied in two dimensions by @xcite and in three dimensions by @xcite .
the jet parameters are @xmath137 , @xmath138 and @xmath139 .
the jet has a classical mach number @xmath140
so the pressure is @xmath141 .
the parameters for the medium are @xmath142 , @xmath143 and @xmath144 .
the eos is assumed to be ideal with @xmath109 .
the jet is injected from the low - z boundary centered at @xmath145 with radius @xmath146 .
outflow boundary conditions are used at other part of the boundary .
figure [ fig : jet ] shows the result at @xmath147 for the three runs : hll - ppm with three levels of refinement ( hll - ppm - l3 ) , hll - plm with three levels of refinement ( hll - plm - l3 ) and hll - plm with four levels of refinement ( hll - plm - l4 ) .
the top grid resolution is @xmath132 zones .
thus the first two runs have an equivalent resolution of @xmath128 zones while the last one has @xmath133 zones .
for the first two , turbulence in cocoon is not fully developed so the cocoon is still symmetric even in 3d .
the hll - ppm - l3 run has slightly more turbulent cocoon due to the higher spatial reconstruction order of ppm .
thus the hll - ppm - l3 jet propagates slightly slower than the hll - plm - l3 jet . on the other hand , for the hll - plm - l4 jet ,
the resolution is @xmath148 cells per beam radius , comparable to the resolution used in the two - dimensional study by @xcite .
the cocoon turbulence is much more developed in this case , as in the 2d case @xcite .
consequently , the hll - plm - l4 jet propagates slower than the two lower resolution ones .
furthermore , the hll - plm - l4 case does not show axisymmetry because instability quickly develops in the lateral motion and consequently lateral motion also becomes turbulent . since we found the jet - cocoon structure differs significantly at higher resolution run and consequently the jet propagation speed decreases ,
we conclude that even at @xmath133 effective resolution of our three dimensional jet simulations the correct solution remains elusive .
moreover , different solvers give disparate answers .
the idea that long - soft " gamma - ray bursts are associated with the deaths of massive stars has been supported by various observations recently ( see e.g. woosley & bloom ( 2006 ) for a recent review ) .
so it is of great interest to calculate how a relativistic jet propagate through and break out of a massive star .
there has been extensive calculations in 2d recently @xcite .
but very few 3d calculation have been reported in the literature so far .
since there are significant turbulent motion and mixing in the jet - cocoon system in 2d calculation , it is important to model this process in 3d .
we use the projenitor he16a of @xcite , which is a stripped - down helium core with initial mass @xmath149 m@xmath150 .
[ star ] shows the density and pressure profiles for the progenitor model .
the radius of the star at onset of collapse is @xmath151 cm . in our simulation , the mass inside @xmath152 cm
is removed and replaced by a point mass of @xmath153 m@xmath150 .
the jet has power @xmath154 erg s@xmath155 , initial lorentz factor @xmath156 , the ratio of its total energy ( excluding rest mass energy ) to its kinetic energy @xmath157 .
this corresponds to a a jet with initial density @xmath158 g @xmath159 and pressure @xmath160 erg @xmath159 .
the jet is injected parallel to the @xmath161 axis with an initial radius @xmath162 cm .
those model parameters are similar to previous 2d calculations @xcite .
we use a simulation box of @xmath163 cm in order to follow the propagation of the jet after break out .
we use hll - plm as this should be the most stable and reliable scheme to carry out resolution study .
the top grid resolution is @xmath127 .
we run the simulation using both 4 and 5 refinement levels , which correspond to resolution of 5.6 and 11 cells per jet beam radius , respectively . in order to always have high resolution for the jet material , we designed a color field refinement strategy in addition to the standard refinement criterion designed for discontinuities .
more specifically , we use two color fields to keep track of the injected jet material and the stellar material . those two color fields give us a fraction of jet material at every cell .
then whenever a cell contains more than 0.1 percent of jet material , we flag that cell for refinement .
this ensures that we also have high resolution when mixing between jet material and star material happens .
the results for those two runs are show in fig .
[ grbjet ] .
it can been seen that the high resolution run gives qualitatively different jet dynamics .
while in the low resolution run the jet break out of the star successfully , in the high resolution run the jet head bifurcate at the stellar edge .
interestingly , this behavior has also been seen in 2d calculation @xcite .
but since we did not get convergent behavior so far , we can not conclude at this stage whether this behavior is physical or purely numerical .
however , it is safe to conclude that much higher resolution will be needed to model the jet break out in 3d .
in this paper , we have described a new code that solves the special relativistic hydrodynamics equations with both spatially and temporally adaptive mesh refinement .
it includes direct flux reconstruction and four approximate riemann solvers including hllc , hll , llf and modified marquina flux formula .
it contains several reconstruction routines : plm , ppm , third order ceno , and third and fifth order weno schemes .
a modular code structure makes it easy to include more physics modules and new algorithms . from our test problems and two astrophysical applications ,
it is clear that relativistic flow problems are more difficult than the newtonian case .
one key reason is that in the presence of utrarelativistic speed , nonlinear structures such as shocked shells are typically much thinner and thus requires the use of very high spatial resolution .
srhd problems also become difficult to solve accurately when significant transverse velocities are present in the problem as we have shown using several one dimensional problems .
one reason for this difficulty is that in srhd velocity components are coupled nonlinearly via the lorentz factor . in studying astrophysical jet problems ,
we have demonstrated the need of both high resolution achievable only through amr and careful choice of hydrodynamic algorithms .
in addition to validate our amr code , the most important implications of the calculations we have done is that _ in relativisitic flow simulations , resolution studies are crucial_. we thank greg bryan and michael norman for sharing _ enzo _ with the astrophysical community
without which this study could have not been carried out .
we would also like to thank miguel aloy for very helpful comment on the draft , roger blandford and lukasz stawarz for helpful discussions .
furthermore , we also thank ralf kehler for help with optimizing the code and slac s computing servers for maintaining an sgi altix super computer on which the reported calculations were carried out .
this work was partially supported by nsf career award ast-0239709 and grant no . phy05 - 51164 from the national science foundation .
w. acknowledges support by the stanford graduate fellowship and kitp graduate fellowship .
w. z. has been supported by nasa through chandra postdoctoral fellowship pf4 - 50036 awarded by the _ chandra x - ray observatory _ center , and the doe program for scientific discovery through advanced computing ( scidac ) .
oshea , b. w. , bryan , g. , bordner , j. , norman , m. l. , abel , t. , harkness , r. , & kritsuk , a.2004 , in `` adaptive mesh refinement - theory and applications '' , eds .
t. plewa , t. linde & v. g. weirs , springer lecture notes in computational science and engineering , 2004 | astrophysical relativistic flow problems require high resolution three - dimensional numerical simulations . in this paper
, we describe a new parallel three - dimensional code for simulations of special relativistic hydrodynamics ( srhd ) using both spatially and temporally structured adaptive mesh refinement ( amr ) .
we used the method of lines to discretize the srhd equations spatially and a total variation diminishing ( tvd ) runge - kutta scheme for time integration . for spatial reconstruction
, we have implemented piecewise linear method ( plm ) , piecewise parabolic method ( ppm ) , third order convex essentially non - oscillatory ( ceno ) and third and fifth order weighted essentially non - oscillatory ( weno ) schemes .
flux is computed using either direct flux reconstruction or approximate riemann solvers including hll , modified marquina flux , local lax - friedrichs flux formulas and hllc .
the amr part of the code is built on top of the cosmological eulerian amr code _ enzo_.
we discuss the coupling of the amr framework with the relativistic solvers . via various test problems ,
we emphasize the importance of resolution studies in relativistic flow simulations because extremely high resolution is required especially when shear flows are present in the problem .
we also present the results of two 3d simulations of astrophysical jets : agn jets and grb jets .
resolution study of those two cases further highlights the need of high resolutions to calculate accurately relativistic flow problems . |
active and normal ob stars occupy the same region in the hertzsprung - russell diagram ( hrd ) . while the observational characteristics of the different forms of activity are well - defined , a comprehensive understanding of the physical drivers of activity has to be established yet .
quantitative analyses of samples of active and normal stars at high precision are without doubt the starting point for progress to be made .
two types of active ob stars are of interest in the following , _ pulsators _ of @xmath2 cephei - type and slowly - pulsating b stars ( spbs ) , and _
magnetic stars_. their surface characteristics will be compared to those of apparently normal objects .
crucial improvements over previous work are achieved by the introduction of sophisticated modelling and a novel analysis technique , which allows in particular abundance uncertainties to be reduced to about 10 - 20% ( 1-@xmath3 statistical ) , in contrast to typical values of a factor @xmath02 in literature .
we focus on surface abundances of c , n and o , which may be altered already on the main sequence due to mixing with cno - cycled matter in the course of the evolution of rotating massive stars ( * ? ? ?
* ( e.g. maeder & meynet 2000 ) ) .
the aim is twofold : i ) to test whether mixing signatures correlate with the presence of magnetic fields ( * ? ? ?
* ( morel et al . 2008 ) ) , and ii ) to test whether the occurrence of pulsations is related to cno mixing .
our observational sample consists of 29 bright and apparently slowly rotating early b - type stars in nearby ob associations and in the field .
high - resolution ( @xmath4=40 - 48000 ) and high - s / n spectra ( @xmath5=250 - 800 ) with wide wavelength coverage were obtained with the echelle spectrographs foces , feros and fies , and from the elodie archive .
details of the observations and the data reduction are discussed by nieva & przybilla ( in prep . ) and nieva , simn - daz & przybilla ( in prep . ) .
the spectra were analysed using a hybrid non - lte approach ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ( nieva & przybilla 2007 , 2008 ; przybilla et al
. 2008b ) ) and a sophisticated analysis methodology .
state - of - the - art atomic input data were used in the modelling .
in contrast to previous work , multiple hydrogen lines , the helium lines , multiple metal ionization equilibria and the stellar energy distributions were reproduced _
simultaneously _ in an iterative approach to determine the stellar atmospheric parameters .
chemical abundances for a wide range of elements were derived from analysis of practically the entire observable spectrum per element .
the rewards of such a comprehensive , but time - consuming procedure were unprecedentedly small statistical error margins and largely reduced systematics ( * ? ? ?
* ( nieva & przybilla 2010 ) ) .
normal stars are marked by circles , @xmath2cep - type pulsators by dots ( candidates by thick circles ) and spbs by diamonds .
the symbol size encodes the n / c mass ratio , varying between @xmath00.3 ( cosmic value , smallest symbols ) and @xmath01.2 .
evolution tracks for rotating ( full lines , @xmath6=300kms@xmath7 ) and non - rotating stars ( dotted lines ) are overlaid ( * ? ? ?
* ( meynet & maeder 2003 ) ) .
the tracks were computed with @xmath8=0.02 , while the sample stars have @xmath8=0.014 , implying a downward shift of the theoretical zams.,width=415 ] the positions of the sample stars in the @xmath9-plane are displayed in fig.[fig1 ] .
all stars are relatively unevolved objects on the main sequence in the mass range of @xmath07 - 20@xmath10 . while the bulk of the objects are normal stars , a small number are known pulsators , either of @xmath2 cephei - type or spbs .
three stars are magnetic , @xmath2cep ( * ? ? ?
* ( henrichs et al . 2000 ) ) , @xmath11cas ( * ? ? ?
* ( neiner et al . 2003 ) ) and @xmath12sco ( * ? ? ?
* ( donati et al . 2006 ) ) , one is a candidate magnetic star , @xmath13cet ( * ? ? ?
* ; * ? ? ?
* ( schnerr et al . 2008 ; hubrig et al .
the majority of the stars has cno abundances compatible with cosmic values ( * ? ? ?
* ( przybilla et al . 2008b ; nieva & przybilla , in prep . ) ) , about a third of the objects shows signs of mixing with cno - cycled matter .
we will concentrate on this aspect and its relation to the presence of magnetic fields in the following .
fundamental parameters of the sample stars and also further details on the modelling and analysis are discussed by nieva & przybilla ( these proceedings ) .
observational indications of superficial abundance anomalies for carbon , nitrogen , and oxygen ( and the burning product helium ) in ob - type stars on the main sequence are known for a long time . they have found a theoretical explanation by evolution models for rotating stars ( * ? ? ?
* ; * ? ? ?
* ( e.g. maeder & meynet 2000 ; heger & langer 2000 ) ) , where meridional circulation and turbulent diffusion provide the means to transport cno - cycled matter from the stellar core to the surface .
the interaction of rotation with a magnetic dynamo may increase the transport efficiency even further ( * ? ? ?
* ( maeder & meynet 2005 ) ) .
models are indicated by the full line , the triangles marking ratios reached at the end of core - h burning for @xmath6 equivalent to 30 , 50 , 70 and 90% of the star s breakup velocity ( * ? ? ?
* ( bottom to top , ekstrm et al . 2008 ) ) .
the dotted line describes a magnetic 15@xmath1 model of ( * ? ? ?
* maeder & meynet ( 2005 ) ) with @xmath6=300kms@xmath7 .
the long - dashed line corresponds to a slope 3.77 calculated analytically for the nuclear path of the cn - cycle and initial cno abundances as in the evolution models , while the grey area spans the full range of theoretical slopes using different references for solar abundances and the cosmic abundance standard ( upper envelope).,width=415 ] the changes of cno abundances relative to each other follow a simple pattern . at the beginning of hydrogen burning ,
the @xmath14n enhancement comes from the @xmath15c destruction via the cn cycle , and the oxygen content remains about constant .
a fraction of this processed material is mixed with matter of pristine composition in the course of the subsequent evolution .
a simple calculation shows that the mixing signature follows the nuclear path with slope d(n
/ c)/d(n / o)@xmath164 when abundances are expressed in mass fraction , see @xcite for details .
stellar evolution models follow this trend and make quantitative predictions concerning the enrichment with nuclear - processed matter .
the mixing signature gets more pronounced with increasing rotational velocity and increasing mass ( not shown here ) , and magnetic models predict the strongest mixing , see fig .
the high - quality analysis of our sample stars facilitates the predicted trend to be recovered for the first time ( * ? ? ?
* ( overplotting data from previous studies would fill all areas in the n / o - n / c plane , see fig.1 of przybilla
2010 , with the error bars exceeding the range of fig .
[ fig2 ] in many cases ) ) .
about half of the sample clusters around ( n / [email protected] , n / [email protected] ) , which corresponds to unaltered cosmic values , several more stars show signatures apparently compatible with the rotating models , and five stars show a very high degree of mixing . of these
all but one are known ( candidate ) magnetic stars , suggesting a magnetic nature for the remaining one , @xmath17tau , as well .
note that the agreement between observations and theory would be even better if the models would account for cosmic ( * ? ? ?
* ( @xmath8=0.014 , przybilla et al . 2008b ) ) instead of solar abundances ( * ? ? ?
* ( @xmath8=0.02 , grevesse & noels 1993 ) ) .
helium abundances for all the sample stars are compatible with cosmic / solar values . .
the same symbols as in fig .
[ fig2 ] are used , the tracks for @xmath6 of 30 , 50 , 70 and 90% breakup velocity ( bottom to top ) are separated in this diagram .
several additional stars with their mixing signature indicating a possible magnetic nature are also identified .
the difference in metallicity between the models ( @xmath8=0.02 ) and the sample stars ( @xmath8=0.014 ) implies a shift of the tracks to the left in this case.,width=415 ] however , while fig.[fig2 ] provides a crucial test for the quality of the analyses it does not facilitate to decide which models are more realistic than others , or to identify the mixing mechanism(s ) .
tighter constraints can be put via the @xmath18-n / c diagram ( see fig.[fig3 ] ) , which relates the mixing signature to the evolutionary state of the sample stars .
indeed , the majority of objects follows the predicted trends for rotating stars , showing only mild enrichment ( n / [email protected] ) during their evolution from the zero - age main sequence ( zams , at the highest gravities ) to the giant stage ( lowest gravities ) .
truely fast rotators ( seen pole - on ) seem to be absent , which is likely a relic of one of the selection criteria for the sample assembly ( stars with low @xmath20 ) .
the magnetic stars stand out again with pronounced mixing signatures ( n / [email protected] ) , complemented by a group of stars at intermediate values of n / [email protected] - 0.7 .
some of the signatures found in the more evolved magnetic stars like @xmath2cep and @xmath11cas may be explained by evolution models accounting for a magnetic dynamo .
however , these models do not predict the magnetic field to reach the stellar surface .
the magnetic fields are probably of fossil origin instead .
it is likely that the enhanced transport efficiency ( when compared to models accounting for rotation only ) is retained in that case , as the basic physics behind is not sensitive to the origin of the magnetic field .
however , this has to be confirmed by detailed modelling . other objects , in particular @xmath12sco , challenge present stellar models seriously .
the observed mixing signature of @xmath12sco may be explained by homogeneous evolution , but this has also to be confirmed by further computations .
homogeneous evolution would require a highly - efficient spin - down mechanism , as the star is a truely slow rotator at present ( * ? ? ? * ( donati et al .
2006 ) ) , like several other magnetic stars with pronounced mixing signature ( * ? ? ?
* ( morel et al . 2008 ) ) .
one may speculate on magnetic breaking due to angular - momentum losses by a magnetically confined line - driven stellar wind or magnetic coupling to the accretion disc during the star - forming process in the case of a fossil field . even though the topic is not understood theoretically in a comprehensive way , spin - down times of the order of 1myr (
* ; * ? ? ?
* ( ud - doula et al . 2009 ; townsend et al . 2010 ) ) or even less ( * ? ? ?
* ( mikulek et al . 2008 ) ) are reported for some magnetic massive stars , possibly leading to the required slow rotation already close to the zams . after having established the connection of pronounced mixing signatures with the presence of magnetic fields , one may turn the argumentation and use the chemical fingerprint as a selection criterion for the search of magnetic fields in massive stars .
based on their mixing signatures we therefore suggest the following objects for follow - up spectropolarimetric observations : @xmath17tau as a high - potential target and @xmath22orib , @xmath23ori , pt pup and @xmath24eri ( * ? ? ? *
( weak constraints for the latter two were presented by hubrig et al .
2009 ) ) .
alternative explanations of the mixing signatures may be required in the case of non - detections .
case a mass - transfer in close binary systems
i.e. near the end of the core h - burning phase of the primary is one of the most promising mechanisms .
this can produce stars highly enriched in cno - cycled products close to the zams .
such stars would be fast rotators because of angular momentum transfer accompanying the mass overflow .
the observational identification of such systems may be challenging , though , as the companion of the visible ob star is likely a low - mass helium star in a wide orbit , see @xcite for details .
+ finally , we want to comment briefly on the question whether the occurrence of pulsations is in any way related to cno mixing .
the pulsations of @xmath2cep stars and spbs are excited via the @xmath25-mechanism because of a subsurface opacity bump due to the ionization of iron - group elements ( * ? ? ?
* ( pamyatnykh 1999 ) ) . yet
, an additional factor is required to explain the occurrence of pulsators and non - pulastors in the same regions of the hrd .
additional changes in the opacities due to modified cno abundances as a consequence of mixing with nuclear - processed matter may be a possibility .
however , our results do not support this hypothesis , pulsations occur independently of mixing signatures , see fig .
theory predicts surface abundances of massive stars to follow a tight relation when mixing with cno - cycled material occurs . the nuclear path in the n / o - n / c plane
is solely determined by the initial cno abundances .
this provides a powerful criterion for judging the quality of quantitative analyses .
few early - type massive stars can be expected to deviate from this nuclear path , either due to atomic diffusion ( which is normally not operational in ob - type stars ) or because of pollution with material from advanced nuclear burning phases , e.g. accreted in the course of a binary supernova ( * ? ? ?
* ; * ? ? ?
* ( przybilla et al . 2008a ; irrgang et al . 2010 ) ) . only observational data passing this test
is well - suited for being used in the verification of different stellar evolution models .
the high - quality analysis of our sample stars indicates that most objects can be explained well by existing models for rotating stars .
it also facilitates the identification of a connection of pronounced mixing signatures with the presence of magnetic fields initially suggested by @xcite at high confidence . in reverse ,
pronounced mixing signatures may be used to define high - potential targets for spectropolarimetric surveys that aim at the detection of magnetic stars .
several sample stars are suggested for follow - up observations .
further studies , observational as well as theoretical , are required to develop a comprehensive understanding of chemical mixing in massive stars .
high - quality analyses have to be extended to larger samples of stars , covering a wide range of rotational velocities .
further contraints to stellar evolution models can be obtained by inclusion of stars at different metallicity , like e.g. objects in the magellanic clouds .
once the ongoing surveys as the mimes project ( wade et al . , these proceedings ) will detect more magnetic stars it will become feasible to develop a better overview on the rle that magnetic dynamos and/or fossil fields play in the course of the evolution of massive stars .
+ we would like to thank a. maeder , g. meynet , v. petit , s. simn - daz and g. wade for fruitful discussion in preparation of this work .
np acknowledges travel support by the _ deutsche forschungsgemeinschaft _ , project number pr 685/3 - 1 . | we report on the abundances of helium , carbon , nitrogen and oxygen in a larger sample of galactic massive stars of @xmath07 - 20@xmath1 near the main sequence , composed of apparently normal objects , pulsators of @xmath2-cephei- and spb - type , and magnetic stars .
high - quality spectra are homogeneously analysed using sophisticated non - lte line - formation and comprehensive analysis strategies .
all the stars follow a previously established tight trend in the n / c - n / o ratio and show normal helium abundances , tracing the nuclear path of the cno - cycles quantitatively .
a correlation of the strength of the mixing signature with the presence of magnetic fields is found . in conjunction with low rotation velocities
this implies that magnetic breaking is highly efficient for the spin - down of some massive stars .
we suggest several objects for follow - up spectropolarimetry , as the mixing signature indicates a possible magnetic nature of these stars . |
the evolution of the cosmic structure has been an important subject in cosmology . in the post - decoupling era
the neutral atoms ( the baryons ) ceased to feel the pressure from the photons .
so the only force that fought against the gravitational pull was its own thermal pressure . due to the hubble expansion ,
the baryon temperature gradually decreased and so was its thermal pressure .
the situation changed drastically at the subsequent reionization epoch , during which baryons turned into fully ionized plasmas and were heated up to @xmath3 by the photo - ionization process @xcite .
such reheating process raised the jeans scale to the size of the galaxies @xcite and as a result the course of the structure formation was significantly altered .
traditionally the study of the baryon evolution has been conducted by staying within the gas hydrodynamics framework while changing the mean molecular weight between the pre- and the post - reionization epochs to account for the effects of ionization .
however , it is known that gas hydrodynamics , which assumes local charge neutrality , can not exhibit crucial collective phenomena that is unique to plasma , such as landau damping .
it is therefore natural to wonder whether relevant plasma collective effects in cosmic evolution might have been overlooked . in 2007 , one of us ( chen ) and lai first considered the plasma effects on the baryon structure formation in the post - reionization epoch , where the conventional baryon acoustic waves at small scales are replaced by the ion acoustic waves in plasmas @xcite .
a new formulation based on the maxwell - einstein - boltzmann equations for pure hydrogen ions was introduced in that paper . to obtain the baryon power spectrum , the authors employed the instantaneous jeans wavenumber @xmath4 and interpolated the solutions from the opposite limits of the wavenumber @xmath1 .
interesting as it is , this new plasma hydrodynamic formulation has some shortcomings , however . in particular , the spectrum in the intermediate regime of @xmath1 , which is the most crucial regime of interest , was unresolved . to properly account for the impact of the hubble expansion
, it was suggested that the relevant scale should be the integrated and smoothed ` filtering wavenumber ' @xmath2 rather than the instantaneous jeans wavenumber @xmath4 @xcite .
motivated by the reasons , in this paper we analytically re - solve the baryon density spectrum under the landau damping from the corrected plasma hydrodynamic equation of hydrogen - helium plasmas .
having the solved spectrum , we are able to investigate the impacts of the landau damping effect on the evolution of baryon overdensity and on the filtering wavenumber @xmath2 due to the decaying amplitude of baryon oscillation modes .
the layout of the paper is the following : in sec.[ion ] we briefly review the concept of ion acoustic wave and its associated new sound speed for baryons . in sec.[spectrum ] we introduce the landau damping of the baryon density spectrum .
this is carried out through the analytical solution of the plasma hydrodynamics equation .
then in sec.[res ] we investigate the influence of landau damping on the baryon spectrum and derive a modified filtering wavenumber @xmath5 in place of @xmath2 .
finally we compare the spectra obtained from plasma and gas approaches , respectively , and discuss the possible ways to distinguish the two models .
an ordinary sound wave in a neutral gas is a pressure wave that propagates through the thermal collisions among neighboring particles . without collisions ,
the sound wave would not occur .
however for an ion acoustic wave in a fully ionized plasma , an extra repulsive , non - collisional , electrostatic pressure exists due to the imperfect debye shielding at finite temperature .
hence an ion acoustic wave can travel through the medium without collisions @xcite . based on the chen - lai formulation @xcite and the fluid equations for matter overdensites @xcite
, we obtain the coupled fluid equations for hydrogen - helium baryon ( ignoring the electron mass ) : @xmath6 where the subscripts @xmath7 and @xmath8 indicate the baryon and the dark matter , respectively , @xmath9 is the fractional overdensity , @xmath10 and @xmath11 are the baryon and dark matter fractions and @xmath12 is the sound speed for ionic baryons . according to plasma physics , the sound speed is defined by @xmath13 where @xmath14 @xmath15 is the helium ion charge and @xmath16 is the combined residual electrostatic potential of hydrogen and helium . here
we have assumed the temperature @xmath17 because the time scale for h - he plasma to reach thermal equilibrium is short compared with that of the duration of the reionization epoch @xcite . in a typical plasma @xmath18 for isothermal compression condition
is often assumed since the thermal velocities of electrons were @xmath19 times larger than the fluid sound speed , which enables electrons to interact with each other over several wavelengths within one oscillation @xcite . hence invoking the primordial helium abundance
@xmath20 @xcite deduced from the big bang nucleosynthesis ( bbn ) , we can express the ion sound speed @xmath12 in terms of the effective mean molecular weight @xmath21 , which combines @xmath22 and @xmath23 , @xmath24 with @xmath25 for @xmath26 and @xmath27 for @xmath28 . note that in contrast @xmath29 and @xmath30 for @xmath26 and @xmath31 , respectively , for the neutral gas .
in general , the ion fluid leads to a lower sound speed by a factor of @xmath32 , which corresponds to @xmath33 increase of the jeans wavenumber @xmath4 .
it is known that plasma waves suffer the landau damping when the thermal velocity of the plasma particles is slightly below the phase velocity velocity .
residing on the down - slope of the maxwellian particle velocity distribution , the plasma wave tends to lose energy to more particles and gain energy from less particles .
this results in the decline of the plasma wave amplitude . learning from that
, we expect the similar effect on the small - scale baryon oscillations , the ion acoustic modes .
mathematically , landau damping is a consequence of the singularity @xmath34 in the vlasov equation , which is absent in the fluid theory . to account for the landau damping effect in the fluid model
, one may separately calculate the landau damping rate and then multiply it to the imaginary part of the plasma fluid solution . to analytically study the baryon density spectrum under the landau damping effect , we start with the coupled fluid equations , eq .
( [ dmh ] ) and eq .
( [ baryon ] ) , with @xmath35 and @xmath36 for simplicity . replacing the variable @xmath37 by @xmath38 via the relation @xmath39 in the matter - dominant era
, we arrive at @xmath40 where @xmath41 is the co - moving jeans wavenumber of the ionic baryon .
the solution of eq .
( [ dma ] ) gives @xmath42 when dropping the decaying mode , while the solution of eq .
( [ baryona ] ) depends on the thermal history of reionization .
we adopt the temperature evolution of the intergalactic medium ( igm ) , @xmath43 , which results from the balance between the photo - heating process and the hubble expansion cooling process @xcite .
in fact the recent measurements indicate a second reheating process at @xmath44 due to the ionization of he@xmath45 @xcite . to simplify the case ,
we assume a single reheating process and the temperature evolution @xmath46 which leads a constant jeans wavenumber @xmath47 , where the subscript @xmath48 stands for the ending time of the reionization process .
the equation with arbitrary thermal evolution can be analytically solved via green s function method @xcite . inserting the solution of eq .
( [ dma ] ) , eq .
( [ baryona ] ) can be rewritten as @xmath49 with @xmath50 , @xmath51 and @xmath52 .
note that @xmath53 . since we are interested in the relation between the baryon and the dark matter evolutions , it is convenient to express the baryon density in terms of that of the dark matter . using the initial conditions : 1 ) @xmath54 and 2 ) @xmath55 at @xmath56
, we obtain the normalized baryon density spectrum : @xmath57+\frac{1}{1+\kappa^2 } , \label{sol } \end{aligned}\ ] ] where @xmath58 and @xmath59 as @xmath60 , @xmath61 becomes imaginary .
we may therefore rewrite eq .
( [ sol ] ) as @xmath62+\frac{1}{1+\kappa^2}\nonumber\\ & & \equiv b_i(\alpha,\kappa)+\frac{1}{1+\kappa^2 } , \label{baryq}\end{aligned}\ ] ] where @xmath63 and @xmath64 denotes the oscillatory part of @xmath65 .
if @xmath66 or equivalently @xmath67 , @xmath65 asymptotically approaches @xmath68 .
we see the amplitude of the baryon oscillations decreases as @xmath69 . because @xmath70 , the thermal velocity of the ions is close to the sound speed .
we then expect that the landau damping would occur to the baryon fluid oscillations . by invoking the nonlinear empirical formula which relates the real and the imaginary parts of the ion acoustic oscillation frequency @xcite
, we introduce the landau damping effect to eq .
( [ baryq ] ) by multiplying the oscillation term with an exponential factor , @xmath71 where @xmath72 this means the ion acoustic wave would be landau - damped to one e - fold of its initial amplitude within just @xmath73 of one oscillation period .
such severe damping effect causes the baryon oscillations to be totally smoothed out at late times .
@xmath74{solo3.eps } \end{array}$ ] @xmath74{sol3.eps } \end{array}$ ]
@xmath74{filtering.eps } \end{array}$ ] @xmath74{early.eps } \end{array}$ ] with the landau damping introduced , we now calculate the complete baryon density spectrum by combining eq . ( [ sol ] ) and ( [ baryld ] ) .
figure [ solo ] shows the result ( in solid curve ) at @xmath75 with @xmath76 .
the dashed and dotted curves correspond to the results obtained from the neutral gas hydrodynamics and plasma physics without applying the landau damping .
we see that the dotted curve shifts from the dashed one by @xmath33 due to the increase of the jeans wavelength mentioned in sec . [ ion ] . when suffering the landau damping , the modes of baryon overdensity at large @xmath1 diminish , leading to a suppression of the baryon density spectrum in the intermediate regime of @xmath1 .
such a suppression , however , accidentally compensates the shift of the spectrum .
thus the filtering wavenumber @xmath2 proposed by gnedin and hui @xcite to characterize the baryon structure would not be suitable when the landau damping effect is included . in an expanding universe ,
the jeans scale changes with temperature and therefore with time . on the other hand
, it takes about one hubble time for baryons to respond to the changed pressure
. therefore the filtering wavenumber @xmath2 was introduced @xcite to account for the migration of the jeans wavenumber over the time scale relevant to the acoustic oscillations , which therefore depends on the entire previous thermal history of the universe @xcite . in the matter - dominant era @xmath2
is related to @xmath4 by @xcite @xmath77.\ ] ] when the baryon fluctuations suffer the landau damping effect , a correction to @xmath2 must be taken .
the original @xmath2 was deduced by supposing the solution @xmath78 under the small @xmath1 approximation ( @xmath79 ) . with the exponential damping factor
included , we find it reasonable to directly assume @xmath80 , where @xmath5 is the modified filtering wavenumber .
assuming @xmath81 , we have @xmath82 upon taking logarithm and then square - root on both sides , we obtain @xmath83 the right - hand side of the above expression approaches a constant for @xmath84 $ ] at @xmath85 , which falls in the intermediate regime of @xmath1 .
hence the modified filtering wavenumber @xmath5 can be obtained by substituting an arbitrary @xmath86 within this range .
figure [ sol3 ] is the same plot as fig.[solo ] but with an additional dotted curve that illustrates the well - fitted function of @xmath87 . in fig .
[ filtering ] we compare the two characteristic scales @xmath2 ( dashed ) and @xmath88 ( solid ) at different redshifts .
the two curves converge to each other at low redshifts , while in general @xmath5 is smaller than @xmath2 until @xmath89 .
in summary , the baryon overdensities has three stages of evolution : 1 ) before the photon - baryon decoupling , where baryons were in a plasma state but were tightly coupled to and dominated by the photons ; 2 ) during the dark age , where baryons formed neutral gases and grew with dark matter overdensities ; and 3 ) after reionization , where baryons turned into a plasma again and followed the plasma hydrodynamics .
although before the decoupling baryons were in a plasma state , the photon - baryon fluid was dominated by the photon pressure rather than the combination of the baryon thermal and electrostatic pressures .
this resulted in a sound speed that is much larger than the ion thermal velocity .
the plasma effects were negligible in that epoch .
this aspect has been widely recognized .
what appears less recognized is the influence of the plasma collective effects in the post - reionization epoch . considering the plasma effects during this epoch , we point out that the baryons would experience two major impacts : the change of the sound speed and the landau damping of the baryon oscillations .
numerically speaking , the first effect is relatively minor compared with the second . since @xmath90 , we expect that the landau damping would be quite effective in dissipating the amplitude of baryon oscillation modes .
this might influence the subsequent galaxy formations history . according to eq .
( [ baryq ] ) , the amplitude of baryon oscillations scales as @xmath69 . therefore the discrepancy between the gas and the plasma approaches is most significant in the early stage of structure formation immediately after the reionization .
figure [ early ] is a comparison of the baryon density spectra based on the gas ( in dashed line ) and the plasma ( in solid line ) approaches at different redshifts ( @xmath91 and @xmath92 ) with @xmath76 .
if the baryons behave as a neutral gas , then we expect a sizable signature of baryon acoustic oscillations in the matter power spectrum below the scale @xmath93 after the photo - reheating process .
conversely , if the baryons act more like a plasma , then we should find little sign of oscillations but instead the landau damping of it , whose rate maybe measured via tracing the spectrum back to the early time .
so far the observations of the matter power spectrum have not yet reached the sensitivity for such small scales at high redshifts @xcite
. it would be interesting if this regime can be explored in the near future , so that we can learn more about the nature of the cosmic structure formation .
this research is supported by taiwan national science council under project no .
nsc 97 - 2112-m-002 - 026-my3 , by taiwan s national center for theoretical sciences ( ncts ) , and by the us department of energy under contract no .
de - ac03 - 76sf00515 .
j. miralda - escude and m. j. rees , _ mon . not .
r. astron .
soc . _ * 266 * , 343 ( 1994 ) .
j. schaye , t. theuns , m. rauch , g. efstathiou and w. l. w. sargent , _ mon . not .
r. astron .
soc._*318 * , 817 - 826 ( 2000 ) , [ arxiv : astro - ph/9912432 ] | it has been suggested by chen and lai that the proper description of the large scale structure formation of the universe in the post - reionization era , which is conventionally characterized via gas hydrodynamics , should include the plasma collective effects in the formulation .
specifically , it is the combined pressure from the baryon thermal motions and the residual long - range electrostatic potentials resulted from the imperfect debye shielding , that fights against the gravitational collapse . as a result , at small - scales
the baryons would oscillate at the ion - acoustic , instead of the conventional neutral acoustic , frequency . in this paper
we extend and improve the chen - lai formulation with the attention to the landau damping of the ion - acoustic oscillations .
since @xmath0 in the post - reionization era , the ion acoustic oscillations would inevitably suffer the landau damping which severely suppresses the baryon density spectrum in the regimes of intermediate and high wavenumber @xmath1 . to describe this landau - damping phenomenon more appropriately ,
we find it necessary to modify the filtering wavenumber @xmath2 in our analysis
. it would be interesting if our predicted landau damping of the ion - acoustic oscillations can be observed at high redshifts . |
obviously , the shape of the luminosity function is defined by a number of factor which can vary from galaxy to galaxy . of these effects ,
most important should be the effects of binary evolution . despite of a number of population synthesis models developed
, there is no a clear prediction about the evolution of the luminosity distribution of x - ray binaries with time . based on simple arguments
, one might expect that these effects are mostly pronounced at the high luminosity end of the xlf .
indeed , a luminosity of @xmath34 erg / s requires a mass accretion rate of @xmath35 m@xmath36/yr , which can be sustained by a low mass star for less than @xmath37 yrs @xcite .
this value is significantly shorter than the life time of a galaxy .
it is yet shorter for the most luminous lmxb systems with @xmath38 erg / s . in order to have the presently observed shape of the luminosity distribution , with a moderate fraction of sources with @xmath39 erg / s , in the stellar systems of the age of @xmath40 yrs
, a continuous replenishment of the high luminosity sources is required .. ] such replenishment can be maintained for example due to binary systems with initially less massive companion stars , reaching the x - ray active phase at later times .
clearly , evolution of the luminosity function with time must be present and it should be more pronounced at the high luminosity end of xlf .
consequently , no universal luminosity function in the strict sense should be expected , from which the source populations observed in different galaxies are drawn . however , the results of the present study suggest that there are no significant variations of the shape of the luminosity distribution of low mass x - ray binaries in majority of nearby galaxies they are broadly consistent with having the same shape
. this fact might be somewhat surprising in the view of the arguments presented above .
an obvious explanation is that the nearby galaxies constituting our sample have similar age of the stellar population , with sufficiently long time passed since the last starburst events . on the other hand
, the departures from the average xlf possibly observed in ngc1553 and ngc1291 ( section [ sec : individual ] ) , if real , might be a manifestation of the xlf dependence on the age and star formation history of the host galaxy . in interpreting the absence of pronounced systematic variations of the shape of the luminosity distribution
, one should take into consideration the sensitivity limitations of the present analysis .
indeed , in order to minimize contamination by the background sources and interference of the ccd boundaries , we analyzed the central parts of the galaxies , within @xmath2 effective radii for the early type galaxies , and inner bulges for spirals .
this resulted in a rather limited number of sources in our sample , @xmath41 in total , and , consequently , to a limited sensitivity to possible xlf variations .
the amplitude of the statistical uncertainties is illustrated by the shaded area in fig.[fig : xlf_comb_uxlf ] . with
increased sample and detection sensitivity , more complex behavior and more subtle effects can be discovered .
we should note , however , that the analyzed annuli encompass from @xmath42 to @xmath43 of the total near - infrared luminosity ( and stellar mass ) of the galaxies .
typically , more than a half of the remaining stellar mass is contained in the central parts of the galaxies , inside the inner boundary of the analyzed annuli .
on the other hand , for many galaxies the outer boundary was sufficiently close to the maximum value allowed by cxb contribution , especially for the luminosity growth curves .
therefore a significant increase of the number of x - ray binaries suitable for analysis would require an adequate improvement in the sensitivity and an accurate treatment of the incompleteness effects in the centers of the galaxies . due to typically larger distances to the elliptical galaxies , the low luminosity end of the xlf was studied using the data of only two late type galaxies , m31 and the milky way .
generalization of these results on the galaxies of all morphological types involves certain degree of extrapolation of the data .
on the other hand , the old stellar systems in the ellipticals and spirals are similar in many respects and such an extrapolation might be justified .
as was shown by @xcite , in a very broad range of luminosities , @xmath44 , the luminosity function of hmxbs is consistent , to the first approximation , with a single slope power law with the differential slope of @xmath45 .
the average luminosity function of lmxbs , on the contrary , has a complex shape .
it is a power law with the differential slope of @xmath3 at low luminosities , steepens significantly at @xmath46 and has a cut - off at @xmath47 ( fig.[fig : xrb_xlf ] ) .
different shape of the luminosity functions of the high and low mass binaries reflects , obviously , the difference in the accretion regimes .
the majority of high mass systems are wind accretors .
it has been shown by @xcite , that the expected xlf slope is @xmath48 , i.e. close to the observed value .
note , that postnov s derivation was based on the assumption , that the mass distribution of the donor stars in hmxbs can be described by the salpeter imf .
the low mass systems , on the other hand , are close binaries fed via roche lobe overflow .
the slope of the average lmxb xlf at @xmath49 is qualitatively similar to that obtained by @xcite from x - ray binaries population synthesis , @xmath50 . at lower luminosities ,
however , the population synthesis model predicts significantly more sources than actually observed . as discussed by @xcite ,
the number of low luminosity sources is substantially reduced , if the irradiation of the donor star by the x - ray emission from the primary is taken into account .
indeed , their illustrative @xmath51 distribution with simplified account of the irradiation effect is significantly flatter below @xmath52 and is qualitatively similar to the observed average xlf of lmxbs .
the different shape of xlfs of hmxbs and lmxbs can be used to diagnose on - going star formation in relatively nearby galaxies ( @xmath53 mpc for the chandra angular resolution ) .
two possibilities can be exploited .
the cut - off in the hmxb luminosity function occurs at @xmath54 times higher luminosity than in lmxbs ( fig.[fig : xrb_xlf ] ) . in our sample of @xmath41 sources in the old stellar systems with total stellar mass of @xmath55 m@xmath56 two brightest sources , in ngc1553 ,
have luminosity @xmath57 and @xmath58 erg / s ( see sections [ sec : nir_x_gcurve ] and [ sec : ngc1553 ] ) .
these two sources excluded , the luminosity of the other sources does not exceed @xmath59 erg / s ( fig.[fig : xlf_cum ] ) .
the brightest high mass x - ray binaries , on the other hand , have luminosities up to @xmath60 erg / s .
detection of even one source with luminosity of the order of @xmath61 erg / s or greater implies that the star formation process is taking place . from the number of such luminous sources the star formation rate can be estimated , as it is directly proportional to the sfr @xcite : @xmath62 where sfr refers to the formation rate of stars more massive than @xmath63m@xmath36 .
this , obviously , can be used for sufficiently high star formation rates , exceeding @xmath54 m@xmath56/yr .
non - detection of luminous sources , on the other hand , immediately constrains the star formation rate , e.g. absence of sources with @xmath64 or @xmath65 erg / s results in upper limit on the star formation rate of sfr@xmath66 and sfr@xmath67 m@xmath36/yr respectively ( 90% confidence level ) . at lower star formation rates one can use the fact , that the xlf slopes at the low luminosity end are substantially different , @xmath68 and @xmath69 for lmxbs and hmxbs respectively ( fig.[fig : xrb_xlf ] ) .
therefore detection of a population of compact sources with sufficiently steep @xmath70 distribution at @xmath71 erg / s might indicate on - going star formation .
conversely , @xmath72 distribution in the low luminosity limit would indicate that the x - ray emission is dominated by an old stellar population .
relative contributions of low and high mass x - ray binaries to the observed population of compact x - ray sources is defined by the ratio of the stellar mass to the star formation rate . due to different shapes of the luminosity functions ( fig.[fig : xrb_xlf ] )
it also depends on the considered luminosities range . using the calibrations of @xmath12sfr and @xmath73 relations obtained by @xcite and in the present paper ,
the ratio of the total luminosities can be estimated : @xmath74 in the low sfr limit , sfr@xmath75 m@xmath36/yr , the luminosity of high mass x - ray binaries will be subject to effects of statistics ( section [ sec : stat ] ) and the ratio of the most probable values is given by : @xmath76 the two above dependences intersect at sfr@xmath77 m@xmath36/yr , the ratio of the number of sources depends non - monotonically on the threshold luminosity @xmath78 . for the sources brighter than @xmath34 erg /
s it equals to : @xmath79 the coefficient in the above formula is @xmath80 for @xmath81 erg / s respectively .
note , that the last number is based on extrapolation of the luminosity functions beyond the luminosity range covered by data . given the shape of the lmxb xlf ( fig.[fig : xrb_xlf ] ) , the effects of statistics are not very pronounced in the considered range of the stellar mass , @xmath82 , and do not affect significantly the @xmath73 relation ( thick solid line and shaded area around it in fig.[fig : x_mass ] ) .
they are more noticeable when considering average luminosity of x - ray sources in a galaxy and lead to an artificial ( unphysical ) dependence of the average luminosity on the stellar mass ( fig.[fig : lx2nx_vs_mass ] ) . as illustrated by fig.[fig : lx2nx_vs_mass ] , the observed dependence agrees well with the theoretical prediction based on the shape of the average xlf .
the deviation of x - ray growth curves from the distribution of the near - infrared light observed in the disks of spiral galaxies has been attributed in section [ sec : gcurves_spir ] to the contribution of high mass x - ray binaries . in fig.[fig : xlf_m101 ] we further support this suggestion by comparing the x - ray luminosity functions of the bulge and disk population in m101 .
the observed difference in the number of sources and the shapes of their luminosity distributions agree both qualitatively and quantitatively with that expected from the `` universal '' luminosity functions of high and low mass x - ray binaries . also shown in fig.[fig : xlf_m101 ] is the expected contribution of the low mass x - ray binaries in the disk of m101 , predicted from its near - infrared luminosity and using the x - ray / nir ratio derived for the bulge .
the smallness of the expected number of lmxbs in the disk is in a good agreement with the fact , that outside the bulge of m101 , the sources exhibit strong concentration towards the spiral arms , with the source density in the inter - arm regions being close to the density of background sources @xcite .
the disk of m101 has a rather small ratio of the stellar mass to the star formation rate , @xmath83 yrs .
correspondingly ( section [ sec : l2m_m2sfr ] ) , the lmxb contribution to the disk population of the compact x - ray sources is small and the luminosity function of the disk sources is close to that of hmxbs . in a general case , the population of the compact x - ray sources can be a superposition of the high and low mass x - ray binaries with comparable relative contributions , resulting to a complex shape of the luminosity distribution , coinciding neither with lmxb nor with hmxb xlf .
this might explain a variety of the shapes of the luminosity functions observed by chandra in different regions of nearby late type galaxies ( e.g. * ? ? ?
* ) , as illustrated by right panel of fig.[fig : xlf_m101 ] , showing the luminosity distribution in a large region of m81 , including both bulge and disk population . due to rather large value of the @xmath84 yrs , comparable fractions of low and high mass x - ray binaries are expected .
the total luminosity function of all sources can be presented as a sum of lmxb and hmxb contributions ( fig.[fig : xlf_m101 ] ) , in agreement with different luminosity distributions of the sources near the spiral arms and in the inter - arm regions observed by chandra @xcite .
the @xmath12 growth curves of two galaxies , ngc1553 and ngc1291 , show significant deviations from the distribution of the nir light ( fig.[fig : gcurves_lx ] ) .
the luminosity distributions for these two galaxies also exhibit largest deviations from the average xlf , although consistent with the latter within statistical uncertainties , according to the kolmogorov - smirnov test ( table [ tab : ks_test ] ) .
these two galaxies might indeed present an example of the population age dependent variations in the shape of the luminosity distribution . in both cases
the deviations from the average lmxb xlf occur at the high luminosity end as might be expected for such variations .
below we assess the statistical significance of the observed deviations and discuss alternative interpretations .
comparison of @xmath11 and @xmath12 growth curves for this s0 galaxy indicates possible variations of the luminosity distribution with the distance from the nucleus .
this is confirmed by fig.[fig : xlf_ngc1553 ] , presenting luminosity functions for the inner and outer regions . in the inner ring ,
the luminosity distribution of the sources is consistent with that of other early type galaxies ( solid smooth line marked `` lmxb '' ) . in the outer ring , on the contrary , it deviates significantly from the average ( long - dashed line ) , both in the slope of the distribution and in the luminosity of the brightest sources .
although the luminosity function of all sources from @xmath85 is consistent with the average lmxb xlf ( table [ tab : ks_test ] ) , the outer ring taken separately gives the value of the kolmogorov - smirnov probability of @xmath86 .
in particular , 3 sources with @xmath87 erg / s are observed at the radial distance @xmath88 from the nucleus .
although their luminosities are not extreme still by a factor of @xmath54 below that of the brightest ulxs in star forming galaxies , they nevertheless stand out in our sample .
the nature of these sources is unclear .
they do not coincide with any of the known globular clusters in the galaxy @xcite . at the time of writing
, there were no repeated observations of ngc1553 , therefore their transient nature can not be established .
the shape of the xlf is reminiscent of that observed in the disks of the spiral galaxies ( cf .
fig[fig : xlf_m101 ] ) .
hypothetically it could be understood as a superposition of the lmxb and hmxb populations ( solid line in fig.[fig : xlf_ngc1553 ] ) .
however , such an explanation would require star formation on the level of @xmath89 m@xmath36/yr .
although on - going star formation in ngc1553 can not be entirely excluded
it forms an interacting pair with ngc1549 @xcite , the required value of sfr exceeds by an order of magnitude the star formation rates detected in some of the s0 galaxies ( e.g. * ? ? ?
moreover , the value of the far - infrared flux observed by iras , @xmath90 and @xmath91 jy ( ned , http://nedwww.ipac.caltech.edu ) , constrains the present star formation rate in ngc1553 by @xmath92 m@xmath36/yr .
another possibility is that the field of ngc1553 is contaminated by a fluctuation in the density of the cxb sources . an indirect evidence in favor of this interpretation is provided by the detection of another very bright source in the periphery of ngc1553 located at @xmath93 from the nucleus ( the source no.38 in the table 1 of @xcite ) .
its luminosity is @xmath94 erg / s assuming that it is a member of galactic population of the compact sources . an optical counterpart has been found in the dss plate @xcite .
the blue magnitude , listed in the usno-2 catalog , @xmath95 , would correspond to the luminosity @xmath96 l@xmath36 if the source was located at the distance of ngc1553 .
obviously , this source is a background agn ( a foreground object in the galaxy is excluded by the @xmath97 ) .
as the expected number of such bright background objects is @xmath98 , its presence might indicate an enhancement in the number density of cxb sources in the filed of ngc1553 .
detection of optical counterparts of other apparently luminous sources would prove this hypothesis , however optical observations inside @xmath99 effective radii of the galaxy are complicated by its optical emission . note , that in order to account for the number of bright sources observed in the outer ring ( fig.[fig : xlf_ngc1553 ] ) , a significant over - density of the cxb sources , by a factor of @xmath100 , on arcmin angular scale is required , which is probably too large to be explained in terms of the average angular correlation function of cxb sources @xcite . in conclusion , we note , that although ngc1553 clearly stands out in our sample from the point of view of the @xmath12 growth curve and the shape of the luminosity function , its growth curve for the number of sources and x / m@xmath1 ratios are within the dispersion of the values observed in other nearby galaxies ( table [ tab : xray ] , figs.[fig : gcurves_nx ] , [ fig : x_mass ] , [ fig : x2mass_vs_type ] ) .
we give in table [ tab : xray ] the x - ray and near - infrared parameters for both the inner ring and the entire @xmath85 region .
ngc1553 data were not included in constructing the average luminosity function ( fig.[fig : xrb_xlf ] ) and determination of its best fit parameters ( table [ tab : xlf ] ) .
its influence on the shape of the average luminosity function is illustrated in the right panel in fig.[fig : xlf_ngc1553 ] , showing that associated changes are well within the statistical and systematic uncertainties .
the deviations of the @xmath12 growth curve from the nir profile are significantly less pronounced and are within about @xmath101 .
the examination of the luminosity distribution in different annuli did not reveal any peculiarities , similar to those observed in ngc1553 .
the overall luminosity distribution ( fig.[fig : xlf_comb_uxlf ] ) agrees very well with the average lmxb xlf at low luminosities , @xmath102 erg / s , but appears to have a too abrupt cut - off above this value . to characterize it quantitatively
, we note that the average xlf , normalized to the total number of sources at the the adopted completeness limit , @xmath103 erg / s , predicts @xmath104 sources above the luminosity of the brightest observed source , @xmath105 erg / s . for a poisson distribution with expectation value of 2.7 ,
the probability to draw zero is 6.7% , i.e. the case of ngc1291 presents @xmath101 deviation .
this value of probability is not small enough to claim presence of statistically significant deviations from the average xlf
. the luminosity function of lmxb sources in the milky way and x - ray / m@xmath1 ratios are by a factor of @xmath2 lower than the main group of galaxies on all plots ( fig .
[ fig : xlf_cum ] , [ fig : x_mass ] , [ fig : x2mass_vs_type ] ) . as the x - ray / m@xmath1 ratios for the milky way were derived in a different way than for other galaxies , they are subject to different systematic uncertainties .
we discuss below various factors , affecting its x - ray / nir ratios .
the stellar mass of the milky way was estimated using the k - band growth curve and mass - to - light ratio of m31 , which is sufficiently similar to our galaxy in the morphological type . in the growth curve analysis we used the low mass x - ray binaries located at @xmath106 , where the origin of the cartesian coordinate system is located at the galactic center and the x - axis
is directed towards the sun @xcite .
thus , to the accuracy defined by the completeness of the lmxb catalog , our analysis should include @xmath107 of the galaxy .
correspondingly , the nir growth curve of m31 was scaled down by a factor of 2 , resulting in the k - band luminosity of the half of the galaxy @xmath108 l@xmath36 ( calculated from the best fit total k - band magnitude of m31 , @xmath109 , table [ tab : sample ] ) .
this value is consistent with accuracy better than @xmath110 with that expected from the the total k - band luminosity of the milky way , @xmath111 l@xmath36 , obtained from 3d modeling of the dirbe data @xcite .
the k - band mass - to - light ratio obtained from the optical color of m31 is @xmath112 .
this value is close , but not identical to the value obtained by @xcite for the solar neighborhood , @xmath113 .
it should be mentioned , though , that the latter value was derived for the disk population of the galaxy and might not be appropriate for the older stellar population of the bulge .
note , that higher mass - to - light ratio would increase the stellar mass estimate , decrease of the x - ray / mass ratio , and therefore would result in a larger discrepancy between the milky way and other galaxies .
an important factor is the completeness of the lmxbs catalog .
@xcite used the sources with known distances and reliable optical identification . comparing with the stellar mass distribution in the galaxy
, they concluded , that their sample is reasonably complete above @xmath114 erg / s and within @xmath115 kpc from the sun . in computing the final luminosity function and total x - ray luminosity of the galaxy they introduced luminosity dependent volume correction using the mass model of the galaxy . on the other hand we computed the x - ray / nir ratios using the sources located at @xmath106 in the range of projected galactocentric distances @xmath116 , which , strictly speaking , is not identical to the definition of the completeness region in @xcite .
it is not clear , whether this procedure could result in an underestimate of the milky way luminosity by a factor of @xmath2 . on the other hand ,
the value of the x - ray - to - mass ratio , derived in this paper , @xmath117 erg / s/@xmath118 is close but somewhat lower than the one obtained by @xcite , using different method , @xmath119 erg / s/@xmath118 .
finally , @xcite used the source luminosities averaged over 5 year period of the asm / rxte observations .
depending on the properties of the collective x - ray light curve of the galaxy this might lead to a difference between their average value and the value to be most probably observed in a snapshot .
the calibration of the @xmath120sfr and @xmath73 relations obtained by @xcite and in the present paper allows us to estimate the total energy output of x - ray binaries throughout the life time of a galaxy .
the total energy output of hmxbs in the chandra passband can be estimated as : @xmath121 where @xmath122 is the time dependent star formation rate of stars more massive than 5 m@xmath36 and describes the star formation history of the galaxy , @xmath118 its present stellar mass , @xmath123 fraction of the total mass of stars formed during the life time of the galaxy , which presently resides in stars ( of all masses ) , @xmath124 is the imf - dependent mass fraction of stars more massive than 5 m@xmath36 and accounts for the fact , that @xmath12sfr was calibrated for formation rate of stars @xmath125 m@xmath36 : @xmath126 for the `` extended '' miller - scalo imf @xcite , @xmath127 for @xmath128 , and @xmath129 for @xmath130 , @xmath131 . studying the intermediate redshift ( up to @xmath132 ) starburst galaxies observed by chandra in the hubble deep field north @xcite showed that the calibration of the @xmath133sfr ratio based on the local galaxies is valid for these distant galaxies as well . for this reason the @xmath133sfr ratio can be taken outside the integration in eq.([eq : e_hmxb ] ) with @xmath134 erg / s per m@xmath36/yr and assuming @xmath135 , the total energy output of hmxbs is @xmath136 the energy output of lmxbs is : @xmath137 unlike for high mass x - ray binaries , eq.([eq : e_hmxb ] ) , the @xmath12mass relations was measured for nearby galaxies only .
therefore the last equality in the above equation relies on the extrapolation of this relation to high redshifts . as the evolution effects might play an important role in the case of lmxbs , this estimate is significantly less robust than that for hmxbs .
however , the details of the evolution of the luminosity distribution of lmxbs are yet unexplored , and even a crude estimate might be of certain interest . using @xmath138 erg / s per @xmath7 m@xmath36 and assuming @xmath139 yrs : @xmath140 which is close to the value obtained for hmxbs . as low mass x - ray binaries were brighter in the past , this number is a lower limit on their energy output in the chandra passband . for the typical value of the stellar mass in the rich clusters of galaxies , @xmath141 m@xmath36 ,
the expected x - ray luminosity of low mass x - ray binaries is @xmath142 erg / s .
taken at face value , this is a negligible fraction , at the level of @xmath143 few per cent , of the luminosity of the x - ray emitting gas , @xmath144 erg / s .
however , due to different energy spectra of x - ray gas and of x - ray binaries , contribution of the latter might become significant in certain energy ranges .
it is well known that the hard tails of comptonized radiation might be present in the spectra of x - ray binaries .
depending on the spectral state , the spectrum of the comptonized component extends to @xmath145 kev ( low state ) or to the @xmath146 mev energies in the high state of black hole binaries .
importantly , in the low state , the hard spectral component is observed in the spectra of both black hole and neutron star binaries ( e.g. * ? ? ?
* ; * ? ? ?
the hard x - ray spectral component might amplify the relative contribution of x - ray binaries at high x - ray energies , @xmath147 kev and higher .
an accurate calculation of this contribution would require knowledge of the x - ray binaries luminosity distribution in hard x - rays , far beyond the chandra bandpass . for a crude ,
an order of magnitude estimate we note , that in the low spectral state , corresponding to the luminosities below @xmath148 erg / s , the luminosity above @xmath149 kev is at the very least equal to that emitted in the chandra bandpass . given the shape of the lmxb luminosity function , the sources with @xmath150 erg / s contribute @xmath28 to the combined x - ray luminosity of lmxbs .
thus , only due to x - ray binaries in the low spectral state the luminosity above @xmath151 kev should be of the order of @xmath152 erg / s .
this is , obviously , a lower limit to the actual value of the total hard x - ray luminosity due to lmxbs . on the other hand , for typical temperatures of the x - ray gas in the clusters of galaxies , @xmath153 kev , of the order of @xmath154 of the total luminosity
is emitted above @xmath151 kev , i.e. @xmath155 erg / s which is comparable to the lower limit on the luminosity of lmxbs . at higher energies
the situation becomes even more favorable for lmxbs . as the emissivity of x - ray gas falls off with radius quicker , than the number density of the galaxies ,
the relative contribution of lmxbs in hard x - ray energy domain will be further enhanced in the outer parts of the clusters .
we should note , however , that the hard x - ray emission may be dominated by the low luminosity agns , depending on their actual frequency and the luminosity distribution .
we analyzed population properties of compact x - ray sources in nearby external galaxies of various morphological types and of the lmxb sources in the milky way .
we focused our analysis on the old stellar systems early type galaxies and the bulges of spiral galaxies .
this ensures that the compact x - ray sources in the external galaxies are dominated by low mass x - ray binaries . in the case of the milky way we explicitly selected lmxb sources based on the results of @xcite .
our findings can be summarized as follows . 1 . for all galaxies the azimuthally averaged spatial distribution of the number of lmxbs follows closely the distribution of the near - infrared light ( fig.[fig : gcurves_nx ] ) .
our analysis covered the central parts of the galaxies , out to @xmath2 effective radii in early type galaxies and @xmath156inside the inner bulge in spiral galaxies . with the exception of ngc1553 and , to lesser
extend , ngc1291 ( section [ sec : individual ] ) , the same is true for the distribution of their combined luminosity in the chandra bandpass , @xmath157 kev ( fig.[fig : gcurves_lx ] ) .
+ in the disks of spiral galaxies , on the contrary , the distribution of the number and combined luminosity of the compact sources deviates significantly from the nir profile .
the deviation is caused by the contribution of hmxbs , as evidenced by the example of the milky way ( fig.[fig : gcurves_mw ] ) and by the comparison of the luminosity functions of the disk and bulge in m101 and m81 ( fig.[fig : xlf_m101 ] ) .
considering galaxies as whole , the total number of lmxbs and their collective luminosity are directly proportional to the stellar mass of the host galaxy ( fig.[fig : x_mass ] ) , the latter calculated from the k - band luminosity using color - based correction of the mass - to - light ratio ( eq.([eq : m2l ] ) ) .
+ there is statistically significant dispersion between values of x / m@xmath1 ratios , with the fractional rms of @xmath28 and @xmath29 for @xmath20 and @xmath19 respectively ( fig.[fig : x2mass_vs_type ] ) .
the accuracy of the color - based correction of the k - band mass - to - light ratio in the present analysis is insufficient to establish reality of the observed dispersion .
neither it is sufficient to confirm or rule out the possibility of a morphological type - dependent trend in the x / m@xmath1 ratios .
if present , its effect does not exceed a factor of @xmath2 ( fig.[fig : x2mass_vs_type ] ) .
+ on average , the number of sources with @xmath158 erg / s and their combined luminosity are related to the stellar mass as follows : @xmath159 the precise values of the coefficients in these formulae depends somewhat on the weighting method ( section [ sec : x_nir_ratios ] ) .
the x - ray/@xmath33 ratios show clear dependence on the morphological type ( fig.[fig : lx2lnir_vs_type ] ) .
the average values are : @xmath160 where the first and the second number in the parenthesis correspond to average values for late and early type galaxies respectively .
the average values for all galaxies are @xmath161 sources and @xmath162 erg / s for @xmath11 and @xmath12 respectively .
these numbers obviously depend on the content of our sample .
+ even stronger dependence on the morphological type should be observed at shorter wavelength , e.g. in the b - band , often used to characterize relation between x - ray and optical properties of the galaxies .
the shape of the luminosity function of the lmxbs is similar in different galaxies ( fig.[fig : xlf_cum ] , [ fig : xlf_comb ] , [ fig : xlf_universal ] ) .
the data on individual galaxies , with a possible exception of ngc1553 ( section [ sec : individual ] ) , are broadly consistent with the average luminosity function of lmxbs , derived using all galaxies from our sample ( fig.[fig : xlf_comb_uxlf ] , table [ tab : xlf ] , [ tab : ks_test ] ) .
the average lmxbs xlf has a complex shape and differs significantly from that of hmxbs ( fig.[fig : xrb_xlf ] , table [ tab : xlf ] ) .
it is consistent with a power law with differential slope of @xmath3 at low luminosities , gradually steepens above @xmath163 and has a rather abrupt cut - off at @xmath47 . in the @xmath164 luminosity range it is approximately represented by a power law with differential slope of @xmath165 .
the normalization of the luminosity function found in individual galaxies is proportional to the stellar mass ( fig.[fig : xlfnorm_vs_type ] ) .
the luminosity of the brightest sources in our sample of old stellar systems ( with total stellar mass of @xmath55 m@xmath56 ) does not exceed the value of @xmath166 erg / s .
the maximum luminosity of hmxb sources is by a factor of @xmath54 larger .
this can be used to diagnose on - going star formation .
relative contributions of low and high mass x - ray binaries to the population of compact sources in a galaxy is defined by the ratio of its stellar mass to the present value of star formation rate , according to the formulae given in section [ sec : l2m_m2sfr ] . 7 .
calibration of the @xmath73 relation derived in the present paper allows one to use the x - ray luminosity of a galaxy due to low mass x - ray binaries as an independent stellar mass indicator .
applicability of this calibration to distant galaxies at larger redshifts is yet to be established and might require proper account for the x - ray binary evolution effects .
the total energy outputs of lmxbs and hmxbs in the chandra passband during the life time of a galaxy are of the same order , @xmath167 erg . in the case of lmxbs , this number does not take in to account the binary evolution effects and , most likely , is a lower limit . | using results of chandra observations of old stellar systems in eleven nearby galaxies of various morphological types and the census of lmxbs in the milky way , we study the population of low mass x - ray binaries and their relation to the mass of the host galaxy .
we show that the azimuthally averaged spatial distribution of the number of lmxbs and , in the majority of cases , of their collective luminosity closely follows that of the near - infrared light . considering galaxies as a whole
, we find that in a broad mass range , @xmath0 , the total number of lmxbs and their combined luminosity are proportional to the stellar mass of the host galaxy . within the accuracy of the light - to - mass conversion
, we can not rule out the possibility of a weak dependence of the x / m@xmath1 ratio on morphological type .
however , the effect of such a dependence , if any , does not exceed a factor of @xmath2 .
the luminosity distributions of lmxbs observed in different galaxies are similar to each other and , with the possible exception of ngc1553 , are consistent with the average luminosity function derived from all data .
the average xlf of lmxbs in nearby galaxies has a complex shape and is significantly different from that of hmxbs .
it follows a power law with a differential slope of @xmath3 at low luminosities , gradually steepens at @xmath4 and has a rather abrupt cut - off at @xmath5 .
this value of the cut - off luminosity is significantly , by an order of magnitude , lower than found for high mass x - ray binaries .
galaxies : fundamental parameters galaxies : elliptical and lenticular galaxies : spiral x - rays : galaxies x - rays : binaries [ cols="<,>,>,>,>,>,^,>,>,^ " , ] + a observed ; b corrected for the effects of statistics ; + @xmath6 in units of src per @xmath7 m@xmath8 ( l@xmath8 ) ; @xmath9 in units of @xmath10 erg / s per @xmath7 m@xmath8(l@xmath8 ) ; + @xmath11 and @xmath12 number and total luminosity of x - ray sources brighter than @xmath13 erg / s [ tab : x2mass ] the number of sources with luminosity above @xmath13 erg / s and their total luminosity is plotted as a function of the stellar mass in fig.[fig : x_mass ] . early and late type galaxies are plotted by different symbols . to further illustrate that there is no significant difference between early and late galaxies we divide the early type galaxies into smaller annuli ,
whose mass is comparable to the typical masses of bulges in spiral galaxies .
these are plotted in fig.[fig : x_mass ] as open symbols .
we fit the @xmath11mass and @xmath12mass relations with a power law @xmath14 using all galaxies from our sample ( table [ tab : xray ] , solid symbols in fig.[fig : x_mass ] ) .
we use @xmath15 minimization , which is strictly applicable in the limit of a large number of sources per galaxy .
this is correct for most of the galaxies from our sample ( table [ tab : xray ] ) , therefore we do not expect any significant bias in the best - fit slope .
the best fit values of the slope are consistent with unity : @xmath16 and @xmath17 for the luminosity and the number of sources respectively .
we therefore fix the slopes at @xmath18 and consider average @xmath19 and @xmath20 ratios . to estimate these we use the unweighted average @xmath21 .
the results are given in table [ tab : x2mass ] separately for early and late type galaxies and for all galaxies .
note that the average ratios given in table [ tab : x2mass ] are not identical to those inferred from eq.([eq : x_nir_from_uxlf ] ) , obtained integrating the average xlf , especially for @xmath19 ratio .
this discrepancy is due to difference between the expectation values of @xmath21 and @xmath22 in the presence of systematic errors .
the systematic errors are present in both @xmath23 , due to transformation to the same @xmath24 based on the average xlf , and in @xmath25 , as they are derived from near - infrared luminosity using an average color - based correction to mass - to - light ratio .
the fact that we are dealing with rather small samples ( e.g. 4 galaxies of late morphological type ) also plays a role . for the same reason , yet different values would be obtained if one used weighted estimates for @xmath26 , e.g. @xmath15 minimization technique .
these factors should be taken into consideration when comparing the @xmath27 ratios .
we show in fig.[fig : x2mass_vs_type ] the @xmath19 and @xmath20 ratios for individual galaxies as a function of the morphological type .
as can be seen , there is statistically significant dispersion , of the order of @xmath28 and @xmath29 from the mean value for @xmath20 and @xmath19 respectively .
a larger dispersion for @xmath19 can probably be explained by the larger correction factors for the number of sources and , respectively , the stronger dependence of the correction factors on the details of the individual xlfs .
fig.[fig : x2mass_vs_type ] and table [ tab : x2mass ] indicate a possible dependence of the @xmath20 ratio on the morphological type .
this can be in part caused by the statistical effects ( section [ sec : stat ] ) .
indeed , the bulges of spiral galaxies have smaller mass than early type galaxies , consequently , smaller numbers of luminous lmxb sources and , therefore , are more subject to the effects of statistics . that these effects can contribute to the observed trend
is demonstrated by the last column in table [ tab : x2mass ] where average @xmath20 ratios , corrected for the statistical effects , are listed . an additional argument in favor of this explanation is that a similar trend is not observed for the @xmath19 , ratio ( fig.[fig : x2mass_vs_type ] , table [ tab : x2mass ] ) . on the other hand , the @xmath19 ratios are more sensitive to the details of the individual luminosity functions and therefore are less robust .
another factor might be insufficiently accurate calibration of the near - infrared mass - to - light ratios . as discussed in section [ sec : nir_mass ] , the color - based correction was obtained under the assumption of the universal initial mass function and might fail , if the morphological type dependent variations in the imf are present @xcite .
such variations can not be excluded a priori , given the broad range of morphological types in our sample . yet
another source of uncertainty is that in computing the @xmath31 ratios integrated colors of the galaxies were used .
as the galaxies do show color gradients @xcite , whose amplitude is type dependent and is largest for late types , @xmath32 , this introduces additional uncertainty in the relative values of the @xmath31 ratio derived for early and late type galaxies . to conclude , due to a number of uncertain factors involved , the possibility of a weak dependence of the x - ray / mass ratios on the morphological type can be neither rejected nor confirmed with sufficient confidence . however ,
if present , it does not exceed a factor of @xmath2 .
interestingly , the x - ray/@xmath33 ratios do show clear dependence on the morphological type ( fig.[fig : lx2lnir_vs_type ] ) .
this is obviously caused by dependence of the near - infrared mass - to - light ratio on the morphological type of the galaxy .
the correlation of the x - ray to optical light ratios with the morphological type should be significantly more pronounces at shorter wavelength , e.g. in the b - band , due to significantly stronger type dependent variations of the mass - to - light ratio in the optical band @xcite . |
as intermediate mass stars ( @xmath018 m@xmath10 ) exhaust the helium in their cores during the late stages of stellar evolution , they begin their ascent of the asymptotic giant branch ( agb ) .
these stars consist of an inert c / o core surrounded by concentric shells of helium and hydrogen . during the early agb ( e - agb ) phase , the evolution is driven by shell hydrogen burning .
stars are typically oxygen - rich during the e - agb phase , with a photospheric c / o ratio @xmath11 . eventually , the helium shell ignites , beginning the thermally - pulsing ( tp - agb ) phase .
subsequent evolution on the agb consists of relatively long periods of hydrogen burning , punctuated at regular intervals by brief helium shell flashes ( thermal pulses ) that rapidly convert the helium to carbon , dramatically increasing the star s brightness for a brief period @xcite .
these thermal pulses drive large - scale convective zones in the stellar interior which can dredge - up " nuclear - processed material , most notably newly created carbon , to the stellar surface , leading to changes in the observable surface chemistry of the star .
this is the third dredge - up for intermediate mass stars , and the result is that for stars with @xmath12 m@xmath10 , the surface c / o ratio changes to become @xmath13 at which time the star is referred to as carbon - rich . for stars with @xmath14 m@xmath10 ,
the temperature at the base of the convective shell is sufficiently high ( @xmath15 k ) to burn the c into n , a process known as hot - bottom burning ( hbb ) , " and the star remains o - rich .
in addition , hydrodynamic instabilities drive shock waves through the stellar interior , levitating the outer layers of the star into cooler regions where dust grains can condense .
radiation pressure then drives the dust ( and the gas to which it is collisionally coupled ) into the interstellar medium ( ism ) .
these pulsations cause agb stas to exhibit complicated , multi - periodic variations in brightness @xcite .
in addition , this process can cause an agb star to lose mass at rates up to @xmath16 m@xmath10 yr@xmath17 @xcite .
this makes agb stars one of the dominant sources of dust in the universe , and a key component in the chemical evolution of galaxies . toward the end of its agb lifetime
, this mass loss can completely enshroud a star with a thick circumstellar dust shell .
these extreme " agb stars are nearly undetectable in the optical , and because the light from the star is dominated by thermal emission from the circumstellar shell , can not be classified as either o - rich or c - rich based on their near - ir photometry @xcite .
most of them are believed to be c - rich , but the brightest should be so massive that hbb has left them o - rich @xcite . without spectroscopic confirmation , we do not propose classifications for specific sources of this class in this paper .
the _ surveying the agents of a galaxy s evolution _
( sage ) survey @xcite is an unbiased @xmath18 survey of the large magellanic cloud ( lmc ) using the irac ( 3.6 , 4.5 , 5.8 , and 8.0 @xmath7 m ) and mips ( 24 , 70 , and 120 @xmath7 m ) , instruments aboard the spitzer space telescope @xcite , intended to trace the life - cycle of the baryonic matter in the lmc .
this wavelength regime makes sage extremely well suited to studies of evolved stars .
the lmc makes an ideal target for studies of stellar populations , as its high galactic latitude minimizes the foreground contamination , and its distance of @xmath050 kpc @xcite makes it possible to resolve individual stars while simultaneously being able to neglect its 3-dimensional structure .
mira variables ( a subset of agb stars ) have been known to follow a linear relationship in magnitude , @xmath19 space for nearly a century ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
@xcite extended the known dataset of agb variables to shorter periods and established that mira - type variables pulsate in the fundamental mode , and that other , higher order pulsational modes exist as well . expanding the number of data points to @xmath01000
, @xcite identified 4 parallel linear sequences in period - luminosity (
p - l ) space , and proposed mechanisms for each .
@xcite , using a sample of @xmath030000 variable evolved stars with period information derived from the _ optical gravitational lensing survey _
( ogle ; * ? ? ?
* ) , further refined these sequences ( see section [ sec : p - l ] ) . also using ogle data
, @xcite identified a new p - l sequence in the secondary , non - dominant variation in the light curves of evolved stars , and proposed this as a means of separating stars dimmer than the tip of the red giant branch ( trgb ) into stars on the first ascent of the rgb and those evolving along the agb . with a dataset of comparable size derived from the macho survey @xcite
, @xcite further investigated the multi - periodic nature of stars on the sequences identified by @xcite . using early results from the sage survey
, @xcite extended work on the p - l relation in agb stars by investigating the dependence of the calculated slope of the relation on wavelength .
our survey combines the sage project s photometry archive and the macho variability dataset to produce the largest sample of variable evolved stars , with photometry measured in multiple near - ir wavelengths , to date .
the unprecedented size of our dataset , combined with our extensive spectral coverage , reveals new patterns and details in the relationship between luminosity , variability , and spectral energy distribution amongst evolved stars .
this paper is organized as follows : [ sec : data ] details our source selection process , the identification of the multiple sequences observed in period - luminosity space , and the division of our sample into 6 photometrically determined categories . [ sec : results ] presents the quantitative and qualitative effects observed within our sample
. [ sec : disc ] compares our current study to previous work in the field , and
[ sec : conclusions ] summarizes the conclusions of the current work .
the sage survey observed the lmc in two epochs , three months apart . from each epoch , two source lists , an archive and a catalog , were generated .
the catalog stresses reliability over completeness and requires more stringent data quality standards ( higher signal to noise , less crowding in the field ) for inclusion of a source than the archive .
the standards for source inclusion in both archive and catalog are detailed in @xcite . in 2009 ,
the sage team released the sage mosaic photometry catalog and archive to the spitzer science center . by combining data from all epochs of the sage survey ,
the mosaic photometry catalog and archive are more complete than the individual epochs , and photometric errors are reduced .
the mosaic photometry data are well documented in the sage data products description document .
the sage mosaic photometry catalog contains @xmath06.5 million sources , and the archive contains @xmath07 million .
we extract a list of 32744 sources from @xcite s catalog of evolved long period variables ( lpvs ) in the lmc with a counterpart in the sage mosaic photometry archive .
we utilize the archive in order to maximize sample size .
we use a matching radius of 2 , and keep only the closest sage source to a given macho source .
we do not include the @xmath011000 sources @xcite identify as the one - year artifact , " which have an artificial period of 365 days caused by macho s annual observing schedule .
the database merging process can create duplicate entries for some sources , due to slight ( @xmath20 ) shifts in position between sage epoch 1 , epoch 2 and the single frame mosaic photometry . by requiring that every source in our list have a unique macho field.tile.sequence identifier , we find @xmath0500 of these duplicate matches in our dataset and cull them from our final list .
based on the definitions used by @xcite , we assign stars to one of 6 roughly parallel sequences in period-@xmath1 space ( fig .
[ fig : pl_seq ] ) .
our definitions are detailed in appendix [ app : p_l ] .
we are left with a final dataset of 30747 evolved stars with well - defined macho periods and high - quality near ir photometry from 2mass and sage .
table [ tab : pops ] compares the number of sources in the present study to those in @xcite , @xcite , and @xcite .
our entire dataset is available online .
table [ tab : cat ] presents the photometric information for a few sources from our sample as a guide to the format of the online table .
@xcite identified 5 parallel period - luminosity sequences in the macho dataset and proposed underlying physical mechanisms for them .
in order of increasing period , the sequences of pulsating stars were named a , b , and c , with sequences d and e exhibiting variation due to an unknown mechanism .
sequence c was identified as pulsation in the fundamental mode , sequence b as the first and second overtones , and sequence a as the third overtone . @xcite and @xcite split sequence b into two sequences ( c@xmath21 and b ) representing pulsation in the first and second overtone respectively .
@xcite retained the names of sequences d and e from @xcite , but renamed the pulsation sequences 14 in order of _ decreasing _ period , in order to easily accommodate shorter period sequences , and align the naming convention with increasing pulsation overtone .
we follow the naming convention of @xcite because of this parallel with theory . numbering the sequences in order of decreasing period
means that higher sequence numbers generally correspond to higher order pulsational modes , aligning the empirical naming convention and a theoretical explanation more gracefully .
@xcite has provided convincing evidence that sequence e consists of ellipsoidal binaries , systems in which the red giant member has filled its roche lobe .
recent work ( e.g * ? ? ?
* ; * ? ? ?
* ) has demonstrated that the variability seen on sequence d can not be due to binarity , but its actual cause remains unknown .
we keep the alphabetic names for these sequences to separate them from the stellar pulsation sequences .
table [ tab : seq_conv ] summarizes these conventions , and figure [ fig : pl_seq ] illustrates these sequences .
the precise definitions of these sequence are discussed in appendix [ app : p_l ] . before being photometrically classified , all sources were de - reddened in the _ j _ , _ h _ and @xmath1 bands following @xcite .
specifically , we use corrections of @xmath22 , @xmath23 , and @xmath24 mag . we define agb candidates using 2mass and sage photometry @xcite .
we identify a source as an agb candidate by requiring that @xmath25 we classify sources as low- or moderately - obscured oxygen - rich and carbon - rich agb candidates based on their near - ir photometry .
specifically , a source is classified as an o - rich candidate if it lies leftward ( blueward ) of the line @xmath26 in the @xmath1 vs. @xmath27 color - magnitude diagram ( cmd ) , and a c - rich candidate if it lies rightward ( redward ) of this line @xcite .
extreme agb stars are defined using their position in the [ 3.6 ] vs. @xmath28[3.6 ] cmd @xmath29 > 3.1 \quad\mbox{and}\quad [ 3.6 ] < 10.5 \label{eqn : x_def}\ ] ] or the irac [ 8.0 ] vs. [ 8.0]@xmath30[3.6 ] cmd if a source lacks a valid _ j _ band magnitude .
these sources are thought to be highly evolved agb stars , enshrouded by thick circumstellar dust shells with high rates of mass loss .
the trgb is located at @xmath1 @xmath31 ( e.g. * ? ? ? * ) or _
i _ = 14.54 @xcite , and is typically identified by a clear fall - off in population density ( e.g. * ? ? ?
however , while the trgb is justified on physical grounds as a maximum luminosity for rgb stars , it does not represent a physically motivated minimum luminosity for stars on the agb .
examination of a plot of @xmath1 magnitude vs. variability period @xcite , shows that the sudden decrease of population density brighter than @xmath1 @xmath31 indicative of the trgb only appears on period - luminosity sequences 3 and 4 .
we thus classify a star as an rgb candidate if it is fainter than the @xmath1 or _
i _ band trgb cutoff , and it lies on either sequence 3 or 4 .
stars dimmer than the trgb on sequences 1 and 2 are classified as o - rich agb candidates .
@xcite proposed a means of separating agb stars dimmer than the trgb on sequence 3 or 4 from the dominant rgb population on these sequences .
their method relied on analysis of non - dominant periods and a distinct p - l sequence with shorter periods than sequence 4 ( this sequence would be called sequence 5 in our nomenclature ) .
however , we do not detect this sequence in either the dominant or secondary pulsation periods of our sample , and thus we do not employ this method .
the use of periods in addition to the dominant two periods available to us grants @xcite finer resolution in separating rgb stars from agb candidates dimmer than the trgb , but the consistency of our results with those of other authors gives us confidence in the broad correctness of our classification .
figure [ fig : pl_chem ] reproduces figure [ fig : pl_seq ] , but this time color coded according to the photometrically determined chemical classification we assign to the stars in our sample .
table [ tab : pops_me ] details the number of each class of star on each p - l sequence
. figure [ fig : seq_mag_hist ] , a histogram of [ 3.6 ] magnitude color - coded by p - l sequence , provides further justification for our classification .
sequences 3 and 4 are dominated by a large population dimmer than @xmath32\approx 12 $ ] . since there is no theoretical expectation for agb stars to pile up at the trgb
, we interpret this as evidence for a large rgb population on sequences 3 and 4 .
sequences 1 and 2 do show evidence of a distinct population of stars at this brightness ( more pronounced in sequence 2 ) , but neither shows the dramatic rgb population found in sequences 3 and 4 .
we interpret this as a small amount of rgb contamination in sequences 1 & 2 .
all four sequences show a stellar population with a peak at @xmath32 \approx 11.2 $ ] , coincident with our photometrically determined o - rich agb candidate population .
in addition , the stars photometrically classified as c - rich stars are visible in sequences 1 and 2 as a peak in the magnitude distribution brighter than @xmath32\approx 10 $ ] .
this population is also visible in sequence 3 but absent from sequence 4 .
these trends in the luminosity functions of the p - l sequences mirror and lend support to the photometrically determined chemical classifications detailed in table [ tab : pops_me ] . highly evolved agb stars and young stellar objects ( ysos ) have similar mid - ir colors .
@xcite suggest that the region of the [ 8.0 ] vs. [ 8.0]@xmath30[24 ] cmd defined by @xmath33-[24 ] > 2.2 \quad\mbox{and}\quad [ 8.0 ] > 11 - 1.33 \times ( [ 8.0]-[24])\ ] ] constitutes a region in color - magnitude space dominated by ysos .
as all of our sources have well determined macho periods , none of our sample are ysos . of our @xmath030000 stars , we find that only 150 sources ( 109 rgb , 3 extreme agb and @xmath040 o - rich and c - rich agb stars ) fall in this region .
we note that the optical wavelengths used by the macho survey biases our dataset to under - sample this region of cmd space , but the extremely small fraction of our sample that falls into this region lends support to the criteria proposed by @xcite .
our sample consists of 17059 agb candidate stars : 12172 sources photometrically classified as o - rich , 4455 as c - rich , and 432 as extreme agb candidates . based on near - ir photometry and variability period , 13688 sources are classified as stars on the rgb .
the @xmath1 magnitude is often used as a proxy for luminosity for agb stars ( e.g. * ? ? ?
* ; * ? ? ?
the inclusion of spitzer irac and mips data at wavelengths out to 24 @xmath7 m reveals new features in the period magnitude diagram , especially among the reddest sources , the extreme agbs , whose seds peak at 3 @xmath7 m or redward @xcite .
fig [ fig:4band ] compares the 2mass @xmath1 band , the irac @xmath3 , @xmath4 , and @xmath5 @xmath7 m bands , along with the mips @xmath6 @xmath7 m band as luminosity proxies , all plotted against period as in figs .
[ fig : pl_seq ] & [ fig : pl_chem ] .
the @xmath1 band , in the upper left , is a reasonable approximation for the brightness of the low - obscuration rgb , o - rich and c - rich candidates , but comparison with the other panels in fig .
[ fig:4band ] , shows that this approximation fails for the extreme agb candidates .
the rgb , o - rich and c - rich candidates appear in roughly the same positions , relative to to one another , in the 2mass and irac bands .
the extreme sources , however , clearly have their luminosity underestimated by the @xmath1 band .
these sources , the most luminous in our sample @xcite , appear no more luminous than the c - rich candidates at this wavelength .
the irac bands , on the other hand , correctly place the extreme candidates at the brightest end of sequence 1 , the fundamental mode pulsators .
the 24 @xmath7 m band is sensitive to emission from cool dusty envelopes .
the observed flux from the sources surrounded by the most optically thick dust shells is dominated by the emission from their outermost , coolest layers .
for this reason , they are able to produce significant 24 @xmath7 m fluxes ( see fig .
10 , * ? ? ?
the rgb / agb stars with optically thin dust shells are either faint or undetected at 24 @xmath7 m
( for example , @xmath34% of the rgb stars in our sample have valid measured 24 @xmath7 m fluxes ) because their seds peak at much shorter wavelengths .
the 24 @xmath7 m flux is therefore not representative of the intrinsic brightness of these sources .
this poor estimate for intrinsic brightness causes the sequence structure to break down entirely .
using the method of weighted least - squares , linear models of the form @xmath35[eq : pl ] were fit to each of the sequences as a whole , and to each subgroup ( o - rich , etc . ) within each sequence .
periods and amplitudes are from the dominant mode of variability derived by @xcite from macho blue - band lightcurves .
periods were considered free of errors , and because the sage observations are not phase corrected , magnitudes were weighted by the inverse of the quadrature sum of the 1-@xmath36 photometric error and one - half the macho peak - to - peak variability amplitude .
table [ tab : fit ] lists the linear fit parameters thus determined for each of the 6 sequences , using the irac [ 3.6 ] magnitude as the luminosity proxy .
previous work has indicated that mira variables exhibit much lower amplitude variation in the ir than in the optical ( lattanzio & wood ( 2003 ) , in : * ? ? ? * ) , and thus using one - half of the macho amplitude in the error term might overestimate the true scatter in the data due to intrinsic variation . to investigate this , we extracted a sample of @xmath0100 sources from the catalog of @xcite with well - determined macho periods between 150 and 200 days . at these periods
, the three month cadence of the sage observations could sample these stars anywhere between maximum and minimum brightness .
thus the range of variation in brightness between epoch 1 and 2 observations could represent the full mid - ir peak to peak amplitude of variation .
the macho amplitudes for the o - rich and c - rich agb candidates are statistically very similar to the variation seen in this smaller sample .
the small sample from @xcite has a median mid - ir variation of @xmath00.2 mag between sage epoch 1 and epoch 2 .
the median optical amplitudes for all o - rich and c - rich sources together are 0.2 and 0.3 magnitudes respectively .
in addition , we compared the distribution of the macho variability amplitudes to the distribution of the residual scatter ( table [ tab : fit ] ) about the p - l fits .
the residual scatter of @xmath00.5 magnitude about the best fit line ( see table [ tab : fit ] ) is due to the intrinsic variability of the sources .
the sage mosaic photometry is constructed from the two epochs of the sage survey , but lacks the time - stamp information necessary for the phase - correction necessary to compute a true mean magnitude for these variable sources .
some of our sources have been observed near maximum brightness , while other were near their minimum brightness .
the residual scatter about the p - l relations gives us an indication of the scale of the ir variability of our sample .
both quantities followed similar distributions , and were of the same scale .
this justifies our use of the macho optical variability amplitude as a measure of the ir variability of our sources for the estimation of the uncertainty in the ir mean magnitudes in the p - l relations .
figure [ fig : seq1 ] shows the linear fits to the stars in sequence 1 in the [ 3.6 ] vs. period diagram .
sequence 1 is consistent with stellar pulsation in the fundamental mode @xcite , and was chosen because it is the only sequence which contains a significant number of detected extreme agb candidates .
the @xmath2 @xmath7 m irac band was chosen because it is the most sensitive of the sage bands @xcite .
the different trends followed by the different types of evolved star candidates is apparent .
the plotted linear fits are : [ cols="<,^ " , ]
the stars in our sample are assigned to one of 6 p - l sequences using cuts in @xmath1-period space , with a small subset also utilizing _
j_@xmath1 color information .
only the primary period of these multi - periodic objects is used to establish sequence membership .
the definitions for the sequences are shown graphically in figure [ fig : app_1 ] , and described mathematically in table [ tab : pl_def ] .
our definitions are based on those used by @xcite , but are more restrictive ( especially in sequence d , which exhibits a very large scatter about the p - l relation ) to better trace the population density
cores " visible to the eye .
this emphasis sacrifices some sample size , but we note that the scatter in our p - l relations ( of order @xmath00.15 mag ) is dominated by the intrinsic variability of the sources and would not be significantly reduced by introducing more sources by using the secondary periods to assign the un - sequenced sources to one of the existing p - l sequences .
we note that ultimately , our p - l sequence definitions are a judgment call , and slightly different definitons used by other authors are equally valid and can affect the measured p - l fits .
in particular , the trgb cut at a constant @xmath37 leads to a sharp parallelogram shape at the brightest end of the rgb population .
these corners cause the rgb and o - rich agb p - l fits to be less steep than the fit to all the stars in a sequence , regardless of class , which is not affected by this selection bias ( table [ tab : fit ] ) . beginning with these definitions of the p - l sequences , careful study of the panels of figure [ fig:4band ]
led us to introduce some additional modifications to properly classify a few anomalous populations .
a group of 100 c - rich agb candidates , classified as being part of sequence 1 under criterion 1@xmath38 in table [ tab : pl_def ] , was found to remain sub - luminous to the rest of sequence 1 at all wavelengths longer than @xmath1 as well .
they are identified and eliminated by removing c - rich agb candidates from sequence 1 with periods @xmath39 days and @xmath1 @xmath40 .
finally , we identified a population of 60 extreme agb candidates so dust - enshrouded that at @xmath1 they lay along the main body of sequence d , but at @xmath2 @xmath7 m and longer they were clearly part of sequence 1 .
they were properly classified by adding all extreme agb candidates assigned to sequence d or to no sequence with periods between @xmath41 and @xmath42 days and irac [ 3.6 ] @xmath43 to sequence 1 .
note that both of these populations also require their members to be photometrically classified as either c - rich or extreme agb candidates , following the definitions detailed in [ sec : class ] .
lllll sequence 1 & @xmath1 @xmath44 & @xmath1 @xmath45 & @xmath1 @xmath46 & @xmath1 @xmath47 + sequence 1 & @xmath1 @xmath48 & @xmath1 @xmath49 & @xmath1 @xmath50 & @xmath1 @xmath51 + sequence 2 & @xmath1 @xmath44 & @xmath1 @xmath52 & @xmath1 @xmath53 & @xmath1 @xmath54 + sequence 3 & @xmath1 @xmath55 & @xmath1 @xmath52 & @xmath1 @xmath56 & @xmath1 @xmath57 + sequence 4 & @xmath1 @xmath58 & @xmath1 @xmath45 & @xmath1 @xmath59 & @xmath1 @xmath60 + sequence d ( top ) & @xmath1 @xmath61 & @xmath1 @xmath49 & @xmath1 @xmath62 & @xmath1 @xmath63 + sequence d ( bottom ) & @xmath1 @xmath64 & @xmath1 @xmath65 & @xmath1 @xmath66 & @xmath1 @xmath67 + sequence e ( top ) & @xmath1 @xmath68 & @xmath1 @xmath49 & @xmath1 @xmath69 & @xmath1 @xmath70 + sequence e ( bottom ) & @xmath1 @xmath64 & @xmath1 @xmath71 & @xmath1 @xmath72 & @xmath1 @xmath73 [ tab : pl_def ] | we combine variability information from the _ massive compact halo objects _ ( macho ) survey of the large magellanic cloud ( lmc ) with infrared photometry from the spitzer space telescope _ surveying the agents of a galaxy s evolution _ ( sage ) survey to create a dataset of @xmath030000 variable red sources .
we photometrically classify these sources as being on the first ascent of the red giant branch ( rgb ) , or as being in one of three stages along the asymptotic giant branch ( agb ) : oxygen - rich , carbon - rich , or highly reddened with indeterminate chemistry ( extreme " agb candidates ) .
we present linear period - luminosity relationships for these sources using 8 separate infrared bands ( _ j _ , _ h _ , @xmath1 , @xmath2 , @xmath3 , @xmath4 , @xmath5 , and @xmath6 @xmath7 m ) as proxies for the luminosity .
we find that the wavelength dependence of the slope of the period - luminosity relationship is different for different photometrically determined classes of agb stars .
stars photometrically classified as o - rich show the least variation of slope with wavelength , while dust enshrouded extreme agb stars show a pronounced trend toward steeper slopes with increasing wavelength .
we find that o - rich agb stars pulsating in the fundamental mode obey a period - magnitude relation with a slope of @xmath8 when magnitude is measured in the @xmath2 @xmath7 m band , in contrast to c - rich agb stars , which obey a relation of slope @xmath9 . |
the canonical metal - rich globular cluster ( @xmath1 > -0.8 $ ] ) has a horizontal branch ( helium core burning phase of stellar evolution ) which is compressed against the red giant branch to form a `` red clump '' ( e.g. , 47 tuc , carney et al . 1993 ; m 71 , hodder et al .
standard stellar evolution theory and synthetic horizontal branches confirm this morphology ( e.g. , lee et al .
1990 , dorman 1992 ) .
however , it is becoming clear that the situation is more complicated than this canonical picture would have us believe .
first , it is evident that a significant fraction of the rr lyrae variable stars ( horizontal branch stars lying in the pulsation instability strip , hereafter rrl ) in the field near the sun are metal - rich .
preston ( 1959 ) first recognized this fact , and the recent metallicity distribution of layden ( 1994 ) shows that 22% of the field rr lyrae within 2 kpc of the sun have [ fe / h ] @xmath4 , and 9% have [ fe / h ] @xmath5 dex .
furthermore , the kinematics of these stars identify them with the galaxy s thick disk population , and a few rrl with abundances approaching solar may be produced by the thin disk ( layden 1995a ) .
the standard horizontal branch star picture does not explain the presence of stars hot enough to lie in the instability strip at these ages and metallicities .
in addition , several metal - rich globular clusters contain one or more rrl candidates ( cf .
suntzeff et al .
note however that the often - quoted example , v9 in 47 tucanae , has an anomalously high luminosity and long period compared with field rrl of similar abundance ( carney et al .
nevertheless , layden ( 1995b ) showed that the number of rrl per unit progenitor luminosity was roughly equal between the metal rich thick disk field and globular cluster populations .
this comparison was hampered by the incomplete search for rrl among the population of metal - rich globulars .
a second important complication to the canonical picture of metal - rich globular cluster horizontal branches ( hbs ) is the recent observation of blue hbs in the cores of two globulars , and in several open clusters . _
hubble space telescope
_ observations by rich et al .
( 1997 ) showed the globulars ngc 6388 and 6441 ( @xmath1 = -0.60 $ ] and @xmath6 , armandroff & zinn 1988 ) to have long blueward extensions to their otherwise red hbs .
they initially suggested that age or dynamical effects could be responsible for the blue hbs .
however , no deep photometry has been published for these clusters to constrain their ages , and calculations by rich et al .
( 1997 ) indicated that stellar interactions are too infrequent to be the culprit .
they left open the question of the physical cause of the phenomenon .
blue hb stars in younger , more metal - rich open clusters have also been observed ( e.g. , ngc 6791 , kaluzny & udalski 1992 ) .
theoretical studies are also focusing on the blue hb star problem .
sweigart & catelan ( 1998 ) presented three scenarios involving enhanced helium abundance and/or stellar rotation designed to explain the blue hb observations of rich et al .
they emphasized that observational tests including the existence and periods of rr lyrae variables , and the relative numbers of hb and red giant branch stars ( the @xmath7-ratio ) , could discriminate between their scenarios .
they suggest that the blue hb extensions seen by rich et al .
( 1997 ) are extreme cases of the `` second parameter effect '' , in which the color distribution of stars along globular cluster hbs is seen to depend on metallicity and a second parameter , possibly age , helium or cno abundances , or core rotation .
thus , ample motivation exists to study the horizontal branches and rr lyrae content of metal - rich globular clusters . with this in mind , we have undertaken a variable star survey of twelve of the metal - rich globulars listed by suntzeff et al .
( 1991 ) as having poor or no variable star searches . in this paper
, we report our results for the first cluster , ngc 6441 .
if dynamical effects are involved , the high luminosity and central concentration of this cluster ( harris 1996 ) make it among the most likely to harbor rrl .
this prospect is enhanced by the presence of the blue hb stars observed by rich et al .
the best existing ground - based photometry of ngc 6441 is the photographic work of hesser & hartwick ( 1976 ) , which reached only the brightest stars on the horizontal branch . in 2 of this paper
, we describe the observation and data reduction procedures we employed . in
3 we present color - magnitude diagrams of the cluster and nearby field . in
4 we identify @xmath250 variable stars in the cluster , derive their photometric properties , and for some present their mean light curves . in 5 we derive the foreground reddening of ngc 6441 and discuss in detail the properties of the cluster s rr lyrae variables and compare them to the predictions of the sweigart & catelan ( 1998 ) models .
we present our conclusions in 6 , and in the appendix we note details on some of the variable stars .
we obtained time - series images of ngc 6441 using the direct ccd camera on the 0.9-m telescope at cerro tololo inter - american observatory during two runs in may and june of 1996 ( 3 and 8 usable nights , respectively ) .
the tek#3 2048 ccd provided a 13.5 arcmin field of view with 0.4 arcsec pixels .
we used filters matched to the ccd to reproduce the johnson @xmath8 and kron - cousins @xmath9 bandpasses .
we processed the raw images by following the usual procedure for overscan subtraction and bias correction , and we used twilight sky frames to flat - field the data . in each pointing toward the cluster
, we obtained a short exposure ( 30100 sec ) @xmath0 frame pair and a long exposure ( 250300 sec ) @xmath0 pair .
this provided two independent magnitude estimates of the hb stars at each observational epoch , and extended the dynamical range of the observations .
such pointings were obtained 03 times each night .
the time interval between pointings was at least 2 hours , so a typical rr lyrae ( 0.30.8 day periods ) would undergo a significant magnitude change . in total , we obtained 15 pointings toward ngc 6441 . the seeing varied between 1.2 and 3.0 arcsec ( 1.7 arcsec median ) .
figure 1 shows a short - exposure @xmath8-band image of the cluster . the bright star to the west of the cluster
was masked out in all the frames . on a photometric night in may 1996
, we obtained a @xmath0 image pair of ngc 6441 , together with a large number of landolt ( 1992 ) photometric standards .
we also secured an off - cluster @xmath0 pair centered 13.3 arcmin north of the cluster center , outside the tidal radius ( 7.8 arcmin , harris 1996 ) designed to determine the photometric properties of the fore / background starfield .
the standards covered a large range in color ( @xmath10 ) and airmass ( @xmath11 ) , and were obtained at roughly hourly intervals throughout the night .
this time ( @xmath12 ) coverage confirmed that the entire night was photometric , and enabled us to seek and correct for small , slow variations in extinction . in total ,
63 individual standards were observed at 13 independent epochs .
the instrumental aperture magnitudes of these stars ( @xmath13 ) , in bandpasses @xmath14 or @xmath9 , were fit with the equation @xmath15 where @xmath16 is the standard magnitude and @xmath17 ( @xmath18 ) are the fitted coefficients .
the time - dependent terms produced corrections of 0.01 mag or less .
the rms scatter about the adopted fits were 0.010 and 0.013 mag in @xmath8 and @xmath9 , respectively .
the on- and off - cluster @xmath0 pairs obtained on this night were reduced using the daophot ii point spread function ( psf ) fitting photometry package ( stetson 1994 ) .
the interactivity of this package enabled us to select @xmath2170 psf stars well - distributed spatially over the frame .
this was critical in accurately characterizing the significant radial variation in psf across the frame . for each frame , the psf was iteratively improved by fitting and removing faint neighbor stars .
the final psf was applied to the image using allstar .
a second application of allstar yielded a frame with all the stars subtracted except a set of @xmath2220 bright , isolated stars .
this frame was used to derive aperture corrections by subtracting the psf - fit daophot magnitudes of the stars from their aperture magnitudes ( derived using the same software and parameters as those used in measuring the standard stars ) .
the aperture correction was then applied to all the stars measured in the frame , and the standard star relations were applied , yielding standard @xmath0 magnitudes for thousands of stars on the on- and off - cluster fields .
the on - cluster @xmath0 data provides a set of secondary standards for calibrating the other on - cluster images .
the remaining images of ngc 6441 were reduced using a version of the dophot psf - fitting package ( schechter , mateo , & saha 1993 ) which allows for spatial variations in the psf .
instrumental magnitudes were obtained for each frame and combined into instrumental @xmath0 pairs , and then transformed to standard @xmath0 magnitudes via fits to the secondary standards described above .
we grouped the @xmath0 frame pairs as follows : good - seeing frames with ( @xmath19 ) long , and ( @xmath20 ) short exposure times , and poor - seeing frames with ( @xmath21 ) long , and ( @xmath22 ) short exposure times .
this grouping strategy minimized mis - matches due to variable depth and seeing . within each group , we spatially matched stars from the different frames , and combined the photometry for each star using an error - weighted mean .
for the error in each combined stellar magnitude , we adopted the standard error of the mean for the individual measures , which proved to be a more conservative estimate than computing the error directly from the dophot errors .
we combined the mean photometry from groups @xmath19 and @xmath20 ( the long- and short - exposure frames with good seeing ) using an error - weighted mean to produce a single dataset with maximum dynamic range . outside a radius of 1100 pix ( 7.33 arcmin ) from the cluster center ,
elongated star images sometimes registered as multiple stars , compromising the mean photometry and complicating variable star detection ( see 4 ) .
figure 2a shows the ( @xmath23 , @xmath8 ) color - magnitude diagram ( cmd ) for 22,871 stars on the on - cluster frame lying within 1100 pix of the cluster center .
only stars detected on four or more frames , and having combined errors in @xmath8 and @xmath23 of @xmath24 mag and @xmath25 mag , are shown .
table 1 lists the combined photometry for all the bright stars ( @xmath26 mag ) within 1100 pix of the cluster center .
figure 2b shows the corresponding cmd for the off - cluster field .
this diagram represents data from a single @xmath0 frame pair , so the errors are larger .
poor photometry , galaxies , and cosmic rays were rejected by performing 3-@xmath27 cuts on the daophot values of `` @xmath28 '' and `` sharp '' as functions of magnitude . only stars with @xmath29 mag and @xmath25 mag are shown ( daophot errors ) .
several features are apparent in these cmds .
first , both the cluster and field have large numbers of extremely red giant stars .
second , in both cmds one can detect the red clump stars of the bulge at @xmath30 and @xmath31 mag .
this agrees well with photometry of the nearby field mm-5b from the ogle group ( stanek et al .
note that the red clump in figure 2b is shifted redward and faintward relative to that in figure 2a , suggesting a mean reddening difference between the two fields .
third , a prominent feature of the on - cluster cmd is the ngc 6441 red clump , which forms a swath of stars near @xmath32 and @xmath33 mag , running parallel to the reddening vector ( we adopt @xmath34 throughout this paper ) .
finally , the on - cluster cmd shows a population of stars bluer and somewhat brighter than the cluster red clump .
such stars are not present on the off - cluster cmd , and represent the first ground - based view of the extended blue horizontal branch seen by rich et al .
( 1997 ) .
figures 3a and 3b zoom in on the hb regions of these cmds .
the tight distribution of red clump stars in figure 3a paralleling the reddening vector suggests that there is significant differential reddening across the face of the cluster .
this also explains why a tight red giant branch ( rgb ) is not seen ; differential reddening spreads the rgb stars roughly perpendicular to the un - reddened rgb fiducial .
we will consider the reddening problem in more detail in 5.2 .
another important feature is highlighted in figure 3c , where stars from figure 3a lying between 40160 arcsec of the cluster center are displayed .
the excess of blue hb ( bhb ) star candidates seen in figure 2a is evident , indicating that they too are members of the cluster .
the cluster s red clump and red giant branch are better defined in this panel as well .
figure 3d shows another effort to isolate ngc 6441 stars from the field . here
, we have statistically subtracted the cmd of off - cluster stars located between 40160 arcsec of the frame center from the stars in figure 3c . before subtracting , we corrected for a 0.11 mag reddening difference between the fields ( the off - cluster field has the higher reddening ) , which we estimated from the shift in color histograms of stars having values of the reddening - independent parameter @xmath35 between 12.1 and 13.3 ( stanek et al .
1994 ) . for each star in the off - cluster cmd , we removed from the on - cluster cmd the star nearest in color - magnitude space . though the subtraction is over - aggressive for the faint stars ,
the brighter sequences ( including the blue hb , red clump , and red giant branch ) are more clearly shown .
the lack of faint , extremely blue hb stars in figure 3c , relative to their abundance in the cmd of rich et al .
( 1997 ) , comes as a mild surprise . even in cmds where we relax the criteria on the photometric errors
, these stars do not appear at @xmath36 mag .
this suggests that these stars may be radially concentrated toward the cluster center , but we can not state this unequivocally without detailed completeness modeling . a simpler , more direct comparison with the rich et al .
( 1997 ) @xmath37 data will soon be available from the deep , ground - based @xmath37 photometry of the cluster recently obtained by horace smith et al .
( private communication ) . does the prominence of the less extreme bhb ( i.e. , brighter than @xmath38 mag ) vary with radial distance from the cluster center ?
rich et al . ( 1997 ) found no evidence for a change in the number ratio of bhb to red clump stars with radius over their small _ hst _ field of view ( @xmath2100 arcsec , with the pc chip centered on the cluster ) , and interpreted this as evidence against the production of bhb by dynamical effects .
we can extend this search to larger radii using our @xmath0 photometry .
for this purpose , we defined bhb stars as having @xmath39 and @xmath40 mag , and red clump stars as having @xmath41 and @xmath42 mag .
we defined the number fraction @xmath43 where the @xmath21 and @xmath44 subscripts indicate stars counted from the cluster and field photometry sets , respectively . for stars in the radial annulus 50100 arcsec
, we obtained @xmath45 , while for radii of 100240 arcsec , we obtained @xmath46 . ,
@xmath47 , @xmath48 , @xmath49 for the inner annulus . for the outer annulus , they are 21 , 7 , 598 , and 77 , respectively . ]
the blue hb stars are a factor of two more prominent in the inner annulus .
the fokker - planck simulations of rich et al .
( 1997 ) indicated that bhb stars produced from dynamically - stripped rgb stars should be concentrated within 4 core radii ( 26 arcsec , harris 1996 ) .
the lack of bhb concentration in the rich et al .
( 1997 ) observations , and our observed concentration at larger radii do not appear to be consistent with simple pictures of bhb star production through tidal stripping of normal rgb stars by close encounters in the dense cluster core .
in many clusters , relative radial population gradients are attributed to the effect of mass segregation .
the relaxation time for ngc 6441 at the half - mass radius , @xmath20.5 gyr at 38 arcsec ( harris 1996 ) , is much longer than the typical lifetime of an hb star ( @xmath20.1 gyr ) , but mass segregation could be at play if the bhb stars are members of binary systems .
bailyn ( 1995 ) reviewed two scenarios for producing extreme bhb stars from binaries : the merger of helium white dwarf binaries , and mass transfer in wide binaries in which the hydrogen envelope of a red giant evolving close to the rgb tip is donated to the secondary star .
bailyn et al . ( 1992 ) found the latter scenario to be particularly appealing in explaining the radial distribution of extreme bhb stars in the globular @xmath50 centauri , and the scenario may be relevant to ngc 6441 as well .
it explains the radial concentration of bhb far from the cluster center , and the broad color distribution of the bhb population ( the latter due to a range in mass loss , produced by a range in orbital separations ) . on the negative side , the scenario does require ngc 6441 to have a much larger population of binaries with the appropriate orbital periods ( a few hundred days ) than metal - rich clusters with `` normal '' red - clump hbs .
observationally , there are few constraints on the distribution of binary periods in globular clusters .
theoretically , the primordial period distribution of a cluster s binary stars could be linked to the angular momentum of the gas cloud which formed the cluster .
also , dynamical evolution of the binary populations will vary from cluster to cluster ( e.g. , bailyn et al . 1992 ) .
thus there seems to be no argument against the existence of large variations in the binary period distribution from one cluster to another .
finally , the wide - binary scenario makes a prediction : most bhb stars should be members of a binary system , with radial velocity variations of a few tens of km s@xmath51 .
such observations should be obtainable using multi - fiber spectrographs on the new generation of large telescopes .
we used the welch & stetson ( 1993 ) variability index , @xmath52 , to characterize the likelihood that a given star is a variable .
the value of @xmath52 is computed for each star from the individual @xmath8 and @xmath9 magnitudes and their errors . for a non - variable star , @xmath52 approaches zero as the number of observations becomes large , whereas for a variable , @xmath52 approaches a constant value . to minimize the effects of seeing and exposure depth ( i.e. , the blending of two nearby stars into a single measurable image as the seeing and/or signal - to - noise conditions worsen ) , we employed the frame groupings described in 2 and computed @xmath52 for each star detected more than four times in each group .
groups @xmath53 and @xmath54 were useful in detecting variable stars near the cluster ( @xmath55 pix ) , and groups @xmath56 and @xmath57 were useful outside @xmath58 pix . the morphology of the points in the ( @xmath52 , @xmath8 ) plane led us to classify stars as probable variables ( @xmath59 ) and possible variables ( @xmath60 ) . when plotted on the cluster cmd , it became clear that three classes of variables are present : extremely red variables ( mira , semi - regular , and irregular variables , hereafter lpvs ) , blue variables ( possible rr lyrae stars ) , and a few other variables scattered around the cmd .
we focused on these stars as follows : all stars with @xmath61 mag and @xmath62 ( lpv candidates ) , all stars with @xmath63 mag and @xmath62 ( rrl candidates ) , and all other stars with @xmath64 .
we enhanced our search for cluster rrl stars by supplementing the rrl candidates with stars in the subset @xmath65 mag , @xmath66 , and @xmath67 .
we produced analogous lists from the data in groups @xmath54 , @xmath56 , and @xmath57 , and merged them into a final list of 115 variable star candidates .
we then plotted the @xmath8 and @xmath9 magnitudes as a function of time for each candidate .
the lpv candidates were characterized by constant or slowly - varying @xmath8 and @xmath9 magnitudes , with distinct jumps between the may and june data .
two candidates near the cluster center were sufficiently crowded , and their light curves scattered , that we could not unambiguously characterize them as lpvs ; their positions and magnitudes are recorded in table 5 ( stars sv6 and sv7 ) .
the remaining lpv stars are considered in 4.2 . among the 56 rrl candidates ,
33 had time - magnitude plots which suggested that a close pair of non - variable stars in some frames had been resolved and in others was detected as a single star ; when the photometry from the various frames was matched ( based on position ) , the blended photometry was systematically brighter than the resolved photometry , and a large value of @xmath52 resulted .
inspection of the images almost always confirmed the blended - neighbor hypothesis .
we discuss the remaining 23 candidates in detail in 4.3 . of the 16 variable star candidates scattered over the cmd ,
all but one were found to be false detections of critically - resolved , non - variable pairs .
the single true variable has a color ( @xmath68 ) consistent with an lpv near maximum light , so we have grouped it with the other lpv stars in table 2 ( v35 ) . clearly , the large number of critically - resolved stars in this dense stellar field presents a complication to the identification of variable stars , but one that is surmountable given careful attention .
the _ catalogue of variable stars in globular clusters _ ( hogg 1973 ) , as updated by clement ( 1997 ) , lists twelve variable stars in ngc 6441 .
no photometric or variable - type information is given .
stars v7 and v8 lie outside our field of view . only variables
v1 , v6 , and v9 were detected as variables in our study .
examination of the remaining stars photometry showed that v2 , v3 , v5 , and v10 are all lpvs , but were absent from enough frames due to saturation and other effects to sidestep detection .
v4 showed no sign of variability in our observations , but at @xmath69 , is red enough to be an lpv ( see 4.2 ) .
v11 and v12 are much fainter , bluer stars which hesser & hartwick ( 1976 ) suspected as being rr lyrae stars ; our photometry shows no sign of variability in these stars ( see table 5 and the appendix ) .
data for the 31 lpv stars ( i.e. , stars which varied on timescales longer than a few days ) found near ngc 6441 are presented in table 2 .
the first column gives the variable star designation ( in roughly decreasing value of the @xmath52 variability index ) .
we name our newly - discovered variables v13 , v14 , etc .
the second and third columns give the position of the star in the @xmath70 and @xmath71 coordinate frame of figure 1 ( in units of pixels ) .
the right ascension and declination offsets from the cluster center ( in arc seconds , with positive values indicating east and north ) can be computed from the pixel positions as follows , @xmath72 the remaining columns give the magnitude - mean @xmath9 magnitude ( @xmath73 ) , the mean color ( @xmath74 ) , and the number of observations ( @xmath75 ) during the may and june observing runs . in @xmath240% of the stars , we observed a significant change in brightness or color during one or both observing runs .
the previously identified lpvs not detected in our study ( see 4.1 ) are included at the bottom of this table . clearly , the data here do not fully sample the lpvs light cycles , and thus do not accurately represent the stars mean photometric properties .
still , the locations of the stars on the cmd , shown in figure 4 , approximately represent these mean positions .
furthermore , the limited phase coverage leads to lpv detection incompleteness ; the observed @xmath8 and @xmath9 magnitude differences of some true lpvs ( e.g. , v4 ) were insufficient to result in @xmath52 values large enough to trigger detections .
finally , saturation and other image complications left some stars with too few observations to fulfill all the variability detection criteria .
thus , it is likely that many more of the stars with @xmath61 are lpvs , in particular faint , red lpvs near the cluster center . in an effort to determine the types of variable stars present among the 23 blue variable star candidates discussed above , we attempted to find the period of each star and create the star s mean light curve .
classical period finding techniques such as fourier transforms or phase dispersion minimization perform poorly , if at all , when the data are sparse and irregularly spaced .
we therefore developed the following new method . in cases where a star s variable type is suspected ( e.g. , an rr lyrae ) , it is possible to assume a period and fit the folded light curve with an appropriate template light curve .
if the assumed period is near the true period , the scatter about the fitted template will be small . by folding the time - magnitude data by a sequence of periods typical of the variable type , and fitting the template and computing the scatter ( or @xmath76 ) at each period , one can easily identify @xmath76 minima indicative of possible periods .
this method is similar to that described by stetson ( 1996 ) . in practice
, we used the template - fitting procedure described in layden ( 1998 ) .
we fit each star with three templates ; @xmath77 and @xmath78 from layden ( 1998 , table 4 ) are typical of rra and rrb type variables , respectively , while a cosine curve was used to represent rrc and contact binary lcs .
we fit the templates over periods ranging from 0.2 to 1.0 days , increasing the period by 0.1% from one step to the next . if no suitable minima were found , we searched on the period interval 13 days .
periods longer than 3 days are evident in the time - magnitude plots . since the templates are seldom a perfect match for the actual lcs , we refined the periods by folding the time - magnitude data by a sequence of periods near the @xmath76 minimum , plotting the resulting lcs , and chosing the cleanest lc .
this resulted in tighter lcs and improved periods .
we note that this technique should be applicable to any type of variable with a characteristic lc . indeed , by including template lcs from a large number of variable types
, it should be possible to determine both period _ and _ variable star type from the template giving the smallest @xmath76 value . of the 23 blue variable candidates
, we were able to determine periods for 11 rr lyrae stars ( v36v46 ) and two contact binaries ( v48 and v50 ) .
light curves for two detached binaries ( v49 and v51 ) and one long period star ( v6 , a cepheid ? ) were obtained by examination of the time - magnitude plots .
the @xmath8 and @xmath9 light curves of these 15 stars are presented in figure 5 .
possible periods were obtained for five additional stars ( sv1sv5 , see table 5 ) .
we were unable to determine with certainty whether these stars were low - amplitude rr lyrae or contact eclipsing binaries ( see figure 6 ) .
we were unable to obtain periods or lcs for the remaining two blue variable candidates , which we list in table 5 as sv8 and sv9 .
table 3 lists the photometric properties of the 11 rr lyrae stars .
columns include the location of the stars in the coordinate frame of figure 1 ( @xmath79 and @xmath80 ) , the period in days , the intensity - mean @xmath8 and @xmath9 magnitudes ( @xmath81 and @xmath82 ) , the @xmath8-band amplitude ( @xmath83 ) , the mean color during the phase interval 0.50.8 ( @xmath84 ) , and a brief comment . for several stars ,
a few discrepant data points were omitted from the computations , ( e.g. , v37 ) , while for several other stars , only good - seeing data were used ( e.g. , v42 and v43 ) .
these data will prove useful in determining the nature of these unusual stars in the following section .
table 4 reports the positions of the four eclipsing binaries from figure 5 , along with their periods , @xmath8 and @xmath9 band maxima and amplitudes , and a comment .
table 5 lists the positions , magnitude - mean @xmath8 magnitudes and @xmath23 colors of the uncertain and suspected variables , along with a brief comment .
the time series photometry for each of the variables is given in table 6 .
the columns include the variable star name , the heliocentric julian date of mid - observation ( _ hjd _ ) , the @xmath8-band magnitude and its error ( @xmath85 ) , the @xmath9-band magnitude and its error ( @xmath86 ) , and a quality code ( @xmath87 ) which corresponds to the symbols used in the light curves : @xmath88 indicates a long exposure under good seeing conditions ( @xmath89 ) , @xmath90 indicates a short exposure in good seeing ( @xmath91 ) , @xmath92 indicates a long exposure in poor seeing ( _ solid triangle _ ) , and @xmath93 indicates a short exposure in poor seeing ( @xmath94 ) .
figure 7 shows the finder charts for each variable star .
the positions of the variable stars in the cmd and their radial locations with respect to the centroid of the cluster light provide useful information regarding the stars membership in the cluster .
figure 4 shows the variable stars placed in the cmd , while figure 8 shows the cumulative radial distributions of the stars in various classes .
the squares in figure 8 give the total sky - subtracted @xmath8-band photon counts inside the radius @xmath7 , normalized by the amount at @xmath95 arcsec .
the crosses give the analogous counts expected if a uniform stellar distribution were spread across the usable pixels of the ccd .
these curves represent the _ approximate _ distributions expected for cluster member and field stars , respectively .
the _ actual _ observed distribution for a given class of star will deviate from these expectations due to brightness and crowding effects .
the lines in figure 8 are the normalized cumulative radial distributions of various classes of stars , selected from the cmd as follows .
red clump stars ( _ solid line _ ) lie in the parallelogram bounded by @xmath41 and @xmath96 , blue hb stars ( _ dotted line _ ) lie in the rectangle @xmath97 and @xmath98 , while the lpv ( _ long dash _ ) , rr lyrae ( _ short dash _ ) , and eclipsing binaries ( _ dash - dot line _ ) were taken from tables 2 , 3 , and 4 , respectively .
the radial locations of the suspected variables in table 5 are indicated along the top of the figure .
first , consider the four well - established eclipsing variables ( table 4 ) .
these stars are scattered about the cmd ( figure 4 ) and their location relative to the cluster is consistent with a uniform distribution across the ccd field ( figure 8) .
thus , it is reasonable to assume that these stars are members of the foreground stellar populations , probably the old disk ( rucinski 1997 ) .
next , consider the rr lyrae stars from table 3 . in the cmd , eight of these stars
are clustered near @xmath99 = ( 1.1 , 17.3 ) , just blueward of the prominent cluster red clump , and just redward of the blue hb stars . the radial distributions of the rr lyrae , red clump , and blue hb stars all resemble that of the cluster light ( figure 8) . of these three classes of stars ,
the red clump stars are the least radially concentrated , as expected if a significant number of field red giants contaminate the counting box .
the bhb stars are more centrally concentrated , and the rr lyrae stars seem even more so , though the number of stars involved is small and the expected background contamination varies between the three stellar classes .
note that all three of these distributions drop precipitously within 1 arcmin of the cluster center , indicating that our detection completeness decreases at smaller radii , reaching zero inside @xmath230 arcsec .
it is highly probable that these eight rr lyrae belong to the cluster . of the other three rr lyrae from table 3 , v36 is almost certainly a foreground star .
its location 297 arcsec from the cluster , its position in the cmd just blueward of the bulge red clump , and its light curve parameters ( see 5.1 ) are all consistent with it being a member of the galactic bulge .
the nature of v41 and v44 is less clear , and discussion of these stars is reserved for 5.3 .
the lpv stars from table 2 have a distribution in figure 8 very similar to that of the cluster light between 1 and 4 arcmin . inside 1 arcmin ,
incompleteness sets in , though not as quickly as for the fainter hb stars .
suspected variables sv6 and sv7 are probably cluster lpvs whose photometry has been compromised by stellar crowding .
thus , most of the lpvs in table 2 probably belong to the cluster . in the cmd ,
most of the lpvs form a fairly tight , linear sequence running from @xmath99 = ( 2.6 , 15.6 ) to ( 5.0 , 18.0 ) .
a set of five lpvs appears to parallel this sequence @xmath20.8 mag brighter .
these five stars ( v3 , v5 , v9 , v30 , and v31 ) have a cumulative radial distribution consistent with a uniform foreground population . given that the distance modulus difference between the cluster and the bulge is @xmath20.8 mag , based on the offset in @xmath8 magnitudes of the red clumps and rrl stars , it seems reasonable to conclude that these five bright lpvs belong to the bulge , rather than the cluster . the population of the two lpvs @xmath21.0 mag fainter than the cluster lpv sequence is unclear .
their magnitudes and large radial distances from the cluster ( @xmath100 arcsec for v25 and 322 arcsec for v28 ) suggest they are members of the background field .
it is conceivable that they belong to the sagittarius dwarf galaxy , but ( @xmath101 ) the distance modulus between sagittarius and the bulge is @xmath22.5 mag ( sarajedini & layden , 1995 ) , 0.7 mag too large for the observed offset , and ( @xmath102 ) ngc 6441 is over 10 degrees from the nearest part of sagittarius in the map of ibata et al .
complete light curve coverage of these variables would enable estimates of their distances , while spectroscopy would indicate whether their radial velocities are consistent with an origin in the galaxy or in sagittarius . finally , we consider the variables of uncertain type from table 5 . of these five stars , sv2 , sv3 , and sv5
lie in the clump of bona - fide cluster rrl in figure 4 , and so we tentatively consider them to be rrl as well . their radial locations ( figure 8) are consistent with those of the other rrl . while sv4 has the correct @xmath8 magnitude , it is 0.3 mag bluer than the other rrl
whether sv1 is a c - type rr lyrae or a contact binary , its location in the cmd clearly indicates that it belongs to the foreground field .
the period - amplitude diagram is a useful , reddening - independent tool for comparing rr lyrae variables . for typical rrl in the field and in metal - poor globular clusters ,
a star s position in this plane depends on its metallicity . in figure 9
, we show the metal - rich field ab - type rrl ( periods and amplitudes from bookmeyer et al .
1977 , metallicities from layden 1994 ) . at a given amplitude ,
the more metal - rich stars have systematically shorter periods .
the ab - type rrl in the globular cluster ngc 6171 ( dickens 1970 ; @xmath1 = -1.04 $ ] , harris 1996 ) overlie the field rrl of similar metallicity , while that cluster s c - type rrl congregate near ( @xmath103 , @xmath83 ) = ( 0.5 , 0.5 ) .
the rr lyrae stars in ngc 6441 are systematically displaced to longer periods and smaller amplitudes .
the bulge rrl v36 is marked by the large asterisk at @xmath104 , a position suggesting a metallicity of @xmath1 = -1.0 $ ] to -1.2 , consistent with the distribution of metallicities in bulge rrl ( walker & terndrup 1991 ) .
the large asterisks at lower amplitude mark the anomalously red variables v41 and v42 .
the open circles mark the suspected c - type rrl sv2 , sv3 , and sv5 .
the solid circles mark the remaining rrl in table 3 .
these results are consistent with the positions of the three previously - known long - period , metal - rich rrl : v9 in 47 tuc ( carney et al . 1993 ) , and v28 ( ab - type ) and v27 ( c - type ) in ngc 6388 ( silbermann et al . 1994 ) . however , our data have more than tripled the number of members of this interesting sub - class of variable . also , the positions of sv3 , sv5 , and perhaps sv2 support our interpretation that these stars are c - type rrl in ngc 6441 . using a variety of methods , hesser & hartwick ( 1976 ) estimated the interstellar reddening toward ngc 6441 to be @xmath105 mag .
they noted that , based on the appearance of the palomar sky survey prints , `` ... the obscuration in the immediate vicinity of ngc 6441 is approximately uniform ... '' .
however , the large uncertainty they quote reflects the large range of solutions produced by the different techniques , @xmath106 .
furthermore , the large angular extent over which some of their estimates were obtained may produce deceptive results if significant patchiness exists on scales of several arcminutes or less .
zinn ( 1980 ) and reed et al .
( 1988 ) obtained @xmath107 and 0.49 , respectively , from integrated light measurements .
the rr lyrae variables we have detected toward ngc 6441 provide an excellent opportunity to determine the reddening in a small region around the cluster .
mateo et al .
( 1995a ) showed that the de - reddened @xmath23 colors of typical rr lyrae near minimum light ( phases 0.50.8 ) is roughly constant .
we determine @xmath108 for the eight stars with @xmath1 > -1.3 $ ] listed by mateo et al .
( 1995a ) .
we use the apparent minimum - light colors of the 11 rrl in table 3 , @xmath84 , to determine the reddening , @xmath109 along each line of sight ( see table 7 ) .
these values are converted into @xmath110 and @xmath111 using the relations of dean et al .
( 1978 ) . of these stars , only v36 is a `` typical '' rrl , judging from its position in the period - amplitude diagram and its expected location in the galactic bulge .
indeed , one could question whether the @xmath112 colors of bulge rrl are the same as the local stars used to calibrate the relation .
it is reassuring that v36 gives @xmath113 mag , in excellent agreement with the existing reddening values .
does the relationship work for the atypical , long - period variables in ngc 6441 ?
unfortunately , @xmath23 colors do not exist for the two other known members of this class , v9 in 47 tuc and v28 in ngc 6388 .
however , the data from carney et al .
( 1993 ) on v9 indicates that the analogous relation employing @xmath114 , period , and metallicity ( blanco 1992 and references therein ) _ does _ work , yielding @xmath115 . for comparison , harris s ( 1996 ) compilation lists @xmath116 . with this encouragement , and excluding the two anomalously red and bright stars v41 and v44 , we find the mean @xmath110 of the nine rrl toward ngc 6441 to be 0.45 mag ( 0.05 mag rms ) .
again , our reddening estimate is in excellent agreement with the existing estimates .
the rms value is somewhat larger than expected from the estimated uncertainties ( @xmath20.03 mag for a single star ) .
this suggests there may be some differential reddening across the face of the cluster .
the extension of the cluster red clump along the reddening vector and the absence of a well - defined red giant branch provide similar evidence ( see 3 ) .
to search for further evidence of spatially - variable reddening , we considered individually the cmds of stars lying in eight equal - sized sectors of an annulus with radius 40100 arcsec centered on the cluster ( the angles @xmath117 and @xmath118 in table 7 define the bounds of each sector , and are measured counter - clockwise from a line radiating in the + @xmath119 direction from the cluster center in figure 1 ) . for each cmd , we isolated the red clump stars using @xmath120 and produced their generalized color histogram ( see figure 10 ) .
several curves are substantially displaced towards redder colors , indicating the presence of differential reddening .
sector 4 is particularly noticeable .
the median @xmath23 color of each histogram ( @xmath121 ) , and the number of stars involved ( @xmath122 ) , are shown in table 7 .
the widths of the individual histograms are not drastically larger than the color width of the red clump in 47 tuc ( @xmath20.2 mag , armandroff 1988 ) , suggesting that the effects of differential reddening within a given sector are small , though red giant branches in the cmds are still noticeably broadened ( especially in sector 4 ) .
the small - scale peaks in the figure 10 histograms are due to shot noise , indicating that further spatial segregation of the cluster is not warranted . in an effort to recover the de - reddened cmd of the cluster
, we corrected the differential reddening in the individual sectors to a representative mean ( @xmath123 , the average of sectors 1 , 2 , 5 and 6 ) , and then corrected the entire dataset for the average rrl reddening ( @xmath124 , based on the rrl stars in or near those sectors ) .
we assumed @xmath34 .
the de - reddened cmd is shown in figure 11 .
the principal sequences seem to have tightened up somewhat , compared with figure 3c . however , the slope of the red clump stars persists , confirming the rich et al .
( 1997 ) _ hst _ photometry .
sweigart & catelan ( 1998 ) note that this intrinsic slope rules out older cluster age or extreme mass loss on the giant branch as causes of the extended blue hb seen in ngc 6441 .
the nature of the variables v41 and v44 is unclear .
their lcs suggest they are rr lyrae stars , or possibly another class of pulsating variable .
they are both @xmath20.7 mag brighter , and @xmath20.3 mag redder , than the other eight cluster rr lyrae .
both are located close to the cluster center ( 63 and 57 arcsec , respectively ) , so are likely to be members .
one possibility is that they are type ii cepheids belonging to the cluster .
however , we searched for periods between 1 and 5 days and found no reasonable minima , and the time - magnitude plots indicate shorter period variations .
anomalous cepheids have periods between 0.36 and 1.6 days , absolute magnitudes between 0.5 and 2 mag brighter than rrl , and colors comparable to rrl ( nemec et al .
based on the first two criteria , v41 and v44 could be anomalous cepheids
. however , both stars are significantly redder than the rrl in ngc 6441 .
furthermore , most anomalous cepheids are found in dwarf spheroidal galaxies , metal - poor systems with a component of younger stars ( @xmath125510 gyr , mateo et al .
bono et al .
( 1997a ) confirm theoretically that the minimum mass for the formation of anomalous cepheids is @xmath21.8 @xmath126 at @xmath1 = -1.7 $ ] , and this minimum increases quickly with increasing metal abundance .
it seems unlikely that ngc 6441 possesses a population of stars young enough to produce hb stars of this mass .
it may also be possible to produce anomalous cepheids through binary star coalescence , but at the metallicity of ngc 6441 , the mass limit appears incompatible with turnoff stars of globular cluster age ( @xmath127 ) . given these complications , and that we are not aware of the existence of any metal - rich examples of this type of variable , it seems unlikely that v41 and v44 are anomalous cepheids .
another possibility is that v41 and v44 are rr lyrae stars in the foreground bulge .
the similarity of their @xmath8 magnitudes to that of v36 supports this idea , but their much redder @xmath23 colors would require a large amount of differential reddening . the two stars _ do _ lie near each other , 21 arcsec apart in sector 3 ; perhaps they lie behind a small , dense foreground dust cloud . to test this
, we extracted a subset of the stars in table 1 which lie within 20 arcsec of the stars centroid , and plotted their cmd .
it shows no displacement relative to the cmd of the other stars in sector 3 , nor to figure 3c , indicating that heavy differential obscuration is _ not _
responsible for the anomalous colors of these stars .
furthermore , the central location of v41 and v44 argues strongly for cluster membership .
if they belong to the field , the probability of both variables lying inside 63 arcsec radius is @xmath128 .
a third possibility is that v41 and v44 are cluster rrl whose photometry has been contaminated by another star .
if the rrl components are typical of the other eight cluster rrl , then the contaminating stars must have @xmath129 and @xmath130 , roughly like those of cluster red clump stars .
the positions of v41 and v44 in the period - amplitude diagram ( figure 9 ) are consistent with this interpretation .
the observed amplitudes are small compared with the other rrl in ngc 6441 with the same period , but the amplitudes become consistent with these stars when corrected for the effect of the photometric companion just described : the @xmath8-band amplitude of v41 becomes @xmath20.7 mag , and of v44 becomes @xmath21.2 mag .
finally , the values of the `` sharp '' parameters produced by daophot are typical of stars of their brightness , indicating that the angular separations of the contaminating stars are small compared to the seeing disk .
perhaps v41 and v44 are members of physical binaries with red clump stars .
this would be compatible with the wide - binary hypothesis for bhb star formation discussed in 3 . in the case of v41 and v44
, the mass - enhanced secondary has evolved onto the red hb , whereas for the other rrl and bhb stars , the secondary is still on the main sequence , and contributes little to the combined light ( @xmath131 mag ) .
we conclude that v41 and v44 are probably rr lyrae members of ngc 6441 , but appear brighter and redder than the other cluster rrl because their photometry has been contaminated by the light of unresolved red clump companions . images with higher spatial resolution and detailed spectroscopy may confirm this conclusion ( unfortunately , v41 and v44 were just off the _ hst _ images of rich et al . 1997 ) .
the light curves of `` normal '' c - type rr lyrae in clusters and the field have a characteristic shape , roughly sinusoidal , but with the region of maximum brightness skewed toward smaller phases .
that is , the phase interval of `` rising light '' between minimum and maximum brightness is shortened , @xmath132 . among the 31 c - type rrl in the _ general catalog of variable stars _ ( kholopov 1985 ) with @xmath8-band photometry , @xmath133 ranges between 0.32 and 0.46 , with a mean of 0.39 and rms of 0.04 .
figure 6 shows that the lc shapes of the c - type rrl in ngc 6441 depart somewhat from this norm .
their maxima appear skewed to longer periods , and @xmath1340.5 .
the minima may be uncharacteristically sharp as well , especially in the case of sv3 .
clearly , better lcs ( more observations with smaller errors ) for more stars are required to verify this phenomenon . when the observations are confronted with stellar pulsation models ( e.g. , bono et al .
1997b ) , this effect may prove to be a valuable tool for constraining the parameters responsible for bright , blue hb extensions in ngc 6441 and ngc 6388 ( e.g. , sweigart & catelan 1998 ) .
the light curve shapes of the ab - type rrl in ngc 6441 appear to fall within the range of shapes exhibited by `` normal '' ab - type rrl .
rich et al .
( 1997 ) concluded that neither age nor dynamical effects , individually , could account for the extended blue hb in ngc 6441 , though they could not rule out a combination of the two .
sweigart & catelan ( 1998 ) used synthetic hb models to explore the effects of changing parameters which may produce blue hb extensions .
they concur that neither age nor enhanced rgb mass loss can produce the sloped hb , and hence luminous bhb extension , seen in ngc 6441 .
they considered three additional scenarios , ( 1 ) high initial helium abundance , ( 2 ) core rotation , which at the rgb tip allows both enhanced mass loss and formation of an over - massive core , and ( 3 ) envelope helium abundance enhancement from deep mixing during the rgb phase . in their brief report
, sweigart & catelan ( 1998 ) could not present detailed results of their models .
we employ three means of comparing their models and @xmath37 synthetic cmds with our @xmath0 photometry : ( a ) the general distribution of stars in color and magnitude on the hb , ( b ) the number ratio of hb stars to rgb stars brighter than the mean rrl luminosity , @xmath7 , and ( c ) the pulsation characteristics of the rrl .
first , we compare the overall appearance of our observed cmd with the synthetic cmds of sweigart & catelan ( 1998 , their figure 2 ) for their various scenarios .
the steep slope of our red clump and its relation to the gentler - sloping bhb are best matched by the scenarios with core rotation and helium - mixing , and moderate mass loss on the rgb ( see their figures 2c and 2e ) .
given the uncertainties introduced by differential reddening , foreground confusion , and by comparing the observed @xmath0 cmd with the synthetic @xmath37 cmd , we can not make a more quantitative statement at this time .
measuring the @xmath7-ratio in our photometry is complicated by the high degree of contamination by stars in the foreground and by stars on the cluster asymptotic giant branch ( agb ) . using as a guide a field - subtracted version of figure 11 , we defined regions in which to count the number of bhb , red clump , and red giant stars .
these regions are shown in figure 11 .
we counted the number of stars in each region on the unsubtracted , de - reddened cluster cmd ( figure 11 ) and the equivalent de - reddened field cmd .
the counts are given in table 8 .
we add nine rrl to the bhb counts and ten lpv stars to the rgb counts .
we used the luminosity functions of bertelli et al .
( 1994 ; @xmath135 , @xmath136 , age = 14 gyr ) to estimate that 64% of the stars in the rgb region are first ascent red giants ( 36% are agb stars ) . combining the bhb and red clump counts , we obtain @xmath137 , where the error is due solely to poisson statistics .
systematic errors incurred by inappropriately defined counting regions are surely higher .
if the counts in each region are uncertain by 25% , the systematic error on our estimate of @xmath7 is 0.66 . even with this level of uncertainty , our counts strongly disfavor the high initial helium abundance scenarios of sweigart & catelan ( 1998 ) .
their simulations with main sequence helium abundances of 0.38 and 0.43 produce sloping bhb extensions , but have @xmath138 and 3.9 , respectively , about 3-@xmath27 from our observed value results in overestimating @xmath7 .
the true value of @xmath7 is therefor less than 1.60 , and the disagreement with the model with high initial helium abundance is stronger . ] . using the relations of buzzoni et al .
( 1983 ) , our value of @xmath7 implies a helium abundance of @xmath139 , where the errors are based on the larger , systematic errors quoted in @xmath7 . using their @xmath140 parameter
, we obtain @xmath141 .
we caution that stellar interactions may alter the relative numbers of hb and rgb stars , and thus affect the estimate of @xmath71 .
the pulsation properties of the cluster rrl stars provide other constraints on the sweigart & catelan ( 1998 ) models .
their figure 3 shows the positions of their model bhb stars ( helium mixing scenario with intermediate mass loss ) in the temperature - period diagram .
the dotted region in our figure 9 encloses those models .
we transposed the region from the ( @xmath142 , @xmath143 ) plane to the ( @xmath143 , @xmath83 ) plane using equation 1 of catelan ( 1998 ) , which relates @xmath144 with @xmath1 $ ] and the @xmath145-band amplitude , @xmath146 .
we converted from @xmath145 to @xmath8 amplitudes using the rrl in ngc 6171 ( dickens 1970 ) : @xmath147 ( rms = 0.04 ) mag , and we assumed @xmath1 = -0.53 $ ] dex .
clearly , the model provides a reasonable , though not perfect , match to the rrl stars in ngc 6441 , ngc 6388 , and 47 tuc .
it will be interesting to compare the observations with the temperature - period predictions of the other scenarios .
we can provide a measurement of the distribution of stars along the hb which may prove useful in constraining future synthetic hb models , and in comparing the hb of ngc 6441 with those of other globulars .
the usual statistic is @xmath148 , where @xmath145 and @xmath7 are the numbers of hb stars on the blue and red sides of the instability strip , and @xmath8 is the number of rrl variables .
we adopt the same cmd regions as above except that the section of the bhb region redward of @xmath149 ( the mean de - reddened mag of the rrl in table 7 , see counts in row labeled `` rhb '' of table 8) is now added to the red clump counts to produce @xmath7 . subtracting the counts in the off - cluster cmd from those in the on - cluster cmd , we obtain @xmath150 , @xmath151 , and @xmath152 stars , so @xmath153 ( poisson ) @xmath154 ( systematic ) .
the hb is predominantly red , and only @xmath217% of hb stars are in the blue extension .
it is tempting to use our photometry to measure the distance to ngc 6441 .
however , the long periods observed for its rrl suggest that they are more luminous than typical rrl , so the absolute magnitudes usually assumed for those stars may be inappropriate
. moreover , the different models presented by sweigart & catelan ( 1998 , their figure 2 ) have absolute magnitudes near the instability strip which range between 0.35 and 0.75 mag ( we have omitted the high initial helium scenarios ) .
the absolute magnitudes of the red clumps present a smaller range of 0.800.95 mag .
combined with a de - reddened apparent red clump magnitude of @xmath155 mag from figure 11 , we find that ngc 6441 lies between 11 and 13 kpc from the sun , on the far side of the galactic bulge .
we present the first ground - based ccd photometry of the metal - rich globular cluster ngc 6441 .
our @xmath0 color - magnitude diagram shows that the extended blue horizontal branch discovered by rich et al .
( 1997 ) using the _ hubble space telescope _ exists even at large radii from the cluster center .
such hot hb stars are not expected in a cluster of this metallicity .
our time - series photometry enabled us to detect a large number of variable stars .
we increased the number of red , long - period variables from nine to 32 .
we argue that five are members of the foreground bulge and that two are members of the background field , or possibly the sagittarius dwarf galaxy .
the remaining 25 are members of ngc 6441 .
we confirm the variability of v6 , which may be a foreground cepheid .
long - term monitoring of these variables is encouraged .
we also discovered 11 ab - type rr lyrae variables , along with several probable c - type rrl .
based on their radial distances from the cluster center and positions in the cmd , we argue that eight of the ab - type rrl and 23 of the c - type rrl are members of ngc 6441 .
one rrl belongs to the foreground bulge , and two appear anomalously bright and red .
we suspect the latter are photometric blends between cluster rrl and red clump stars , and could be physical binaries .
we also discovered four eclipsing binary stars in the foreground field , and several stars which we could not classify with certainty .
we present a method for determining variable star periods in which the data are folded by a sequence of periods , and at each period , are fit with a set of light curve templates .
the @xmath76 of the fits are then plotted as a function of the employed periods .
minima in this plot indicate probable periods . using this method , we determined periods for the rrl , and produced mean light curves in @xmath8 and @xmath9 .
the 811 rrl in ngc 6441 have periods that are systematically longer than field rrl of comparable metallicity .
rrl in the metal - rich globulars 47 tuc ( v9 ) and ngc 6388 ( v27 and v28 ) have similar properties .
thus , a new sub - class of rrl appears to be emerging . using the observed @xmath23 colors of the rrl at minimum light
, we determined the reddening toward ngc 6441 to be @xmath156 mag .
two points suggest that these long - period rrl have the same @xmath157 colors as the `` normal '' rrl , ( @xmath19 ) we obtain @xmath158 mag from the bulge variable v36 , and ( @xmath20 ) the analogous relation for @xmath159 at minimum light works for the long - period rrl v9 in 47 tuc .
evidence for small changes in the amount of reddening across the face of the cluster come from the star - to - star dispersion in the rrl @xmath110 values , and from differential shifts along the reddening vector of cmds generated from small regions around the cluster .
the south - east portion of the cluster has @xmath160 mag higher than the rest of the cluster , though there is evidence for patchy reddening on a scale of several arcmin or less .
we employ a differentially de - reddened color - magnitude diagram to count stars in various regions along the horizontal and red giant branches .
we find the ratio of hb stars to red giants to be @xmath161 , indicating a normal helium abundance of @xmath162 . about 17% of the hb stars lie blueward of the red clump , and @xmath163 , indicating a predominantly red hb .
we also find that the bhb stars are more centrally concentrated than the cluster red clump stars .
this could result from mass segregation if the bhb stars are members of binary systems .
furthermore , the bhb stars could be _ produced _ by mass transfer in wide binary systems , as described by bailyn ( 1995 ) .
the anomalously bright and red rrl v41 and v44 , which we have discovered in ngc 6441 , could represent binaries consisting of an rrl and a red clump star .
spectroscopic monitoring of these and the other bhb stars for binary motions would provide stronger evidence for or against the binary hypothesis of bhb star formation .
we also consider a different line of hypotheses , involving variations in the physical properties of single stars , which might explain the presence of the bhb stars high surface temperatures .
as noted by sweigart & catelan ( 1998 ) , the intrinsic slope of the hb , and in particular of the red clump , can not be reproduced by standard models invoking extreme cluster age or red giant branch mass loss .
similarly , our measurement of the hb - to - giant star ratio , @xmath7 , argues strongly against the sweigart & catelan ( 1998 ) model predictions for a scenario which involves high initial helium abundance .
in contrast , the observed periods and amplitudes of our rrl are in reasonable agreement with their model predictions for a scenario in which the helium abundance in the atmosphere has been enhanced through deep mixing on the rgb .
however , the results of the sweigart & catelan ( 1998 ) models appear to be rather sensitive to the assumed mass loss parameters ( mean and distribution ) , which are at present poorly constrained by either observation or theory .
thus , it remains unclear whether the existing observations can isolate a unique set of physical parameters which explain the production of the extended blue hb .
we note one additional piece of observational evidence which may be brought to bear on this problem .
the shapes of the c - type rrl light curves in ngc 6441 appear to differ significantly from their `` normal '' counterparts .
they have longer phase intervals between minimum and maximum light , and the minima appear to be sharper . perhaps by adjusting the physical parameters in stellar pulsation models ( e.g. , bono et al .
1997b ) , similar light curves can be produced .
if so , the intersection of plausible parameters from evolution and pulsation theory could further isolate the physical differences between `` normal '' clusters and those like ngc 6441 .
we encourage additional work on this problem both from the theoretical and observational perspectives .
we thank mrcio catelan , mario mateo , ata sarajedini , horace smith , and an anonymous referee for helpful comments and suggestions .
support for this work was provided by nasa through hubble fellowship grant hf-01082.01 - 96a awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa under contract nas 5 - 26555 .
support was also provided by the national science and engineering council of canada through a grant to d.l.w .
* v11 and v12 : * hesser & hartwick ( 1976 ) identified these stars as possible rr lyrae variables ( rrl ) .
we obtained an rms scatter about the mean values in table 5 of 0.03 mag in @xmath8 and @xmath20.08 mag in @xmath9 for both candidates .
their @xmath52 variability indices are 0.0 and 3.2 , respectively .
our data thus show no sign of variability in these stars .
indeed , they are @xmath164 mag bluer than the _ bona fide _ rrl stars we have discovered ( see figure 4 ) .
* sv1 : * definitely variable , but the class of variable is uncertain .
the data fold with minimal scatter at @xmath165 days to produce the light curve ( lc ) of a c - type rrl , and at @xmath166 days to produce the lc of a contact eclipsing binary .
the star is @xmath20.8 mag brighter than the member rrl . *
sv2 : * definitely variable , but the class of variable is uncertain .
the data fold with minimal scatter at @xmath167 days to produce an rrl lc , and at @xmath168 days to produce the lc of a contact eclipsing binary .
the rrl lc is too sharp at the extrema for a typical rrc , and the rise - time is too long for a typical rrab .
however , the star falls among the other cluster rrl in the cmd .
* sv3 : * definitely variable , but the variability class is uncertain .
the data fold well at @xmath169 days , yielding a contact eclipsing binary .
a less convincing fold is achieved with @xmath170 days to produce an rrc lc , but the minimum is too sharp and the maximum is skewed to late phases . the star s position in figure 4 suggests it is an rrl member of ngc 6441 . *
sv4 : * definitely variable , but the variability class is uncertain .
the @xmath8-band data fold well with @xmath171 days , yielding an rrc lc , though the @xmath9-band curve has significant scatter . a period of @xmath172 days
yields a contact eclipsing binary with an eccentric orbit .
the color is 0.3 mag bluer than the _ bona fide _ cluster rrl .
* sv5 : * definitely variable , but the variability class is uncertain .
the data fold well with @xmath173 days , yielding an rrc lc , though the @xmath9-band curve has significant scatter . a period of @xmath174 days
yields a contact eclipsing binary with considerable asymmetry .
the star s position in figure 4 suggests it is an rrl member of ngc 6441 .
rrrcccc 1 & 1267.64 & -27.67 & 17.001 & 0.169 & 1.071 & 0.192 2 & 1336.58 & -25.93 & 18.611 & 0.043 & 1.150 & 0.059 3 & 949.77 & -25.65 & 17.553 & 0.151 & 1.567 & 0.171 4 & 1225.50 & -25.00 & 18.059 & 0.021 & 1.206 & 0.030 5 & 1216.89 & -24.55 & 18.799 & 0.041 & 1.135 & 0.069 lrrcccccc v13 & 1005 & 1185 & 12.98 & 3.49 & 5 & 12.59 & 3.05 & 18 v1 & 890 & 1112 & 13.76 & 4.97 & 6 & 12.85 & 4.20 & 17 v14 & 432 & 1183 & 12.89 & 2.70 & 5 & 13.27 & 3.17 & 20 v15 & 1156 & 1609 & 13.03 & 3.85 & 5 & 12.79 & 3.54 & 18 v16 & 1360 & 1147 & 13.16 & 3.65 & 5 & 12.90 & 3.12 & 18 v17 & 1182 & 937 & 13.19 & 2.88 & 5 & 13.43 & 3.05 & 19 v18 & 1463 & 49 & 12.78 & 3.90 & 5 & 13.04 & 4.09 & 19 v19 & 1065 & 1729 & 13.32 & 6.60 & 4 & 12.83 & 6.28 & 13 v20 & 477 & 1510 & 13.03 & 2.70 & 5 & 13.29 & 2.95 & 21 v21 & 1423 & 1401 & 13.02 & 4.53 & 5 & 12.81 & 4.39 & 17 v22 & 1268 & 1076 & 12.95 & 2.95 & 5 & 12.86 & 2.82 & 16 v23 & 979 & 1129 & 12.80 & 3.46 & 5 & 12.68 & 3.29 & 16 v24 & 480 & 962 & 13.07 & 3.06 & 5 & 13.17 & 3.19 & 17 v25 & 1431 & 1347 & 14.68 & 4.10 & 7 & 14.44 & 4.18 & 19 v26 & 1271 & 747 & 12.76 & 3.75 & 5 & 12.67 & 3.54 & 17 v27 & 823 & 396 & 12.98 & 3.02 & 5 & 12.93 & 2.97 & 17 v28 & 1092 & 190 & 14.33 & 2.44 & 7 & 14.29 & 2.37 & 21 v9 & 1073 & 1119 & 12.52 & 2.84 & 3 & 12.23 & 2.53 & 12 v29 & 1486 & 501 & 12.46 & 4.80 & 3 & 12.81 & 4.97 & 17 v30 & 354 & 658 & 12.38 & 3.08 & 3 & 12.26 & 2.93 & 12 v31 & 804 & 1581 & 12.14 & 3.02 & 3 & 12.25 & 3.12 & 12 v32 & 1062 & 1045 & 12.94 & 2.76 & 3 & 12.87 & 2.64 & 12 v33 & 936 & 986 & 13.00 & 2.56 & 5 & 13.15 & 2.70 & 13 v34 & 601 & 1580 & 13.31 & 3.26 & 4 & 13.23 & 3.14 & 19 v35 & 877 & 1017 & 13.52 & 2.19 & 5 & 13.39 & 2.11 & 19 v6 & 933 & 877 & 13.56 & 1.66 & 5 & 12.79 & 1.40 & 18 v2 & 918 & 938 & 12.09 & 2.70 & 3 & 13.32 & 4.25 & 4 v3 & 124 & 1212 & 11.60 & 3.49 & 3 & 12.37 & 4.74 & 12 v4 & 853 & 1440 & 13.00 & 2.88 & 5 & 13.00 & 2.90 & 18 v5 & 504 & 425 & 12.29 & 3.93 & 3 & 11.54 & 2.98 & 4 v10 & 819 & 1149 & 12.72 & 3.03 & 4 & 12.92 & 3.39 & 18 lrrllllll v36 & 1703 & 1299 & 0.5078 & 16.54 & 15.50 & 1.16 & 1.16 & bulge v37 & 926 & 705 & 0.6132 & 17.38 & 16.41 & 1.04 & 1.10 & v38 & 994 & 605 & 0.7347 & 17.35 & 16.33 & 0.76 & 1.10 & v39 & 716 & 847 & 0.8330 & 17.53 & 16.36 & 0.72 & 1.28 & v40 & 1088 & 1139 & 0.6490 & 17.22 & 16.21 & 1.11 & 1.18 & v41 & 923 & 1131 & 0.7345 & 16.65 & 15.26 & 0.38 & 1.46 & see appendixv42 & 1011 & 765 & 0.8140 & 17.39 & 16.33 & 0.57 & 1.14 & v43 & 897 & 861 & 0.7730 & 17.38 & 16.34 & 0.52 & 1.10 & v44 & 975 & 1138 & 0.6090 & 16.58 & 15.23 & 0.64 & 1.45 & see appendixv45 & 1179 & 1339 & 0.5028 & 17.15 & 16.11 & 0.73 & 1.09 & v46 & 1377 & 1198 & 0.9050 & 17.31 & 16.18 & 0.42 & 1.18 & lrrllllll v47 & 836 & 1765 & 0.703 & 16.20 & 15.02 & 1.47 & 1.10 & detatched v48 & 158 & 1171 & 0.6674 & 15.23 & 14.22 & 0.32 & 0.28 & contact v49 & 365 & 768 & 1.010 & 16.52 & 15.45 & 0.36 & 0.20 & detatched?v50 & 838 & 1410 & 0.4335 & 17.85 & 16.55 & 0.55 & 0.55 & contact ?
crrccl sv1 & 1364 & 1501 & 15.81 & 0.99 & see appendix sv2 & 536 & 1057 & 17.22 & 1.03 & see appendix sv3 & 580 & 1244 & 17.25 & 1.07 & see appendix sv4 & 1268 & 1129 & 17.24 & 0.72 & see appendix sv5 & 912 & 796 & 17.37 & 0.97 & see appendix sv6 & 1062 & 940 & 16.89 & 2.59 & crowded , lpv ?
sv7 & 1034 & 969 & 15.13 & 2.49 & crowded , lpv ?
sv8 & 643 & 563 & 17.21 & 1.18 & blended image , rr lyrae ?
sv9 & 1359 & 93 & 17.10 & 1.17 & blended image , eclipsing binary ?
v11 & 1120 & 755 & 17.86 & 0.48 & non - variable , see appendix v12 & 1123 & 867 & 17.29 & 0.69 & non - variable , see appendix rcccccr v1 & 258.6055 & 17.231 & 0.010 & 12.962 & 0.005 & 3 v1 & 258.6006 & 17.195 & 0.021 & 12.901 & 0.004 & 4 v1 & 258.8444 & 17.201 & 0.016 & 12.920 & 0.005 & 2 v1 & 258.8511 & 17.164 & 0.010 & 12.909 & 0.004 & 1 v1 & 259.5721 & 17.142 & 0.013 & 12.906 & 0.005 & 4 ... & & & & & & v1 & 265.8469 & 17.371 & 0.063 & 12.847 & 0.007 & 3 v1 & 266.7073 & 16.844 & 0.021 & 12.785 & 0.010 & 4 v2 & 258.8444 & 17.352 & 0.033 & 13.217 & 0.009 & 2 v2 & 225.6801 & 14.754 & 0.017 & 12.078 & 0.012 & 2 v2 & 260.7804 & 17.512 & 0.034 & 13.300 & 0.012 & 2 cccccccc 1 & 0 & 45 & 1.391 & 64 & v36 & 0.588 & 0.456 & & & & & v46 & 0.612 & 0.474 2 & 45 & 90 & 1.429 & 63 & v40 & 0.604 & 0.468 & & & & & v45 & 0.516 & 0.400 3 & 90 & 135 & 1.472 & 59 & v41 & 0.884 & 0.683 & & & & & v44 & 0.874 & 0.676 4 & 135 & 180 & 1.524 & 95 & - & & 5 & 180 & 225 & 1.460 & 82 & v39 & 0.707 & 0.548 6 & 225 & 270 & 1.401 & 79 & v37 & 0.529 & 0.410 & & & & & v38 & 0.531 & 0.412 & & & & & v42 & 0.572 & 0.444 & & & & & v43 & 0.532 & 0.413 7 & 270 & 315 & 1.409 & 92 & - & & 8 & 315 & 360 & 1.393 & 90 & - & & * figure 1 : * a 30 sec exposure image of ngc 6441 in the @xmath8-band .
north is down and east is to the left .
the field is @xmath175 across .
pixel coordinates range from ( @xmath70 , @xmath71 ) = ( 1 , 1 ) at the lower - left to ( 2048 , 2048 ) at the upper right , with the cluster center lying at ( 1014 , 1000 ) .
the bright star hr 6630 has been masked out of the image .
a larger scale image is available at http://bethe.bgsu.edu/ layden / astro / preprints / preprints.html .
* figure 2 : * color - magnitude diagrams of stars toward ( a ) ngc 6441 , and ( b ) a field centered @xmath176 north of the cluster . in both panels , only stars with small errors ( @xmath29 mag and @xmath25 mag ) are shown . in ( a ) ,
only stars lying within @xmath177 of the cluster center are shown , while in ( b ) only stars within @xmath178 are shown .
the line segment represents the reddening vector ( @xmath179 mag ) . *
figure 3 : * color - magnitude diagrams for stars with moderate errors ( @xmath180 mag and @xmath181 mag ) and located ( a ) within @xmath177 of ngc 6441 , ( b ) within @xmath178 of the center of the adjacent field , ( c ) between 40 and @xmath182 of ngc 6441 . in ( d ) , the points in ( c ) are shown after the statistical subtraction of the cmd generated from off - cluster stars located in the same radial zone .
* figure 4 : * the ngc 6441 color - magnitude diagram from fig 3a with the variable stars marked as follows : long - period variables from table 2 ( _ solid squares _ ) ; rr lyrae stars from table 3 ( @xmath89 ) ; eclipsing binaries from table 4 ( @xmath94 ) ; variables of uncertain class ( sv1sv5 ) from table 5 ( @xmath183 ) ; and the remaining suspected variables from table 5 ( @xmath184 ) . *
figure 5 : * mean light curves for the blue variable candidates which yielded clear periods .
each panel shows the @xmath8 ( lower curve ) and @xmath9 ( upper curve , in some panels offset vertically by the indicated amount for convenience of display ) light curve for the indicated star and period . in all panels ,
the minor tick marks indicate intervals of 0.2 mag .
* figure 6 : * mean light curves for the blue variable candidates which yielded two possible periods . in each panel , the left - most set of light curves ( @xmath8 below ; @xmath9 above , possibly shifted vertically ) employs the shorter period and has the characteristics of a c - type rr lyrae , while the right - most set employs the longer period , and resembles a contact binary .
* figure 7 : * finder charts for the variable stars .
each image is @xmath185 square , and is oriented with north at the top and east to the left . for the lpvs ( v1v35 plus sv6 and sv7 ) ,
@xmath9-band images are used , while @xmath8-band images are used for all other stars .
* figure 8 : * the cumulative radial distributions of stars of various classes , normalized to the total number inside @xmath177 radius for : red clump stars ( _ solid line _ , rc ) ; blue hb stars ( _ dotted line _ , bhb ) ; rr lyrae stars ( _ short - dashed line _ ,
rrl ) ; lpv stars ( _ long - dashed line _ ) ; and eclipsing binaries ( _ dash - dot line _ , ecl ) .
the @xmath8-band light distribution ( sky subtracted , in intensity units ) indicates the distribution expected for cluster members ( @xmath186 , cl ) .
a uniform illumination over the useable ccd pixels indicates the distribution expected for a field star population ( @xmath184 , fld ) .
the radial locations of the suspected variable stars , sv1sv9 , are marked along the top , with the probable rr lyrae stars in italics .
* figure 9 : * the period - amplitude diagram for metal - rich rr lyrae stars .
the small symbols include ab - type field rrl with @xmath1 > -0.5 $ ] ( @xmath184 ) , field rrl with @xmath187 < -0.5 $ ] ( @xmath183 ) , and ab- and c - type rrl in ngc 6171 ( dickens 1970 , @xmath186 ) .
large symbols include : ngc 6441 variables v36 , v41 , and v44 ( @xmath183 ) ; sv2 , sv3 , and sv5 ( @xmath91 ) ; and the eight other rrl from table 3 ( @xmath89 ) ; 47 tuc variable v9 ( rrab , _ solid square _ ) ; ngc 6388 variables v28 ( rrab , _ solid triangle _ ) and v27 ( rrc , @xmath94 ) from silbermann et al .
the dotted region represents one of the model predictions of sweigart & catelan ( 1998 , from their figure 3 ) .
* figure 10 : * generalized color histograms of red clump stars in the eight sectors numbered in the figure ( see table 7 ) .
each histogram is a sum of unit - area gaussian curves placed at the @xmath23 color of each star and having a width equal to the color error of the star .
the shifts between sectors indicate the presence and degree of differential reddening . * figure 11 : * the differentially de - reddened color - magnitude diagram of ngc 6441 .
variable stars are indicated as follows : long period variables ( _ filled square _ ) , ab - type rrl ( @xmath89 ) , suspected c - type rrl ( @xmath183 ) .
the dotted boxes define the regions used to count blue hb stars ( bhb ) , red clump stars ( rc ) , and red giant branch stars ( rgb ) in sec . | we present time - series @xmath0 photometry of the metal - rich ( @xmath1 = -0.53 $ ] ) globular cluster ngc 6441 .
our color - magnitude diagram shows that the extended blue horizontal branch seen in _ hubble space telescope _
data exists in the outermost reaches of the cluster .
about 17% of the horizontal branch stars lie blueward and brightward of the red clump . the red clump itself slopes nearly parallel to the reddening vector .
a component of this slope is due to differential reddening , but part is intrinsic .
the blue horizontal branch stars are more centrally concentrated than the red clump stars , suggesting mass segregation and a possible binary origin for the blue horizontal branch stars .
we have discovered @xmath250 new variable stars near ngc 6441 , among them eight or more rr lyrae stars which are highly - probable cluster members .
comprehensive period searches over the range 0.21.0 days yielded unusually long periods ( 0.50.9 days ) for the fundamental pulsators compared with field rr lyrae of the same metallicity .
three similar long - period rr lyrae are known in other metal - rich globulars .
with over ten examples in hand , it seems that a distinct sub - class of long - period , metal - rich rr lyrae stars is emerging .
it appears that these stars have the same intrinsic colors as normal rr lyrae . using the minimum - light color of the rr lyrae
, we determine the mean cluster reddening to be @xmath3 mag , with a significant variation in reddening across the face of the cluster .
the observed properties of the horizontal branch stars are in reasonable agreement with recent models which invoke deep mixing to enhance the atmospheric helium abundance , while they conflict with models which assume high initial helium abundance .
the light curves of the c - type rr lyrae seem to have unusually long rise times and sharp minima .
reproducing these light curves in stellar pulsation models may provide another means of constraining the physical variables responsible for the anomalous blue horizontal branch extension and sloped red clump observed in ngc 6441 . |
as reported by @xcite the interstellar medium ( hereafter ism ) is very poor in crystalline solids .
for instance , @xcite well reproduce the interstellar absorption band using a mixture composed of @xmath0% of amorphous pyroxene and @xmath1% of amorphous olivine by mass , leading to a crystalline fraction of the interstellar silicates around @xmath2% .
the proto - solar nebula is supposed to have been formed from material coming from the ism . as a consequence , the primordial dust in the solar system
should be composed of amorphous solids except grains which have undergone either a thermal annealing in high temperature regions ( _ i.e. _ around @xmath3 k ) close to the star ( _ i.e. _ @xmath4 au ) or a sequence starting by a vaporization or a melting and finishing by a re - condensation . besides this , comets are presumed to have formed in the cold outer part of the solar nebula along the lines of the scenario recalled below .
+ the dissipation of the early gaseous and dusty protoplanetary disk is presumed to have left a disk of icy planetesimals beyond @xmath5 au . the edgeworth - kuiper belt ( hereafter kb ) and the oort cloud ( hereafter oc ) are the remains of this disk and are the well accepted two main cometary reservoirs @xcite .
a corpus of studies , based on dynamics , show that icy asteroids originally located between @xmath5 au and @xmath6 au have either participated to the formation of giant planets or have been ejected far away because of the planet migration .
the majority of the kb objects have probably been accreted at their current distances ( _ i.e. _ @xmath7 au ) ; but as suggested by results of the nice model , an orbital resonance of saturn and jupiter may have partially filled the kb with objects initially located between @xmath8 au and @xmath6 au @xcite . furthermore , the planetesimals of the inner part of the primordial disk ( _ i.e. _ located in the interval @xmath9 au ) have been likely ejected and finally participated to the formation of the outer regions of the kb and the oc . in summary ,
the crystals have been incorporated into icy planetesimals when they occupied zones from @xmath5 au to @xmath10 au and beyond ; prior to be scattered to the kb and oc .
of course , the transport processes studied in this paper should deposit crystalline dusts at least around @xmath11 au , but the deeper the penetration beyond @xmath11 au will be , the more credible the proposed mechanism will be . the formation of the comets is also believed to have occurred early during the solar system formation .
for instance @xcite showed that the formation of comets could have been completed in @xmath12 years .
+ @xcite have found that crystalline olivine is a major component of the silicates in comet halley and @xcite detected crystalline silicates in hale - bopp observations .
_ stardust _ samples of comet 81p / wild 2 include large single mineral crystals and x - ray microscopic analysis leads to a crystal mass fraction @xmath13 larger than @xmath14% @xcite .
more generally , comets have a ratio @xmath13 exceeding @xmath15% @xcite .
this discrepancy between the crystallinity of the ism grains and the of cometary grains one is the mark of a radial transport process and/or a specific physical phenomenon occurring in the accretion disk .
several transport processes have been proposed to explain the presence of these refractory material in comets : annealing by shock waves in the outer solar nebula @xcite , radial mixing by turbulent diffusion @xcite or mixing in a marginally gravitationally unstable ( mgu ) disk ( see the series of papers by boss and co - authors : * ? ? ?
* could be a starting point ) , transport by photophoresis @xcite .
some authors @xcite built models including the vertical disk structure and found a radial outflow of dust grains due to pressure gradients within the protoplanetary disk .
the x - wind mechanism , which has been advocated for the redistribution of cais throughout the nebula @xcite , could have also brought a contribution .
+ in this paper , we focus on photophoretic process for which we employ a 1 + 1d accretion disk undergoing a time - dependent irradiation .
indeed , the proto - sun , at the epoch during which it was surrounded by a disk of gas and dust , was evolving through the so - called t tauri phase . the stars belonging to the t tauri class ,
are known to possess an accretion disk and are evolving along the pre - main sequence ( hereafter pms ) tracks where their luminosity could be much higher than that of the sun .
this high luminosity could have an important influence on photophoresis and might change the disk structure or its dynamics by irradiation .
this is why we were interested in modeling an accretion disk evolving consistently with a proto - sun .
it is of particular interest to assess if the assumption of the existence of a central gap ( as hypothesized in previous works , see * ? ? ?
* ) in the disk is still required .
it should be noticed that @xcite have taken into account the pms luminosity of the proto - sun in the context of jupiter formation .
+ the presence of crystals in comets is not the only issue for which photophoresis is supposed to play a role , @xcite have proposed that photophoresis could accelerate the giant planets formation and @xcite have investigated the involvement of this effect in the formation of mercury - like planets and in the metal depletion in chondrites . detailed theoretical and experimental studies on chondrules photophoretic properties have been also conducted @xcite .
+ concerning the photophoretic transport itself we have essentially adopted the approach described in @xcite and @xcite , as recalled in sec .
[ model ] .
the main features of our disk model are presented in the same section , and additional details are given in [ append ] .
section [ res ] is devoted to our results , while in sect .
[ discu ] we discuss some aspects of the problem such as the influence of aggregate properties , in particular that of thermal conductivity as suggested by @xcite , and the role of gas dust opacity .
conclusions can be found in sect .
[ concl ] .
we have developed an original implementation of a 1 + 1d irradiated protoplanetary accretion disk model based on the equation introduced by @xcite and @xcite @xmath16 which governs the secular variations of the disk surface density @xmath17 ( kg m@xmath18 ) as a function of the heliocentric distance @xmath19 .
the numerical method employed to solve eq .
( [ eq_evol ] ) is described in [ secevol ] .
the average turbulent viscosity @xmath20 ( m@xmath21 s@xmath22 ) , depending on a free parameter @xmath23 , is computed in the frame of the @xcite formalism after the integration of the vertical structure equations ( see [ vertstruct ] for details ) .
our disk model is based on a generalization of the method introduced by @xcite ( hereafter pt99 ) and @xcite .
the photoevaporation @xmath24 is provided by the simple prescription published by @xcite : @xmath25 for @xmath26 and @xmath27 for @xmath28 , where @xmath29 is taken to be 5 au and @xmath30 is an adjustable parameter . a rigorous treatment of the radiative transfer in all its complexity is beyond the scope of the present work , and represents probably a physical and numerical challenge . instead
, the effect of the irradiation by the central star has been taken into account by modifying the temperature boundary condition at the disk external surface .
this way , the temperature at this surface , denoted @xmath31 , is given by @xmath32 where @xmath33 is the background temperature ( usually @xmath34 k ) and @xmath35 comes from the following equation , derived in @xcite , @xmath36^{1/4 } \\
\end{array}\ ] ] where @xmath37 and @xmath38 are respectively the stellar radius and effective temperature at the age @xmath39 ; @xmath40 represents the disk pressure height at the heliocentric distance @xmath19 and at time @xmath39 .
this prescription has also been adopted by @xcite who have emphasized its significance that we recall here : the term ( 1 ) corresponds to the radiative flux that would be intercepted by a flat disk , the term ( 2 ) is an estimation for the effect of the disk flaring .
following @xcite and @xcite we fixed @xmath41 at its equilibrium value , namely @xmath42 . in order to validate our model ,
we have performed several tests .
for instance , we have verified that we reproduce the mid - plane temperature @xmath43 , the surface density @xmath17 and the shape factor @xmath44 obtained by pt99 ( see e.g. figs . 2 and 3 in pt99 ) .
we also obtained a good agreement between our computations and the pt99 ones for the accretion rate @xmath45 ( m@xmath46 yr@xmath22 ) as a function of @xmath17 for various turbulence parameters @xmath23 and distances @xmath19 to the sun . in their study , pt99 did not take into account the effect of irradiation .
notice that , this is why we verified that our computations are compatible with results found by @xcite , even if the treatment of the irradiation is not exactly the same .
this way , following the same procedure adopted by @xcite , we also verified that our computations are compatible with results found by @xcite , who used a more sophisticated approach concerning irradiation .
+ both stars and protoplanetary disks change with time .
the stars evolve under the influence of gravitation and nuclear reactions , while disks lose their mass either by accreting on the star and/or evaporating to the interstellar space .
thus , we have decided to include both evolutions in a model in which the proto - sun and the disk interact via irradiation .
+ young main sequence stars may exhibit a debris disk , that has a very tenuous or even non - existent gaseous component @xcite .
in contrast , t tauri stars have their spectroscopic features ( as the excess uv or ir radiation ) well explained by the presence of a surrounding , optically thick , gaseous accretion disk @xcite .
the objects of this class are pre - main sequence stars , located in the hertzsprung - russell diagram between the `` birthline '' and the early main sequence , where the stellar luminosity is provided by the hydrogen burning in their core with the secondary elements at equilibrium .
the `` birthline '' is a line in the hr diagram corresponding to a threshold below which stars become visible to the observer , they therefore begin their quasi - static contraction and move to the main sequence .
we have adopted a pms evolutionary track computed by @xcite who have employed the well - tested and developed stellar evolutionary code franec @xcite together with up - to - date input physics @xcite .
+ solid line : pre - main sequence evolutionary track computed by @xcite , chosen parameters are representative for the sun : @xmath47 and @xmath48 are respectively the initial heavy elements and the helium mass fractions , @xmath49 is the solar calibrated mixing length theory parameter and @xmath50 denotes the initial mass fraction of deuterium .
this stellar track has been computed with the mixture of heavy elements provided by @xcite ( labeled as05 ) . dashed line : the stellar `` birthline '' from @xcite .
dotted line : the zams .
circles are displayed each 1 myr and the sun position on the track has been marked with a diamond.,width=10 ] the starting time @xmath51 of the evolution of our accretion disk has been taken where the `` birthline '' ( from * ? ? ? * ) intersects the proto - solar pms stellar track .
the solar models have been calibrated on the present solar luminosity ( @xmath52 ) , radius ( @xmath53 ) , for the surface chemical composition ( @xmath54 , see @xcite ) .
we used an iterative procedure that consists in varying the initial helium abundance ( @xmath55 ) , metallicity ( @xmath56 ) , and mixing length parameter ( @xmath57 ) , in order to reproduce at the age of the sun ( @xmath58 ) the present @xmath52 , @xmath53 , and @xmath59 within a tolerance of , at most , @xmath60 .
then , we obtained for the proto - sun the initial helium and metal mass fraction @xmath61 , @xmath62 , and @xmath63 ( see fig .
[ pms ] ) . in their study of samples of stars belonging to six young stellar clusters
, @xcite have measured the jhkl infrared excess fractions of stars . for each cluster
, these ratios can be regarded as the fractions of stars surrounded by a disk .
for the selected clusters , age determinations are available in the literature .
this way , @xcite have shown that , not surprisingly , the fraction of stars accompanied by a disk decreases with age and should be negligible for an age around @xmath64 myr ( hereafter 1 myr@xmath65 years ) .
this limitation is confirmed by @xcite and references therein . according to these works
, protoplanetary disks older than @xmath64 myr should not exist or be very rare .
our 1 + 1d disk model requires the pre - computation of tables of turbulent viscosity @xmath20 depending on various parameters among which the luminosity of the star @xmath66 and its effective temperature @xmath67 . consequently , along the pms track we took a set of couples
@xmath68 corresponding to ages @xmath69 myr with @xmath70 at the `` birthline '' . hence , for each @xmath39 , tables of mean viscosity were constructed using @xmath71 or equivalently @xmath72 and playing with a set of parameters values : @xmath73 and @xmath74 ; @xmath45 ranging from @xmath75 yr@xmath22 to @xmath76 yr@xmath22 ; and @xmath77 au .
a dedicated subroutine allows interpolation in mentioned tables of @xmath78 .
+ the dust grains transported through the disk are aggregates of small particles @xcite and are very approximately considered as spherules of radius @xmath79 .
hereafter , the terms `` particles '' , `` grains '' and `` aggregates '' will be synonyms ; while `` monomer '' will be reserved for the components of aggregates . + following the approach developed by @xcite and @xcite we assume that the gas flow conditions are described by the knudsen number , @xmath80 , which is defined as @xmath81 , where @xmath82 is the mean free path of the gas molecules .
+ an expression of the photophoretic force @xmath83 , valid for free molecular regime ( @xmath84 ) and for continuum regime ( @xmath85 ) , has been proposed by @xcite ( see their eq .
( 31 ) ) @xmath86 we are not going to recall the expression of each term of eq .
( [ phforce ] ) , instead we invite the interested reader to consult previous works @xcite . nonetheless , we specify some important points .
first of all , given that the collision / scattering cross section of molecules is not known , the free mean path @xmath82 is estimated from the value of the dynamic viscosity @xmath87 provided by @xmath88 ( see * ? ? ?
* ) with @xmath89 pa s , @xmath90 the temperature and @xmath91 k. since the viscosity @xmath87 for a dilute gas is also given by @xmath92 ( with @xmath93 the number of molecules per unit of volume and @xmath94 the average mass of these molecules , * ? ? ?
* ) , we compute @xmath95 where @xmath96 is the mid - plane density and @xmath97 the thermal velocity of gas molecules .
+ the parameter @xmath98 that appears in eq .
( [ phforce ] ) measures the thermal relaxation of aggregates , we have ( see * ? ? ?
* ) @xmath99 , with @xmath100 the thermal conductivity of gas and @xmath101 the effective thermal conductivity given by the following expression , @xmath102 where @xmath103 is the emissivity ( assumed to be 1 ) and @xmath104 is the heat conductivity of aggregates . as usual , @xmath105 denotes the stefan - boltzmann constant .
although minerals , in massive and not porous form , have thermal conductivity around @xmath106 w.m@xmath22 k@xmath22 ( see for instance * ? ? ?
* ) , sketch of the photophoretic transport model discussed in this paper .
the disk structure has been computed between @xmath107 au and @xmath108 au .
the existence of an inner gap of radius @xmath109 is hypothesized.,width=9 ] aggregates due to their porosity are believed worse thermal conductors than massive dust grains . in their experimental investigations of laboratory analogs , @xcite found thermal conductivity ranging from @xmath110 to @xmath111 w m@xmath22 k@xmath22 .
we have fixed @xmath104 to @xmath112 w m@xmath22 k@xmath22 in order to facilitate comparisons with earlier researches having adopted this value @xcite . nevertheless , in sect
. [ discu ] we will explore the influence of @xmath104 .
the density of dust particles have been fixed to @xmath113 kg m@xmath114 ; this value has been adopted by @xcite and @xcite .
this density value is based on an olivine density of @xmath115 kg m@xmath114 @xcite , and on an aggregates filling factor of @xmath116% @xcite .
the influence of @xmath117 is briefly discussed in sect .
[ influ_rhop ] .
the photophoretic force ( see eq .
[ phforce ] ) is not the only one taken into account ; the force @xmath118 due to residual gravity @xcite and the radiative pressure force are also considered .
these forces are respectively provided by @xmath119 @xmath120 where @xmath121 is the mass of an aggregate , @xmath122 and @xmath123 are the density and the pressure of the gas ; @xmath124 represents the radiative flux ( w.m@xmath18 ) at time @xmath39 , at the location of the test particle .
the speed of light is noted @xmath125 .
the radiative flux @xmath124 , as it appears in eq .
( [ phforce ] ) and ( [ frad ] ) is derived from the star luminosity @xmath126 .
similarly to the previous works @xcite , we have hypothesized the existence of an inner gap of radius @xmath127 , optically thin enough for particles to see the proto - sun , but still containing the gas content of the disk structure .
this assumption is supported by an increasing number of observational evidences .
for instance , @xcite show that observations of the pre - main sequence star coku tau/4 suggest the presence of an accretion disk with an inner hole cleared of small dust grains .
the infrared imaging survey , conducted by @xcite with the spitzer space telescope , indicates that around 10% of the stars of their sample , owning a disk , exhibit spectral features explained by the existence of an optically thin inner disk .
in addition , @xcite studied three disks , and detected an inner gap depleted in dust particles but containing gas . up to now , it is not clear which physical mechanism could be the origin of such inner cavities or gaps in protoplanetary disks ( see for instance * ? ? ? * ) .
this is the reason why we have only postulated their existence .
the model is sketched in fig .
[ sketch ] .
the radiative transfer along the line of sight between particles and proto - sun is treated in a very simplified way .
firstly , neglecting the absorption by the gas , the radiative flux @xmath128 ( w m@xmath18 ) is evaluated at the distance @xmath19 reached by particles at the time @xmath39 @xmath129 this reflects the sphericity of stellar emissions .
in a second step , the equation @xmath130 is integrated at fixed @xmath39 from @xmath131 to @xmath132 using the boundary condition @xmath133 ; in this manner the effect of gas absorption is combined with the pure geometrical decrease .
the opacity @xmath134 is due to rayleigh s scattering for which @xcite derived @xmath135 ( [email protected]@xmath22 ) .
it has to be emphasize that , using eq .
( [ kappar ] ) , the disk is assumed to have been cleared of dust by the photophoretic transport between @xmath127 and @xmath137 ( see fig .
[ sketch ] ) .
we also point out that the employment of the opacity law provided by @xcite ( see [ vertstruct ] ) is not more relevant , since it assumes the absence of dust depletion caused by the photophoretic transport . with an overestimated contribution of dust grains , @xcite opacity law yields to the annihilation of the photoretic migration .
thus , the approach adopted here is clearly an idealized situation and the derived results have to be considered as maximized effects .
we will come back to this opacity issue in a dedicated section of the discussion .
the use of eq .
( [ kappar ] ) means that the effect of the thermal radiation field of the gas itself and the effect of the photon multi - scattering have also been neglected .
finally , the aggregates are drifted ( in the gas frame ) in the radial direction with the velocity ( see fig .
[ sketch ] ) @xmath138 where @xmath139 , the coupling time of particles with gas , is provided by @xmath140 the correction factor @xmath141 is expressed as a function of the knudsen number @xmath80 @xcite @xmath142 in protoplanetary disks , the mid - plane temperature decreases with the distance to the central star .
this is why we have checked that the thermophoretic force given by @xcite , and recalled in @xcite , remains negligible compared to the forces contributing to eq .
( [ vdrift ] ) . for a given age of the disk ,
the distance @xmath132 reached by particles is obtained by integrating the difference between the drift velocity @xmath143 and the accretion velocity @xmath144 , we have @xmath145 more rigourously than @xcite , we derived @xmath144 from the disk model , that provides naturally the speed of the gas which is falling to the star ( see fig .
[ sketch ] ) @xmath146 @xcite .
the concept of the minimum mass of solar nebula ( hereafter mmsn ) dates back to the end of the seventies @xcite , and the initial total mass of the solar nebula is now expected to be approximately a few mmsn ( _ e.g. _ * ? ? ?
* found @xmath106 mmsn ) .
we have chosen @xmath147 mmsn as a typical value , that allows easy comparison with past investigations @xcite .
the proto - planetary disks are believed to be the place of turbulent mixing produced by magneto - rotational instabilities @xcite and the associated @xmath23-parameter introduced by @xcite has admitted values within the interval @xmath148 ; we have selected @xmath149 as our nominal value which also facilitates comparisons . the literature reported the observation of accretion disks having a central hole with a radius of @xmath150 au ( see for instance * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ; for this reason we first assumed the existence of such an inner gap of radius @xmath151 au .
this strong hypothesis , already made by @xcite and @xcite , favors clearly the transport driven by photophoresis since it decreases the radiative flux attenuation between the proto - sun and the dust grains .
+ the results of our first calculations can be seen in fig .
[ firstresults](a ) , they are very similar to those plotted in fig . 4 of @xcite ; for which equivalent parameters and hypothesis have been chosen .
we recall that , in this case , the lifetime of the protoplanetary disk has been adjusted to @xmath152 myr ( see * ? ? ?
* ; * ? ? ?
* ) by tuning the photoevaporation parameter @xmath30 . in moudens _
_ s work the disk is not irradiated , and the solar luminosity was provided by @xcite for the sun at the zams .
for each simulation of a disk evolution , we stopped the calculation when the total mass of the disk reached @xmath153% of its initial mass .
this criterion will be applied in all simulations discussed in the rest of the present work .
+ in fig .
[ firstresults](b ) an identical disk structure is kept but the photophoresis force is computed in a slightly different way : the solar luminosity is no longer taken constant at its zams value , but instead it follows the luminosity variations along the pms tracks plotted in fig .
in addition , the gas opacity which depends on the stellar effective temperature @xmath67 ( see eq .
[ kappar ] ) , is computed consistently .
not surprisingly , the aggregates are very efficiently pushed outwards , because the luminosity at the birthline is about @xmath154 times larger than that at the zams .
interestingly , particles with a radius of @xmath155 m reach @xmath106 au after @xmath156 myr and do not move further away even after @xmath152 myr . + keeping fixed the inner gap at @xmath157 au , and letting unchanged all the other parameters , we adopted a protoplanetary disk model computed with a time - dependent stellar luminosity given by the already mentioned pms track .
the subsequent positions of the aggregates are represented in fig .
[ firstresults ] ( c ) .
it is clearly visible that the use of an irradiated disk damps the particle movement .
this behavior could be caused either by a higher accretion velocity or by smaller forces . in order to disentangle the various effects
, we have focused our analysis on the movement of particles of radius @xmath60 m , which appear to be typical at the examination of fig .
[ firstresults ] ( b ) and fig . [ firstresults ] ( c ) .
the panel ( a ) of fig .
[ expla ] indicates that the accretion velocity undergone by the considered particle , along its path in the disk mid - plane , is larger ( in absolute value ) when the disk structure is irradiated . beside this
, we have checked that during the period of interest , _
i.e. _ before @xmath158 myr , the photophoretic force has the dominant contribution to the drift velocity ( see eq .
[ vdrift ] ) . to catch the effect of the disk structure
, we have normalized the photophoretic force by the received flux of light ( see eq .
[ phforce ] ) .
the resulting ratio @xmath159 is either not significantly different or slightly higher , respectively when irradiation is accounted or when it is not ( see fig .
[ expla].b ) .
therefore , the larger accretion velocity appears to weaken the efficiency of the photophoretic force during the t tauri phase of the sun . [ cols="<,^,^,^,^,^,^",options="header " , ] [ forcesat2au ] at the external surface , the force due to the residual gravity is lower than at the disk mid - plane ; all the three forces have a similar order of magnitude ( _ i.e. _ @xmath160 n ) although the radiative pressure force dominates slightly the others .
the major difference between situations at external surface and mid - plane consists in the role played by @xmath161 at the surface
[ forcesat2au ] gives precise value of forces at @xmath157 au from the proto - sun .
we can remark that the photophoretic force does not undergo a large change between the two locations : the effect of the higher radiative flux at the surface seems to be compensated by the increasing scarcity of gas molecules contributing to the photophoresis .
finally , the global resulting force ( _ i.e. _ given by the sum @xmath162 ) appears to be not so different at disk mid - plane and at external surface ( _ i.e. _ it ranges between @xmath163 for our test particle ) . as a consequence , the quantitatively significant difference between mid - plane and external surface comes from factor of @xmath164 in gas
grain coupling time @xmath139 ( see fig.[fphetc ] panels ( g ) and ( h ) for @xmath165 au ) .
this ratio of 4 orders of magnitude is due to local thermodynamic conditions .
we recall that @xmath166 ( see eqs .
[ tau ] and eq .
[ cc ] ) in a low density environment .
moreover , in low density regions @xmath87 becomes very low implying large @xmath167 values , while the knudsen s number @xmath80 tends to be very large , and for @xmath168 we have @xmath169 , so that @xmath141 also increases .
the net result is a strong rise of @xmath139 .
the physical meaning of these large value of @xmath139 is a weak coupling between the aggregates and the gas .
+ in summary , at the external boundary of the dense region of the protoplanetary disk the photophoretic force may not be the dominant force , particularly if the thermal conductivity @xmath170 is much higher than @xmath112 w m@xmath22 k@xmath22 .
the fact that the aggregates particles are weakly coupled to the gas allows any small force to generate a transport process . in such circumstances ,
particles can sediment , be pushed outward by stellar wind bursts , taken away by photoevaporation , etc .
a study of such transport processes is far beyond the scope of the present paper and required a minimum modelization of the disk s `` atmosphere '' .
the 1 + 1d disk model used here implements the classical @xmath23-viscosity prescription that allows a description of the physical evolution of the disk as it undergoes mass and angular momentum transport . to this standard approach , we have added time - dependent irradiation which is consistent with the evolution of the proto - sun along the pre - main sequence phase .
since this phase corresponds to the t tauri period of our star , which is known to be contemporary of the existence of a circumstellar dusty and gaseous accretion disk , our model represents a noticeable improvement of the proto - solar nebula modeling .
this is particularly true in our context , in which we examine the influence of proto - sun radiation on the transport of dust driven by photophoresis .
+ by adopting the same prescription used in previous works for the opacity of the gas , we have found that the high luminosity of the sun considered in its t tauri phase favors the migration of dust grains to the outer parts of the protoplanetary disk . however , the effect is not as high as if the irradiation would not change the disk properties .
indeed , the irradiation enhances the computed turbulent viscosity which in turn increases the accretion velocity .
consequently , the particles are slowed down whereas the disk lifetime is reduced .
this latter effect reduces the need for an _ ad hoc _ photoevaporation in order to get ages in agreement with astronomical observations . unfortunately , since the dust particles have migration velocities that depend on their size , a trail of small grains should be left along the line of sight .
the smallest aggregates show an important contribution to the gas dust opacity .
this leads to a strong extinction which could eliminate the photophoretic effect .
this point is one of our most important results and , models involving a consistent treatment of the opacity as a function of dust content , are highly desirable .
noticeably , external optically thin disk regions could provide an environment favorable to stellar photons driven transport processes .
in such a case , the net quantity of dust delivered to the comets formation zone could not ba large enough to explain the observations , because of the vertical stratification of grain distribution .
+ besides this , using gas temperature distribution provided by our model , we confirm that amorphous grains can not be annealed into regions beyond @xmath150 au , due to arguments based on phase transition kinetics at gas temperature .
more importantly , we have identified the heat conductivity of aggregates as a crucial parameter . using realistic estimations and
published experimental results for the thermal conductivity , we have shown that the photophoretic strength can be considerably reduced and could yield to situations in which photophoretic migration through the mid - plane could be marginal even if the disk is irradiated by a bright pms star .
+ laboratory experimentations conducted by @xcite on ice aggregates , trapped in a cell under the combined effects of photophoresis and thermophoresis , show that rotation induced by photophoretic forces does not change the strength of photophoresis force , but ignores the influence of a turbulent flow . @xcite and @xcite have discussed , respectively the possible effects of the turbulence and of the particles rotation . however , they considered turbulence and rotation as independent processes .
@xcite , who discussed turbulence as a factor affecting the mean radial motion of dusts , concluded that turbulence essentially does not prevent the outward migration .
@xcite have looked to the influence of rotation of particles on themselves , induced for instance by particle to particle collisions .
we have to keep in mind that dusts are embedded in a gaseous environment , to which they are strongly coupled ( at least in the mid - plane ) . following a simple picture ,
a turbulent flow is made of a cascade of eddies , with a distribution of sizes and lifetimes .
in such a flow , aggregates can meet eddies with rotational motion in a plane that contains the radial direction . in this configuration ,
if the thermal relaxation timescale of the particle is larger than the overturning timescale of the turbulent eddy , the temperature distribution at the surface of the particle could be uniformized , leading to the removal of any photophoretic force . the global effect on dust migration through
the disk depends on the intrinsic properties of dust ( thermal properties and aerodynamics drag ) and on statistical properties of the involved turbulent flow .
the net result will be a convolution of the statistical distribution of grains properties ( bulk material nature , porosity , size , etc ) and of the properties of the cascade of turbulent eddies .
+ in a first approach , the influence of the turbulence on dust trajectories could also be investigated using a particle - tracking technics as already employed in other contexts by @xcite or @xcite .
thanks to this method , applied in a 2d or better a 3d geometry , we could follow the trajectories of particles within the disk and even those that could be launched upwards , pushed outward by the radiation pressure and possibly fall back onto the nebula at different locations . a similar scenario has been already studied by @xcite in their investigations concerning the outward transport of cais during fu - orionis events .
the vertical transport of grains , caused by convection and thermophoresis could be also included together with the existence of a quiet `` dead zone '' in which the turbulent activity should be very low .
our model of accretion disk is basically based on a generalized version of the procedure originally published by @xcite .
more specifically , this is a 1 + 1d model for which the turbulence is treated in the frame of the well known @xmath23 formalism @xcite . during its temporal evolution ,
the disk is irradiated by the star which also evolves along its pre - main sequence evolutionary track .
both static model and disk temporal evolution programs have been implemented from scratch in ` fortran 2008 ` , and parallelized using open multi - processing ( openmp ) .
we have nicknamed the whole package ` evad ` .
the following sections provide a detailled description of our model .
the vertical structure of the disk is governed by the equations ( see * ? ? ?
* ; * ? ? ?
* ) ( already labeled pt99 ) , @xcite @xmath171 @xmath172 @xmath173 where @xmath123 , @xmath174 and @xmath90 are respectively the pressure , the vertical radiative flux and the temperature ; @xmath175 represents the altitude above the mid - plane and @xmath176 is the keplerian angular velocity .
the density is denoted @xmath177 while @xmath178 is the opacity of the disk s material taken in @xcite .
following pt99 the boundary condition at the external disk surface are given by @xmath179 where @xmath180 with @xmath181 ( the disk surface mass density ) and @xmath182 the vertically averaged viscosity .
the pressure at external surface is @xmath183 where @xmath184 is the disk mid - height , @xmath185 is the optical depth above the disk ( following pt99 , we have taken @xmath186 ) and @xmath187 is the opacity at the external surface .
the temperature @xmath31 can be obtained by solving for given values of @xmath23 and @xmath188 @xmath189 where @xmath190 is given by @xmath191 where @xmath35 is given by eq .
( [ tirr ] ) , @xmath33 is the background temperature , _
i.e. _ the temperature of the medium in which the disk is immersed .
we have chosen @xmath34 k. the mid - height of the disk being not known a priori , we have to solve a two boundary value problems ( tbvp ) .
while some authors work with relaxation algorithms @xcite ; we have prefered an algorithm based on shooting methods @xcite turning our tbvp into an initial values problems ( ivp ) : the equations of the vertical structure are integrated from the surface of the disk to the mid - plan . the 5th - order runge - kutta method with adaptive step length described in @xcite , already used in previous other works
( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , has been employed to perform the vertical integration . to reduce the effects of the stiffness of the set of equations , we have opted for after several tests the set of variables
@xmath192 , @xmath193 and @xmath194 ( also adopted by * ? ? ?
the equation @xmath195 is then solved by a root finding method .
finally , the vertically averaged turbulent viscosity @xmath196 is tabulated as a function of disk surface density @xmath197 , the steady state accretion rate @xmath198 , the turbulent viscosity parameter @xmath23 and the age of the system @xmath39 ( _ i.e. _ providing @xmath38 and @xmath37 which are directly involved in the calculation . during the integration of the eq .
( [ eq_evol ] ) , these pre - built tables are interpolated `` on - the - fly '' by a dedicated routine involving advanced b - splines technics @xcite
. the temporal evolution of the disk is governed by eq .
( [ eq_evol ] ) , combined with boundary conditions similar to those used by @xcite , which is a non - linear equation ( @xmath20 depends on the solution @xmath17 ) .
unfortunately , there is no standard numerical method for such equation . although explicit finite difference method is often used ( _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) the time - step must meet the courant - friedrichs - lewy condition ( see * ? ? ?
* ) that limits the value of the time - step to small values . instead of an explicit finite difference scheme
, we have chosen a fully implicit scheme which has the great advantage of being unconditionally stable at least in the case of linear equations .
the obtained set of non - linear finite difference equations is solved using a multidimensional newton raphson algorithm .
our approach is similar to the methods used by @xcite and @xcite , respectively in the contexts of stellar structure or accretion disc outburst .
we used a decentralized finite difference formula @xmath199 where @xmath200 is a chosen function of @xmath19 .
the initial distribution of matter being of the form @xmath201 ; we have found convenient to adopt a function like @xmath202 where @xmath203 is an adjustable real parameter that tunes the distribution of points through the disk .
finally , the evolution equation has been rewritten using @xmath204 and @xmath205 .
121 natexlab#1#1url # 1`#1`urlprefix , y. , mordasini , c. , benz , w. , winisdoerffer , c. , apr .
2005 . models of giant planet formation with migration and disc evolution .
a&a 434 , 343353 . , r. l. , bernstein , g. m. , malhotra , r. , mar .
2001 . the edge of the solar system .
apjl 549 , l241l244 . ,
a. , lauretta , d. s. ( eds . ) , 2010 .
protoplanetary dust astrochemical and cosmochemical perspective , 2nd edition .
cambridge university press . ,
m. , grevesse , n. , sauval , a. j. , sep .
2005 . the solar chemical composition . in : barnes , iii , t. g. , bash , f. n. ( eds . ) , cosmic abundances as records of stellar evolution and nucleosynthesis .
336 of astronomical society of the pacific conference series .
p. 25 .
, s. a. , hawley , j. f. , jan .
instability , turbulence , and enhanced transport in accretion disks .
reviews of modern physics 70 , 153 . , k. r. , lin , d. n. c. , jun
. 1994 . using fu orionis outbursts to constrain self - regulated protostellar disk models .
apj 427 , 9871004 . , s. , chernyak , v. , fomyagin , g. , 1993 .
physics of fluids 5 , 2043 . , g. , wu , y. , jan .
formation of narrow dust rings in circumstellar debris disks .
apj 655 , 528540 . , t. , 2011 .
the evolution of gas and dust in protoplanetary accretion disks .
thesis , phd thesis , 2011 . ,
j. , gundlach , b. , mhle , s. , trigo - rodriguez , j. m. , jun . 2014 .
comets formed in solar - nebula instabilities ! - an experimental and modeling attempt to relate the activity of comets to their formation process 235 , 156169 . , j. , schrpler , r. , sep .
structure and mechanical properties of high - porosity macroscopic agglomerates formed by random ballistic deposition .
physical review letters 93 ( 11 ) , 115503 . , j. , wurm , g. , kempf , s. , poppe , t. , klahr , h. , kozasa , t. , rott , m. , henning , t. , dorschner , j. , schrpler , r. , keller , h. u. , markiewicz , w. j. , mann , i. , gustafson , b. a. , giovane , f. , neuhaus , d. , fechtig , h. , grn , e. , feuerbacher , b. , kochan , h. , ratke , l. , el goresy , a. , morfill , g. , weidenschilling , s. j. , schwehm , g. , metzler , k. , ip , w .- h . ,
sep . 2000 .
growth and form of planetary seedlings : results from a microgravity aggregation experiment .
physical review letters 85 , 2426 . ,
d. , gautier , d. , hersant , f. , hur , j .-
m . , robert , f. , mar .
turbulent radial mixing in the solar nebula as the source of crystalline silicates in comets .
384 , 11071118 . ,
a. p. , apr .
2008 . mixing in the solar nebula : implications for isotopic heterogeneity and large - scale transport of refractory grains .
earth and planetary science letters 268 , 102109 . ,
j. , alencar , s. h. p. , harries , t. j. , johns - krull , c. m. , romanova , m. m. , 2007 .
magnetospheric accretion in classical t tauri stars .
protostars and planets v , 479494 . ,
d. , tsou , p. , alon , j. , alexander , c. m. o. . , araki , t. , bajt , s. , baratta , g. a. , bastien , r. , bland , p. , bleuet , p. , borg , j. , bradley , j. p. , brearley , a. , brenker , f. , brennan , s. , bridges , j. c. , browning , n. d. , brucato , j. r. , bullock , e. , burchell , m. j. , busemann , h. , butterworth , a. , chaussidon , m. , cheuvront , a. , chi , m. , cintala , m. j. , clark , b. c. , clemett , s. j. , cody , g. , colangeli , l. , cooper , g. , cordier , p. , daghlian , c. , dai , z. , dhendecourt , l. , djouadi , z. , dominguez , g. , duxbury , t. , dworkin , j. p. , ebel , d. s. , economou , t. e. , fakra , s. , fairey , s. a. j. , fallon , s. , ferrini , g. , ferroir , t. , fleckenstein , h. , floss , c. , flynn , g. , franchi , i. a. , fries , m. , gainsforth , z. , gallien , j .- p .
, genge , m. , gilles , m. k. , gillet , p. , gilmour , j. , glavin , d. p. , gounelle , m. , grady , m. m. , graham , g. a. , grant , p. g. , green , s. f. , grossemy , f. , grossman , l. , grossman , j. n. , guan , y. , hagiya , k. , harvey , r. , heck , p. , herzog , g. f. , hoppe , p. , hrz , f. , huth , j. , hutcheon , i. d. , ignatyev , k. , ishii , h. , ito , m. , jacob , d. , jacobsen , c. , jacobsen , s. , jones , s. , joswiak , d. , jurewicz , a. , kearsley , a. t. , keller , l. p. , khodja , h. , kilcoyne , a. l. d. , kissel , j. , krot , a. , langenhorst , f. , lanzirotti , a. , le , l. , leshin , l. a. , leitner , j. , lemelle , l. , leroux , h. , liu , m .- c . , luening , k. , lyon , i. , macpherson , g. , marcus , m. a. , marhas , k. , marty , b. , matrajt , g. , mckeegan , k. , meibom , a. , mennella , v. , messenger , k. , messenger , s. , mikouchi , t. , mostefaoui , s. , nakamura , t. , nakano , t. , newville , m. , nittler , l. r. , ohnishi , i. , ohsumi , k. , okudaira , k. , papanastassiou , d. a. , palma , r. , palumbo , m. e. , pepin , r. o. , perkins , d. , perronnet , m. , pianetta , p. , rao , w. , rietmeijer , f. j. m. , robert , f. , rost , d. , rotundi , a. , ryan , r. , sandford , s. a. , schwandt , c. s. , see , t. h. , schlutter , d. , sheffield - parker , j. , simionovici , a. , simon , s. , sitnitsky , i. , snead , c. j. , spencer , m. k. , stadermann , f. j. , steele , a. , stephan , t. , stroud , r. , susini , j. , sutton , s. r. , suzuki , y. , taheri , m. , taylor , s. , teslich , n. , tomeoka , k. , tomioka , n. , toppani , a. , trigo - rodrguez , j. m. , troadec , d. , tsuchiyama , a. , tuzzolino , a. j. , tyliszczak , t. , uesugi , k. , velbel , m. , vellenga , j. , vicenzi , e. , vincze , l. , warren , j. , weber , i. , weisberg , m. , westphal , a. j. , wirick , s. , wooden , d. , wopenka , b. , wozniakiewicz , p. , wright , i. , yabuta , h. , yano , h. , young , e. d. , zare , r. n. , zega , t. , ziegler , k. , zimmerman , l. , zinner , e. , zolensky , m. , dec
comet 81p / wild 2 under a microscope .
science 314 , 1711. , d. e. , 1978 .
microparticle studies by sampling techniques . in : mcdonnell , j. ( ed . ) , cosmic dust .
john wiley & sons , pp .
295336 . , h. , ryan , e. v. , jun .
the identification of crystalline olivine in cometary silicates .
apj 341 , 10591066 . , j. k. , jan . 1992 .
accretion disks in active galactic nuclei - vertically explicit models 385 , 94107 . , f. j. , oct . 2007 .
outward transport of high - temperature materials around the midplane of the solar nebula .
science 318 , 613. , f. j. , sandford , s. a. , apr . 2012 . organic synthesis via irradiation and warming of ice grains in the solar nebula .
science 336 , 452. , r. , friedrichs , k. , lewy , h. , 1928 .
ber die partiellen differenzengleichungen der mathematischen physik .
mathematische annalen 100 , 3274 . ,
minimum mass solar nebulae and planetary migration .
apj 698 , 606614 . , j. , 1910 .
proceedings of the royal society of london series a 83 , 357 . ,
j. n. , davis , s. s. , dobrovolskis , a. r. , dec . 2003 . blowing in the wind .
ii . creation and redistribution of refractory inclusions in a turbulent protoplanetary nebula 166 , 385402 . , j. n. , estrada , p. r. , davis , s. s. , feb .
utilitarian opacity model for aggregate particles in protoplanetary nebulae and exoplanet atmospheres .
apjs 210 , 21 . , p. , cant , j. , calvet , n. , lizano , s. , jun .
accretion disks around young objects . i. the detailed vertical structure .
apj 500 , 411427 . , p. , hartmann , l. , calvet , n. , franco - hernndez , r. , forrest , w. j. , sargent , b. , furlan , e. , uchida , k. , green , j. d. , watson , d. m. , chen , c. h. , kemper , f. , sloan , g. c. , najita , j. , mar .
2005 . the truncated disk of coku tau/4 621 , 461472 . , c. , 1985 . a practical guide to splines , 3rd edition .
springer . , s. , prada moroni , p. g. , marconi , m. , ruoppo , a. , aug .
2008 . the franec stellar evolutionary code .
apss 316 , 2530 . , m. , valle , g. , deglinnocenti , s. , prada moroni , p. g. , apr . 2012 . the pisa stellar evolution data base for low - mass stars .
a&a 540 , a26 . ,
s. e. , willacy , k. , bodenheimer , p. , turner , n. j. , beichman , c. a. , apr .
ice lines , planetesimal composition and solid surface density in the solar nebula 200 , 672693 . , p. p. , 1971 .
the evolution of low mass stars .
mnras 151 , 351 . , m. , loeffler , m. j. , raut , u. , baragiola , r. a. , may 2010 .
radiation - induced amorphization of crystalline ice 207 , 314319 . ,
j. , dougados , c. , cabrit , s. , jul . 2006 . which jet launching mechanism(s ) in t tauri stars ?
453 , 785796 .
, l. , alibert , y. , mordasini , c. , benz , w. , apr . 2012 .
effects of disk irradiation on planet population synthesis .
a&a 540 , a107 .
, j. , king , a. , raine , d. , 1992 .
accretion power in astrophysics .
cambridge uni . press .
, san diego . ,
, oct . 2001 .
radial mixing in protoplanetary accretion disks .
i. stationary disc models with annealing and carbon combustion 378 , 192213 . , j. m. , 1985 . the chemical and physical evolution of interstellar dust . physica scripta volume t 11 , 1426 .
, b. , blum , j. , jun . 2012 .
outgassing of icy bodies in the solar system - ii : heat transport in dry , porous surface dust layers 219 , 618629 . , c. , blum , j. , zsom , a. , ormel , c. w. , dullemond , c. p. , apr .
the outcome of protoplanetary dust growth : pebbles , boulders , or planetesimals ? .
i. mapping the zoo of laboratory collision experiments .
a&a 513 , a56 . , jr . , k. e. , lada , e. a. , lada , c. j. , jun . 2001 .
disk frequencies and lifetimes in young clusters .
apj 553 , l153l156 . ,
m . , menou , k. , dubus , g. , lasota , j .-
hure , j .-
, aug . 1998 .
accretion disc outbursts : a new version of an old model .
mnras 298 , 10481060 . ,
d. e. , desch , s. j. , feb .
annealing of silicate dust by nebular shocks at 10 au .
apjl 565 , l109l112 . ,
c. , 1981 .
structure of the solar nebula , growth and decay of magnetic fields and effects of magnetic and turbulent viscosities on the nebula .
progress of theoretical physics supplement 70 , 3553 . ,
k. , simmons , g. , 1969 .
earth and planetary science letters 6 , 359368 . ,
r. , guillot , t. , nov .
evolution of protoplanetary disks : constraints from dm tauri and gm aurigae .
a&a 442 , 703725 . , a. m. , andrews , s. m. , wilner , d. j. , meyer , m. r. , carpenter , j. m. , qi , c. , hales , a. s. , casassus , s. , hogerheijde , m. r. , mamajek , e. e. , wolf , s. , henning , t. , silverstone , m. d. , sep . 2010 . structure and composition of two transitional circumstellar disks in corona australis .
aj 140 , 887896 . , j .-
2000 . on the transition to self - gravity in low mass agn and yso accretion discs .
a&a 358 , 378394 . , d. k. , harper , m. h. , felder , r. l. , 1995 . aerosol science and technology 22 , 202 . , l. , sui , n. , feb . 2010 .
the evolution of the solar nebula i. evolution of the global properties and planet masses .
apj 710 , 11791194 . , c. , gail , h .-
radial mixing in protoplanetary accretion disks .
vi . mixing by large - scale radial flows .
a&a 415 , 11771185 .
, m. s. , wooden , d. h. , aug .
the composition of dust in jupiter - family comets inferred from infrared spectroscopy 57 , 11331145 . , f. , vriend , w. j. , tielens , a. g. g. m. , jul .
the absence of crystalline silicates in the diffuse interstellar medium .
apj 609 , 826837 . , j. e. , trivedi , n. c. , barker , j. m. , krukowski , s. , 2006 .
industrial minerals & rocks ( book - cd set ) , 7th edition , 7th edition .
society for mining , metallurgy , and exploration , englewood , co 80112 . , m. , blum , j. , skorov , y. v. , trieloff , m. , jul . 2011 .
thermal conductivity measurements of porous dust aggregates : i. technique , model and first results 214 , 286296 . , o. , wurm , g. , sep . 2005 .
photophoresis and the pile - up of dust in young circumstellar disks .
apj 630 , 10881092 . , o. , wurm , g. , mousis , o. , petit , j .-
m . , horner , j. , alibert , y. , feb .
the photophoretic sweeping of dust in transient protoplanetary disks .
a&a 462 , 977987 . ,
c. , remusat , l. , bernard , s. , brearley , a. j. , leroux , h. , sep . 2013 .
amorphization and d / h fractionation of kerogens during experimental electron irradiation : comparison with chondritic organic matter 226 , 101110 . , p. , gail , h
.- p . , henning , t. , jul .
dust evaporation in protostellar cores .
apj 447 , 848 .
d. r. , e. , 2005 .
handbook of chemistry and physics , 1st edition . , d. n. c. , papaloizou , j. , apr . 1980 .
on the structure and evolution of the primordial solar nebula 191 , 3748 . , s. s. , wooden , d. h. , harker , d. e. , kelley , m. s. , woodward , c. e. , murphy , j. r. , mar . 2013 .
absorption efficiencies of forsterite .
i. discrete dipole approximation explorations in grain shape and size 766 , 54 . , c. , wurm , g. , teiser , j. , friedrich , j. m. , bischoff , a. , dec . 2013 .
photophoretic strength on chondrules .
1 . modeling 778 , 101 . , h. f. , 2012 .
the effect of thermophoresis on the particle deposition on a cylinder .
thesis , ntnu trondheim norwegian university of science and technology . ,
d. , pringle , j. e. , sep .
1974 . the evolution of viscous discs and the origin of the nebular variables .
mnras 168 , 603637 . , p. , donn , b. , jun .
aerodynamic properties of fractal grains - implications for the primordial solar nebula 329 , l39l41 . , f. , meyer - hofmeister , e. , feb . 1982 .
vertical structure of accretion disks 106 , 3442 . ,
j. a. , chen , x. , taam , r. e. , feb .
the vertical structure and stability of accretion disks surrounding black holes and neutron stars 421 , 668676 . ,
s. , osaki , y. , 1983 .
disk - instability model for outbursts of dwarf novae time - dependent formulation and one - zone model 35 , 377396 . ,
s. , tuchman , y. , wheeler , j. c. , aug . 1990 .
structure and evolution of irradiated accretion disks .
ii . dynamical evolution of a thermally unstable torus 359 , 176 . ,
a. , mousis , o. , petit , j .-
m . , wurm , g. , cordier , d. , charnoz , s. , jul .
photophoretic transport of hot minerals in the solar nebula .
a&a 531 , a106 . ,
o. , petit , j .-
m . , wurm , g. , krauss , o. , alibert , y. , horner , j. , may 2007 .
photophoresis as a source of hot minerals in comets .
a&a 466 , l9l12 . , m. j. , charnley , s. b. , sep . 2011 . the chemical composition of comets - emerging taxonomies and natal heritage .
ara&a 49 , 471524 . , t. , nakagawa , y. , feb . 1994 .
formation , early evolution , and gravitational stability of protoplanetary disks .
apj 421 , 640650 . , r. c. , westphal , a. j. , gainsforth , z. , butterworth , a. l. , fakra , s. c. , marcus , m. a. , nov . 2009 .
nebular mixing constrained by the stardust samples .
meteoritics and planetary science 44 , 16751681 . , m. , encrenaz , t. , roques , f. , selsis , f. , casoli , f. , 2009 .
planetary systems detection , formation and habitability of extrasolar planets , 3rd edition .
springer - verlag , berlin , heidelberg . , c. w. , spaans , m. , tielens , a. g. g. m. , jan .
dust coagulation in protoplanetary disks : porosity matters .
a&a 461 , 215232 . ,
t. c. , dec . 2006 .
the origin of nitrogen atmospheres on earth and titan .
agu fall meeting abstracts , a5 + . ,
j. , faulkner , j. , lin , d. n. c. , nov . 1983 . on the evolution of accretion disc flow in cataclysmic variables .
ii - the existence and nature of the collective relaxation oscillations in dwarf nova systems .
mnras 205 , 487513 . , j. c. b. , terquem , c. , aug .
critical protoplanetary core masses in protoplanetary disks and the formation of short - period giant planets .
apj 521 , 823838 . , i. , tachibana , s. , feb . 2010 . the clearing of protoplanetary disks and of the protosolar nebula .
. 263298 . , a. , cassisi , s. , salaris , m. , castelli , f. , sep .
2004 . a large stellar evolution database for population synthesis studies .
i. scaled solar models and isochrones .
apj 612 , 168190 . , k. m. , blake , g. a. , van dishoeck , e. f. , smette , a. , ireland , m. j. , brown , j. , sep .
spectroastrometric imaging of molecular gas within protoplanetary disk gaps .
apj 684 , 13231329 . , m. a. , christensen , p. r. , mar .
thermal conductivity measurements of particulate materials 1 . a review .
jgr 102 , 65356550 . , w. , teukolsky , s. , vetterling , w. , flannery , b. , 1992 .
numerical recipes in fortran 77 . cambridge university press . , j. e. , 1981 . accretion discs in astrophysics .
ara&a 19 , 137162 . , f. , 1967 .
berkeley physics course : statistical physics .
berkeley physics course .
mcgraw - hill . , a. , czerny , b. , ycki , p. t. , pojmaski , g. , may 1999 . vertical structure of accretion discs with hot coronae in active galactic nuclei 305 , 481491 . ,
s. p. , lin , d. n. c. , sep .
the global evolution of the primordial solar nebula .
apj 308 , 883901 . , h. w. , 1935 .
18 , 1 . , n. i. , sunyaev , r. a. , 1973 .
a&a 24 , 337 .
, f. h. , shang , h. , lee , t. , mar .
1996 . toward an astrophysical theory of chondrites .
science 271 , 15451552 . ,
a. , hartmann , l. , calvet , n. , megeath , s. t. , muzerolle , j. , allen , l. , dalessio , p. , mern , b. , stauffer , j. , young , e. , lada , c. , feb . 2006 .
disk evolution in cep ob2 : results from the spitzer space telescope 638 , 897919 . ,
j. , 1984 .
accretion in cataclysmic binaries .
iv - accretion disks in dwarf novae 34 , 161189 . , s. w. , palla , f. , 2004 .
the formation of stars , wiley edition . new york . , k. d. , lin , d. n. c. , aug . 2000 .
formation of icy planetesimals in a turbulent solar nebula 146 , 525540 . ,
j. , dodson - robinson , s. e. , jul . 2013 .
photophoresis boosts giant planet formation .
a&a 555 , a98 . ,
c. , grady , c. a. , goto , m. , wisniewski , j. p. , janson , m. , henning , t. , fukagawa , m. , honda , m. , mulders , g. d. , min , m. , moro - martn , a. , mcelwain , m. w. , hodapp , k. w. , carson , j. , abe , l. , brandner , w. , egner , s. , feldt , m. , fukue , t. , golota , t. , guyon , o. , hashimoto , j. , hayano , y. , hayashi , m. , hayashi , s. , ishii , m. , kandori , r. , knapp , g. r. , kudo , t. , kusakabe , n. , kuzuhara , m. , matsuo , t. , miyama , s. , morino , j .-
nishimura , t. , pyo , t .- s .
, serabyn , e. , shibai , h. , suto , h. , suzuki , r. , takami , m. , takato , n. , terada , h. , tomono , d. , turner , e. l. , watanabe , m. , yamada , t. , takami , h. , usuda , t. , tamura , m. , aug .
imaging of a transitional disk gap in reflected light : indications of planet formation around the young solar analog lkca 15 .
apjl 718 , l87l91 .
, a. g. g. m. , waters , l. b. f. m. , bernatowicz , t. j. , dec . 2005 .
origin and evolution of dust in circumstellar and interstellar environments . in : krot , a. n. , scott , e. r. d. , reipurth , b. ( eds . ) , chondrites and the protoplanetary disk .
341 of astronomical society of the pacific conference series .
p. 605 .
, e. , deglinnocenti , s. , prada moroni , p. g. , dec .
@xmath206li surface abundance in pre - main sequence stars .
testing theory against clusters and binary systems .
a&a 548 , a41 . , e. , prada moroni , p. g. , deglinnocenti , s. , sep .
2011 . the pisa pre - main sequence tracks and isochrones .
a database covering a wide range of z , y , mass , and age values .
a&a 533 , a109 . , k. , gomes , r. , morbidelli , a. , levison , h. f. , may 2005 .
origin of the orbital architecture of the giant planets of the solar system .
nature 435 , 459461 . , n. j. , choukroun , m. , castillo - rogez , j. , bryden , g. , apr . 2012 . a hot gap around jupiter s orbit in the solar nebula .
apj 748 , 92 . , j. , wurm , g. , feb . 2012 . the implications of particle rotation on the effect of photophoresis .
mnras 420 , 183186 . , d. , armitage , p. j. , jan .
outward migration of extrasolar planets to large orbital radii .
mnras 347 , 613624 . ,
i. , blum , j. , dec .
photophoresis of dust aggregates in protoplanetary disks .
a&a 548 , a96 . ,
m. , gail , h .- p .
radial mixing in protoplanetary accretion disks .
ii . time dependent disk models with annealing and carbon combustion 385 , 181204 .
, s. j. , sep . 1977 . the distribution of mass in the planetary system and solar nebula .
space sci .
51 , 153158 . , s. j. , sep . 1977 .
the distribution of mass in the planetary system and solar nebula .
astrophysics and space science 51 , 153158 . , s. j. , jun . 1997 .
the origin of comets in the solar nebula : a unified model 127 , 290306 . , d. h. , butner , h. m. , harker , d. e. , woodward , c. e. , jan
mg - rich silicate crystals in comet hale - bopp : ism relics or solar nebula condensates ?
icarus 143 , 126137 . , d. h. , harker , d. e. , woodward , c. e. , butner , h. m. , koike , c. , witteborn , f. c. , mcmurtry , c. w. , jun .
silicate mineralogy of the dust in the inner coma of comet c/1995 01 ( hale - bopp ) pre- and postperihelion .
apj 517 , 10341058 . , g. , haack , h. , jul .
outward transport of cais during fu - orionis events .
meteoritics and planetary science 44 , 689699 . , g. , krauss , o. , feb .
2006 . concentration and sorting of chondrules and cais in the late solar nebula 180 , 487495 . , g. , teiser , j. , bischoff , a. , haack , h. , roszjar , j. , jul .
experiments on the photophoretic motion of chondrules and dust aggregates indications for the transport of matter in protoplanetary disks 208 , 482491 . ,
g. , trieloff , m. , rauer , h. , may 2013 .
photophoretic separation of metals and silicates : the formation of mercury - like planets and metal depletion in chondrites .
apj 769 , 78 . , y. , shankland , t. j. , linhardt , s. , rubie , d. c. , langenhorst , f. , klasinski , k. , 2004 .
thermal diffusivity and conductivity of olivine , wadsleyite and ringwoodite to 20 gpa and 1373 k. physics of the earth and planetary interiors 143 , 321336 . , j. b. , 2011 .
thermophoresis of a spherical particle : reassessment , clarication , and new analysis . aerosol science and technology 45 , 927948 . , m. e. , zega , t. j. , yano , h. , wirick , s. , westphal , a. j. , weisberg , m. k. , weber , i. , warren , j. l. , velbel , m. a. , tsuchiyama , a. , tsou , p. , toppani , a. , tomioka , n. , tomeoka , k. , teslich , n. , taheri , m. , susini , j. , stroud , r. , stephan , t. , stadermann , f. j. , snead ,
c. j. , simon , s. b. , simionovici , a. , see , t. h. , robert , f. , rietmeijer , f. j. m. , rao , w. , perronnet , m. c. , papanastassiou , d. a. , okudaira , k. , ohsumi , k. , ohnishi , i. , nakamura - messenger , k. , nakamura , t. , mostefaoui , s. , mikouchi , t. , meibom , a. , matrajt , g. , marcus , m. a. , leroux , h. , lemelle , l. , le , l. , lanzirotti , a. , langenhorst , f. , krot , a. n. , keller , l. p. , kearsley , a. t. , joswiak , d. , jacob , d. , ishii , h. , harvey , r. , hagiya , k. , grossman , l. , grossman , j. n. , graham , g. a. , gounelle , m. , gillet , p. , genge , m. j. , flynn , g. , ferroir , t. , fallon , s. , ebel , d. s. , dai , z. r. , cordier , p. , clark , b. , chi , m. , butterworth , a. l. , brownlee , d. e. , bridges , j. c. , brennan , s. , brearley , a. , bradley , j. p. , bleuet , p. , bland , p. a. , bastien , r. , dec .
mineralogy and petrology of comet 81p / wild 2 nucleus samples .
science 314 , 1735. , a. , ormel , c. w. , gttler , c. , blum , j. , dullemond , c. p. , apr .
2010 . the outcome of protoplanetary dust growth : pebbles , boulders , or planetesimals ?
ii . introducing the bouncing barrier .
a&a 513 , a57 .
we acknowledge ulysse marboeuf , together with james owen and philippe rousselot for scientific discussion .
we also warmly thank pierre morel for useful advices concerning numerical aspects , and we are grateful to panayotis lavvas for his scientific comments and for reading the manuscript .
we express our grateful thanks to jeff cuzzi who provided us his opacity code .
simulations have been executed on computers from the utinam institute of the universit de franche - comt , supported by the rgion de franche - comt and institut des sciences de lunivers ( insu ) .
we thank skou diakit who help us to parallelize the _ _ e__v_ad _ source code and to deal with the utinam institute cluster .
finally , we thank the anonymous reviewers who improved the clarity of the paper with their remarks and comments . | there is a growing body of evidences for the presence of crystalline material in comets .
these crystals are believed to have been annealed in the inner part of the proto - solar nebula , while comets should have been formed in the outer regions .
several transport processes have been proposed to reconcile these two facts ; among them a migration driven by photophoresis .
the primarily goal of this work is to assess whether disk irradiation by a pre - main sequence star would influence the photophoretic transport .
to do so , we have implemented an evolving 1 + 1d model of an accretion disk , including advanced numerical techniques , undergoing a time - dependent irradiation , consistent with the evolution of the proto - sun along the pre - main sequence .
the photophoresis is described using a formalism introduced in several previous works . adopting the opacity prescription used in these former studies , we find that the disk irradiation enhances the photophoretic transport : the assumption of a disk central hole of several astronomical units in radius is no longer strictly required , whereas the need for an _ ad hoc _ introduction of photoevaporation is reduced .
however , we show that a residual trail of small particles could annihilate the photophoretic driven transport via their effect on the opacity .
we have also confirmed that the thermal conductivity of transported aggregates is a crucial parameter which could limit or even suppress the photophoretic migration and generate several segregation effects .
comets : composition , dust , origin solar nebula |
modelling the light - front hadron wave - function is a challenge as long as the exact wave - function from quantum cromodynamics is not yet avaliable .
the wave - function contains physical information complementary to the spectrum . at scales below 1 gev
, the wave - function is described by effective degrees of freedom , and the lowest fock - state component of the light - front hadron wave function is composed by the minimal number of constituent quarks necessary to give the quantum numbers . within this general framework ,
a light - front qcd - inspired model was recently applied to the pion and other mesons @xcite .
a reasonably description of the pion structure as well as the masses of the vector and pseudo - scalar mesons was found .
this model , without confinement , describes the vector meson as weakly bound system of constituent quarks .
the spin does not play a dynamical role besides justifying the contact term coming from the hyperfine interaction . in this way
the one - gluon - exchange interaction is simplified to two components only : a contact term and a coulomb - type potential .
the contact term is essential to collapse the constituent quark - antiquark system to form the pion , while the vector meson is dominated by the coulomb - type potential .
the contact interaction brings to the model the physical scale of the pion mass which determines the masses of the other pseudo - scalar and vector mesons@xcite . here
, we follow closely the work of ref.@xcite and revise the extension of the concepts coming from the effective qcd - model , applied to mesons @xcite , to study the spin 1/2 low - lying states of the nucleon , @xmath4 , @xmath5 and @xmath6@xcite .
we use a flavor independent effective interaction between the constituent quarks , a property necessary to describe the masses of these baryons@xcite .
the three - quark relativistic dynamics of the @xmath3 system is formulated within the light - front framework in a truncated fock - space@xcite which is stable under kinematical boost transformations@xcite and yields a wave - function covariant under kinematical boosts @xcite .
we use only the contact interaction , which in this case provides to the model the physical scale of the mass of the nucleon ground state , while the spin is averaged out .
the binding energy of the baryon is calculated using three - quark faddeev equations@xcite as a function of the mass of one of the constituent quarks @xmath7 , while the bare strength of the effective contact interaction and the mass of the quark @xmath8 are kept constant . this light - front model with a contact force @xcite has been applied to the proton and described its mass , charge radius and electric form factor up to 2(gev / c)@xmath9 @xcite .
recently , it was also applied to study the dissolution of the nucleon at finite temperature and baryonic density @xcite .
this work is organized as follows . in sec.ii
, we present the coupled integral equations for the faddeev components of the vertex of the three - quark light - front bound - state wave function from a flavor independent contact interaction derived in ref.@xcite . in sec .
iii , we present the numerical results for the masses and binding energies of the nucleon , @xmath4 , @xmath10 and @xmath6 , obtained from the numerical solution of light - front integral equations . a key point in this section is the assignement of the constituent quark masses which are done using the experimental values of the vector meson ground states masses . in sec .
iv , we present our summary .
the light - front hyperplane is defined by the time @xmath11 and the position coordinates on the hypersurface are defined by @xmath12 and @xmath13 @xcite .
the coordinate @xmath14 is the light - front time and the momentum , @xmath15 , corresponds to the light - front energy .
the momentum coordinates @xmath16 and @xmath17 , are the kinematical momenta canonically conjugated to @xmath18 and @xmath19 , respectively .
the dynamics of the model is defined by a two - body contact interaction between the constituent quarks at equal light - front times@xcite .
considering the contact interaction between the quarks , the baryon - qqq vertex is described by two terms written as @xmath20 in the baryon rest frame , with @xmath21 or @xmath22 , where @xmath23 is the bjorken momentum fraction and @xmath24 is the baryon mass .
the separable structure of the contact interaction implies that the vertex function depends only on the kinematical variables of the spectator quark in the process of the interaction of the other two .
as we are averaging over spin , only two independent vertex functions are necessary to describe the baryon wave function , i.e. , @xmath25 and @xmath26 .
the coupled faddeev - bethe - salpeter equations for a heavy - light - light three quark system ( @xmath3 ) in the light - front are derived in the ladder approximation@xcite , and are a generalization of the weinberg equation@xcite to three particle systems .
their diagrammatical representation are given in figures 1 and 2 .
the first light - front equation in which the quark @xmath22 is the spectator while the pair @xmath27 interacts is represented in figure 1 , from where one reads : @xmath28 the second light - front equation , which is represented diagrammatically in figure 2 , the quark pair @xmath29 interacts while the quark @xmath8 is the spectator .
the second equation is given by : @xmath30 \ .
\label{mass32}\end{aligned}\ ] ] the maximum value for @xmath31 is chosen to keep the mass squared of the @xmath27 or @xmath29 subsystem real , i.e. , @xmath32 and @xmath33 , respectively .
these constraints in the spectator quark phase - space come through the theta functions in the integrations of eqs .
( [ mass3 ] ) and ( [ mass32 ] ) . for @xmath33 one
has @xmath34 , and @xmath35 . for @xmath32 one
has @xmath36 , and @xmath37 . for equal particles , eq.([mass3 ] ) , reduces to the one derived in ref.@xcite . the baryon four - momentum
is given by @xmath38 , the light and heavy quark masses are @xmath39 and @xmath40 , respectively .
the masses of the virtual two - quark subsystems are @xmath41 and @xmath42 due to the conservation of the total four - momentum .
the two - quark scattering amplitudes @xmath43 and @xmath44 are the solutions of the bethe - salpeter equations in the ladder approximation for a contact interaction between the quarks @xcite . in this approximation the scattering amplitude , which is the infinite sum of the powers of the product of the `` bubble''-diagram with the bare interaction strength , is given by the geometrical series : @xmath45 where @xmath21 or @xmath22 and @xmath46 is the bare interaction strength , and @xmath47 in which @xmath48 is the total four - momentum of the quark pair and @xmath49 .
the four - dimensional integration of the function @xmath50 is the `` bubble''-diagram is performed in light - front variables .
first , the virtual propagation of the intermediate quarks is projected at equal light - front times @xcite , by analytical integration over @xmath51 in the momentum loop .
then , using the frame in which @xmath52 is zero and introducing the invariant quantity @xmath53 , one obtains : @xmath54 we suppose that the light - quark pair system has a bound state , which allows to define the two - quark scattering amplitude . this physical condition has been used in ref .
therefore , the bound state pole of the light - quark scattering amplitude , @xmath55 is found when @xmath56 is equal to the mass of the bound @xmath57 pair , @xmath58 which demands that @xmath59 .
this is suficient to render finite the scattering amplitudes @xmath60 and @xmath61 .
using that the bare strength of the effective contact interaction between the constituent quarks @xmath8 and @xmath62 does not depend on flavor , the final equation for the two - quark scattering amplitude is : @xmath63 the log - type divergence of @xmath64 is removed by the subtraction in eq.([mass11 ] ) .
the solution of the coupled integral equations ( [ mass3 ] ) and ( [ mass32 ] ) for a relativistic system of three constituent quarks with a pairwise zero range interaction , gives the baryon mass as a function of the quark mass @xmath65 .
the physical inputs of the model are the constituent quark mass @xmath66 and the diquark bound state mass .
the result for the mass of the ground state baryon @xmath67 alllows to calculate the binding energy , defined by @xmath68 .
however , the quarks are in fact confined in the hadron and to compare the model with data , one has to define an experimental quantity which could be compared to the model binding energy . for that purpose ,
we use that the low - lying vector mesons are weakly bound systems of constituent quarks while the pseudo - scalars are more strongly bound within the same model@xcite .
therefore , we suppose that the masses of the constituent quarks can be derived directly from the vector meson masses as : @xmath69 where the masses of the mesons are taken from ref.@xcite . using the constituent quark masses from eq .
( [ mconst ] ) and the experimental values of the baryon masses @xcite , we can attribute a binding energy to the low - lying spin 1/2 baryons , as given below : @xmath70 in figure 3 , we plot the binding energies of the low - lying pseudo - scalar mesons , defined as @xmath71 , where @xmath72 and @xmath73 are the masses of the vector and pseudoscalar low - lying mesons respectively , against the mass of the corresponding pseudo - scalar meson . also in the figure is shown the values of the binding energies of the spin 1/2 baryons ( @xmath74 , @xmath4 , @xmath10 and @xmath6 ) from eq .
( [ bind ] ) as a function of the corresponding baryon mass .
the systematic behaviour of the defined binding energy for the hadron is qualitative the same independent of the quark content .
the results from the numerical solution of the coupled equations ( [ mass3 ] ) and ( [ mass32 ] ) are obtained for a fixed @xmath75 gev ( we kept the value found in ref.@xcite ) which together with the nucleon mass of @xmath76 gev implies in @xmath77 gev . for the given value of the diquark mass @xmath78 and different values of @xmath40
, we obtain the binding energy for the spin 1/2 baryons @xmath4 , @xmath10 and @xmath79 as a function of @xmath40 ( @xmath80 ) . for the baryon mass above @xmath81 gev
, the bound @xmath3 system of the light - front model goes to the diquark threshold .
this gives the saturation value of @xmath82 seen in this figure .
the theoretical results are in excelent agreement with the baryon data , consequently the dynamical assumption of the flavor independence of the effective interaction is indeed reasonable . in figure 4 , we shown the baryon binding energy as a function of @xmath83 , again we observe the agreement between the model and the attributed experimental values for the binding energies and constituent quark masses .
the binding energy of the constituent quarks forming the the low - lying spin 1/2 baryonic states of the nucleon , @xmath4 , @xmath84 and @xmath6 , obtained from the experimental values of baryon masses and constituent quark masses derived from the low - lying vector mesons masses was studied within a light - front model .
the effective interaction between the constituent quarks was chosen of a contact form and spin was averaged out .
the motivation of chosing this particular interaction was two - fold : it was necessary to bring to the effective qcd - inspired model of the light - mesons the pion mass scale , and to give the observed splitting between the pion and rho meson spectrum@xcite , on one side , and on the other side , it was succesfull in describing the proton mass and radius simultaneously@xcite . in the present study
the contact interaction was used because it allows to introduce in the model the minimal number of physical scales necessary to describe the low - lying spin 1/2 baryons .
therefore , the relativistic three - quark model of the baryon defined on the light - front , with a falvor independent interaction , has as inputs the constituent quark masses and the diquark mass in the light sector , which defines the strength of the contact interaction .
this model allowed a surprising reproduction of the trend and magnitude of the binding energies as a function of the distint quark mass .
as the present light - front model is still very schematic , we believe that our conclusion of the flavor dependence of baryonic masses may still hold in a more realistic model , and support the extension of the qcd - inspired model applied previously only to mesons@xcite also to baryons .
acknowlegments : efs and jpbcm thank the brazilian funding agencies fapesp ( fundao de amparo a pesquisa do estado de so paulo ) and tf thanks fapesp and cnpq ( conselho nacional de pesquisa e desenvolvimento of brazil ) .
low - lying hadron binding energy @xmath85 of the low - lying hadrons as a function of the corresponding ground state mass .
experimental data for pseudoscalar mesons from table i@xcite ( full squares ) .
experimental data for the low - lying spin 1/2 baryons comes from table ii@xcite ( empty circles ) .
the results of the light - front model from the solution of eqs.[mass3 ] and [ mass32 ] are shown by solid line .
low - lying baryon binding energy @xmath85 as a function of the constituent quark mass @xmath86 .
experimental data for the low - lying spin 1/2 baryons comes from table ii@xcite ( empty circles ) and the constituent quark masses are given in table i. the results of the light - front model from the solution of eqs.[mass3 ] and [ mass32 ] are shown by solid line . | the ground state masses and binding energies of the nucleon , @xmath0 , @xmath1 , @xmath2 are studied within a constituent quark qcd - inspired light - front model .
the light - front faddeev equations for the @xmath3 composite spin 1/2 baryons , are derived and solved numerically .
the experimental data for the masses are qualitatively described by a flavor independent effective interaction . |
geometry and the quantum mechanics of particles have an uneasy relationship which is why general relativity is hard to incorporate into quantum field theory .
string theory is an ambitious and remarkable attempt at unification with many successes but the final theory has proved mysterious and elusive .
einstein and bohr fought a long battle on this front and bohr was generally deemed to have won , with the copenhagen interpretation of quantum mechanics accepted . but einstein s belief in the role of geometry made a partial come - back with the adoption of gauge theories as models of particle physics .
a more modest and limited role for geometry in nuclear physics was proposed by skyrme @xcite with the solitonic model of baryons , i.e. proton , neutron and nuclei , now known as skyrmions .
these have been shown to be approximate models of the physical baryons occurring in gauge theories of quarks and gluons , and have been extensively studied @xcite with considerable success . in this paper
we explore a geometric model of particles which is inspired by skyrme s idea but with potential applications to both baryonic and leptonic particle physics .
our model differs from the skyrme model in that it uses riemannian geometry rather than field theory , and so it is closer in spirit to einstein s ideas .
another key difference is that we absorb the kaluza - klein idea of an extra circle dimension to incorporate electromagnetism .
however , we exchange the roles of electricity and magnetism relative to the standard kaluza - klein approach , so that the extra circle dimension is magnetic rather than electric , and it is the electric charge that is topologically quantised by the famous dirac argument . we will also , initially , ignore time and dynamics , focusing on purely static models .
our geometric models will therefore be 4-dimensional riemannian manifolds .
we require the manifolds to be oriented and complete but not usually compact : we use non - compact manifolds to model electrically charged particles , and compact manifolds for neutral particles .
both baryon number and electric charge will be encoded in the topology , with baryon number ( at least provisionally ) identified with the signature of the 4-manifold . ] and electric charge with minus the chern class of an asymptotic fibration by circles . in particular
, the number of protons and the number of neutrons will therefore be determined topologically .
the manifolds will also have an ` asymptotic ' structure which captures their relation to physical 3-space .
the non - compact manifolds we consider have an asymptotic region which is fibred over physical 3-space , so no additional structure is required . for the compact ( neutral ) models ,
however , we fix a distinguished embedded surface @xmath2 where the 4-manifold @xmath3 intersects physical 3-space .
we also call @xmath4 the inside of @xmath3 . for single particles , the symmetry group of rotations should fix all of the above data
. it should act isometrically on the 4-manifold and preserve the asymptotic structure . in the non - compact cases
this means that it should be a bundle map in the asymptotic region , covering the usual @xmath5-action on physical 3-space . in the compact cases
it should preserve the distinguished surface @xmath2 . in order to capture the fermionic nature of the particles considered in this paper we also require spin structures on the non - compact manifolds and on the inside ( in the sense defined above ) of compact manifolds .
moreover , the lift of the rotation group action to the spin bundle ( over the entire manifold in the non - compact case and over the inside in the compact case ) should necessarily be an @xmath6-action , and this is what we mean by saying that our models are fermionic .
a key restriction on ( the conformal classes of ) our manifolds is that they are self - dual .
recall that the riemann curvature is made up of the ricci tensor plus the weyl tensor @xmath7 , which is conformally invariant . in dimension 4
, @xmath7 is the sum of self - dual and anti - self - dual parts w = w^+ w^- .
a 4-manifold @xmath3 is said to be self - dual if @xmath8 .
these manifolds are precisely those which have twistor spaces in the sense of penrose @xcite .
the latter are 3-dimensional complex manifolds @xmath9 with a real @xmath10-fibration over @xmath3 .
the complex structure of @xmath9 ( together with a real involution which is the antipodal map on each @xmath10 ) encodes the entire conformal structure of @xmath3 .
even the einstein metric can be captured by complex data on @xmath9 .
a simply - connected self - dual 4-manifold , for which the ricci tensor is also zero , is a hyperkhler manifold , whose structure group reduces to @xmath6 .
it is a complex khler manifold for an @xmath10-family of complex structures , but for any of these complex structures , and for the complex orientation , it would be anti - self - dual . since we want self - dual manifolds we choose the opposite orientation .
while some of our particles , including the proton , will be modelled by hyperkhler manifolds , we do not want to be so restrictive .
instead we will only require our 4-manifold models of particles to be self - dual and einstein , so there can be a non - zero scalar curvature .
our model for the neutron is of this type , distinguishing it from the proton .
the neutron will of course also have electric charge zero .
we should point out that reversing the orientation of a 4-manifold turns a self - dual manifold into an anti - self - dual one .
this should be interpreted as giving the geometric model of an anti - particle .
the existence of anti - particles follows from @xmath11 invariance , and our models are compatible with this .
self - dual 4-manifolds are , in many ways , the 4-dimensional analogue of riemann surfaces , with @xmath12 replacing @xmath13 in homology .
in particular there are theorems @xcite which assert that such manifolds admit connected sums although , unlike in the case of riemann surfaces , there are restrictions on when this is possible .
such connected sums model composite objects , like nuclei .
although we focus at present on static particles we do envisage a deformation theory , using the moduli space of self - dual manifolds , which could underlie particle interactions . fortunately a lot is now known about self - dual 4-manifolds with many metrics explicitly calculated .
this makes it possible to put forward some definite models for the proton and neutron .
even though our ideas are inspired by skyrme s theory of baryons , it turns out that geometric models of leptons , i.e. the electron and ( electron-)neutrino , are even simpler , and we shall describe them too in the class of self - dual manifolds .
thus , somewhat surprisingly , our framework of self - dual manifolds allows us to describe baryons and leptons in a unified fashion . the language and spirit of our model for particles is close to that of general relativity and suggests the possibility of a unification with gravity , but we do not address this issue here . in particular , we do not specify an action functional .
instead , we focus on how our model describes general features of particles such as their various quantum numbers .
we are aware that a description of elementary particles as 4-dimensional riemannian manifolds is radically different from established treatments in terms of quantum field theory .
what we aim to show in this paper is that such a geometric approach is possible , and that it has some surprising and attractive features , such as the possibility of describing the electron and the proton in one framework . while we do propose definite identifications of certain 4-manifolds with specific particles in sects .
3 - 5 of this paper , these should be seen as illustrations of the geometric approach , not necessarily as final proposals .
the paper is organised as follows . in sect . 2 we outline the genesis of our geometric models of particles , starting with the skyrme model of baryons .
electrically charged particles are necessarily described by non - compact 4-manifolds in our approach , and in sect .
3 we explain how to model the electron and proton in terms of the taub - nut and atiyah - hitchin manifolds , respectively .
neutral particles are described by compact 4-manifolds , and this is discussed in sects .
we propose @xmath0 as a model for the neutron and @xmath1 as a model for the neutrino .
these are the simplest choices , but we also discuss some more sophisticated versions . in sect .
6 we describe how our particle models glue into empty space , and how the particles may interact with each other .
7 contains an outline of how our geometric models capture the spinorial nature of the particles they describe . in sect .
8 we give the dictionary which translates topological properties of 4-manifolds into the electric charge and baryon number of particles , and discuss in some detail how these charges are related to fields and densities used in conventional lagrangian models of particle physics .
9 contains our conclusion and some ideas for follow - up work .
conventions and calculations are collected in appendices a , b and c.
we begin by spelling out in detail how the skyrme model suggests our 4-manifold model .
the skyrme model is based on a group - valued field from @xmath14 , [ skyrmefield ] u : ^3 g , where the lie group @xmath15 is usually taken to be @xmath6 , and @xmath16 as @xmath17 . the degree of @xmath18 as a map @xmath19 is identified with baryon number .
the minima of the skyrme energy , for each baryon number , are called skyrmions .
skyrmions are free to rotate , both in physical space and through conjugation by elements of @xmath6 .
quantising this motion gives the skyrmions spin and electric charge .
the proton and neutron , for example , are distinct quantum states of the essentially unique skyrmion of degree 1 . in @xcite
it was shown how to generate such skyrme fields naturally by starting with an @xmath6 yang - mills gauge field on @xmath20 and calculating the holonomy along the 4th direction .
suitable asymptotic behaviour on @xmath20 guarantees a well - defined map @xmath18 .
although this construction does not preserve the respective energy functionals it does provide a good way of using instantons on @xmath20 ( i.e. self - dual gauge fields ) to construct approximate minima of the skyrme energy .
it also identifies instanton number with the skyrme degree .
see also the recent papers @xcite where the difference between the yang - mills and skyrme energy functionals is interpreted as due to an infinite tower of mesons .
since the yang - mills energy functional in dimension 4 is conformally invariant we could replace the decomposition
^4 = ^3 ^1 by s^4 s^2 = h^3 s^1 , where @xmath21 is hyperbolic 3-space .
in fact we can vary the curvature of @xmath21 provided we rescale the circle @xmath22 the opposite way , so that large circles correspond to almost flat @xmath21 .
we can now fix a gauge field on @xmath23 and take the holonomy round the circles .
there are some technicalities ( due to base - points ) which we shall ignore but basically we expect to end up with a skyrmion on @xmath21 , an idea which has been explored in @xcite
. now replace @xmath24 by any riemannian 4-manifold @xmath3 which is asymptotically fibred by circles over @xmath14 .
this is the kind of kaluza - klein 4-manifold we are going to consider .
an @xmath6 gauge field on @xmath3 would then give a skyrme field on ` the quotient of @xmath3 by @xmath22 ' .
since we do not want to assume there is a global circle fibration , this skyrme field will only be defined asymptotically outside some ` core ' . but an oriented 4-manifold has two natural @xmath6 bundles over it , the two spin bundles @xmath25 and @xmath26 ( assuming @xmath3 is a spin - manifold , i.e. @xmath27 ) . picking one of these , say @xmath25 , we then get from the connection on @xmath25 a natural construction of an asymptotic skyrme field on @xmath14 .
this is roughly the genesis of our idea to model particles by 4-manifolds , but the topology of these asymptotic skyrme fields does not quite fit , and would not give integer baryon numbers as defined in our model .
the reason lies in a fundamental difference between the topology of gauge bundles which can have arbitrary instanton number ( or second chern class ) and the topology of tangent bundles , where there are divisibility theorems .
for example the first pontrjagin class of a 4-manifold ( compact and oriented ) is divisible by 3 and then gives the signature .
an example is @xmath0 which has signature 1 and pontrjagin class 3 ( times the generator of @xmath28 ) . in the skyrme model
the basic idea is that baryon number is identified with the degree of the map @xmath18 in , or equivalently with the instanton number ( or second chern class ) of the @xmath6 bundle over @xmath20 .
this differs from the 4-manifold model we want to explore , where baryon number is identified with the signature of the 4-manifold .
the signature is additive under taking connected sums of 4-manifolds @xcite , and this captures the additivity of baryon number for composites of particles , for example , fusion of nuclei .
the integrality of the signature is linked to it being an index of an elliptic operator .
this means we are in the realm of k - theory rather than cohomology .
a consequence of this change of viewpoint is that the geometry of the 4-manifold model is important for us , but we will not try to define a global 3-dimensional skyrme field @xmath18
. recall that in the skyrme model , baryon number is cohomological and electric charge arises at the quantum level . for our 4-manifold model ,
electric charge is cohomological , arising , as already explained , from the first chern class of the asymptotic @xmath22-fibration , while baryon number as just indicated should be seen as an index . to sum up our discussion , we see that our model goes beyond the skyrme model in aiming to understand topologically both the basic integer physical invariants , baryon number and electric charge .
the two models are different , but possibly dual in a suitable sense .
we hope to explore this in detail at a later stage .
models for the basic particles should exhibit a high degree of symmetry and we expect the rotation group @xmath5 of @xmath14 , or its double cover @xmath6 , to act as isometries . for electrically charged particles , we take our geometric models to be non - compact hyperkhler manifolds . we also assume that the volume grows with the third power of the radius , to allow for an interpretation of the asymptotic region in terms of physical 3-space .
as recently shown in @xcite , this forces the hyperkhler manifold to be alf .
we are therefore looking for rotationally symmetric and complete alf hyperkhler manifolds .
there are just two possibilities : 1 . the taub - nut manifold @xcite depending on a positive parameter @xmath29 ( interpreted as mass in the gravitational context ) . for brevity
we denote it by tn .
the atiyah - hitchin manifold , the ( simply - connected double cover of the ) moduli space of centred @xmath6-monopoles of charge two @xcite . for brevity we denote it by ah .
note that we could also single out tn and ah among non - compact , complete and rotationally symmetric hyperkhler manifolds by demanding that the @xmath6- ( or @xmath5- ) action rotates the complex structures , see our discussion following .
this turns out to play a role in recovering the usual rotation action on physical 3-space in the asymptotic region of our geometric models , as discussed in sect .
[ spinsect ] . both tn and ah can be parametrised in terms of a radial coordinate @xmath30 and angular coordinates on @xmath6 ( for tn ) or @xmath31 ( for ah ) .
details are given in appendix [ ahtndetails ] . in terms of the left - invariant 1-forms defined in ,
the metrics of both tn and ah can be written as @xmath32 with the functions @xmath33 satisfying the self - duality equations [ dual ] = ( b - c)^2 -a^2 , , where @xmath34 means we add the two further equations obtained by cyclic permutation of @xmath35 .
we adopt the convention f(r)=- , where ( for reasons that will emerge later ) the radial coordinate @xmath30 has the range @xmath36 for tn and @xmath37 for ah .
the self - duality equations become @xmath38 this system has solutions in terms of elementary functions [ tncoef ] a(r)=b(r ) = r c= , with parameters @xmath39 , associated to the tn manifold .
the topology is that of @xmath20 , and as @xmath40 the metric tends to the flat metric . for @xmath41
the manifold is asymptotic to an @xmath22 fibre - bundle over @xmath14 with the length of the circle being @xmath42 .
there is a @xmath43 symmetry acting along the fibres , with just one fixed point at the origin , @xmath44 .
the whole isometry group is @xmath45 .
as @xmath40 the @xmath43-action becomes the scalar action on @xmath46 .
the complex orientation of @xmath46 determines the orientation of tn as a self - dual manifold ; this is opposite to the orientation given by any of the complex structures in the hyperkhler family , see appendix [ signstnah ] for a discussion . at infinity ,
the @xmath43-action gives the standard hopf line bundle over @xmath47 with chern class @xmath48 ; details are given in appendix [ conventions ] .
the tn metric with coefficient functions has the following behaviour under scaling by non - vanishing real numbers @xmath49 : r r , mm , ^2 , ds^2 ^2 ds^2 .
we use rescaling by @xmath50 to set @xmath51 , and rescaling by @xmath52 to set @xmath53 from now on .
this amounts to picking a unit of length for the radial coordinate @xmath30 and to fixing an overall scale for the metric .
our choice is motivated by the asymptotic form of the ah metric , to be discussed below .
note that , with this choice , the length of the asymptotic circle , in the length units chosen , is @xmath54 .
the solution which gives rise to ah has the _ asymptotic _ form , for large @xmath30 , [ ahcoef ] a(r ) ~b(r ) ~r c(r ) ~- .
these asymptotic expressions , a tn metric with @xmath55 , also satisfy .
however @xmath56 is not actually equal to @xmath57 , and @xmath30 only extends down to @xmath58 . for @xmath30 near @xmath58 ,
is a poor approximation .
instead , one finds the leading terms [ ahsmallr ] a(r ) ~2(r- ) , b(r ) ~+ 1 2 ( r- ) , c(r ) ~-+1 2 ( r- ) , which we will need later in this paper .
the manifold ah is the complement of @xmath59 ( the real projective plane ) embedded in @xmath0 , and the complex orientation of @xmath0 determines the orientation of ah as a self - dual manifold . as for tn
this is opposite to the orientation given by any of the complex structures in the hyperkhler family ; see our discussion in appendix [ signstnah ] .
ah has an @xmath5 symmetry with just one 2-dimensional orbit at @xmath60 , which is a minimal 2-sphere .
we refer to this minimal 2-sphere , which is the totally imaginary conic in @xmath0 and determined by @xmath61 in the homogeneous coordinates introduced in sect .
[ cp2sect ] , as the core .
asymptotically , the manifold is fibred by circles . as further discussed below ,
neither the circles nor the base space of this asymptotic fibration are oriented because of a @xmath62-identification , given explicitly in .
the manifold tn is usually interpreted as the geometry of a dirac monopole at the origin @xcite . for us , with electric and magnetic charges reversed , it has to be interpreted as an electrically charged particle .
since the signature of tn is zero ( we discuss this further in sect . [ enfluxsection ] ) the particle is leptonic .
we therefore interpret tn as a model for the electron . down on @xmath14 , after factoring by @xmath43 , any 2-sphere surrounding the origin has an electric flux emerging from it due to the electron , which carries charge @xmath63 .
this implies that there is a sign change in going from the chern class to the electric charge .
the manifold ah has the opposite asymptotic behaviour with a sign change for @xmath29 and an orientation change ( see appendix [ conventions ] ) and so would lead us naturally to expect electric charge + 1 . also , the topology at the core is different , with a 2-sphere instead of a point . as a result ,
ah has signature 1 ( again discussed further in sect .
[ enfluxsection ] ) and looks like the model we want for the proton ( rather than the positron ) .
however things are not quite that simple , as we shall now explain .
the ` asymptotic boundary ' of ah is not @xmath64 as for tn but the boundary of a tubular neighbourhood of @xmath59 in @xmath0 , which is @xmath64 divided by a cyclic group of order 4 .
moreover the base of this unoriented circle fibration is @xmath59 , not @xmath10 , and is non - orientable .
this means that the 3-manifold which is the base of the asymptotic fibration is not @xmath14 , and has a fundamental group of order 2 .
it is not orientable .
this might seem to be a disaster , but we shall argue that , while unexpected , it is not as bad as it looks . the most convincing argument in its favour is to show that the electric charge is well - defined and equal to + 1 as hoped .
this is done in detail in appendix [ conventions ] .
the lack of orientability in physical 3-space should be thought of as follows .
the lack of orientation in @xmath14 locally is compensated by a corresponding ambiguity in the sign of electric charge ( non - orientability of the circle fibres ) .
physically the geometric orientation is not felt .
having put forward a definite proposal for the proton we now have to face the neutron .
since the neutron has no electric charge any non - compact model would need to have a trivial asymptotic circle fibration .
the 4-manifold should have signature 1 and it should resemble the ah model of the proton in its @xmath5 orbit structure . however , the latter requirement rules out asymptotically trivial circle bundles over physical 3-space since the generic @xmath5 orbits would be 2-dimensional in that case .
we therefore consider compact 4-manifolds .
in fact , there is an obvious choice which is just the complex projective plane @xmath0 with its fubini - study metric ( and its natural complex orientation ) .
this is a self - dual manifold of positive scalar curvature .
@xmath0 has even more symmetry than we need , since it is acted on by @xmath65 .
the rotation group @xmath5 sits inside as the subgroup that preserves the real structure given by complex conjugation .
this preserves @xmath59 as in the case of the proton , so we fix this @xmath59 as the distinguished surface where the 4-manifold intersects physical 3-space , thus breaking the symmetry to @xmath5 . the global @xmath65 symmetry might give us a link to quarks , but this remains to be explored .
the fubini - study metric is often written in coordinates which exhibit the invariance under @xmath66 .
this brings out the parallels with tn @xcite but is not the symmetry we want in our neutron model . for the interpretation of @xmath0 as a neutron and for a comparison with the ah model of the proton
, we need to write the fubini - study metric in coordinates adapted to the @xmath5-action , which is discussed in @xcite and @xcite .
we write the results of @xcite in the conventions used in our discussion of the ah metric . in terms of homogeneous coordinates @xmath67 ( with the identification @xmath68 , @xmath69 )
the fubini - study metric on @xmath0 is ds^2 = . for calculations we can fix @xmath70 and parametrise [ ourchoice ] = e^i r _ 0 , where @xmath71 can , in turn , be parametrised in terms of euler angles as shown in appendix b.1 .
the reference vector @xmath72 ( which depends on one parameter ) should be a unit vector and we can assume , by adjusting the phase @xmath50 if necessary , that its real and imaginary parts are orthogonal .
for our purpose , it is convenient to single out the 1-axis and pick [ referencevec ] _ 0 = 0 + a_2 + i a_3 , a_2 ^ 2 + a_3 ^ 2=1
. then we find the following expression for the fubini - study metric in terms of the left - invariant forms on @xmath5 ( for details of an analogous calculation we refer the reader to @xcite ) : [ cp2first ] ds^2 = d a_2 ^ 2 + da_3 ^ 2 + ( a_3 ^ 2-a_2 ^ 2)^2 _ 1 ^ 2 + a_3 ^
2_2 ^ 2 + a_2 ^ 2 ^ 2_3 . parametrising [ anglepick ] a_2=(+ ) , a_3= ( + ) , , we obtain , finally , the fubini - study metric in the form [ cp2second ] ds^2 = 1 4 d ^2 + ^2_1 ^ 2 + ^2(+ ) _ 2 ^ 2 + ^2 ( + ) ^2_3 . there is a simple interpretation of the geometry of @xmath0 and its orbit structure in terms of orientated ellipses up to scale @xcite , which is useful for comparison with the ah metric .
as already exploited above , we can adjust the phase in the homogeneous coordinate @xmath73 ( no longer fixed to satisfy @xmath74 ) so that the real vectors @xmath75 and @xmath76 are orthogonal : if @xmath77 this is automatic , and if @xmath78 we multiply by a unit complex number to set im@xmath79 and re@xmath80 ( we pick the negative sign to agree with the choice made in above ) .
we can interpret @xmath76 and @xmath75 as the major- and minor - axis of an ellipse .
this ellipse is only determined up to scale ( we can still rescale @xmath81 by any positive real number ) but it is orientated .
the totally degenerate case @xmath82 is excluded by the definition of homogeneous coordinates , but circles ( @xmath83 or @xmath84 ) and lines ( @xmath85 ) can occur . in terms of our parametrisation ,
the reference vector and the definition of @xmath86 in , we see that for @xmath87 the ellipse is a circle and for @xmath88 it degenerates to a line . for generic values of @xmath86 ,
the @xmath5-orbit is @xmath31 , with the @xmath62 generated by the 180@xmath89-rotation about the 1-axis , but for @xmath87 the orbit is a 2-sphere and for @xmath90 the orbit is @xmath59 .
this is the same as the orbit structure of ah compactified by an @xmath59 at infinity , although the metric is of course different .
we also note that the khler form = i takes the simple form [ cp2harmonic ] = _ 2_3- d_1 , which should be compared to the expression given in @xcite in coordinates adapted to the @xmath45 symmetry of @xmath0
. the form @xmath91 is invariant under the 180@xmath89-rotation about the 1-axis and hence well - defined on the generic @xmath5-orbits .
it is manifestly closed , but not exact : for @xmath92 we can write @xmath93 but this expression is not valid on the exceptional @xmath5-orbit where @xmath87 , since @xmath94 is not well - defined there .
the khler form is self - dual with respect to the complex orientation ( with volume element @xmath95 ) . since it is closed , it is also harmonic .
the existence of a non - exact harmonic , self - dual form on @xmath0 follows from the fact that the signature of @xmath0 is 1 . in view of our interpretation of signature as baryonic charge one might expect there to be a baryonic interpretation of @xmath91 .
we return to this question when discussing the ah model of the proton further in sect .
[ enfluxsection ] .
if the @xmath0-model for the neutron turns out to be too nave , there is a more sophisticated variant which could be explored .
this arises from the sequence of self - dual einstein manifolds @xmath96 , for @xmath97 , studied by hitchin @xcite .
the manifolds @xmath96 for even @xmath98 are all defined on the same space as our proton model , namely @xmath0 but with @xmath59 removed .
@xmath99 is @xmath0 with the fubini - study metric and all the metrics on @xmath96 , @xmath100 and even , are incomplete on the open set in @xmath101 , but can be completed to metrics on @xmath0 with a conical singularity of angle @xmath102 along @xmath59 . for odd @xmath103
the manifolds are defined on @xmath1 , with @xmath104 being @xmath1 with its standard metric .
this time the metrics of @xmath96 , @xmath105 and odd , are incomplete on the open set in @xmath106 , but can be completed to metrics on @xmath1 with a conical singularity of angle @xmath107 along @xmath59 .
the sequence of conical manifolds @xmath96 for even @xmath98 and starting with @xmath0 , has decreasing scalar curvature and converges to ah as @xmath108 .
it may turn out that some other value of @xmath109 gives a better model for the neutron than @xmath110 . note that for @xmath100 the conical singularity breaks the symmetry down to @xmath5 . even for @xmath110
we shall see later that other factors break the symmetry in this way .
hitchin also pointed out @xcite that the family @xmath96 can be extended to real parameter values . for any real @xmath111
, @xmath96 is related to the moduli space of centred @xmath6 monopoles over hyperbolic space of curvature @xmath112 , where @xmath113 .
when @xmath109 is not an integer , the conical angle is not a rational multiple of @xmath58 and @xmath96 is not an orbifold .
consequently , the explicit methods of @xcite do not then apply . nonetheless , having @xmath109 as a real parameter gives useful room for manoeuvre in modelling the neutron and may provide contact with conventional nuclear models .
in particular , @xmath114 may play a role as a small parameter that controls the breaking of isospin symmetry .
the forthcoming paper @xcite contains a signature formula for riemannian manifolds with conical singularities , like the hitchin manifolds @xmath96 .
we summarise that result in sect .
[ sigsect ] .
having put forward models of the electron , proton and neutron , it is then natural to look for a similar model of the neutrino
. since it has no electric charge it should , like the neutron , be modelled by a compact manifold .
it should have symmetry similar to the @xmath45 symmetry of the electron and should have positive curvature .
it should have zero baryon number , that is , vanishing signature . just as @xmath0 is the most obvious model for the neutron ,
the standard 4-sphere , @xmath1 , is the most obvious model for the neutrino .
again this has more symmetry than we need , @xmath115 instead of @xmath5 . just as a distinguished @xmath59 in @xmath0 picks out the smaller @xmath5 symmetry ,
so a distinguished @xmath10 in @xmath1 is needed to cut down the symmetry of the neutrino . to exhibit the symmetry , we parametrise @xmath1 in terms of vectors @xmath116 and @xmath117 satisfying the constraint + = 1 .
the metric is then ds^2= dd + d d .
the group @xmath118 acts in the obvious way on the pair of vectors @xmath119 and preserves the metric . in order to compare with other metrics discussed in this paper
we parametrise @xmath120 with @xmath121 $ ] , @xmath122 and the usual range for the polar coordinates @xmath123 on @xmath10 , and find the expression [ s4metric ] ds^2= d^2 + ^2 ( _ 1 ^ 2 + _ 2 ^ 2 ) + ^2 d ^2 .
here we used that @xmath124 in terms of the left - invariant 1-forms defined in .
the generic @xmath118 orbit is @xmath125 , but this collapses to @xmath22 when @xmath87 and to @xmath10 when @xmath126 .
note that @xmath1 is conformally flat , so the weyl tensor vanishes and is trivially self - dual .
@xmath1 has no middle - dimensional homology , so the signature is zero , and hence our model neutrino has zero baryon number , as required . since @xmath1 also has an orientation - reversing isometry , our model seems to suggest that the neutrino coincides with the anti - neutrino . for this and other reasons ( see later ) our choice of @xmath1 is very tentative and provisional . as with the @xmath0 model for the neutron it should be regarded at present as a prototype . to address the symmetry breaking issue , and several others
, we will in the next section discuss how our various models are supposed to fit into conventional 3-space .
so far our models are abstract objects , 4-manifolds on their own , which are supposed to model four basic particles of nature . how are we to view them in the real world ?
let us begin with the easiest case , that of the electron . thought of originally as the dirac monopole , the idea is well - known .
we consider kaluza - klein space as a riemannian 4-manifold with a circle action .
away from matter this space is assumed to be a circle bundle over @xmath14 . outside a given region in @xmath14 which is electrically neutral
the bundle is assumed to be ( topologically ) just the product space .
if a region is electrically charged the circle bundle over the boundary is supposed to have a chern class equal to minus the charge .
if we start from the vacuum , then inserting one electron amounts to attaching a truncated version of tn to the boundary .
this truncation turns the idealized model into a more realistic model of a particle .
if other particles are present , tn will be an approximation to the precise metric .
this approximation is some measure of the force exerted on the electron . in a dynamic theory
, forces should emerge from the equations , a task for the future .
next let us move on to the proton .
this is similar , using a truncated version of ah .
however , as pointed out earlier , the circle fibration is now not oriented so that the asymptotic 3-space is not @xmath14 ( unless our region contains equal numbers of protons and anti - protons ) .
the model for the neutron is compact , so there is no way to attach it to a boundary .
instead we propose that our model neutron ( a copy of @xmath0 ) intersects our 4-space in a surface .
this surface should project to a surface in 3-space which is the ` boundary ' of the neutron as seen by an observer .
since we want to keep the neutron similar to the proton ( except for the charge ) it seems reasonable to take this surface to be a copy of @xmath59 .
but since the charge is now zero the circle bundle over the surface should be trivial . this could arise as follows .
pick a point in @xmath14 and blow it up to give an @xmath59 , so that we are modifying 3-space in this neighbourhood .
keep the circle bundle trivial , so making the charge zero .
lift this @xmath59 into the total space of the circle bundle and let the @xmath0 neutron model intersect 4-space in this @xmath59 , which is the distinguished @xmath59 in @xmath0 ( hence breaking the larger symmetry of @xmath0 ) . in this construction the metric on @xmath0 need not be changed .
the only change that is needed is the change in metric on the background 4-space got from the blowing up process in 3-space .
finally we come to the neutrino . from the other examples it is clear what is required .
this time we just take a 3-ball in @xmath14 with boundary a 2-sphere .
the circle bundle over it is trivial and we lift the sphere to the total 4-space .
we now require the @xmath1 neutrino model to intersect our 4-space in the chosen 2-sphere .
the choice of the 2-sphere in @xmath1 again breaks the symmetry , down to @xmath118 .
the 2-sphere need not be a great ( geodesic ) sphere and this provides a parameter to play with .
note that in this case we could reinterpret the picture as the surgery that kills off the circle and leaves a 2-sphere .
this means that for the electron , the proton and the neutrino we can still think of our ` space ' as a 4-manifold .
but this does not seem to work for the neutron where we have to settle for a 4-space with intersecting components like a complex algebraic surface with double curves .
the four models that we have proposed for the four basic particles should be geometrically related in some way to account for the process of beta decay in which a neutron breaks up into a proton , an electron and an anti - neutrino .
the opposite asymptotic behaviour of ah and tn is a good start but the difference in the asymptotic fundamental groups presents a problem .
this suggests that the model of the neutrino should somehow bridge the gap and it argues against the simplicity of the 4-sphere .
we hope to pursue this question .
in all our models we have a natural action of the symmetry group of rotations ( @xmath5 or @xmath6 ) preserving the metric and the ` asymptotics ' , the details of which differ according to the cases . for the neutral , compact models we interpret ` asymptotic ' to mean the behaviour near the distinguished surface where the 4-manifold intersects 3-space , which is either an @xmath59 or an @xmath10 . for
the electrically charged , non - compact models we have an asymptotic fibration by circles over physical 3-space ; rotations preserve this fibration and induce an @xmath5-action on the base . the hyperkhler structures on tn @xcite and ah @xcite
can be used to construct cartesian coordinates on physical 3-space with the physically required transformation properties under spatial rotations . here
we make essential use of the fact that the complex structures on tn and ah transform as vectors under rotations , as explained at the beginning of sect .
[ protonelectron ] . for a model to represent a particle of spin 1/2 we must include the data necessary to lift the rotation group action to an @xmath6 action , and to construct its 2-dimensional representation . to achieve this for non - compact (
electrically charged ) models we require a spin structure on the 4-manifold while for compact ( neutral ) models we only require a spin structure on the inside , obtained by removing a distinguished surface @xmath2 from the 4-manifold . in this short section
we explain why , in each of the models considered in this paper , the lift of the rotation group action to the spin bundle is an @xmath6-action .
we do not attempt to construct naturally associated spin 1/2 representations , but comment on how this may be done . for the tn model of the electron
, there is nothing to do since the rotation group action on tn is an @xmath6-action . for the neutron model
we view the required data as the compactification of the proton model ah , not just topologically , but also with the action of the rotation group .
in particular the @xmath59 at infinity is part of the data .
we now simply require the extra data of a spin structure on the inside @xmath127 , i.e. , on ah .
it might be thought that , since the spin structure on ah is unique , there is nothing gained by the additional data , but this is to ignore the interaction with the symmetries
. we must now require the rotation group to lift to the spin bundle , and this may require us to pass to @xmath6 , in which case we label the model as fermionic .
otherwise we call it bosonic . to see that , in principle ,
either case could occur , consider the two manifolds @xmath128 and @xmath129 .
we take the left - translation action of @xmath5 on @xmath130 and the standard action on @xmath131 . for @xmath130 ,
the tangent bundle is trivial and we can choose the trivial spin bundle ( though since @xmath130 is not simply - connected there is another choice ) .
the @xmath5-action extends without going to @xmath6 , so @xmath130 is bosonic in our terminology . for @xmath131 the fixed point at the origin of @xmath14
means that we can only lift to the spin bundle after passing to @xmath6 , so @xmath131 is fermionic . for ah we have to show that , with its action of @xmath5 , it is fermionic .
there are several ways to do this .
perhaps the simplest ( in line with the example @xmath130 above ) is to note that the action is not free and that the isotropy group of a point on the core is @xmath132 .
to lift even this subgroup to the spin bundle over the fixed point requires us to go to the double cover .
a comparison between our models for the electron and proton is illuminating . as we pointed out , @xmath6 acts on the electron but only @xmath5 acts on the proton .
thus the fermionic natures of our two models differ . in one case
it is inherent in the symmetry while in the other it is geometric or topological .
our tentative model for the neutrino is just the round 4-sphere , with a distinguished 2-sphere at infinity given by the decomposition @xmath133 and the corresponding action of @xmath5 .
the inside , got by removing @xmath10 , has an infinite cyclic fundamental group so there are infinitely many spin structures , but only one extends to the whole of @xmath1 . if we pick this , then it is easy to see that the lift of @xmath5 to the spin structure requires us to pass to @xmath6 ( for example we can use an @xmath132 fixing a point at infinity and argue as with the proton ) . thus our model of the neutrino is fermionic .
our discussion so far shows that our geometric approach furnishes fermionic models , but it does not establish that they necessarily give spin 1/2 .
this requires constructing the 2-dimensional representation of @xmath6 and relating it to the asymptotic region .
we expect that the required 2-dimensional representations can be constructed in terms of eigenspaces of the dirac operator on our model manifolds , possibly twisted by a @xmath43-bundle with curvature proportional to one of the harmonic 2-forms discussed in sects .
[ cp2sect ] and [ enfluxsection ] .
so far we have focussed on topological and geometrical features of our models and explained how they describe general properties of particles like baryon number , electric charge and location in space .
we want to keep an open mind about how our geometric models make quantitative contact with the physics of elementary particles .
in particular , we do not assume that this should necessarily happen in the standard framework of lagrangian field theory , where dynamics , conservation laws and even the quantum theory are all derived from an action functional .
the purpose of this section is to illustrate that our geometric models for particles nevertheless contain natural candidates for the kind of quantities which arise in lagrangian models , like energy density and electric fields .
we show that electric charge as defined in our model can be represented by a harmonic 2-form , thus making contact with the usual description of electric flux
. one important feature of the densities and fields considered in this section is that they are defined on the 4-manifold so that they can only be interpreted as conventional _ spatial _ densities and fields in the asymptotic region of the 4-manifold which canonically projects down to physical 3-space .
we begin with a summary of the topological quantities and their physical interpretation for each of the 4-manifolds considered thus far . for compact manifolds the electric charge is zero and for non - compact manifolds
it is minus the self - intersection number @xmath134 of the manifold @xmath2 representing infinity in their compactification .
the compactification is @xmath0 for both tn and ah , but @xmath135 for tn with self - intersection number @xmath136 , while @xmath137 for ah , with self - intersection number @xmath138 ; see appendix [ signstnah ] for details .
the baryon number is identified with the signature of the 4-manifold . for a non - compact oriented manifold
, the signature is defined as the signature of the image of the compactly supported cohomology in the full cohomology @xcite .
the topology of tn is that of @xmath20 so the signature vanishes .
the signature of ah is 1 .
this follows from the fact that its 2-dimensional homology is generated by the core 2-sphere , and that the self - intersection number of this 2-sphere is positive ( in fact equal to + 4 ) .
the same argument applies to the sequence of hitchin manifolds @xmath96 for @xmath98 and even , reviewed in sect .
[ hitchinsection ] , which are all compact and topologically equivalent to @xmath0 . in table [ manifoldlist ] we list the electric charge and baryon number as well as the euler characteristic @xmath139 and the squared @xmath140-norm @xmath141 of the riemann curvature for the four 4-manifolds mainly discussed in this paper .
the euler characteristic is homotopy invariant , so can be computed for tn and ah by noting that the former retracts to a point and the latter to a 2-sphere .
.geometric properties of 4-manifolds and their physical interpretation [ cols="^,^,^,^,^",options="header " , ] for compact riemannian 4-manifolds there are formulae for the euler characteristic and signature in terms of integrals over the 4-manifold involving the riemann curvature , which we shall review below . in the non - compact cases ,
these bulk contributions need to be supplemented by boundary integrals and ( for the signature ) a subtle spectral contribution ( @xmath142-invariant ) , see e.g. @xcite and @xcite for a summary . for manifolds with conical singularities like the sequence of hitchin manifolds reviewed in sect .
[ hitchinsection ] , a signature formula was recently found @xcite , which we also review below .
we now discuss the bulk contributions , relegating most detailed calculations to appendix [ riemanncalculations ] .
writing @xmath143 for the riemann tensor as in appendix [ ahtndetails ] , we define the squared @xmath140-norm of @xmath143 ( for compact and non - compact manifolds ) as [ l2defined ] dual of @xmath144 .
the form that integrates to the first pontrjagin class on a compact manifold is @xmath145 and the form which integrates to the signature @xmath146 in the compact case is s= p_1 , so that @xmath147 the form that integrates to the euler characteristic @xmath139 in the compact case is @xmath148 note that , with our conventions and for compact einstein manifolds @xcite , ( m ) = _ m e = ||r||^2 , which determines @xmath141 for @xmath0 and @xmath1 in terms of their topology .
the riemann curvature on a 4-manifold may also be thought of as a mapping of 2-forms .
exploiting the fact that the space @xmath149 of 2-forms on an oriented 4-manifold decomposes into @xmath150-eigenspaces @xmath151 of the hodge star operator @xmath152 , we get a corresponding decomposition of the riemann curvature into irreducible pieces @xcite r = w^+ + & b + * b & w^-+ .
here @xmath153 are the self - dual and anti - self - dual parts of the weyl tensor , @xmath154 is the scalar curvature and @xmath155 amounts to the tracefree part of the ricci curvature .
then the signature of a compact manifold can also be expressed as @xcite ( m ) = ( ||w^+||^2 - ||w^-||^2 ) . for a self - dual manifold @xmath156
so , up to the factor @xmath157 , @xmath146 is given by the @xmath140-norm of the weyl tensor @xmath158 , and is non - negative . since the metrics on both tn and ah are hyperkhler , @xmath155 and @xmath154 vanish , so the full riemann curvature is self - dual .
therefore the bulk contribution to both the signature and the euler characteristic can be expressed in terms of the @xmath140-norm of the riemann curvature , which equals the @xmath140-norm of @xmath158 . in appendix [ riemanncalculations ] , we show that riemann curvature , we deduce that the bulk contributions to the euler characteristic are both in agreement with the topological results listed above : _ tne = = 1 , _
ah e= = 2 .
the bulk contributions to the signatures , on the other hand , turn out to be fractional : [ bulkahtn ] _
tn s= = , _ ah s = = . as shown in @xcite ,
the fall - off of the spin connection and curvature imply that the boundary integrals do not contribute in the limit to either the euler characteristic or the signature .
however , the fractional values of the bulk integrals for the signature show that there must be a non - zero contribution from the @xmath142-invariant . in @xcite ,
a signature formula is derived for riemannian 4-manifolds with conical singular metrics . as reviewed in sect .
[ hitchinsection ] , the hitchin manifolds @xmath96 are of this type , and @xmath159 can be identified with ah .
the general formula derived in @xcite for the signature of a riemannian 4-manifold @xmath3 whose metric has a conical singularity of angle @xmath160 along a surface @xmath2 is ( m)=_m p_1 - 1 3 ( 1-)x^2 , where @xmath161 is the form which integrates to the pontrjagin class , and @xmath134 is the self - intersection number of @xmath2 in @xmath3 . since the signature @xmath146 is interpreted as the baryon number in our model , it is tempting to interpret the integrand of the bulk contribution to the signature as a baryon number density .
our calculations above show that this can not be the whole story , since the signature @xmath146 also receives a contribution from the @xmath142-invariant of the boundary .
we nevertheless compute and plot the integrands of the bulk contributions to the signature below .
since overall factors are not important here , we look at the squared @xmath140-norm of the riemann curvature . in the notation of appendix [ ahtndetails ] , the integrand of in both the tn and ah case can be written as df_1_2_3 = dv , where @xmath162 is the volume element , so that the combination [ curvkernel ] may be interpreted as a density .
we plot this density for tn and ah in fig .
[ curvdens ] .
it is finite at , respectively , the origin and the core , and for large @xmath30 falls off like @xmath163 . in the ah case we compute the functions @xmath35 and hence @xmath164 numerically ; in the tn case we have the explicit formula = .
we now construct fields on tn and ah which carry the electric flux measured by the self - intersection numbers @xmath134 tabulated in table [ manifoldlist ] .
we do this by systematically studying rotationally symmetric harmonic forms on both tn and ah , with most of the details given in appendix [ harmonicsect ] .
one finds @xcite that the only square - integrable harmonic and rotationally symmetric 2-form on tn is , up to an overall arbitrary constant , [ tnelectricfield ] _
3^+ =( _ 1_2 + dr_3 ) = d ( _ 3 ) .
we claim that this form is a harmonic representative of the poincar dual of the surface @xmath135 at infinity in tn , the surface parametrised by @xmath165 and @xmath166 .
our calculation also gives an alternative computation of the electric charge as minus the self - intersection number of @xmath47 in @xmath0 .
let @xmath167 be this self - intersection number , then _
^1 = - _ 3^+ is poincar dual to @xmath47 since _ ^1 _ ^1 = p .
however , by the definition of the self - intersection numbers in terms of cohomology , we also have [ electronintersect ] _ tn _ ^1_^1 = p . evaluating the integral in appendix [ harmonicsect ]
we find p^2 = p , so that @xmath168 , confirming that , with the self - dual orientation , the self - intersection of @xmath47 in @xmath0 is 1 .
since the 2-form @xmath169 is harmonic and since its total flux through infinity equals the electric charge we interpret it as the electric field of the electron .
although we have adopted a viewpoint dual to standard electromagnetism , with a purely spatial 2-form being interpreted as electric rather than magnetic flux , it is interesting that the self - dual 2-form also contains a term which allows for a conventional electric interpretation : when we contract @xmath170 with the vector field @xmath171 along the fibres ( using @xmath172 ) , we obtain a purely radial field which , asymptotically , falls off like @xmath173 .
the integral is the squared @xmath140-norm of the electric field . ignoring overall factors and working with @xmath174
we write ||_3^+||^2 = _
3^+ dv , with @xmath162 defined as in , and interpret the integrand _
3^+(r ) = as an electric energy density . for comparison with the ah case below , we plot the profile function @xmath175 and the energy density @xmath176 in fig .
[ tnenpic ] .
and the energy density @xmath176,title="fig : " ] and the energy density @xmath176,title="fig : " ] turning now to ah , we review in appendix [ harmonicsect ] why there are only two @xmath5-invariant harmonic forms @xmath177 on ah which respect the identification .
they have the structure ^_1=g^_1 _ 2_3 + dg^_1 _ 1 , for functions @xmath178 of @xmath30 which satisfy the ordinary differential equations for @xmath179 in . using the asymptotic formulae and
one shows @xcite that only @xmath180 is finite at the core and decays at infinity .
in fact , the solution decays exponentially fast at infinity , with the leading term proportional to @xmath181 .
we normalise @xmath182 , so that near the core [ harmonicbolt ] g_1^+(r)~1-(r-)^2 , ( r- ) .
the 2-form @xmath183 is not dual to the surface @xmath137 at infinity in ah and has no interpretation in terms of electric flux .
however , it is dual to the core @xmath47 in ah . with
_ core= -_1^+ one finds _ core _
core = p . as a check
, we use this form to compute the self - intersection number of the core . recalling that the volume of @xmath31 is @xmath184 , and using , we have _ ah _ core_core = ( -4 ^ 2 ) ( 0 ^ 2 - 1 ^ 2 ) = , so that = p , confirming @xmath185 for the self - dual orientation ( this agrees with @xcite , where the anti - self - dual orientation gives the opposite sign ) . the harmonic form on @xmath0 is related to the signature of @xmath0 through the fact that the signature of a compact 4-manifold equals the difference of the dimensions of the spaces of self - dual and anti - self - dual harmonic representatives of the second de rham cohomology .
it seems likely that the ( up to scale ) unique bounded and rotationally symmetric self - dual harmonic form @xmath186 on ah is related to the self - dual harmonic form via the sequence of hitchin manifolds reviewed in sect .
[ hitchinsection ] .
since the existence of @xmath186 on ah is linked to the signature of ah being 1 , and since signature represents baryon number in our approach , it may be possible to interpret the detailed structure of @xmath186 in baryonic terms .
the exponential fall - off exhibited by @xmath186 is reminiscent of the proton s pion field in the yukawa description of the nuclear force . in fig .
[ ahbaryon ] we plot the numerically computed profile function @xmath187 and the associated energy density , which turns out to be @xmath188 this is finite at the core and decays exponentially according to @xmath189 for large @xmath30 . and the energy density @xmath190,title="fig : " ] and the energy density @xmath190,title="fig : " ] to find a harmonic 2-form on ah which can play the role of the proton s electric field we need to go to a branched cover of ah , denoted @xmath191 .
the metric on the branched cover is not smooth at the core , but this will not affect the following calculations near infinity . dropping the requirement that forms are invariant under the identification , we find that the closure condition on the 2-forms has only one further solution which is finite at the core and which remains bounded for large @xmath30 .
this is the 2-form ^-_3=g^-_3 _ 1_2 + dg^-_3
_ 3 , with a radial function @xmath192 satisfying .
the solution vanishes at the core as @xmath193 for some constant @xmath194 .
the large @xmath30 behaviour is @xmath195 where @xmath196 is another constant .
this leads to an anti - self - dual form on @xmath191 , which is square - integrable ( as we shall show below ) and which has not been considered previously .
it is our candidate for the electric field of the proton . as a manifold , @xmath191 compactifies to @xmath197 , as discussed in appendices [ branched1 ] and [ branched2 ] .
the surface @xmath137 which compactifies ah to @xmath0 becomes the anti - diagonal @xmath10 in @xmath197 .
choosing @xmath198 in , a harmonic representative of the poincar dual class to this @xmath10 is _ s^2= - _ 3 ^ - . then _
s^2_s^2 = p , and since the volume of @xmath5 is @xmath199 by , we also compute _ _
s^2_s^2 = ( -8 ^ 2)(1 ^ 2 - 0 ^ 2 ) = - .
hence -=p , which is solved by @xmath200 . dividing by 2 , to take us back to @xmath59 , we confirm the self - intersection number of @xmath59 in @xmath0 as @xmath63 , again for the self - dual orientation .
since the 2-form @xmath201 on @xmath191 is harmonic and since its total flux through infinity , suitably interpreted , equals the electric charge , we think of @xmath201 as representing the electric field of the proton .
the comments made after about the possibility of recovering a conventional electric field by contracting the electric 2-form on tn with the vector field along asymptotic fibres apply , suitably modified , to @xmath191 . in fig .
[ ahelfield ] we plot the numerically computed profile function @xmath192 for @xmath202 ( as defined in ) and the associated energy density , which turns out to be @xmath203 this function has a @xmath204 singularity at @xmath60 , and falls off like @xmath205 for large @xmath30 as in the tn case . and the energy density @xmath206,title="fig : " ] and the energy density @xmath206,title="fig : " ]
in introducing and illustrating geometric models of matter in this paper we have concentrated on general , global properties of particles like baryon number and electric charge .
the most striking phenomenological success of our geometric approach is its prediction of precisely two stable electrically charged particles , one leptonic and one baryonic , with opposite electric charges .
an important theme throughout this paper is the implementation of rotational symmetry in our models .
the rotation action preserves the metric and , in the compact cases , the distinguished surface where the 4-manifold intersects physical 3-space .
we have shown that our models are fermionic in the sense that the lift of the rotation action to spin bundles ( over the ` inside ' of the 4-manifold in the compact cases ) is necessarily an @xmath6-action , but we have left the explicit construction of spin 1/2 states for future work .
non - vanishing electric charge and baryon number give rise to harmonic 2-forms on our model manifolds .
these have allowed us to make contact with the conventional description of charged particles in terms of associated fields and fluxes . there are various geometrically natural candidates for measuring energy or mass in our models , but we have not committed ourselves to any particular energy functional in this paper . in the absence of an energy measure , we are not able to make contact with experimental data about particle masses or forces between particles .
we end our summary with some observations about relative scales predicted by our models for charged particles .
the ah model relates three scale parameters : the size of the core ( radius @xmath207 in our geometric units ) , the size of the asymptotic circles ( radius 2 in geometric units ) and the scale @xmath208 implicit in the exponential corrections to the asymptotic form of the coefficient functions for the ah metric , which are proportional to @xmath189 ( see @xcite for a discussion of these corrections in the context of magnetic monopoles ) . interpreting the core radius as the proton radius and the scale in the exponential decay as the compton wavelength @xmath209 of the pion
, we find that our model correctly assigns the same order of magnitude to those two quantities .
the details are not quite right ( experimentally , the proton radius is just over half of @xmath209 ) but this is not surprising given our very rudimentary understanding of how our model relates to experiment . in any case , these considerations suggest that we should pick a length unit of about 1 fermi or @xmath210 m in the ah model .
since the asymptotic fibration by circles arises for all electrically charged particles , we expect the size of the asymptotic circle to have a purely electric interpretation , and we also assume that it is the same for both ah and tn ( this was implicit in the way we fixed scale and units for tn ) .
one natural guess , at least for tn , is to relate the size of the asymptotic circle to the classical electron radius .
our models would then equate the orders of magnitude of the classical electron radius with the compton wavelength of the pion , which is phenomenologically right .
it is an attractive feature of tn as a model for the electron that it has a length scale , but no core structure .
many avenues remain to be explored . in our geometric approach ,
fusion and fission of nuclei as well as decay processes involving both baryons and leptons ( like the beta decay of the neutron ) should have a description in terms of gluing and deformation of self - dual 4-manifolds . in order to study masses and binding energies of particles
we need to pick an energy measure .
this could involve the norms of curvatures and harmonic forms discussed here , but may take a less conventional form .
our ideas about spin 1/2 need to be fleshed out .
the dirac operator on our model manifolds is likely to play a role in combining energy and spin considerations , and seiberg - witten theory on the model manifolds seems relevant , too . in order to move beyond static consideration
, time needs to be introduced .
the list of open issues may seem daunting , but in each case the geometric framework introduced here suggests natural lines of attack .
we hope to pursue them in future work .
* acknowledgment * we thank nigel hitchin and claude lebrun for assistance and jos figueroa - ofarrill for working with us during the early stages of the project .
let @xmath211 be a complex line bundle over a compact oriented surface @xmath2 .
then [ selfintersect ] c_1(l ) [ x ] = x^2 where @xmath134 denotes the self - intersection number of the zero section in the total space of @xmath211 .
this is essentially one of the definitions of the first chern class .
note that both @xmath2 and @xmath211 are oriented , with the fibres having the natural orientation of the complex numbers .
however the self - intersection @xmath134 depends only on the orientation of the total space , since @xmath212 .
let us review the definition of the hopf bundle @xmath213 , the standard line bundle over the complex projective line , paying particular attention to orientations .
we start with @xmath46 , and @xmath47 as the space of complex lines through the origin .
this complex line bundle is defined to be the dual of the line bundle @xmath213 .
the reason for this apparently perverse choice is that @xmath213 has holomorphic sections ( given by @xmath46 ) and its dual @xmath214 does not .
this is clear from the exact sequence of holomorphic vector bundles over @xmath47 : 0 h^ * ^2 h 0 .
since holomorphic intersections are always positive , the first chern class of @xmath213 gives @xmath48 when evaluated on the fundamental class of @xmath47 .
if we compactify @xmath46 by adding a @xmath47 at infinity then the self - intersection number of this line is clearly @xmath48 , so that the normal bundle is @xmath213 .
now assume that @xmath2 is a non - orientable surface in an oriented 4-manifold @xmath3 .
this defines a mod 2 homology class in @xmath3 and so it would appear that its self - intersection number is only defined modulo 2 .
however a more careful look shows that we can define an integer self - intersection .
there are several ( equivalent ) ways to see this .
first we note that the tangent bundle and the normal bundle of @xmath2 in @xmath3 have isomorphic orientation real line bundles @xmath211
. then both the euler class of the normal bundle and the fundamental class of @xmath2 are defined as classes ` twisted ' by @xmath211 .
the evaluation therefore makes sense and defines @xmath134 .
an alternative way is to pass to the double covering of a neighbourhood of @xmath2 in @xmath3 , where @xmath2 acquires an orientation , compute the self - intersection there and divide by 2 .
a third way is to deform @xmath2 into a transverse surface @xmath215 and , near each intersection point of @xmath2 and @xmath215 , choose an orientation of @xmath2 and the corresponding orientation of @xmath215 given by the deformation .
now compute the local intersection numbers of @xmath2 and @xmath215 and sum them up .
the example we want to study is that of the real projective plane @xmath216 inside the complex projective plane @xmath217 .
we give @xmath3 its standard complex orientation .
we plan to show that @xmath218 . to do this we will use the double branched covering of @xmath0 given by the product of two copies of the complex projective line @xmath47 .
we can identify @xmath0 with the symmetric product of the two copies , branched along the diagonal @xmath219 .
this map is holomorphic and compatible with the action of @xmath220 .
now introduce a metric on @xmath47 identifying it with a 2-sphere , and consider in @xmath221 the graph @xmath222 of the antipodal map .
the image of @xmath222 in @xmath0 is the standard embedding of @xmath59 into @xmath0 . outside the branch locus
@xmath219 the map from the product of the two 2-spheres to @xmath0 is a double covering so we can compute @xmath134 as half of @xmath223 .
this is most easily done by observing that , in the cohomology of the product , with @xmath224 and @xmath225 the two generators of @xmath226 , the classes of @xmath219 and @xmath222 are d = x + y , d = x - y . since @xmath227 and @xmath228 it follows that d^2 = 2 ( d)^2 = -2 ( while @xmath229 agreeing with the fact that @xmath219 and @xmath222 are disjoint ) . dividing by 2 to get back down to @xmath0 we end up with @xmath218 as stated . as a check
we can also use a deformation of @xmath222 in @xmath230 by choosing a different metric on @xmath47 giving a different antipodal map . in particular , consider the 2-sphere as the boundary of the unit ball in @xmath14 and move the origin so that antipodal points are now the end points of lines through the new origin .
the line joining the two origins defines a unique intersection point of the two copies of @xmath59 in @xmath0 . since this line acquires different orientations from the two origins the intersection number is @xmath63 .
a third sign check is to observe that the normal bundle to the diagonal in the product of the two 2-spheres is the tangent bundle with euler number 2 , while for the antidiagonal ( graph of the antipodal map ) it is the cotangent bundle with euler number @xmath231 .
it is enlightening to analyze the @xmath5 action on the spaces in the example above .
in addition we can introduce the branched double covering @xmath232 given by complex conjugation @xcite .
this is branched over @xmath59 .
thus we have maps s^2 s^2 ^2 s^4 , which are compatible with the @xmath5 actions . here
@xmath1 is best thought of as the space of real @xmath233 symmetric matrices of zero trace and fixed norm .
the generic orbits are , in each case , 3-dimensional but there are two exceptional orbits of dimension 2 . inside the product
@xmath197 they are the diagonal and anti - diagonal ( both @xmath10 ) , in @xmath1 they are both @xmath59 , while in @xmath0 one is @xmath10 and one is @xmath59 .
each of the two maps is branched along a surface : @xmath10 for the first map and @xmath59 for the second . to examine the self - intersection numbers of the exceptional orbits it is useful to note two general rules : 1 .
if @xmath2 in @xmath3 is double covered by @xmath234 in @xmath235 then @xmath236 .
if @xmath2 is a branch locus in the double covering then @xmath237 .
rule a is obvious and rule b is familiar in algebraic geometry , but it applies more generally even in the non - orientable case . using rules a and b we can now check what happens in our two maps . in @xmath1 the two @xmath59s have self - intersection numbers @xmath238 and @xmath231 . in @xmath239
the two 2-spheres also have self - intersection numbers @xmath238 and @xmath231 . in @xmath0 the 2-sphere , which is a conic , has self - intersection number @xmath240 , while @xmath59 has self - intersection number @xmath63 .
we want to analyze various sign conventions in relation to self - dual manifolds , and begin by checking our sign conventions for two examples of complex khler surfaces . referring to our discussion of the signature and euler characteristic in sect .
[ sigsect ] , we note the examples : 1 . the k3 surface with the yau metric , which makes it hyperkhler . with its complex orientation
its signature is @xmath241 and its euler characteristic is 24 . with the opposite orientation ,
the metric is self - dual and the signature is positive .
2 . @xmath0 with the fubini - study metric is self - dual for its complex orientation @xcite , agreeing with the positive signature @xmath48
. we will also be interested in non - compact examples similar to the above .
the argument that a hyperkhler manifold is self - dual for the orientation opposite to the complex orientation ( given by one of the family of complex khler structures ) is purely local .
it just depends on the fact that the bundle of self - dual 2-forms for these complex orientations is flat . more generally , on any khler manifold , the bundle of self - dual 2-forms is the direct sum of the canonical line bundle @xmath242 , its dual and a trivial bundle generated by the khler form .
the first non - compact example that interests us is the taub - nut manifold tn .
this has the isometry group @xmath45 , its topology ( and symmetry ) is that of @xmath46 with the central @xmath43 giving the scalar action .
the other @xmath43 actions , inside @xmath6 , define the complex hyperkhler structures which have the opposite orientation . hence tn is self - dual for the orientation that becomes that of @xmath46 in the limit when the parameter @xmath243 goes to 0 , as discussed after .
the second example is that of the ( simply - connected version of the ) atiyah - hitchin manifold ah , which is an open subspace of @xmath0 got by removing @xmath59 .
as shown in @xcite , it has a complete hyperkhler metric .
this is self - dual for the orientation opposite to the complex orientations given by the hyperkhler metric . but this is just the orientation given by its complex structure as an open subset of @xmath0 .
this can be checked directly , but it is best seen by using the results of hitchin @xcite , reviewed in sect .
[ hitchinsection ] , which give a whole sequence of self - dual manifolds on the same space , starting from the fubini - study metric of @xmath0 and converging to ah . in each case , for tn and ah , we have a hyperkhler manifold acted on by @xmath6 ( the action descends to @xmath5 for ah ) .
the manifolds have a ` core region ' and an ` asymptotic region ' with an asymptotic action of @xmath244 .
in addition to the 2-parameter family of complex structures given by the hyperkhler metric ( and rotated by @xmath6 ) there is another complex structure compatible with the @xmath6 action , but giving the opposite orientation to that of the 2-parameter family .
for tn this complex manifold is just @xmath46 and for ah it is @xmath245 . in both cases
the asymptotic @xmath244 action gives us topologically a disc bundle at infinity and this can be identified with the normal bundle of the surface @xmath2 that naturally compactifies the manifold . for tn
this compactification is @xmath0 with @xmath246 , and for ah it is @xmath0 with @xmath137 .
we have shown that these give rise to opposite signs : @xmath247 as explained in the main text , tn and ah are our geometric models for the electron and the proton , and we identify the self - intersection number of @xmath2 with minus the electric charge .
in this paper we need to parametrise @xmath6 explicitly at various points , and we use generators t_1=i_1 = 0 & i + i & 0 , t_2=i_2 = 0 & 1 + -1 & 0 , t_3=i_3 = i & 0 + 0 & -i , where @xmath248 , @xmath249 , are the pauli matrices ; the commutators are @xmath250=-\epsilon_{abc}t_c$ ] .
we use euler angles @xmath251 , @xmath252 and @xmath253 for parametrising @xmath6 as follows : g(,,)= e^t_3e^t_2 e^t_3= e^i 2 ( + ) & e^i 2 ( - ) + -e^i 2 ( - ) & e^-i 2 ( + ) . defining left - invariant 1-forms @xmath254 on @xmath6 via @xmath255
we compute to find @xmath256 satisfying @xmath257 .
they descend to @xmath5 by simply restricting @xmath258 to @xmath259 .
explicitly , generators of the lie algebra of @xmath5 are @xmath260 .
they also satisfy @xmath261=-\epsilon_{abc}t_c$ ] .
we can then parametrise @xmath5 matrices via [ rotpara ] r(,,)= e^t_3e^t_2 e^t_3 , with @xmath251 , @xmath252 and @xmath262 .
then @xmath263 provides an alternative definition of the left - invariant 1-forms on @xmath5 : one obtains the same expressions as in , but with the range of angles automatically appropriate to @xmath5 .
both tn and ah can be parametrised in terms of euler angles and a radial coordinate . for tn , the angular ranges are @xmath264 , @xmath265 and @xmath266 .
for ah they are @xmath264 , @xmath265 and @xmath267 with the additional @xmath62 identification [ ident ] ( , , ) ( - , + , - ) , which , in the asymptotic region , is the simultaneous reversal of spatial and fibre direction .
the following angular integrals , which enter into various calculations in this paper , are therefore @xmath268 note that @xmath94 is invariant under , but @xmath269 and @xmath270 change sign .
recalling that the metrics on tn and ah are of the bianchi ix form @xmath271 we introduce the tetrad [ tetrad ] ^1=a _ 1 , ^2=b_2 , ^3 = c_3 , ^4=f dr . the self - duality of the riemann tensor computed from with respect to the volume element [ volumedef ] dv = a_1b_2c_3fdr = -fabc dr _ 1 _ 2 _ 3 is then equivalent to the set of ordinary differential equations [ gendual ] = ( b - c)^2 -a^2 + 2bc , , where ` + cyclic ' means we add the two further equations obtained by cyclic permutation of @xmath35 , and @xmath272 is a parameter which has to be either @xmath273 or @xmath208 . in all cases
the resulting metrics are hyperkhler , but in order to obtain metrics whose hyperkhler structures are rotated by the @xmath6 action we need to set @xmath274 .
this is the case for both tn and ah , so @xmath274 in .
one checks the equivalence of the self - duality of the riemann tensor and as follows .
denoting the riemann tensor by @xmath143 , its component 2-forms with respect to are @xmath275 and similar expressions for the components obtained by cyclic permutation of @xmath276 , where [ mfcts ] _ 1= , _ 2= , _
3= and _ 1= , _ 2= , _
3= . when hold , then @xmath277 , @xmath278 , with @xmath274 or @xmath279 , and the self - duality relations r_12=r_34 , r_23=r_14 , r_31=r_24
are easily verified .
one also checks that each of the three independent 2-forms @xmath280 is self - dual with respect to , as expected .
the only solutions of with @xmath274 which give rise to complete manifolds whose generic @xmath6- or @xmath5-orbit is three - dimensional are the tn and ah metrics , whose coefficient functions we discussed in sect .
[ protonelectron ] .
it is important that the coefficient function @xmath281 in is negative for all @xmath30 in the ah manifold .
it implies in particular that the canonical volume element / orientation @xmath162 in , with @xmath282 , [ volume ] dv = -fabc dr_1_2_3 = - drd dd , has opposite signs for tn and ah , in the following sense . assuming the same orientation @xmath283 of the @xmath6 and @xmath5 orbits in tn and , respectively , ah , the radial direction is oppositely oriented in the two cases : the natural radial line element @xmath284 has the same orientation as @xmath285 for tn , but the opposite orientation for ah
. this will be important when evaluating various integrals .
it means that , when using the coordinates @xmath286 we can use the conventional orientations for these coordinates when computing for ah , but should use the opposite orientation when computing for tn .
thus we integrate in the negative @xmath30-direction when calculating on tn .
the squared @xmath140-norm of the riemann tensor is [ curvsquared ] ||r||_m^2 = _ m 2(r_12r_12 + r_23r_23 + r_31r_31 ) , where @xmath3 is tn or ah , and we have used the self - duality of the curvature . in terms of the functions
the integrand can be simplified ( see also @xcite , where this calculation is carried out for the ah metric ) and becomes @xmath287 with f=2(_1+_2+_3 - 1)^2 - 8_1_2_2 .
we denote these functions in the tn and ah cases by @xmath288 and @xmath289 . for tn , we can compute explicitly and find f_tn(r)= -8 . integrating and using as well as our sign convention for the volume form we find for ah , similarly , ||r||^2_ah = ( -4 ^ 2 ) ( f_ah()-f_ah ( ) ) . to complete the calculation we need the behaviour of @xmath289 at infinity and at @xmath58 .
one finds for ah @xmath290 as well as _ 1()=-1 , _ 2 ( ) = , _ 3()= 1 2 , where we took careful limits , using and .
therefore @xmath291 so
the question of computing harmonic 2-forms on the tn and ah manifolds arose in physics in the context of testing s - duality , see @xcite for ah and @xcite for tn .
harmonic forms also play a role as curvatures of line bundles over tn and ah , see @xcite for a discussion of this for ah where the form later used for testing s - duality @xcite is the curvature of an index bundle .
this appendix contains a systematic discussion of these forms and also a harmonic form on the branched cover of ah which has not previously been considered in the literature .
the physical interpretation of the forms as electric and baryonic fluxes is given in sect .
[ enfluxsection ] . on both tn and ah , rotationally symmetric 2-forms can be written in terms of the metric coefficient functions @xmath33 appearing in the respective metrics and functions @xmath292 and @xmath293 of the radial
coordinate @xmath30 as @xmath294 with @xmath295 standing for self - dual and @xmath296 standing for anti - self - dual . introducing the functions [ coeffren ] g_1^= f_1^bc , g_2^= f_2^ca ,
g_3^= f_3^ab , closure of the forms implies the ordinary differential equations @xmath297 for solutions of these differential equations ( we shall discuss below in which cases regular solutions exist ) , the forms can be written @xmath298 all these forms are locally exact , that is , ^_i
= d(g^_i _ i ) , i=1,2,3 , but the 1-forms in brackets may not be globally defined and , when defined , they may not be @xmath140 , so that the corresponding forms @xmath299 are not necessarily @xmath140-exact . substituting the tn coefficient functions ( with @xmath51 and @xmath53 ) , the functions simplify further to -=-=- , -= - .
it is easy to check , and was shown in @xcite and @xcite , and earlier in @xcite and @xcite , that only the equation for @xmath301 has a solution which is both regular at the origin and bounded as @xmath302 . integrating
, one finds explicitly g^+_3(r ) = c , with an arbitrary constant @xmath194 . setting @xmath202 ,
gives the associated normalisable , harmonic form _
3^+ =( _ 1_2 + dr _ 3 ) = d ( _ 3 ) . then , using , _ tn _ 3^+_3^+ = _ ^0 dr(-16 ^2 ) = 16 ^2 .
on ah , the analysis of harmonic forms is analogous .
we write @xmath303 for the solutions of with the radial functions @xmath304 of the ah metric , and write the forms obtained by solving these equations as @xmath305 as for tn , all these forms are formally exact , ^_i = d(g^_i _ i),i=1,2,3 , but the 1-forms in brackets may not be globally defined , as we shall see . note that only the forms @xmath177 respect the identification . however , as explained in the main text we also consider the branched double cover @xmath191 of ah and therefore keep all forms in the discussion .
on ah , we use and to find that the functions appearing in the differential equations have the following behaviour near the core ( i.e. @xmath306 small ) , [ ahbolt ] ~-(r- ) , ~ ~- , and the following behaviour for large @xmath30 , [ ahinfty ] ~ ~-(1- 2 r ) , ~- . for the differential equations on ah
, this means that only the equations for @xmath180 and @xmath307 have solutions which are finite at the core and remain bounded for large @xmath30 .
both of these solutions are discussed and interpreted in the main text .
for any solution of on ah , we note @xmath308 where @xmath309 and @xmath310 stand for any of the forms and coefficient functions in , and we again used .
we also note that the @xmath140-norm of each of these forms can be written as @xmath311 where @xmath162 is the volume element defined in and @xmath86 is a density which we interpret as an energy density in the main text .
its general form is @xmath312 which can be simplified in each of the cases , using the differential equations on ah . c. bouchiat and g. w. gibbons ,
non - integrable quantum phase in the evolution of a spin-1 system : a physical consequence of the non - trivial topology of the quantum state - space , j. phys .
france 49 ( 1988 ) 187199 . c. lebrun , curvature functionals , optimal metrics , and the differential topology of 4-manifolds , in _ different faces of geometry _ , s. donaldson , ya .
eliashberg and m. gromov , eds . , new york , kluwer academic / plenum , 2004 . | inspired by soliton models , we propose a description of static particles in terms of riemannian 4-manifolds with self - dual weyl tensor . for electrically charged particles ,
the 4-manifolds are non - compact and asymptotically fibred by circles over physical 3-space .
this is akin to the kaluza - klein description of electromagnetism , except that we exchange the roles of magnetic and electric fields , and only assume the bundle structure asymptotically , away from the core of the particle in question . we identify the chern class of the circle bundle at infinity with minus the electric charge and the signature of the 4-manifold with the baryon number .
electrically neutral particles are described by compact 4-manifolds .
we illustrate our approach by studying the taub - nut manifold as a model for the electron , the atiyah - hitchin manifold as a model for the proton , @xmath0 with the fubini - study metric as a model for the neutron , and @xmath1 with its standard metric as a model for the neutrino .
empg-11 - 12 + damtp-2011 - 49 * geometric models of matter * * m. f. atiyah * , + school of mathematics , university of edinburgh , + king s buildings , edinburgh eh9 3jz , uk . + [email protected] * n. s. manton * , + damtp , centre for mathematical sciences , + university of cambridge , + wilberforce road , cambridge cb3 0wa , uk .
+ [email protected] + * b. j. schroers * , + department of mathematics , heriot - watt university , + riccarton , edinburgh eh14 4as , uk .
+ [email protected] + august 2011 8 pt |
it is well known that during the fabrication process of superconductor materials , a variety of point defects , such as substitution and interstitial impurities , vacancies , and cation - disorder are involved .
these defects have a large effect on the properties of superconductors @xcite . in type ii superconductors ,
high critical current densities can be achieved by the presence of high - density defects which will provide suitable pinning centers for the magnetic flux lines .
the ideal size of defects for flux line pinning should be comparable to the superconducting coherence length . for cuprates such as yba@xmath2cu@xmath3o@xmath4 ( y-123 ) ,
the coherence lengths are in the order of tens of angstroms while the conventional superconductors have a coherence length of several thousand angstroms .
thus atomic - scale structural inhomogeneities such as point defects and columnar defects can play an important role in flux - line pinning @xcite .
an increasing number of applications of the 123-type high-@xmath5 superconductors use materials other than y-123 .
for example , in many bulk forms and multilayer applications , y is replaced by nd , sm , or other rare - earth elements ( re ) .
doping yba@xmath2cu@xmath3o@xmath4 with ca , sr , or other alkaline earth elements ( ae ) has also been shown to improve bulk and grain boundary transport and other properties @xcite .
since it is well - known that strain effects are important in the studies of point defects , it is expected that studies of strain effects on point defects for series of reba@xmath2cu@xmath3o@xmath4 ( re-123 ) compounds will be important for practical applications of superconductivity . in particular , the concentration and ordering of oxygen vacancies have significant effects on the superconducting properties . for y-123 , the superconducting temperature , @xmath5 , depends on the oxygen stoichiometry . as an example of a generic doping curve in cuprates @xcite , when @xmath6 is larger than around 0.7 , the crystal loses superconductivity .
however , if @xmath6 is smaller than 0.7 , the compound is superconducting .
it is generally believed that higher oxygen content can create more holes in the structure .
when @xmath6 is between 0.1 and 0.7 , the crystal is in the underdoped region , and @xmath5 increases with increasing hole concentration .
when the oxygen content is in the proper range , @xmath5 is above the boiling point of liquid nitrogen .
however , if more holes are created , the @xmath5 value decreases instead , and the crystal is in an overdoped region .
oxygen ordering in the cu - o chains of the 123 structure gives rise to further complex structures @xcite .
even when the average occupancies of oxygen sites remain constant , the occupation at chain and anti - chain sites can vary .
jorgensen et al .
@xcite observed that the superconducting transition temperature in y-123 changes as a function of time following the quenching experiment while the oxygen content is fixed .
this demonstrates that the specific ordering of the oxygen atoms in the basal plane is another important parameter that controls @xmath5 in the y-123 system @xcite .
the ordering process in the cu - o chain is the origin of the structural transition between tetragonal and orthorhombic phases in the 123 structure , which has been extensively studied since the discovery of y-123 @xcite .
structural transition of y-123 is strongly affected by external pressure as well @xcite .
in addition , there are systematic results reported by wong - ng et al .
@xcite showing that the orthorhombic / tetragonal phase transition temperatures in re-123 can be scaled approximately linearly with the ionic radius of re@xmath7 .
the above experimental observations indicate that lattice strain may play an important role in phase transition , which has not been investigated theoretically and systematically so far . in this paper , we plan to study the effects of homo - valent substitutions at y and ba sites , and hydrostatic pressure on the order - disorder transition theoretically .
considering large stress fields due to dislocations around grain boundaries , this study will also provide valuable information for understanding the transport properties in the vicinity of grain boundaries .
first we briefly explain methods used in atomistic simulations .
secondly , we focus on the effects of strain on schottky defects and related phenomena for a series of homovalent substitutions at y and ba sites of yba@xmath2cu@xmath3o@xmath4 .
although the 123 structure is not stable in the ca- and sr - analogs under ambient pressure @xcite , partial substitution of sr on the ba site has been reported @xcite .
finally , strain effects on chain - oxygen order - disorder transition of reba@xmath2cu@xmath3o@xmath4 will be investigated .
since the parent compounds of high-@xmath5 cuprate superconductors can be considered as charge - transfer insulators , ionic bonding can be assumed to have a large contribution to the lattice energy .
methods used to study atomistic phenonmena in conventional oxides can therefore be applied to study cuprates .
the cuprates become superconductors at a proper doping level and temperature .
however , the charge carrier density is very low compared with that of conventional metals . in addition , the charge carriers are confined within copper - oxygen planes .
consequently , the screening effect is not as strong as that of conventional metals .
the lattice energy calculated from ionic models is somewhat over - estimated .
the point defect s energy is also slightly overestimated due to the omission of screening effect in metallic region of the phase diagram , which can be improved by including polarization effects in the shell model @xcite .
many previous theoretical investigations are based on this type of ionic model ( for example , see @xcite ) .
several pair - potential sets of shell model parameters have been determined for y-123 by baetzold @xcite . to systematically study homo - valent substitutions on y- and ba- sites , a consistent set of shell model parameters using the data set for yba@xmath2cu@xmath3o@xmath8 @xcitewas further developed to account for the dependence of the born repulsion of the two ions on their net charges , on outer electronic configurations @xcite , and on the common @xmath9 law " between polarizability and radius @xcite .
the virtual crystal method " is applied here to interpret experimental data of mixing two types of elements on one site .
this method essentially is , for the purpose of calculating average structural and elastic properties , to approximate the mixture of two ions distributed over one sublattice by identical average virtual ions " .
it allows the incorporation of compositional changes at the atomic level , but ignores explicit effects of disorder .
for instance , when ba@xmath10 is partially replaced by sr@xmath10 in experiments , the composition - weighted average value of the two ions radii is taken to approximate that of each virtual divalent ion .
after constructing a consistent interatomic pair potential set @xcite , we used the general utility lattice program " ( gulp ) @xcite , which integrates the above modeling methods at an atomistic scale , to study lattice energy , elastic constants , and lattice dynamic properties . in summary ,
our calculations are based on short - range potentials of the buckingham type , long - range coulomb potentials , and displacement - induced deformations of the electronic charge density in the framework of a shell model .
crystal structure of yba@xmath2cu@xmath3o@xmath8 . ] schottky defects formation energy vs volumetric strain .
the volumetric strain is defined as @xmath11 , where @xmath12 is the volume of y-123 compound .
all volumes are obtained by optimizing cell parameters and internal coordinates to minimize the total energy .
the solid solutions are treated by the virtual crystal method " . ]
schottky defects formation energy vs b@xmath13 of 123 compounds .
b is the bulk modulus and @xmath13 is the mean volume per atom of 123 compounds .
both b and @xmath13 are calculated from the optimized structures with minimal total energy .
the solid solutions are treated by the virtual crystal method " . ]
any deviation in a crystal from a perfect structure is an imperfection .
the simplest imperfection is a lattice vacancy , which is a missing atom or ion , known as a schottky defect .
schottky defects involve multiple " vacancies while preserving electrical neutrality .
regardless of how a schottky defect is created , it is necessary to expend a certain amount of work per atom to take it to the surface .
we calculated point defect energy by mott - littleton approach @xcite .
some vacancies are at anionic sites and others at the cationic sites . for y-213 ,
the defect reaction is given as follows : @xmath14 where v represents a vacancy . according to mass action law
, the equilibrium constant ( k ) at a finite temperature can be written as @xmath15 where c is the equilibrium concentration of vacancies ; @xmath16 is the average concentration of vacancies .
the equilibrium concentration of the vacancy j , @xmath17 , can be computed from boltzmann statistics as @xmath18 where @xmath19 is the number of atom j per unit volume , and @xmath20 is the energy required to take atom j from its lattice site inside the crystal to a site on the surface . substituting the equilibrium concentraions of vacancies into equation ( [ eq : mass ] )
, we obtain the expression for the average concentration of vacancies , @xmath16 , as follows : @xmath21 where @xmath22 is the number of formula units per unit volume , and @xmath23 is given , in terms of the formation energy of individual vacancy ( an ion is removed to infinity instead of the surface of crystals ) and lattice energy @xmath24 , as @xmath25 in re-123 , the absolute value of site potential decreases continuously with increasing ion radius at y - site , which is consistent with applying external tensile pressure .
usually , the change of short - range repulsion is less significant than that of the madelung site potential . in general
, the smaller absolute site potential value is favorable for lowering vacancy energy , and vice versa .
furthermore , an increase of the total volume leads to the expansion of effective relaxation space so that the vacancy energy can further decrease due to the relaxation of the atoms . in the shell model
, the electronic polarization due to the electronic relaxation also lowers the defect energy . in @xmath1 ( ae-123 ) , as the radius of the ion at ae - sites becomes larger , a similar phenomenon is observed . in both @xmath26 and @xmath1 ,
the lower schottky defects formation energies are always associated with the larger volume .
this indicates that the volume of formation of schottky defects is positive , or the volume of the entire crystal expands during the formation of schottky defects . as shown in fig .
( 1 ) , the complex structure of 123 compounds can be roughly divided into six layers along c - axis direction : re - cuo@xmath2(cu2)-bao - cuo(cu1)-bao - cuo@xmath2-re . for the sake of simplicity ,
the distances between planes are represented by the separations of cations projected along c - axis .
the external and internal strains are computed with y-123 as a reference .
replacing y@xmath7 by larger re@xmath7 ions leads to positive external strains , indicating the dimensions of the cell expand as increasing re size .
the more interesting observations are the changes of internal parameters within the unit cell .
the variance of internal strains provide direct information of the changes of layers separations in re - series .
the interatomic distances vary quite differently with ionic radius in the re-123 and ae-123 series @xcite .
for instance , the distance of re - cuo@xmath2 becomes larger with increasing re size , which is expected from the change of short - range repulsion of re - o .
however , even when the entire cell volume increases as re ion becomes larger , some parts of unit cell contract such as interlayer separations between cuo@xmath2(cu2)-bao and bao - cuo(cu1 ) ( or , bond lengths of cu2-o4 and cu1-o4 ) .
consequently , the average energy of the point defect only decreases slightly because of the existing compressive blocks within the structure as plotted in fig .
( 1 ) . in the ae-123 series ,
when ba@xmath10 is ( partially ) replaced by smaller ae@xmath10 , the variances of external and internal strains are of the same sign , reflecting a somewhat even " expansion of the entire structure . the difference between re-123 and ae-123 is reflected by the changes of ae - ae and cu1-cu2 distances . unlike those in re-123 , for ae-123
both distances become larger with increasing ae size .
the different changes of internal strains due to doping at either y or ba sites governs the absolute value of slope of schottky defects of re-123 and ae-123 ( see fig .
( [ fig : schottky ] ) ) . in the previous studies @xcite
, they found interesting correlations between dopant radius and energy of solution in yba@xmath2cu@xmath3o@xmath8 .
the trend of subsituting divalent cation for barium matches well with that in the schottky defect energy of yae@xmath2cu@xmath3o@xmath8 . however , there exist clear different trends between replacing yttrium by trivalent ions in yba@xmath2cu@xmath3o@xmath8 and that in the schottky defect energy of reba@xmath2cu@xmath3o@xmath8 .
the reason is that the parent structure is fixed for former case , so that the trivalent ion larger than yttrium leads to more energy of solution .
one the other hand , the schottky defect has to consider all the atoms in the unit cell . unless the structures change more or less homogeneously " , the energy of solution at one specific site may evolve differently from the schottky defect energy .
recall that the relation between cell volume , bulk modulus of re-123 , and the trend of thermal expansion coefficients are also the manifestation of the complexity of the 123 structure @xcite .
there is an unusual " relation between schottky defects formation energy and @xmath27 ( where b is bulk modulus and @xmath13 is the mean volume per atom ) of re-123 in figure ( [ fig : varotsos ] ) .
as reported by varotsos @xcite , the schottky defects formation energy is proportional to @xmath27 for elemental and binary crystals . while this relation appears to be obeyed by the ae-123 series ,
it is violated by the re-123 series .
the possible reason is still the non - uniform changes of the interatomic distances within the unit cell for re-123 .
we have studied the trend of melting temperature of re-123 by lindemann law " @xcite , and found that the vibrations along c - axis are the most important modes to determine the melting temperatures while the isotropic approximation fails to yield correct results @xcite . here
, we take another approach to investigate the relation between melting temperatures and formation energies of schottky defects .
kurosawa @xcite gave a direct correlation between them , which is listed in table ( [ tab : defectmelt ] ) .
while it appears that a smaller schottky defect formation energy can lower the melting temperature of simple binary significantly , it has not been proved to be true for compounds with complex structure .
note that the predominant point defect could be frenkel defect rather than schottky defect in some complex structures .
however , schottky defect energy provides an unambiguous measure on the average cohesive strength .
for example , despite the schottky defect formation energy of sm-123 is smaller than that of y-123 , sm-123 has a higher melting temperature .
it is necessary to study internal strains rather than volumetric strains in order to obtain a detailed analysis of the thermodynamic properties of compounds with complex structures . in general
, we found that the signs of the slopes ( @xmath28 ) of re-123 and ae-123 in fig .
( [ fig : schottky ] ) are the same .
the difference between these two slopes and the unusual " relationships between schottky defects formation energies and @xmath27 and melting temperatures of re-123 compounds all reflect the complexity of the crystal structure . . the schottky defects formation energy and melting temperature of binary and re-123 compounds .
the data of binary compounds are from ref .
the melting temperature data of re-123 are from ref .
[ cols=">,>,>,>,>,>,>",options="header " , ]
the oxygen content is an important parameter for superconductivity in cuprates . in addition
, the distribution ( ordering ) of oxygen atoms among the atomic sites strongly affects @xmath5 .
the order - disorder transition of chain - oxygen has been extensively studied both experimentally and theoretically @xcite .
raman studies show that there exists a pressure - induced ordering phenomenon @xcite .
the strain effect on this transition has not been investigated theoretically in a systematic fashion . here
we focus on the effects of homo - valent substitutions at y and ba sites , and hydrostatic pressure on the order - disorder transition . in the quasi - chemical approximation
( qca)@xcite , the short - range order is characterized by the fraction number of near - neighbor pair sites occupied by oxygen - oxygen pairs @xmath29 , where n is the number of sites on each of the sublattice @xmath30 and @xmath31 , and @xmath32 is the number of oxygen - oxygen near neighbor pairs .
the long - range order parameter s is defined such that the fractional site occupancy on sublattice @xmath31 is @xmath33 , while that on sublattice @xmath30 is @xmath34 where c is the fractional site occupancy averaged over both sublattices .
[ note that for yba@xmath2cu@xmath3o@xmath4 , c is 0.5 when @xmath6 is zero ; i.e. @xmath6=1 - 2c . ]
the partition function is given by @xmath35 where @xmath36 ; @xmath37 is the probability of pairs @xmath38 ; and @xmath39 is proportional to the total number of ways one can divide n entities into four groups of @xmath40 , and @xmath41 pairs . the configuration energy is denoted as @xmath42 . as usual , we may replace each sum by its maximum term in the summation for the system of a great many assemblies .
hence , we have @xmath43 in which the most probable value @xmath44 of @xmath37 is determined by @xmath45 there are three unknowns for a given value of temperature t and oxygen partial pressure @xmath46 : c , s , p. three
equations involving these three unknowns are obtained by requiring that : 1 ) the chemical potential of oxygen atoms is the same on both sublattices ; 2 ) the chemical potential of oxygen atoms is the same in the solid and in the gas phase(which consists mostly of diatomic molecules but has an equilibrium concentration of atomic oxygen ) ; and 3 ) the free energy of the system is a minimum with respect to the fractional number of oxygen - oxygen pairs .
the following derivations follow the similar procedure as described in @xcite . using above conditions , qca yields @xmath47 \over ( 1-s)[1-c(1+s ) ] } ) ; \\ \ln [ ( { 1-c(1+s ) \over c(1+s)})^3 & ( & { c(1+s)-p \over 1 - 2c+p})^4 ] = \ln [ ( { p_{o_2 } \over \xi})^{1\over 2 } { 1\over ( kt)^{7\over 4 } } ] - { \epsilon + { 1\over 2 e_d } \over kt } ; \\ { v\over kt } & = & \ln ( { c(1+s)-p)(c(1-s)-p ) \over p(1 - 2c+p)});\end{aligned}\ ] ] where @xmath48 is the energy to remove an oxygen atom from the gas and place it in the lattice ; @xmath49 is the repulsion energy between near - neighbour oxygens ; e@xmath50 is the dissociation energy of one oxygen molecule ; @xmath51 is equal to 4.144 @xmath52 10@xmath53 pa(ev)@xmath54 .
the desired values of c , s , and p are obtained for given values of t and @xmath46 , by simultaneously solving the above equations .
we note that the order - disorder transition temperature t@xmath55 is related to the oxygen - oxygen repulsion energy @xmath49 and the value of the average site occupancy c at this temperature by @xmath56 frenkel defects involve an atom displaced from its normal site into an interstitial site . if the interstitial site is chosen as anti - chain site for 123 compounds , this frenkel pair is closely related to chain - oxygen order - disorder transition .
forming a frenkel pair in an otherwise perfect crystal is an elementary disordering event .
as the disordering proceeds , it is important to account for defect - defect interactions . using mott - littleton approach ,
we have computed the isolated frenkel pair formation energy for reba@xmath2cu@xmath3o@xmath8 , yae@xmath2cu@xmath3o@xmath8 , and y-123 under hydrostatic pressure .
we found that this formation energy increases in compressive regions and decreases in tensile regions under hydrostatic pressure .
this is the origin of ordering under stress . from the systematic investigation of the phonon spectral characteristics with the application of pressure @xcite
, it was observed that the changes induced by the hydrostatic pressure have a strong effect on chain ordering .
results of our calculations are consistent with the reported observation .
figure ( [ fig : frenkel ] ) shows the plots of the frenkel pair formation energy versus lattice strain ( volumetric strain ) for both the re - series and the ae - series .
these two plots demonstrate a similar trend and they form a well - connected smooth curve .
the similar behavior of the curve for the re- and ae - series ( with chemical doping ) and the curve for y-123 ( with hydrostatic pressure ) indicates that the oxygen disordering energy is dominated by lattice strain , which is expected if the short range repulsion terms dominate the energy required to form the oxygen interstitial ion .
frenkel pair formation energy vs. volumetric strain for 123 compounds .
the energy corresponds to moving one chain - oxygen into an anti - chain site .
all volumes are obtained by optimizing cell parameters and internal coordinates to minimize the total energy .
the solid solutions are treated by the virtual crystal method " . ]
furthermore , scaling the oxygen - oxygen short - range repulsive energy by the calculated frenkel pair formation energy , the transition temperature between orthorhombic and tetragonal phases is computed based on equation ( [ eq : vt ] ) .
the transition temperature of @xmath57 was used as a reference .
the results are plotted in fig .
( [ fig : tod ] ) with the experimental data taken from @xcite .
there exists observable difference ( c.a .
100 k ) between theory and experiment in nd-123 system @xcite .
it has been determined experimentally that the orthorhombic - to - tetragonal phase transition in the re-123 series take place at an oxygen composition in the range of 6.4 ( er ) to 6.83 ( nd ) , not all at 6.5 ( y ) @xcite .
the change of oxygen content leads to the change of structure such as lattice parameters and atomic positions @xcite , which in turn alters the formation energy of frenkel pairs .
since we set the reference transition temperature to be the value in y-123 case , this corresponds to the transition at oxygen content being 6.5 .
if we track the whole process of order - disorder transition starting from fully oxygenated case , the total oxygen content decreases until the transition is finished .
the formation energy of frenkel pairs decreases also as anisotropy in the ab plane ( defined as @xmath58 ) reduces .
note that the transition occurs at oxygen content being 6.83 in nd-123 system , this indicates the anisotropy in nd-123 remains higher compared with y-123 .
therefore , the formation energy of frenkel pairs decreases less than that in y-123 during the whole process of transition as observed experimently .
this means that in our simple " model the transition temperature for nd-123 is underestimated .
but , the remarkable thing is that using such a simple model , the trend of the theoretical results agree well with the experimental measurements .
it is seen that stress and chemical pressure " can substantially alter the degree of disorder .
this study provides a clear evidence for the effects of strain on order - disorder transition .
previous studies also show there are quite rich microstructures resulting from this orthorhombic - to - tetragonal phase transition : for instance twin structures and related twinning dislocations @xcite , tweed morphology caused mainly by ( 110 ) and ( @xmath5910 ) shear displacements @xcite . in particularly , our results can be applied further to deduce that strain near edge dislocations in low angle grain boundaries @xcite , which can also have a significant effect on the degree of ordering in cuprate materials .
for example , the interactions between point defects and strain fields due to dislocations and/or grain boundaries can affect the distribution of point defects , and the content and degree of order of oxygen atoms @xcite .
chain - oxygen order - disorder transition temperature of reba@xmath2cu@xmath3o@xmath8 .
the theoretical data are calculated by using the transition temperature of @xmath57 as a reference .
the experimental data are from refs .
the error bars indicate the temperature range measured by x - ray diffraction . ]
based on the mott - littleton approach we studied the schottky defect formation energy in the 123 phase as a function of volumetric strain .
generally , a more expanded lattice favors a lower schottky defect formation energy , and vice verse .
the difference of slopes ( @xmath60 ) between re-123 and yae@xmath2cu@xmath3o@xmath4 , the
unusual " relation between schottky defects formation energies and @xmath27 and melt temperatures of re-123 compounds all reflect the complexity of the crystal structure of 123 compounds .
our study also illustrates the importance of strain effects on the orthorhombic / tetragonal phase transition in the re-123 compounds .
we have calculated the formation energy of frenkel pair defects as a function of volumetric strain for reba@xmath2cu@xmath3o@xmath8 and yae@xmath2cu@xmath3o@xmath8 , and for yba@xmath2cu@xmath3o@xmath8 under hydrostatic pressure .
our calculations show good agreement with experimental observations in that pressure favors ordering of the cuo chains .
for example , the frenkel pair formation energy indeed increases significantly ( c.a .
-0.25 ev / 0.01 volumetric strain ) under compression . based on a quasi - chemical approach ,
the orthorhombic / tetragonal transition temperatures for re-123 have been computed by scaling the effective oxygen - oxygen short - range repulsive energy in the cuo chain using the frenkel pair formation energy .
the calculated results agree with experimental data in that the larger the ionic size of re , the lower the orthorhombic / tetragonal phase transition temperature .
hbs is grateful for chinatsu maeda s kind assistance in preparing the manuscript .
the work at nist was partially supported by the us department of energy ( doe ) .
the work at brookhaven national laboratory was performed under the auspices of the division of materials sciences , office of science , u.s .
department of energy under contract no .
de - ac-02 - 98ch10886 . | the energetics of schottky defects in 123 cuprate superconductor series , @xmath0 ( where re = lanthandies ) and @xmath1 ( ae = alkali - earths ) , were found to have unusual relations if one considers only the volumetric strain .
our calculations reveal the effect of non - uniform changes of interatomic distances within the re-123 structures , introduced by doping homovalent elements , on the schottky defect formation energy .
the energy of formation of frenkel pair defects , which is an elementary disordering event , in 123 compounds can be substantially altered under both stress and chemical doping . scaling the oxygen - oxygen short - range repulsive parameter using the calculated formation energy of frenkel pair defects , the transition temperature between orthorhombic and tetragonal phases is computed by quasi - chemical approximations ( qca ) .
the theoretical results illustrate the same trend as the experimental measurements in that the larger the ionic radius of re , the lower the orthorhombic / tetragonal phase transition temperature .
this study provides strong evidence of the strain effects on order - disorder transition due to oxygens in the cuo chain sites . |
early theoretical speculations suggested the appearance of a weakly interacting perturbative quark - gluon plasma state in heavy ion collisions at ultrarelativistic energies .
however , recent data obtained at rhic experiments , and their theoretical analysis , indicate the formation of a strongly interacting deconfined matter @xcite . in fact
, lattice - qcd calculations on equation of state have already shown the strong coupling and the theoretical analysis based on massive quasi - particle picture described quantitatively many properties of this matter @xcite .
thus we expect a relatively dilute , massive , quark and antiquark dominated matter to appear before hadronization .
widely used molecular dynamics simulation @xcite is the ideal tool to study such a particle compound , to investigate non - perturbative features and to determine different properties in the strongly interacting matter . in this simulation
we can study the influence of the color charge fluctuation and pair correlation function can be generated , displaying if the quark matter is gaseous or rather liquid like .
although molecular dynamics models for non - abelian interactions are very much approximate , we can study qualitatively very important microscopic processes , for what other methods do not exist . here
we summarize the main findings of our work , details can be found in ref .
in the model the quarks are the only actively simulated species , the gluons are represented as a background field .
this approach is similar to the case of the classical multicomponent or one component plasma ( ocp ) model , where the motion of the ions is of interest , the electronic background is treated as a constant field @xcite . in the model pairwise interaction of the quarks
is assumed .
the possible realizations of ( single flavored ) two - quark ( qq ) system are composed of symmetric sextet and antisymmetric anti - triplet combinations . the bookkeeping of all possible qq pair - interactions in the system is realized using the so called `` interaction table '' of which an example is shown in table [ table:3 ] .
the particles are assigned a label and all particles have an associated color value .
the elements of the table show the @xmath0 values of the actual qq pair realizations .
the values of the @xmath1 entries are randomly re - distributed periodically ( see below ) .
the total system is color - balanced ( white ) , but neglecting the self - interaction results in the appearance of a net attraction . in the molecular dynamics simulation
the quarks interact with the color - dependent effective pair - potential given by ( [ eq : v ] ) .
periodic boundary conditions are applied together with the ewald summation technique for the proper treatment of the long range interaction .
forces acting on each of the particles ( quarks ) due to all other particles are calculated in every timestep and the equation of motion for each particle is integrated in time @xcite . at the start of the simulation the particles
are randomly distributed in the cubic simulation box with kinetic energies determined by an initial temperature .
an initialization phase is used before any measurement is taken on the system . in this phase the momenta of the particles are scaled back according to the given initial temperature in every timestep .
the simulation time of this phase is chosen to be long enough to reach equilibrium conditions for the unperturbed system over the subsequent measurement period .
the effect of the gluon field on the quark component is taken into account in three ways : ( i ) it represents an infinite heat reservoir , energy can be coupled out of it without any changes in the qluon field ; ( ii ) gluons can transfer color charge , the color state of two quarks may be exchanged ( color rotation ) with a characteristic time given by @xmath2 ; ( iii ) qluons perturb qq - pairs , leading to a new realizations of the pair - interaction ( redistribution of the interaction table ) with a characteristic time @xmath3 .
the input parameters in our simulations are : density @xmath4 quarks/@xmath5 , initial temperature @xmath6 mev , effective quark mass @xmath7 mev / c@xmath8 , interaction coupling strength @xmath9 , time between reassignment of all @xmath1 ( different - color pair elements ) in the interaction table @xmath3 , time between color rotation events @xmath2 and the ratio of pairs color rotated in one event @xmath10 .
the plasma frequency is derived as @xmath11 .
the behavior of the system is now governed by the two , independently chosen `` refreshment '' times @xmath3 and @xmath2 .
figure [ fig1 ] shows the time - evolution of the quark kinetic temperature for different @xmath3 values .
it can be seen , that in the limiting cases @xmath12 and @xmath13 the temperature fluctuates closely to its initial value .
intermediate values of @xmath3 result in increasing temperatures .
the @xmath3-dependence of the temperature is plotted in figure [ fig2](a ) for three snapshots in time .
a remarkable resonant - like behavior of the heating of the system can be observed , where the maximum heating - rate establishes at @xmath3 values around the inverse plasma frequency @xmath14 fm / c .
figure [ fig2](b ) shows the dependence of the resonant - like peak on an additional color rotation rate , where @xmath17 .
we have found that the color exchange increases the energy transfer from the background field to the quark component by 10 to 20% .
figure [ fig3](a ) displays the pair - correlation function ( pcf ) for the @xmath19 fm / c case neglecting color - rotation .
the total pcf is decomposed into contributions of equal - colored- , and different - colored pairs .
the attractive force acting between some of the different - colored pairs result in the appearance of the correlation peak at @xmath20 .
this indicates the onset of clusterization @xcite of particles which gets more pronounced at higher @xmath3 values . focusing only on the ( always ) repulsive equal - colored pairs ,
the pcf can be compared with classical ocp results [ see figure [ fig3](b ) ] .
this comparison shows that the repulsive equal - colored component of the qq plasma is structurally similar to classical ocp in the so called `` gas '' phase ( at @xmath21 values around unity ) , which indicates already strong non - ideality of the qq - plasma investigated here .
fm / c case ( a ) decomposed into equal - colored and different - colored contributions ; and ( b ) compared to classical ocp plasmas . here
fm is the wigner - seitz radius , and @xmath21 is the plasma ( non - ideality ) parameter . ]
m. gyulassy and l.mclerran , nucl .
a * 750 * , 30 ( 2005 ) .
p. lvai , u. heinz , phys .
c * 57 * , 1879 ( 1998 ) .
d. frenkel and b. smit , _ understanding molecular dynamics simulations _
( academic press , new york , 2001 ) . | the properties of a strongly interacting quark plasma are investigated by molecular dynamics method including non - abelian quark - quark potential .
our main goal is to study the thermalization process in this system .
we find an interesting resonance - like behaviour : at a characteristic time close to the inverse plasma frequency the quark plasma is heated up substantially via energy transfer from quark potential energy into one particle kinetic energy .
color rotation mechanism enhances the effectivity of this heating process , leading to a very fast thermalization with high temperature . |
detailed knowledge of dust and its effect on the total bolometric luminosity of galaxies is essential for an unbiased study of the evolution of galaxies . for example , dust is often associated with star formation activity and , hence , any measurement of this requires correction for dust extinction @xcite . furthermore ,
dust is directly correlated with the metallicity of galaxies and their chemical evolution ( e.g. , * ? ? ?
as dust is mainly produced by supernov @xcite or low / intermediate mass asymptotic giant branch ( agb ) stars @xcite , it also provides a clue towards studying stellar evolution in galaxies .
finally , the knowledge of the effect of dust at high redshifts is important in studies of early formation of galaxies , their star formation rates and mass assembly ( e.g. , * ? ? ? * and references therein ) .
dust affects the output energy of galaxies by absorbing and re - processing their uv - to - optical light ( produced by star forming regions or black - hole accretion ) , resulting in its re - emission mostly at far - infrared ( fir , @xmath10 m throughout ) wavelengths .
hence , the ir spectral regime ( and the millimeter spectral regime at higher redshifts ) has been most useful in unveiling the properties of dust in galaxies ( for a review , see * ? ? ? * ) . however , much of this work ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) has relied on shallow data or small number statistics when compared to optical / near - infrared - based studies , even though stacking analysis have been used to lessen this problem ( e.g. , * ? ? ?
while the fir / mm spectral range is sensitive to the cold dust component ( t@xmath11100k ) mostly found in the inter - stellar medium ( ism ) , the hot dust ( t@xmath125001500k ) component , mostly emitting at @xmath1218@xmath1 m , is nearer to the heating source , thus being the front - line medium to absorb direct uv / optical light and trace the most active regions in galaxies . here , we perform a statistical study of the contribution of hot - dust and 3.3@xmath1 m polycyclic aromatic hydrocarbon ( pah ) feature to the observed spectral energy distributions ( seds ) of different populations of galaxies and their evolution with redshift .
the results here can be used to quantify the amount of the uv / optical light obscuration in galaxies undergoing different degrees of star - formation or nuclear activity and its evolution with look - back time .
this is done by isolating the hot - dust component in galaxies and performing a statistical measurement of its luminosity density .
such a study requires wide - area , multi - waveband data .
particularly important , is the availability of deep or medium - deep near- to mid - infrared ( nir , mir ) data ( 18@xmath1 m ) tracing the hot - dust regime .
these requirements are achieved by the observations from the ir array camera ( irac , * ? ? ?
* ) on board the _ spitzer space telescope _
( _ spitzer _ , * ? ? ?
* ) covering the cosmic evolution survey ( cosmos , * ? ? ?
* ) . in section [ sec : samp ]
, the sample used in this study is described , and the redshift estimates are presented in section [ sec : reds ] . in section [ sec : apcorr ] we correct the aperture - to - total flux estimates as a function of redshift .
section [ sec : estimdust ] describes the method used to estimate dust and pah emission , using seds for individual galaxies .
the evolution of dust - to - total and pah - to - total luminosity ratios are studied in section [ sec : d2 t ] .
we present the dust and pah luminosity density functions ( ldfs ) , their dependence on galaxy types and their evolution with redshift in section [ sec : dlfs ] .
the dependency on galaxy type and redshift is also shown for the overall dust and pah luminosity densities in section [ sec : ldz ] .
finally , section [ sec : conc ] summarizes the conclusions . throughout this paper
we use the ab magnitude system , [ 3.6 ] , [ 4.5 ] , [ 5.8 ] , [ 8.0])@xmath13 = ( @xmath14 , [ 3.6 ] , [ 4.5 ] , [ 5.8 ] , [ 8.0])@xmath15 + ( 1.841 , 2.79 , 3.26 , 3.73 , 4.40 ) from @xcite and _ http://spider.ipac.caltech.edu / staff / gillian / cal.html_. ] . a @xmath16cdm cosmology is assumed with h@xmath17kms@xmath18mpc@xmath18 , @xmath19 , @xmath20 .
we use observations from the cosmos field , covering an area of 1.8deg@xmath0 , with available multi - waveband data . the ultraviolet - to - ir coverage ( 0.158@xmath1 m )
is unique and there are , in total , 30 broad , intermediate , and narrow - band filters available ( table 1 in * ? ? ?
* i09 henceforth ) . for the present study ,
we use both the @xmath21-band ( _ subaru _ telescope ) and 3.6@xmath1m - selected catalogs described in i09 .
our sample selection is based on flux cuts applied to the appropriate irac band tracing the rest - frame 3.3@xmath1 m wavelength at a given redshift interval .
we divide the galaxies in this study into early - type , late - type , starburst and active galactic nuclei ( agn ) , based on their seds .
the non - agn classification results from fitting galaxy sed templates to the observed galaxy seds as briefly described in section [ sec : reds ] ( see the detailed description in i09 ) . for each galaxy ,
the best fit template reveals its spectral type .
the agn in our sample are color - color selected with the following criteria : @xmath22>0 $ ] and @xmath23-[8.0]>0 $ ] ( ki , * ? ? ?
whenever the @xmath14-band is not available , sources are considered to be agn when satisfying the @xcite criterion or when brighter than a rest - frame @xmath14 absolute magnitude of m@xmath24 ( for the 211 sources in our sample with unavailable @xmath14-band photometry , 55 are ir agn ) .
no other spectral regime ( x - rays , optical , or radio ) is adopted to identify agn hosts .
the goal of this study is to independently trace ir emission due to hot - dust , induced by different radiation mechanisms .
however , x - rays / optical / radio - selected agn may reveal little of the hot - dust emission induced by the agn itself , with the ir emission being potentially dominated by stellar light or dust emission from the host galaxy .
the ir - selection ensures that the hot - dust emission observed in agn is mostly induced by nuclear activity .
this study focuses on two adjacent spectral regimes at rest - frame wavelengths 3.3@xmath1 m and 4.2@xmath1 m .
these are chosen so that hot - dust emission continuum is traced in both regimes ( corresponding to black - body temperatures of @xmath25k and @xmath4k , respectively ) , as well as the known pah feature at 3.3@xmath1 m . by comparing the two adjacent spectral regimes
, we aim to infer differences between hot - dust and pah emissions and assess their evolution with cosmic - time .
although pahs are not actual dust particles , they are large molecules of carbon rings ( composed of @xmath1250 carbon atoms ) and hydrogen , which act as small dust grains significantly blocking uv radiation , producing broad emission features in the ir seds of galaxies ( for a review , see * ? ? ?
also , pahs comprise a non - negligible fraction of the carbon existing in the universe ( 510% ) and are believed to be closely related to star - formation activity ( * ? ? ?
* and references therein ) .
the 4.2@xmath1 m spectral regime is free from pah emission and traces only ir continuum due to dust heated by energetic radiation fields .
obscured star - formation as well as agn activity can account for such emission ( * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) .
the final samples are first selected based on redshift criteria ( table [ tab : zrange ] ) allowing the same rest - frame wavelength to be traced by each of the irac filters .
the redshift intervals considered in this study are set by the target 3.3@xmath1 m rest - frame wavelength and the widths of the irac filters .
table [ tab : zrange ] shows the adopted redshift intervals for this study , resulting from the specific redshifts where the 3.3@xmath1 m wavelength enters or leaves the 50% throughput limits of an irac filter .
the rest - frame 4.2@xmath1 m wavelength is then traced by the contiguous irac filter at longer - wavelengths relative to the band tracing rest - frame 3.3@xmath1 m . as a result
, the work is limited to the @xmath26 range , where rest - frame 4.2@xmath1 m is still possible to observe with the irac filter - set .
the redshift estimates for individual galaxies are described in section [ sec : reds ] .
crrrrr 3.6 & 0.05 & 0.19 & 0.84 & 1.96.0 & 4.013.0 + 4.5 & 0.21 & 0.52 & 1.25 & 6.511.8 & 14.125.5 + 5.8 & 0.52 & 0.94 & 4.38 & 11.815.0 & 25.532.3 we apply two completeness magnitude cuts to the resulting redshift selected samples .
the first ( upper panels in figure [ fig : magcut ] ) considers apparent magnitudes and is set as the brightest between the magnitude value at which the magnitude distribution starts to drop and the magnitude beyond which the fraction of sources with a redshift estimate ( redshift completeness ) is lower than 0.7 ( section [ sec : reds ] ) .
the second completeness magnitude cut ( lower panels in figure [ fig : magcut ] ) considers rest - frame 1.6@xmath1 m absolute magnitudes ( assumed to be dominated by stellar light alone ) and takes into account the fact that it is not always possible to simultaneously trace the rest - frame 3.3 and 4.2@xmath1 m for one galaxy .
this is driven by the fact that rest - frame 4.2@xmath1 m is always traced by shallower data .
this magnitude cut is set as the brightest between @xmath27 value beyond which the fraction of sources with a hot - dust and pah estimate drops below 0.7 and the value at which the magnitude distribution starts to drop .
hence , we consider all @xmath28 sources with @xmath29<23.56 $ ] and @xmath30 , all @xmath31 sources with @xmath23<23.75 $ ] and @xmath32 , and all @xmath33 sources with @xmath34<22.25 $ ] and @xmath35 .
no source flagged photometrically ( e.g. , due to blending or bad pixels ) was considered in this study , and no incompleteness correction regarding this rejection is attempted . the final number counts for different populations are listed in table [ tab : numb ] .
crrrrr @xmath28 & 2954 & 514(17 ) & 341(12 ) & 2081(70 ) & 18(1 ) + @xmath31 & 11439 & 2383(21 ) & 3418(30 ) & 5545(48 ) & 93(1 ) + @xmath33 & 14435 & 3198(22 ) & 4470(31 ) & 6340(44 ) & 427(3 )
whenever available , a spectroscopic redshift is assigned to galaxies . however , the bulk of the redshift estimates are photometric , based on the broad - to - narrow - band photometry coverage available . briefly , the photometric redshifts are measured using a sed fitting procedure using the _ le phare _ codearnouts / lephare/ + cfht_lephare / lephare.html ] ( s. arnouts & o. ilbert ) . the procedure applied on the cosmos photometry data - set
is described in i09 . in the worst case scenario for our sample ( at @xmath36 and @xmath37 ) ,
the photometric redshift accuracy is expected to be @xmath38 ( where @xmath39 ) and the percentage of outliers ( sources with @xmath40 , denoted by @xmath41 ) is @xmath42 ( i09 ) .
* ; * ? ? ?
* s11 henceforth ) computed photometric redshifts for xmm-_newton _ and _ chandra_-detected sources .
variability effects in x - ray agn hosts are addressed before computing photometric redshifts , and sed templates accounting for agn emission contribution are also adopted .
this method provides photometric redshifts for the x - ray agn sample with an excellent quality , reaching an accuracy of @xmath43 with an outlier fraction of @xmath44 .
one conclusion from this study was that below a given soft x - ray flux ( @xmath45 ) , non - variable and/or morphologically extended x - ray sources do not require sed templates taking into account agn emission .
however , the recipe adopted in s11 should not be applied to x - ray - undetected sources , as intrinsically x - ray - luminous agn may remain undetected due to high obscuration .
appendix [ sec : zphagn ] presents a detailed discussion of the photometric redshift quality for the ir - selected agn sample ( described in section [ sec : gal ] ) .
some sources in the sample will have more than one redshift estimate resulting from different methods .
the final redshift priority sequence ( from highest to lowest ) is as follows : * spectroscopic redshifts with a probability larger than @xmath46 of being correct from the @xmath47cosmos catalog , which considers only @xmath48 sources @xcite ; * spectroscopic redshifts from @xcite , @xcite , @xcite , @xcite , and @xcite , which provide additional spectroscopy for fainter sources ; * photometric redshifts from s11 for xmm-_newton _ and _ chandra _ detected sources ; * photometric redshifts from i09 for non - ir - agn sources ; * combination of photometric redshift solutions using non - agn or agn templates for ir - agn sources as described in the appendix [ sec : zphagn ] . in case
the only redshift measurement is photometric , it is only adopted if the source has @xmath49 ( as suggested by i09 ) and a good quality ( @xmath50 , with unflagged optical wavebands ) .
the incompleteness due to these quality constraints is accounted for while computing source densities ( section [ sec : dlfs ] ) .
this correction is given in figure [ fig : zincomp ] for each of the irac bands used for sample selection ( section [ sec : sampsel ] ) .
the figure shows the variation of the fraction of sources with @xmath50 and @xmath49 depending on observed magnitude in each of the irac - channels .
we do not consider sources fainter than the magnitude beyond which the redshift completeness is smaller than 0.7 ( figure [ fig : magcut ] ) .
the redshift distribution is shown in figure [ fig : zdist ] , highlighting the fraction of the sources with available @xmath51 , with @xmath52 , with @xmath49 , and the total population .
we note that by including only sources with @xmath49 in the study , highly obscured sources may be missed .
nevertheless , figure [ fig : kc24 ] shows that the @xmath53 population is composed mainly of starburst galaxies at @xmath54 ( with the bulk at @xmath55 ) or agn at any redshift . while this work does not address the @xmath54 regime ( table [ tab : zrange ] ) , extremely obscured agn at @xmath26 with @xmath56 will be missed . at this stage
, one can only estimate an upper limit of the @xmath56 agn by considering all sources inside the ki region to be agn ( they can also be @xmath57 starbursts ) .
this yields a 32% rejection of the ki population when applying the @xmath49 cut . as the redshift of the faint ki - selected sources is unknown ,
we add in quadrature a 0.32 factor to the error budget while estimating agn source densities ( see below ) .
this is a conservative value given not all ki sources with @xmath56 are agn at @xmath26 .
when computing measurement - associated errors , we consider in quadrature poisson and @xmath58-induced errors ( @xmath59 and @xmath60 , respectively ) , and a factor ( @xmath61 ) to account for photometric redshift outliers ( we adopt a conservative value of @xmath62 , i09 and appendix [ sec : zphagn ] ) . in the case of the agn sample ,
an additional factor ( @xmath63 ) , accounting for the agn with @xmath56 , ought to be considered .
for instance , the error associated to source density for the agn sample is : @xmath64 , where @xmath65 represents the source density .
we note that the reference irac photometry used to select our sample was measured using an aperture of 1@xmath669 and then scaled to total flux using a correction factor assuming a point source .
however , such assumption is not always verified especially at low redshifts . at such distances ,
the physical size traced by a 1@xmath669 aperture is small ( @xmath67kpc , table [ tab : zrange ] ) , thus probing only the central region of a galaxy .
hence , we apply a statistical correction to the four irac bands depending on redshift and spectral - type ( only for non - agn populations ) by comparing with the 4@xmath661 aperture ( ap4 ) photometry .
figure [ fig : photcorr ] shows the magnitude difference between the two apertures versus the 1@xmath669 aperture ( ap2 ) magnitude .
each panel refers to different redshift bins where the rest - frame 3.3@xmath1 m is traced by the irac bands .
note , however , that the corrections are applied to all irac bands at each redshift bin and are different between bands .
the discordance between the two apertures results from three factors : ( _ i _ ) flux being missed by the ap2 photometry ( shifting the whole distribution up ) ; ( _ ii _ ) contamination by neighboring sources ( producing the scatter to positive y - axis values ) ; ( _ iii _ ) and bright nearby sources affecting the background measurement , which produces a flux underestimate at large apertures ( producing the scatter to negative y - axis values ) .
the correction aims at compensating for missing flux ( _ i _ ) , while being independent of artifacts ( _ ii _ and _ iii _ ) .
only sources with @xmath68ab are considered , so that the correction value is unaffected by the large scatter at fainter magnitudes .
it is clear that the lower the redshift , the larger is the correction .
we finally refer that the agn sample is not corrected for this effect , as the emission from agn sources is assumed to come from the nuclear region , thus not being affected by aperture photometry .
star - formation in the host - galaxy could in principle still produce such bias .
however , we do not observe strong evidence for such a scenario in our ir - selected agn sample ( see section [ sec : d2 t ] ) .
in section [ sec : rfwl ] we have mentioned the two rest - frame bands where hot - dust and pah emission is expected : 3.3 and 4.2@xmath1 m . however , at these wavelengths , one also needs to take into account stellar emission .
figure [ fig : cont33 ] shows the estimated 1.6@xmath1 m luminosities versus the observed hence including both stellar plus non - stellar emission 3.3@xmath1 m ( left hand - side plot ) and 4.2@xmath1 m ( right hand - side plot ) luminosities for each galaxy population considered in this study .
the observe fluxes include both stellar and non - stellar emission .
the regions between the dotted lines represent the locus where the seds dominated by stellar emission alone are expected to fall ( see below ) .
the elliptical data `` cloud '' tends to fall in this `` stellar region '' .
any deviation from this region , clear in starburst and agn populations , is assigned to non - stellar emission ( hot - dust and pah ) .
this is the emission excess we aim to extract . in order to disentangle stellar from non - stellar emission
, we first estimate the stellar emission at 3.3 and 4.2@xmath1 m for each galaxy .
this is done in two steps .
first we consider a reference wavelength where the observed flux is expected to come entirely from stellar emission . using this , we then estimate the stellar emission at 3.3 and 4.2@xmath1 m . the remaining flux is thus assumed to be due to contribution from non - stellar emission alone . the reference wavelength used to estimate
the stellar emission is the stellar bump at rest - frame 1.6@xmath1 m ( @xmath69-band ) .
this emission bump is frequently observed in galaxy seds ( figure [ fig : ellmod ] ) , being indicative of pure stellar contribution , as at @xmath70 m no significant dust emission is expected .
this is because uv / optical ( stellar or agn ) emission is not enough to heat dust at the level for its cooling emission to dominate at such short wavelengths , or because any dust particle or pah molecule in a strong radiation field is dissociated .
hence , at wavelengths below @xmath71 m , only emission from the wien tail of the spectrum due to the hottest dust grains ( e.g. , around thermally pulsating agb , tp - agb , stars or in a dust torus around an agn ) and from scattered agn light is expected .
such emission can still be significant enough at 1.6@xmath1 m to induce a systematic overestimate of the stellar emission , and consequently underestimate the non - stellar contribution . at this stage
, we avoid quantifying such bias either for individual galaxies or for the statistically large sample used here .
the flux at rest - frame 1.6@xmath1 m is obtained through interpolation between the two wavebands which straddle this rest - frame wavelength at the source s redshift .
however , interpolation is likely to underestimate the true rest - frame 1.6@xmath1 m flux value depending on the source redshift and sed shape .
this is evident from figure [ fig : interpol ] where discrepancies between estimated and true value ( always below the 20% level ) are shown for typical early ( red ) and late ( green ) galaxies , blue starbursts ( blue ) , and agn hosts ( magenta ) .
these trends were used to correct the interpolated 1.6@xmath1 m flux for each galaxy at its respective redshift . due to the absence of the @xmath69-band from the i09 catalogue , at @xmath72 the ratio in figure [ fig : interpol ] is @xmath73 , because @xmath74 and @xmath14 passbands were used to interpolate the 1.6@xmath1 m flux .
the final corrected values are the ones shown in figure [ fig : cont33 ] . with the estimated stellar flux at 1.6@xmath1 m , the corresponding stellar contribution at 3.3 and 4.2@xmath1 m is obtained .
this is done with a pure stellar model ( solid line in figure [ fig : ellmod ] , corresponding to a 2gyr old elliptical from @xcite ) , the same used in figure [ fig : cont33 ] to compute the stellar - dominated locus .
the conversion from 1.6@xmath1 m stellar flux to that at 3.3 and 4.2@xmath1 m is slightly redshift dependent , because the considered filters will probe slightly different rest - frame wavelength ranges .
hence , using the pure stellar model ( figure [ fig : ellmod ] ) , a conversion table was generated by convolving that stellar model with the nir filters ( from @xmath74-band to 8@xmath1 m ) at redshift steps of @xmath75 .
we consider the underlying shape of the galaxy sed at 1.64.2@xmath1 m , due to stellar emission alone , to be common to all galaxy populations referred in this study .
this is a fair assumption for a universal initial mass function .
the stellar emission in this spectral regime originates in cold stars , which live longer , thus producing a constant sed shape over a wide range of ages .
such assumption may be affected by strong differential obscuration affecting the rest - frames 1.6@xmath1 m and 3.3 or 4.2@xmath1 m . however
, this will only occur in extremely obscured systems ( e.g. , * ? ? ?
* ) , which are rare .
figure [ fig : ellmod ] supports this assumption , showing how similar the 1.64.2@xmath1 m stellar sed is between a 13gyr old elliptical and a 0.05gyr old starburst ( models from i09 ) .
the next step is to separate the hot - dust emission from the pah emission at 3.3@xmath1 m . in order to do this ,
we use the results from @xcite who find that a gray - body model of temperature 850k and with a dust emissivity index of 1 is optimal to fit galaxies continuum emission in the 35@xmath1 m spectral range . assuming this continuum shape
, we normalize it to the 4.2@xmath1 m dust emission flux to estimate the flux at 3.3@xmath1 m . removing this from the total non - stellar 3.3@xmath1 m emission , one estimates the pah 3.3@xmath1 m emission flux .
this procedure is only applied to non - agn sources , given that @xcite only considered non - agn sources . in the case of agn ,
a dominant power - law continuum shape is more likely .
hence , we estimate the 3.3@xmath1 m continuum emission in agn via interpolation of the two bands straddling the band tracing rest - frame 3.3@xmath1 m , assuming @xmath76 .
henceforth , whenever `` pah emission '' is mentioned , we refer to the 3.3@xmath1 m pah emission flux estimated by the above procedure , while whenever `` hot - dust emission '' is mentioned , we refer to the estimated 4.2@xmath1 m non - stellar emission flux assigned to hot - dust emission alone .
figure [ fig : d2 t ] presents the evolution of the average dust - to - total ( left hand - side ) and pah - to - total ( right hand - side ) luminosity ratios .
it is clear that the average contribution of hot - dust at 4.2@xmath1 m ir - selected agn remains constant with redshift around the 80% level .
this constancy results from using ir agn selection criteria alone .
the goal here is to independently trace different dust - heating mechanisms ( agn activity versus star - formation ) .
note also the extremely low pah contribution in agn , consistent with pah - annihilation by the agn radiation or emission - dilution in the bright agn - dominated sed .
the increase at the lowest redshift interval for the agn population may be explained by the aperture photometry adopted to estimate colors .
with decreasing redshift , a fixed aperture will probe more nuclear regions , enabling the detection of lower luminosity agn .
being less powerful , these agn allow for the existence of pah molecules as opposed to more luminous agn , the dominant population with increasing redshifts , which shows average pah - to - total luminosity ratios close to zero .
while the dust - to - total luminosity ratio for early - type galaxies is consistent within @xmath77 to be constant ( @xmath78 ) with redshift , it increases with increasing redshift for late - type and starburst galaxies .
the starburst and early - type populations present a constant pah - to - total luminosity ratio ( within the errors ) with redshift , while the late - type population shows a decrease at the lowest redshift interval .
the pah - to - dust luminosity ratio seems to vary slightly .
for instance , the ratio for the starburst sample is @xmath79 at @xmath28 , @xmath80 at @xmath31 , and @xmath81 @xmath33 . in the literature ,
the 3.3-pah feature strength is compared to colder - dust ( emitting at 81000@xmath1 m ) instead , and there is no evidence for evolution : @xcite estimate for a sample at @xmath82 an average 3.3-pah to 81000@xmath1 m luminosity ratio of @xmath83 , while @xcite find the ratio to be @xmath84 in a @xmath85 source ( see also * ? ? ?
* ; * ? ? ?
* ) , being both values in agreement with those found for lower - redshift starburst galaxies @xcite
. the fact that we find an increase of the average dust - to - total luminosity ratio with redshift does not necessarily mean the sample is progressively incomplete at higher redshifts ( i.e. , missing pure stellar emission sources , @xmath86 ) . figure [ fig : d2tlum ] ( bottom panel in left hand - side plot ) shows that higher redshift galaxies are indeed dustier in the luminosity range where all samples are complete .
however , below the luminosity completeness cuts ( section [ sec : sampsel ] ) , one can not assess the trends for the high-@xmath47 intervals .
the decrease in the number of galaxies showing no dust emission ( @xmath86 ) with decreasing 1.6@xmath1 m luminosity is likely to be real ( top panel in left hand - side plot ) , as the slope of the relation is constant up to the brightest luminosities .
in this section we measure the hot dust and pah luminosity density functions ( ldfs ) and their evolution with redshift for different populations of galaxies .
figure [ fig : lf33d ] compares the populations at different redshift intervals ( shown in individual panels ) .
the populations are color - coded : total ( black ) , early - type ( red ) , late - type ( green ) , starburst ( blue ) , and agn hosts ( magenta )
. figure [ fig : lf33dz ] shows the evolution with redshift of the ldfs for each population ( shown in individual panels ) . the volume associated with each galaxy
is based on the flux limit of the sample and the @xmath87-correction , derived from the galaxy s own sed ( as given by the observed multi - wavelength photometry ) , and obtained through the @xmath88 method @xcite .
the dust ldfs enable us to evaluate how much dust is contributing to the ir radiation at any given luminosity and redshift for each galaxy population .
note that 1.6@xmath1 m luminosities can be taken as a proxy to stellar mass assuming 1.6@xmath1 m to be dominated by stellar emission .
therefore , one could assume here , as a proxy , the dust contribution as a function of galaxy stellar mass . also , using a common luminosity discriminator , allows for a direct comparison between the dust and pah ldfs .
although agn hosts are rare ( @xmath89 in our sample ) , this population clearly dominates the bright - end of the dust ldfs at least at @xmath90 .
the opposite is seen in the pah ldfs , where the agn population is the weakest contributor at all redshifts .
figure [ fig : lf33dz ] shows evidence for a number - density evolution in dust ldfs for agn . if one assumes the dust - to - total luminosity ratio to be constant in ir - selected agn ( figure [ fig : d2 t ] ) , then one would not expect to see any evolution with redshift at fixed luminosities in the dust luminosity density ( @xmath91 ) assuming a constant number - density . however , this seems to be the case , especially from @xmath33 to @xmath31
. further confirmation of this result may be accomplished by extending this study to deep irac coverages available , e.g. , in the goods fields .
as shown in figure [ fig : lf33dz ] , the major evolution in @xmath91 and pah luminosity density ( @xmath92 ) for the non - agn populations mainly appears at the highest luminosities .
that is , the most massive galaxies present the strongest evolution with redshift in both @xmath91 and @xmath92 . regarding the pah ldfs in figure [ fig : lf33d ] ,
the faint - end is always dominated by the starburst population , followed by the late - type , and then the early - type galaxies . at high luminosities ,
the three non - agn populations seem to contribute at comparable rates .
this , however , does not happen in the dust ldfs , where the early - type population is the weakest contributor , while late - type and starburst populations have the highest contribution .
this is explained in figure [ fig : d2 t ] , which shows that the higher the redshift , the larger is the difference in dust - to - total luminosity ratio between the early - type galaxies and late - type or starburst galaxies , while this is not the case for the pah - to - total luminosity ratios .
the reason why this happens is not straightforward , but assuming the gray - body model scaling should instead more - or - less correlate to 3.3-pah strength , one can conclude the following scenarios : ( i ) extra flux in early - type galaxies at 3.3@xmath1 m unaccounted by pah emission ; ( ii ) or extra emission at 4.2@xmath1 m in starburst and late - type galaxies .
the former could imply higher gray - body temperatures ( @xmath93850k ) in early - types , for instance , induced by low - luminosity agn , which were missed by the adopted agn selection , while ( ii ) could be explained by @xcite , who show evidence for extra emission in the 25@xmath1 m spectral range unaccounted by the same gray - body model we assume plus pah emission .
they also find that this extra emission correlates with star - formation @xcite , thus explaining why we observe the enhancement in starburst and late - type galaxies and not in early - types .
this reasoning also implies that the extra - emission heating mechanisms are in part different from those of the 3.3-pah feature .
one source of extra - emission at @xmath94 could be the br@xmath95 emission line ( 4.05@xmath1 m ) .
if present , it would imply an underestimate of the 3.3-pah flux .
however , the star - forming sample gathered by @xcite shows that the luminosity ratio between br@xmath95 and 3.3-pah feature is @xmath96 , which is not enough to account for the difference pointed out in figure [ fig : lf33d ] .
in figure [ fig : zlum ] , we present the evolution of @xmath91 and @xmath92 since @xmath971 ( left and right hand - side , respectively ) .
we also compare @xmath91 to the cold - dust luminosity density estimated from fir observations ( 70500@xmath1 m , dark gray shaded region , * ? ? ?
* ; * ? ? ?
* ) , linearly translated to obscured star - formation . however , comparing the luminosity density at present time to that at @xmath33 , the drop in @xmath91 is more significant than that of the cold - dust luminosity density , by @xmath98dex . such decrease is driven mostly by the non - starburst populations .
while one may attribute part of the decay to the strong evolution shown by agn hosts , whose fir sed is related to star - formation and not to agn activity , the late - type population also plays a role shaping such difference between hot and cold @xmath91 evolution .
however , even the starburst population shows a different rate of evolution compared to that of the far - ir evolution : the drop in @xmath91 is still greater by @xmath99dex compared to that of cold - dust .
we recall that the faint population , gradually missed with increasing redshifts by our completeness cuts , is likely a significant contributor to the overall @xmath91 and @xmath92 given the higher hot - dust and pah contribution observed in their seds ( figure [ fig : d2tlum ] ) .
this implies that the discrepancy between the hot and cold dust regimes is likely larger .
also , figure [ fig : lf33dz ] shows a slower evolution in @xmath91 and @xmath92 at fainter luminosities . accepting this as true
, one can then consider the faint - end shape of the ldfs ( i.e. , that below the completeness cut ) is unchanged from a redshift interval to a subsequent lower one and estimate a completeness correction .
this is shown in figure [ fig : comp ] for the total and starburst populations . integrating a given ldf at luminosities fainter than the completeness cut will thus provide an estimate of the extra @xmath65 missed by our sample . applying this correction yields extra luminosity density represented as the empty regions shown in figure [ fig : zlum ] ( only for the total and non - agn populations ) .
these corrections are expected to provide upper - limits for the @xmath65 values .
recent work on the 3.3@xmath1 m pah feature with _
spitzer _ and _ akari _ out to @xmath100 does not show an evolution with redshift of the luminosity ratio @xmath101 @xcite .
however , comparing with hot - dust , our study seems to support slight variation of the hot - dust - to - pah luminosity ratio with redshift .
@xcite find evidence for the stellar nurseries ( birth clouds ) to be producing a higher fraction of obscuration than the diffuse inter - stellar medium at high - redshift .
figure [ fig : zlum ] corroborates such finding when comparing our results with the cold - dust luminosity density evolution .
this explains why high - redshift star - forming galaxies are more obscured @xcite , as also pointed out by @xcite . some interpretations for this discrepancy between hot and cold - dust evolution
may then be given : either the reduced star - formation at low redshifts is unable to heat enough quantities of dust for it to dominate at 3.34.2@xmath1 m , thus becoming diluted in the much brighter stellar flux ; or dust may be , with decreasing redshift , gradually farther from uv / optical sources ( e.g. , blown by strong radiation fields from an active nucleus or stellar winds ) , thus decreasing the relative contribution of the hottest spectral dust - components ; or , in the case of the agn population , different evolution of the heating mechanisms at play agn activity dominating the hot - dust regime versus star - formation dominating the cold - dust regime in agn hosts .
table [ tab : fracld ] shows , for each redshift interval , the contributions of each of the galaxy populations to the overall dust luminosity density at rest - frames 3.3 and 4.2@xmath1 m .
these values refer to contributions down to the adopted luminosity completeness cut .
ccrrrr dust & @xmath28 & 2 & 23 & 72 & 3 + & @xmath31 & 6 & 28 & 47 & 19 + & @xmath33 & 3 & 19 & 45 & 33 + + pah & @xmath28 & 6 & 18 & 76 & 0 + & @xmath31 & 12 & 35 & 53 & 0 + & @xmath33 & 15 & 28 & 56 & 1
in this work , the properties of hot dust ( @xmath12690k ) emission was explored using observations tracing rest - frame 4.2@xmath1 m .
this wavelength is also used to infer the dust contribution at rest - frame 3.3@xmath1 m , and help assessing 3.3-pah strength .
our approach considers stellar , dust , and pah emissions separately , as well as the separation of the ir galaxy population into early , late , starburst , and ir - selected agn .
this allows the evaluation of the ir luminosity functions depending on galaxy - type and distance , as well as to estimate how much dust is contributing to the ir emission .
we conclude : * evolution with redshift of the hot - dust luminosity densities resembles that of cold - dust , but it drops more steeply ( at least 0.5dex more ) with decreasing redshift .
the discrepancy is larger if one attempts to correct for sample incompleteness .
the reason for the discrepancy is probably a combination of : with decreasing redshift , the star - formation becomes gradually unable to heat dust to such high temperatures for it to significantly contribute to the galaxy sed at such short wavelengths ; dust is gradually located at increasingly larger distances from the heating sources ( stars or active nuclei ) ; and/or distinct evolution of different heating mechanisms at play in agn hosts .
this trend gives support to the scenario where galaxy obscuration increases with redshift , as already proposed in the literature ( section [ sec : ldz ] ) .
* while the average dust - to - total luminosity ratio is observed to increase with redshift , the average pah - to - total luminosity ratio is consistent with being constant .
the hot - dust - to - pah luminosity ratio also depends on galaxy spectral type .
a plausible explanation may be related to the findings of @xcite , who observe increased extra emission at 25@xmath1 m with increasing star - formation rates ( section [ sec : dlfs ] ) .
* hot - dust and pah ldfs of the non - agn populations show that the hot - dust and pah emission in the most luminous galaxies at 1.6@xmath1 m ( i.e. , the most massive ) decreases faster ( with decreasing redshift ) than in less luminous galaxies ( section [ sec : dlfs ] ) . *
down to our luminosity limit , the agn population comprises only @xmath6 of the total galaxy population at @xmath102 , but its contribution to the overall hot - dust luminosity density increases from 3% at @xmath103 to 33% at @xmath104 , and it clearly dominates the bright - end of the total hot - dust ldfs at @xmath104 ( section [ sec : dlfs ] ) .
the main driver for this is the high dust - to - total luminosity ratio at rest - frame 4.2@xmath1 m found in ir - selected agn : @xmath120.8 ( section [ sec : d2 t ] ) .
* at @xmath9 , there is an increase of the dust - to - total and pah - to - total luminosity ratios with decreasing luminosity , but deeper data may be required to confirm this result ( section [ sec : d2 t ] ) .
the authors thank the cosmos team for providing the photometric and redshift catalogs which make the base of this work .
hm acknowledges the support from fundao para a cincia e a tecnologia through the scholarship sfrh / bd/31338/2006 and conicyt - alma by the post - doc scholarship under the project 31100008 .
hm and ja acknowledge support from fundao para a cincia e a tecnologia through the projects ptdc / cte - ast/105287/2008 , pest - oe / fis / ui2751/2011 and ptdc / fis - ast/2194/2012 .
hm acknowledges the support by ucr while visiting dr .
bahram mobasher as a visitor scholar .
hm acknowledges the frequent use of c , topcat , supermongo , and python .
the photometric redshift ( @xmath58 ) analysis adopted in this work ( and done in i09 ) is based on 0.158@xmath1 m photometry .
however , agn activity may induce emission in this spectral regime , thus affecting the analysis ( * ? ? ?
* s11 ) . here
, we test the @xmath58 quality obtained for sources in our ir - selected agn sample . for the purpose of quality check we adopt the normalised median absolute deviation ( nmad , by definition @xmath105 , * ? ? ?
* ) , which , for a gaussian distribution , is directly comparable to the @xmath106 value quoted in other papers .
also , the fraction of outliers ( sources with @xmath40 ) is denoted by @xmath41 .
the discussion below only considers sources with @xmath107ab . among the ir agn sample , there are 605 sources with good quality spectra . comparing the estimated spectroscopic redshifts with the photometric redshifts estimated by i09 ( @xmath108 ) , the quality achieved is @xmath109 , with @xmath110 .
@xcite also computed in parallel the best - fit solution using agn - like templates ( those considered by s11 ) .
if one adopts these estimates instead ( @xmath111 ) , the quality is improved to @xmath112 and @xmath113 .
this is similar to what is achieved when applying the @xcite method to _ chandra_-detected sources , even though plagued by a higher fraction of outliers .
one can further improve these results by adopting @xmath108 or @xmath111 solutions depending on which presents the lowest @xmath114 value .
this procedure assigns the @xmath111 solution to 66% of the ki - selected sample with spectra , and reduces both the @xmath115 and @xmath41 values to @xmath116 and @xmath117 .
figure [ fig : zagni09 ] compares directly @xmath108 and @xmath111 solutions for ir - selected agn sources .
one can identify three regions in the plots where the two model sets provide quite distinct solutions : ( a ) sources with @xmath118 and @xmath119 ; ( b ) sources with @xmath120 and @xmath121 ; ( c ) sources with @xmath122 and @xmath123 . in these regions , both agn and i09 solutions present on average @xmath124 ( meaning both provide a good fit to the observed data ) , hence the @xmath114 check referred above is irregular in these specific samples . in region
a there are 18 sources with spectroscopy , all but three fall at @xmath125 ( only one at @xmath126 ) .
the x - ray detection rate ( 58 sources among 1724 ) is rather low for a @xmath127 agn population , but acceptable for a higher redshift population including both agn and non - agn sources .
not only the xmm-_newton _ and _ chandra _ coverages gradually miss more agn sources with increasing redshift , but the reader should also be reminded that ki agn - selection is quite contaminated by non - agn sources at @xmath57 , where 63% of the a sample likely falls ( considering the photometric redshifts from i09 ) .
in fact , considering s11 redshift estimates ( @xmath128 ) for the 58 x - ray detected sources , only two of them fall at @xmath7 . these evidences support the scenario where sample a is mostly composed by @xmath129 sources , and we thus assign the photometric redshifts from i09 to sources in sample a. we note however , that 37 of these sources ( 64% ) do require agn - like templates ( s11 ) . the difference between @xmath128 and @xmath111 is likely not just due to variability ( about 30% of those sources are non - variable , i.e. , variability below 0.25 ; s11 ) , but rather a result from not considering luminosity priors ( as used in s11 ) when computing @xmath111 .
in sample b ( selected as sources with @xmath130 , @xmath131 , and @xmath132 ) we do not find such a clear trend , with both spectroscopic measurements and @xmath128 estimates showing equal evidence for a sample composed by sources at @xmath133 and @xmath134 .
the upper histogram in figure [ fig : zagni09 ] shows the redshift uncertainty distribution by comparing @xmath108 and @xmath111 to @xmath135 for sample b ( where @xmath131 ) .
the peak of the @xmath111 distribution at zero is dominated by sources which already show @xmath136 .
hence , we do not modify the assigned photometric redshifts resulting from the comparison between @xmath114 values .
sources found in region c ( selected as @xmath130 , @xmath137 , and @xmath138 ) mostly support the @xmath111 solution .
the peak in the uncertainty distribution in the lower histogram in figure [ fig : zagni09 ] is dominated by sources with @xmath139 , while sources with a @xmath111 uncertainty of @xmath12 - 0.4 mostly have @xmath140 and the @xmath108 uncertainty is around zero .
the remainder sources with large @xmath111 uncertainties have @xmath128 estimates and most show variability .
hence , for the ir - selected agn sample , we adopt the @xmath111 solution for sources with @xmath130 and : @xmath137 and @xmath139 ; or @xmath141 and @xmath142 .
the remainder of the population is assigned the @xmath108 solution .
the photometric redshift quality achieved for the ir - selected agn sample when compared to spectroscopic estimates is @xmath143 and @xmath144 ( figure [ fig : zspph ] ) .
the reader should recall that these @xmath58 estimates become obsolete when a higher priority redshift estimate is available ( such as spectroscopic or s11 @xmath58 estimates , see section [ sec : reds ] ) .
for instance , by assuming the s11 @xmath58 estimates , which take into account luminosity - priors and the effect of variability , the photometric redshift quality is improved to @xmath145 and @xmath146 .
+ aird , j. , et al .
2010 , , 401 , 2531 assef , r. j. , et al . 2011 , , 728 , 56 barlow , m , et al . 2010 , , 518 , 138 blain , a. w. ; smail , ian ; ivison , r. j. & kneib , j .- p .
, , 302 , 632 bouwens , r. j. , et al .
2009 , , 705 , 936 brusa , m. , et al .
2010 , , 716 , 348 bruzual , g. & charlot , s. 2003 , , 344 , 1000 buat , v. , et al 2005 , , 619 , 51 calzetti , d. , kennicutt , r. c. , engelbracht , c. w. , et al .
2007 , , 666 , 870 chary , r .-
r . , & pope , a. 2010 , arxiv:1003.1731 choi , p. i. , yan , l. , i m , m. , et al .
2006 , , 637 , 227 clements , d. l. , rigby , e. , maddox , s. , et al .
2010 , , 518 , l8 croom , s. m. , smith , r. j. , boyle , b. j. , shanks , t. , miller , l. , outram , p. j. , & loaring , n. s. 2004 , , 349 , 1397 da cunha , e. , charlot , s. , & elbaz , d. 2008 , , 388 , 1595 dasyra , k. m. , yan , l. , helou , g. , et al .
2009 , , 701 , 1123 donley , j. l. , koekemoer , a. m. , brusa , m. , et al .
2012 , , 748 , 142 drory , n. , et al .
2009 , , 707 , 1595 dunne , l. , gomez , h. l. , da cunha , e. , et al .
2011 , , 417 , 1510 fazio , g. g. , hora , j. l. , allen , l. e. , et al .
2004 , , 154 , 10 ferrarotti , a. s. & gail , h .- p . , 2006 , , 447 , 553 fu , h. , et al . 2010 , , 722 , 653 garn , t. , sobral , d. , best , p. n. , et al .
2010 , , 402 , 2017 gehrz , r. 1989 , iaus , 135 , 445 greve , t. r. , ivison , r. j. , bertoldi , f. , et al .
2004 , , 354 , 779 greve , t. r. , pope , a. , scott , d. , et al .
2008 , , 389 , 1489 hainline , l. j. , blain , a. w. , smail , i. , et al .
2011 , , 740 , 96 hoaglin , d. c. , mosteller , f. , & tukey , j. w. 1983 , wiley series in probability and mathematical statistics , new york : wiley , 1983 , edited by hoaglin , david c. ; mosteller , frederick ; tukey , john w. , hnig , s. f. & kishimoto , m. 2010 , , 523 , 27 hunt , l. k. 2010 , msais , 14 , 78 ilbert , o. , et al .
2009 , , 690 , 1236 imanishi , m. , nakagawa , t. , shirahata , m. , ohyama , y. , & onaka , t. 2010 , , 721 , 1233 ivison , r. j. , smail , i. , dunlop , j. s. , et al .
2005 , , 364 , 1025 jacobs , b. a. , et al .
2011 , , 141 , 110 kartaltepe , j. s. , sanders , d. b. , le floch , e. , et al .
2010 , , 709 , 572 knobel , c. , lilly , s. j. , iovino , a. , et al .
2012 , , 753 , 121 lilly , s. j. , et al .
2009 , , 184 , 218 magnelli , b. , chary , r. r. , pope , a. , et al .
2008 , , 681 , 258 magnelli , b. , elbaz , d. , chary , r. r. , et al .
2009 , , 496 , 57 mauduit , j .- c .
, lacy , m. , farrah , d. , et al .
2012 , , 124 , 714 mentuch , e. , abraham , r. g. , glazebrook , k. , et al .
2009 , , 706 , 1020 mentuch , e. , abraham , r. g. , & zibetti , s. 2010 , , 725 , 1971 messias , h. , afonso , j. , salvato , m. , mobasher , b. , & hopkins , a. m. 2012 , , 754 , 120 mortier , a. m. j. , serjeant , s. , dunlop , j. s. , et al .
2005 , , 363 , 563 nenkova , m. , sirocky , m. m. , ivezi , . & elitzur , m. 2008 , , 685 , 147 polletta , m. , et al .
2007 , , 663 , 81 popescu , c. c. , tuffs , r. j. , dopita , m. a. , fischera , j. , kylafis , n. d. , & madore , b. f. 2011 , , 527 , a109 pozzetti , l. , et al .
2010 , , 523 , 13 rho , j. et al .
2008 , , 673 , 271 richards , g. t. , et al .
2005 , , 360 , 839 roche , n. d. , dunlop , j. , & almaini , o. 2003 , , 346 , 803 rowan - robinson , m. , babbedge , t. , oliver , s. , et al . 2008 , , 386 , 697 sajina , a. , spoon , h. , yan , l. , et al . 2009 , , 703 , 270 salvato , m. , et al . 2009 , , 690 , 1250 salvato , m. , ilbert , o. , hasinger , g. , et al . 2011 , , 742 , 61 sargent , b. a. , et al . 2010 , , 716 , 878 saunders , w. 1990 , phdt , 353 saunders , w. ; rowan - robinson , m. ; lawrence , a. ; efstathiou , g. ; kaiser , n. ; ellis , r. s. & frenk , c. s. 1990 , , 242 , 318 schmidt , m. 1968 , , 151 , 393 scoville , n. , et al .
2007 , , 172 , 1 scott , s. e. , fox , m. j. , dunlop , j. s. , et al .
2002 , , 331 , 817 siana , b. , smail , i. , swinbank , a. m. , et al .
2009 , , 698 , 1273 silva , l. , granato , g. l. , bressan , a. , & danese , l. 1998 , , 509 , 103 tielens , a. g. 2011 , eas , 46 , 3 trump , j. r. , impey , c. d. , elvis , m. , et al .
2009 , , 696 , 1195 villar , v. , gallego , j. , prez - gonzlez , p. g. , et al .
2008 , , 677 , 169 werner , m. w. , roellig , t. l. , low , f. j. , et al .
2004 , , 154 , 1 yamada , r. , oyabu , s. , kaneda , h. , et al .
2013 , arxiv:1307.6356 | we study the contribution of hot - dust to the luminosity density of galaxies and its evolution with cosmic time . using the _ spitzer_-irac data over an area of 1.8deg@xmath0 covered by the cosmos field
, we estimate the contribution from hot - dust at rest - frame 4.2@xmath1 m ( from @xmath2 up to @xmath3 ) .
this wavelength corresponds to black - body temperature of @xmath4k .
the contribution due to stellar emission is estimated from the rest - frame 1.6@xmath1 m luminosity ( assumed to result from stellar emission alone ) and subtracted from the mid - infrared luminosity of galaxies to measure hot - dust emission . in order to attempt the study of the 3.3@xmath1m - pah feature
, we use the rest - frame 4.2@xmath1 m to infer the hot - dust flux at 3.3@xmath1 m .
this study is performed for different spectral types of galaxies : early - type , late - type , starburst , and ir - selected active galactic nuclei ( agn ) .
we find that : ( a ) the decrease of the hot - dust luminosity density since @xmath5 is steeper ( by at least 0.5dex ) compared to that of the cold - dust , giving support to the scenario where galaxy obscuration increases with redshift , as already proposed in the literature ; ( b ) hot - dust and pah emission evolution seems to be correlated with stellar mass , where rest - frame 1.6@xmath1 m luminous non - agn galaxies ( i.e. , massive systems ) show a stronger decrement ( with decreasing redshift ) in hot - dust and pah emission than the less luminous ( less massive ) non - agn galaxies ; ( c ) despite comprising @xmath6 of the total sample , agn contribute as much as a third to the hot - dust luminosity density at @xmath7 and clearly dominate the bright - end of the total hot - dust luminosity density function at @xmath8 ; ( d ) the average dust - to - total luminosity ratio increases with redshift , while pah - to - total luminosity ratio remains fairly constant ; ( e ) at @xmath9 , the dust - to - total and pah - to - total luminosity ratios increase with decreasing luminosity , but deeper data is required to confirm this result .
future study is necessary to enlighten further the characterization of the different spectral components at play in 25@xmath1 m spectral regime . |
high - temperature superconductivity in cuprates , such as la@xmath1sr@xmath2cuo@xmath3 ( lsco ) and yba@xmath9cu@xmath12o@xmath13 ( ybco ) , appears in the characteristic regime between the insulating antiferromagnetic and overdoped metallic regimes .
considering that the magnetic fluctuations play an important role in the superconductivity of these compounds , describing the dynamical magnetic response in the superconducting regime is an important issue .
it is well understood that the magnetic excitations in undoped la@xmath9cuo@xmath3 ( lco ) can be described by the spin - wave theory with nearest - neighbor , next - nearest - neighbor , and cycle - exchange terms .
@xcite in contrast , neutron scattering experiments have revealed a characteristic magnetic excitation in the superconducting regime @xcite : an incommensurate magnetic signal apparently disperses inward below a certain energy , @xmath14 , and then disperses outward above @xmath14 .
@xcite this excitation , called an `` hour - glass '' excitation , is observed in a wide doping range of lsco from non - superconducting @xmath15 ( ref . ) to slightly overdoped @xmath16 .
@xcite in the underdoped region , the low - energy incommensurability and @xmath14 increase linearly with doping .
@xcite the low - energy incommensurate spin fluctuation appears up to the overdoped region ; however , its cross section decreases linearly with superconducting transition temperature @xmath17 as the superconductivity decreases with overdoping .
@xcite these facts suggest that the hole doping strongly affects the low - energy magnetic excitations and that they are closely related to the superconductivity .
an alternative to neutron inelastic scattering for the observation of single magnons is provided by the recently developed resonant inelastic x - ray scattering ( rixs ) technique at the cu-@xmath0 edge .
@xcite in contrast to rixs with the cu-@xmath18 edge , which is more sensitive to charge excitations triggered by the core - hole potential in the 1@xmath19 orbital , rixs with the cu-@xmath0 edge can trigger a single magnon excitation by the spin - orbit coupling of 2@xmath20 orbitals .
@xcite a systematic study of paramagnons by this technique using the hole - doped ybco @xcite and lsco @xcite family compounds revealed that the dispersion relation along the @xmath10 direction above 150 mev is nearly independent of the hole concentration in a wide doping range from the undoped to the non - superconducting overdoped samples .
this fact is in sharp contrast to the neutron results in the low - energy region . because of the limited wave vector of incident photons at the cu-@xmath0 edge and relatively relaxed energy resolution of 120 mev , compared with neutron inelastic scattering , rixs at the cu-@xmath0 edge is more appropriate for the measurement of magnon excitations above 100 mev dispersing from the @xmath21 position , where the magnetic structure factor for neutron scattering is small .
in contrast , neutron scattering has a fine energy resolution , typically a few mev , which makes it difficult to observe high - energy magnetic excitation in doped samples , which is broad in energy , and consequently , neutron inelastic scattering is suitable for the observation of magnetic excitations below 150 mev , dispersing from the antiferromagnetic ( af ) zone center @xmath8 for the doped samples .
thus , rixs and neutron inelastic scattering measurements are complementary to each other . in this paper , we report high - energy magnetic excitations above 100 mev of overdoped lsco studied by both neutron inelastic scattering and cu-@xmath0 edge rixs using identical crystals for the two techniques .
the overdoped sample was selected such that we can also test the doping independence of the paramagnon dispersion relation .
the neutron scattering results of lsco @xmath4 indicate that the magnetic dispersion relation along the @xmath8 direction up to 250 mev reasonably agrees with the spin - wave dispersion relation of lco , which is consistent with the doping - independence of magnetic dispersion reported using rixs .
the cu-@xmath0 edge rixs measurements of the identical sample reveal that the dispersion along the @xmath10 direction is consistent with the lco spin - wave dispersion and previous rixs results by dean _
et al . _
@xcite using lsco thin films .
in contrast , the paramagnon near @xmath11 appears to be less dispersive and the excitation energy near @xmath11 is lower than the spin - wave energy of lco .
single crystals of lsco with @xmath4 and @xmath22 used for the neutron and cu-@xmath0 rixs measurements were grown using the traveling solvent floating zone method .
the growth and post - annealing conditions are the same as those described in ref . .
single crystals of @xmath4 were used for the neutron measurements , and both @xmath4 and @xmath22 were used for the neutron inelastic scattering and rixs .
the crystal structures of @xmath4 and @xmath6 are tetragonal with space group @xmath23 . in the present study ,
we use the notation based on this tetragonal structure .
therefore , the antiferromagnetic wave vector @xmath8 corresponds to @xmath24 .
neutron scattering experiments were performed using the sequoia chopper spectrometer @xcite at the spallation neutron source ( sns ) at oak ridge national laboratory ( ornl ) .
five single crystals of @xmath4 , with volumes of @xmath25 @xmath26 , were co - aligned using a neutron diffractometer installed at the cg-1b beam port of the high flux isotope reactor ( hfir ) at ornl before the sequoia measurements .
the set of co - aligned crystals was attached to the cold finger of a @xmath27he closed cycle refrigerator and set on the spectrometer .
an incident neutron energy ( @xmath28 ) of either @xmath29 mev or @xmath30 mev was selected by the coarse resolution fermi chopper spinning at 240hz .
for the 250mev and 350mev settings , the t0 chopper was spun at 60hz and 150 hz , respectively . for both conditions the @xmath31 axis of the crystal was oriented parallel to the incident beam .
these experimental conditions provides an energy resolution of 5 20 mev , and an momentum resolution of approximately @xmath32 @xmath33 along the @xmath8 direction for the energy transfer between 100 250 mev .
cu-@xmath0 edge rixs experiments were performed using the axes spectrometer at the id08 beam line of the european synchrotron radiation facility ( esrf ) .
crystals of @xmath4 and @xmath22 were cut into disk shapes with the @xmath31-axis normal to the disk surface .
the @xmath4 crystals were cut from the same crystal rod used for the neutron measurements .
the crystals were pasted on a cu plate and attached to a refrigerator . either the @xmath34 $ ] or @xmath35 $ ] axis was set horizontal to enable the measurement of the paramagnon dispersion along the @xmath10 or @xmath8 direction , respectively , by a horizontal rotation of the sample .
the incident photon polarization was set either horizontal ( @xmath36-polarization ) or vertical ( @xmath37-polarization ) . the combined ( monochromator for incident photons , spectrometer for scattering photons )
energy resolution was @xmath38 mev .
beam aperture from the sample to the detector is approximately 10 mrad , which leads to a momentum resolution of @xmath39 @xmath33 with the photon energy at cu-@xmath0 edge .
the scattering angle was 130@xmath40 .
individual rixs spectra obtained after 5 min of accumulation were summed for a total of 120 min at each @xmath41 value and polarization .
the energy spectra were measured every 5@xmath40 of sample rotation , corresponding to less than 0.05 reciprocal lattice units ( r.l.u . ) . and @xmath42 at selected energies .
red lines are fits to a function containing two lorentzians symmetric to the af zone center .
each lorentzian component is shown by a dashed line .
, width=321 ]
neutron magnetic cross - sections below 100 mev of lsco @xmath4 have already been reported in ref . . here
, we focus on magnetic scattering in the energy range above 100 mev to compare the paramagnon dispersion measured by rixs .
the magnetic excitation signal in the high - energy region becomes broad in energy upon hole doping .
furthermore , the overall spectral weight in the overdoped region is smaller than that in the underdoped samples . therefore , we utilized a relatively poor energy resolution of 5 20 mev , for these measurements .
figure 1 presents contour maps of neutron cross - sections on the @xmath43-plane at energy transfers of @xmath44 , @xmath45 , and @xmath46 mev . in fig .
1 ( a ) , a magnetic signal can be observed near @xmath24 , namely , the @xmath8 position .
this signal appears to become more dispersed from the af zone center as the energy transfer increases .
measured by neutron referred from ref .. in both figures , solid lines are the spin - wave dispersion of non - doped lco referred from ref .. , width=321 ] to draw the magnetic dispersion , we analyzed intensity profiles along the trajectories of @xmath47 and @xmath48 across the two af zone centers , namely @xmath8 and @xmath49 .
these trajectories are indicated by dashed arrows in fig .
1 ( a ) . @xcite we fit the profiles by a two - gaussian function , which is symmetric to the af zone center .
we present select profiles with the results of the fits in fig .
2 as a representative data set .
figures 2 ( a ) ( c ) present profiles around @xmath8 at @xmath44 , @xmath45 , and @xmath29 mev , and figs 2 ( d )
( f ) present profiles around @xmath49 at @xmath44 , @xmath50 , and @xmath51 mev , respectively . here
, it is more clearly demonstrated that the magnetic signals disperse outward as the energy increases consistently at both af zones .
the peak positions and full - width at half - maximum ( fwhm ) values obtained by the fits are summarized in fig .
3 ( a ) . here
, the closed symbols represent the peak positions and the horizontal bars represent the fwhm values . because the analyzed scan trajectories are the @xmath47 directions , the peak positions correspond to the magnetic dispersion relation along @xmath8 from the magnetic zone center .
the solid curved line is the spin - wave dispersion along @xmath8 of lco referred from ref . .
the observed magnetic peak is broad in @xmath41 in this energy range ; however , the peak positions roughly follow the spin - wave dispersion .
this finding is consistent with the doping independence of the paramagnon dispersion observed using rixs for the @xmath10 direction .
it is worth comparing the present data to the data of lsco @xmath52 reported by lipscombe _
et al . _
@xcite in the same energy range ( @xmath53 mev ) .
our results indicate that the half - width at half - maximum ( hwhm ) is @xmath54 @xmath33 for the profiles around @xmath8 and @xmath55 @xmath33 for those around @xmath49 . at this stage ,
the reason for the difference in the different zones is unknown ; however , the values are apparently consistent with that of lsco @xmath52 .
_ reported that the width @xmath56 suddenly increases to @xmath57 r.l.u . above @xmath58
note that the functional form they used to analyze the data is different from ours and that their @xmath56 is larger than our hwhm by a factor of @xmath59 .
thus , the width of lsco @xmath52 corresponding to our hwhm is @xmath60 @xmath33 , which is in reasonable agreement with our values .
next , we compare the dispersion relation . in ref .
, the incommensurability @xmath61 , which approximately corresponds to the peak position in our analyses , along the @xmath10 direction is reported .
the open symbols in fig.3 ( b ) show @xmath61 values of lsco @xmath52 between @xmath62 and @xmath63 mev together with the spin - wave dispersion of lco along @xmath10 @xcite .
in contrast to the @xmath8 direction data , the peak positions along @xmath10 of lsco @xmath52 clearly deviate from the spin - wave dispersion relation , although it appears that the peak position approaches the spin - wave dispersion of lco as the energy increases . in the next section , we present the cu-@xmath0 rixs results and compare these findings with the neutron results . and @xmath22 .
the high intensity at @xmath64 ev is due to the @xmath65 excitation .
dispersive feature below 0.5 ev is the paramagnon .
solid dispersive lines are the spin - wave dispersion of lco .
solid circles indicate the peak positions of neutron magnetic peak in fig .
3 . , width=321 ]
cu-@xmath0 rixs profiles were measured using single crystals of @xmath4 and @xmath22 at several @xmath41-positions between @xmath21 and @xmath10 and between @xmath21 and @xmath8 .
figure 4 presents contour maps of the rixs intensity measured with the @xmath36-polarization configuration .
the intensity is normalized by the integrated intensity of the @xmath65-excitation , which appears at @xmath59 ev as a dispersionless excitation .
the contour maps clearly indicate dispersive modes below @xmath66 mev for both the @xmath10 and @xmath8 directions . in the same figure ,
the spin - wave dispersion of lco is represented by solid lines , and the magnetic peak positions determined by neutron inelastic scattering along the @xmath8 direction are represented by circles .
below , we compare these data in detail . rixs profiles .
data were analyzed by fitting to a function contains two gaussians : one for magnon and the other for the @xmath67 components , and a squared lorentzian describing the tail of the @xmath65 excitation . the @xmath67 component is assumed to have resolution width 0.35 ev , while the magnon component is convoluted by the resolution .
thick solid lines are results of fits , thinner solid lines are magnon components , and dashed lines represent the @xmath67 and @xmath65-tail components . ,
width=321 ] to evaluate the energy of the paramagnon modes , we fit the rixs profiles to a function containing a quasielastic peak ( containing an elastic peak and phonon excitations ) , a paramagnon peak , and the tail of the @xmath65-excitation .
the quasielastic and paramagnon peaks are assumed to be gaussian functions , and the tail of the @xmath65-excitations is assumed to be a squared - lorentzian tail .
the quasielastic component is resolution limited , whereas the magnon component is convoluted with the energy resolution of @xmath30 mev . in a previous rixs study of paramagnons with better energy resolution of @xmath68 mev
, the rixs profiles were analyzed with more peak components such as phonon and two - magnon components . in the present study , the relatively broad energy resolution prevents these lower energy modes from being distinguished ; however , the major paramagnon contribution should be evaluated by the above analyses .
the results of the fits are summarized in fig .
the magnon component is indicated by the solid curves , and the elastic component and tail of the @xmath65-excitation are represented by dashed curves .
the magnon energies determined by the fits of the rixs data are represented by the filled symbols in fig . 6 : circles ( @xmath4 ) , squares ( @xmath6 ) , and diamonds ( @xmath69 referred from ref . )
. it is observed in fig . 6 ( b ) that the present data of @xmath4 along the @xmath10 direction are consistent with the data of @xmath69 reported by dean
_ et al . _
@xcite using a thin - film sample .
this fact indicates that our analyses are valid , and importantly , the magnons are consistently observable both in bulk crystals and thin films .
figure 6 demonstrates that the agreement between the paramagnon energies , represented by solid symbols , and the spin - wave dispersion of lco , represented by solid curves , is excellent for the @xmath10 direction , whereas it is somewhat poor in the @xmath8 direction .
the rixs data in the @xmath8 direction appear to be less dispersive and tend to be located at lower energies for @xmath41 larger than @xmath70 .
direction for @xmath4 ( circles ) and @xmath22 ( closed squares ) , and those along ( b ) the @xmath71 directions for @xmath4 ( circles ) and for @xmath69 thin film ( closed diamonds ) reported in ref .. neutron data in fig . 3 are also shown by open squares .
solid lines are the spin - wave dispersion of lco . the thick dashed and gray lines in the inset indicate q - trajectories of dispersions measured by rixs and neutron , respectively .
, width=321 ] the data measured by neutron inelastic scattering are also shown in fig .
6 as open symbols . the data in fig . 6 ( a ) are those of the present study and the data in fig .
6 ( b ) are those of @xmath52 by lipscombe _
et al . _
@xcite it should be noted that the rixs data show the dispersion relation from the @xmath21 position , whereas the neutron data from the af zone center @xmath8 are shown in the inset of fig .
6 . however , these data sets should be identical as long as the antiferromagnetic correlation exists . in both directions ,
the neutron and rixs data connect with each other , thus indicating the consistency between these two probes and confirming that the spin excitation can be consistently observed by both probes .
we have demonstrated that the magnetic excitations of overdoped lsco single crystals measured by neutron and rixs are qualitatively consistent with each other and that the high - energy magnetic dispersion in the @xmath10 direction follows the spin - wave dispersion relation of lco in the energy range above 150 mev .
this behavior is consistent with the doping independence of magnon dispersion observed by rixs for hole - doped cuprate thin films.@xcite the magnetic dispersion along @xmath8 agrees less with the spin - wave dispersion ; however , the overall energy scales roughly agree with the spin - wave dispersion . the present results are in contrast with the excitations in electron - doped systems . by combining the neutron inelastic scattering and rixs results , ishii _
et al._@xcite and lee _
et al._@xcite have reported that the magnetic excitation near the af zone center becomes steeper as doping increases and the overall excitation energy shifts to higher energies with respect to the undoped system .
in contrast , our rixs data for the @xmath8 direction suggest that the excitation energy near @xmath11 might be lower than the spin - wave energy .
such asymmetry between hole- and electron - doped systems is consistent with the numerical calculation of the hubbard model by jia _ et al._@xcite in addition to the qualitative consistency with the theoretical calculations , our rixs data along the @xmath8 direction in fig . 6 ( a ) agree less with the spin - wave dispersion than those along the @xmath10 direction .
previous rixs studies of cuprates using thin films mostly focused on the @xmath10 direction and are consistent with the spin - wave dispersion
. recently , guarise _ et al._@xcite and dean _ et al._@xcite reported that the rixs profiles of bi - based cuprates along @xmath8 exhibit anomalous softening or dispersionless broad excitation .
the present rixs data along @xmath8 also appear to be less dispersive than those along the @xmath10 direction . in fig .
7 , we present a comparison of the rixs spectra with @xmath36- and @xmath37-polarization configurations at the position where the sample is rotated by 45@xmath40 from the specular position .
the data shows appreciable rixs intensities of the @xmath37-polarization , which is even larger than the intensities of the @xmath36-configuration .
theoretical calculations based on a single cu@xmath72 ion where the valence band has the @xmath73 orbital symmetry indicate that in our experimental geometry , the rixs cross - sections with and without spin - flip processes dominate the @xmath36 and @xmath37 incident polarization configurations , respectively.@xcite therefore , the results in fig . 7 indicate that appreciable charge excitation exists in the same energy range of magnetic excitation
. these charge excitations may affect the magnetic excitations , possibly in different manners between the @xmath8 and @xmath10 directions , causing the observed difference in the magnetic excitations . to address this question ,
more precise rixs measurements with finer energy resolution and polarization analyses are necessary to distinguish the magnetic and charge excitations . and @xmath37 configurations at @xmath74 and @xmath75 corresponding to a 45-degree rotation of the sample from the specular orientation .
, width=321 ] remarkably , fig .
7 shows that the rixs intensity of the @xmath37-configuration is always higher than that of the @xmath36-configuration .
this feature qualitatively resembles the @xmath36 and @xmath37-configuration data of electron - doped nd@xmath1ce@xmath2cuo@xmath3 with @xmath76 and @xmath77 reported in ref . , but not those of the optimally hole - doped system in ref . .
as discussed in ref .
, such a balance of the @xmath36 and @xmath37-configuration spectra suggests that the spin and charge excitations are mixed because of itinerant character .
this finding suggests that in the electron - doped system , the itinerant character becomes stronger even at the optimally doped level and the magnetic excitation changes concomitantly , whereas in the hole - doped system , the itinerant character slowly grows with doping , and the high - energy magnetic excitation hardly changes with doping .
in contrast , our data indicates that the overdoped @xmath4 has itinerant character , but the magnetic excitation still remains mostly at the spin - wave dispersion of lco .
this fact suggests that the asymmetry of the magnetic excitation between electron and hole - doped systems is not simply due to the stronger itinerant character of the electron - doped system .
we have performed neutron and cu-@xmath0 edge rixs measurements of overdoped la@xmath1sr@xmath78cuo@xmath79 using identical single crystals .
the combination of neutron and rixs indicates that both data sets are consistent with each other and that the overall high - energy magnetic excitation agrees with the spin - wave dispersion relation of the parent compound la@xmath80cuo@xmath79 particularly in the @xmath10 direction , which is consistent with the previous rixs studies using thin - films .
we also draw the magnetic excitation above 100 mev in the @xmath8 direction by the complementary use of neutron and rixs . the magnetic excitation for the @xmath41 values smaller than @xmath70 ( r.l.u .
) measured by neutron inelastic scattering follows the lco spin - wave dispersion , whereas that for the @xmath41 values larger than @xmath70 ( r.l.u . ) measured by rixs is apparently less dispersive , and the excitation energy near @xmath11 is smaller than the lco spin - wave excitation energy .
the polarization dependence of the rixs spectra indicates that appreciable charge excitations exist in the same energy range of magnetic excitations , which may affect the magnetic excitation .
authors thank r. kajimoto , m. matsuda , m. matsuura , k. nakajima , t , tohyama and k. yamada for invaluable discussion .
authors also acknowledge l. j. santodonate and m. matsuda for their assistance in using cg1b at ornl .
this work is supported by grant - in - aid for scientific research ( c ) no .
25390132 and no . 35400333 , and by grant - in - aid for scientific research ( b ) no .
part of the research conducted at ornl s high flux isotope reactor and spallation neutron source was sponsored by the scientific user facilities division , office of basic energy sciences , us department of energy .
the rixs experiments were performed using the axes instrument at id08 at the european synchrotron radiation facility ( esrf ) .
we acknowledge esrf for provision of synchrotron radiation facilities and we would like to thank n. b. brookes for the support at id08 .
m. le tacon , g. ghiringhelli , j. chaloupka , m. moretti sala , v. hinkov , m. w. haverkort , m. minola , m. bakr , k. j. zhou , s. blanco - canosa , c. monney , y. t. song , g. l. sun , c. t. lin , g. m. de luca , m. salluzzo , g. khaliullin , t. schmitt , l. braicovich , and b. keimer , nature phys . * 7 * , 725 ( 2011 ) .
m. p. m. dean , g. dellea , r. s. springell , f. yakhou - harris , k. kumer , n. b. brookes , x. liu , y .-
sun , j. strle , t. schmitt , l. braicovich , g. ghiringhelli , i. bozovic , and j. p. hill , nature mat .
* 12 * , 1029 ( 2013 ) .
m. le tacon , m. minola , d. c. peets , m. moretti sala , s. blanco - canosa , v. hinkov , r. liang , d. a. bonn , w. n. hardy , c. t. lin , t. schmitt , l. braicovich , g. ghiringhelli , and b. keimer , phys . rev .
b * 88 * , 020501(r ) ( 2013 ) .
k. ishii , m. fujita , t. sasaki , m. minola , g. dellea , c. mazzoli , k. kumaer , g. ghiringhelli , l. bracivich , t. tohyama , k. tsutsui , k. sato , r. kajimoto , k. ikeuchi , k. yamada , m. yoshida , m. kurooka , and j. mizuki , nature comm .
* 5 * , 3714 ( 2014 ) .
w. s. lee , j. j. lee , e. a. nowadnick , s. gerber , w. tabis , s. w. huang , v. n. strocov , e. m. motoyama , g. yu , b. moritz , h. y. huang , r. p. wang , y. b. huang , w. b. wu , c. t. chen , d. j. huang , m. grecen , t. schmitt , z. x. shen , and t. p. devereaux , nature phys . * 10 * , 883 ( 2014 ) . m. moretti sala , v. bisogni , c. aruta , g. balestrino , h. berger , n. b. brookes , g. m. de luca , d. di castro , m. grioni , m. guarise , p. g. medaglia , f. miletto granozio , m. minola , p. perna , m. radovic , m. salluzzo , t. schmitt , k. l. zhou , l. braicovich , and g. ghringhelli , new j. phys , * 13 * , 043026 ( 2011 ) .
l. braicovich , m. moretti sala , l. j. p.
ament , v. bisogni , m. minola , g. balestrino , d. di castro , g. m. de luca , m. salluzzo , g. ghiringhelli , and j. van den brink , phys .
b * 81 * , 174533 ( 2010 ) . | we have performed neutron inelastic scattering and resonant inelastic x - ray scattering ( rixs ) at the cu-@xmath0 edge to study high - energy magnetic excitations at energy transfers of more than 100 mev for overdoped la@xmath1sr@xmath2cuo@xmath3 with @xmath4 ( @xmath5 k ) and @xmath6 ( non - superconducting ) using identical single crystal samples for the two techniques . from constant - energy slices of neutron scattering cross - sections , we have identified magnetic excitations up to @xmath7 mev for @xmath4 . although the width in the momentum direction is large , the peak positions along the @xmath8 direction agree with the dispersion relation of the spin - wave in the non - doped la@xmath9cuo@xmath3 ( lco ) , which is consistent with the previous rixs results of cuprate superconductors . using rixs at the cu-@xmath0 edge ,
we have measured the dispersion relations of the so - called paramagnon mode along both @xmath8 and @xmath10 directions .
although in both directions the neutron and rixs data connect with each other and the paramagnon along @xmath10 agrees well with the lco spin - wave dispersion , the paramagnon in the @xmath8 direction probed by rixs appears to be less dispersive and the excitation energy is lower than the spin - wave of lco near @xmath11 .
thus , our results indicate consistency between neutron inelastic scattering and rixs , and elucidate the entire magnetic excitation in the @xmath8 direction by the complementary use of two probes .
the polarization dependence of the rixs profiles indicates that appreciable charge excitations exist in the same energy range of magnetic excitations , reflecting the itinerant character of the overdoped sample .
a possible anisotropy in the charge excitation intensity might explain the apparent differences in the paramagnon dispersion in the @xmath8 direction as detected by the x - ray scattering . |
the optically emitting gas of planetary nebulae ( pns ) is illuminated and ionized by the radiation from the central hot star , which is evolving to become a white dwarf ( wd ) .
the shaping of the optical nebula , formed from the envelope of the progenitor asymptotic giant branch ( agb ) star , is a more complicated process , with many open questions ( kastner , soker , & rappaport 2000 ) .
of particular interest is whether a binary companion is required to form non - spherical nebulae , or whether single agb stars can blow axisymmetrical winds and form axisymmetrical pns . according to soker & rappaport ( 2000 ) one of the processes
by which a companion can shape a pn is by accreting from the wind of the agb progenitor ; if an accretion disk is formed around the compact companion , then a jet - like bipolar outflow ( or collimated fast wind , cfw ) can result .
when the jet material is shocked , it heats up and may emit in the x - ray band .
another ingredient in the shaping process is the fast wind from the central star , blown during the post - agb and pn stages ( kwok , purton , & fitzgerald 1978 ) .
the shocked fast wind is expected to emit in the x - ray band . in particular ,
an intermediate - velocity ( @xmath0 ) wind during the post - agb phase may explain many of the observed properties of the x - ray emitting gas in pns ( soker & kastner 2003 ) .
processes behind the extended x - ray emission are tied to the shaping processes of pns ( kastner 2002 ) , and although not all shaping processes will lead to x - ray emission , the x - ray properties of pns may hint at the nature of their progenitor ( soker & kastner 2003 ) . with this goal in mind , we present a re - analysis of the spatially extended x - ray spectra of two pns , bd + 303639 ( pn g064.7 + 05.5 ) and ngc 7027 ( pn g084.9 - 03.4 ) .
material relevant to these pns is summarized in our earlier papers announcing the detection of extended x - ray emission from them ( kastner 2000 ; kastner , vrtilek , & soker 2001 ) , and a study of the x - ray morphologies of these two pns ( kastner 2002 ) .
the observations are described in 2 . in 3 , we describe our spectral analysis , and in 4 we present the results , emphasizing the large differences in the abundances of some elements between the x - ray emitting gas and the optical nebula . in @xmath1
we suggest explanations for these findings , focusing on magnesium ( mg ) in ngc 7027 , and neon ( ne ) in bd + 303639 .
_ chandra _ observed ngc 7027 for 19.0ksec and bd + 303639 for 18.5ksec with the advanced ccd imaging spectrometer ( acis ) ( garmire 1988 ) as the focal plane instrument .
for both observations the science instrument module was translated and the telescope was pointed such that the telescope boresight was positioned near the center of the spectroscopy ccd array ( acis -s ) ; the objects where imaged on the central back - illuminated ccd ( s3 ) , which provides moderate spectral resolution of @xmath24.3 at 0.5 kev and @xmath29 at 1.0 kev .
analyses of these data have already appeared in kastner ( 2000,2001 ) .
data from these observations were reprocessed by the chandra x - ray center ( cxc ) subsequent to the publication of our earlier papers and the analyses presented here were performed on the reprocessed files .
each spectrum was extracted using the _ chandra _ interactive analysis of observations ( ciao ) software within a region judged to contain all the x - ray flux from the nebula .
we note that the current extraction ( ciaov2.2.1 ) reflects upgrades to the calibration system that were made after our previous papers on these systems were published .
the extracted events are aspect - corrected , bias - subtracted , graded , energy - calibrated and limited to grade 02346 events ( asca system ) . for both observations , the background count rate from a large , off - source annulus region ( 30- and 50-pixel radii ) was negligible in comparison to the source count rate .
the low photon statistics and limited energy range characteristic of these spectra make it difficult to derive abundances for these objects on the basis of x - ray spectra alone .
here we model the spectra using , as guidance , abundances derived from studies in the optical , ultraviolet , and infrared .
we used the vmekal ( mewe , lemen , & van den oord 1986 ) code as incorporated into xspec v11.0 ( arnaud , borkowski , & harrington 1996 ) to generate model spectra that are appropriate for optically thin thermal plasma in ionization equilibrium .
when abundance values for elements that are fit by vmekal are not available we set the abundance to solar . in this
we differ from our original papers ( kastner 2000 , 2001 ) where unspecified abundances were set to zero for elements heavier than mg .
the x - ray spectrum of ngc 7027 shows emission that peaks at approximately 0.9 kev and 1.3 kev and drops off at approximately 1.4 kev .
no significant emission was detected below 0.5 kev and above 2.0 kev , hence we restrict our fits to the energy range 0.5 - 2.0 kev .
nebular abundances reported by bernard salas ( 2001 ) using iso and optical observations produced a fair fit ( fig .
1 ) , but do not adequately account for features between 0.6 - 0.7 kev and @xmath21.3 kev . we attribute the feature at 0.65 kev to a blend of o lines .
we note that while the nebula has been reported to be somewhat depleted in o ( bernard salas 2001 ; beintema 1996 ; keyes , aller , & feibelman 1990 ; middlemass 1990 ) , the stellar wind of ngc 7027 is expected to show a greatly enhanced o abundance ( hasegawa , volk , & kwok 2000 ) .
this high o abundance could explain the feature at 0.6 kev .
however , we emphasize that , while kastner ( 2001 , 2002 ) ascribe the x - ray emission to the action of a fast wind , no such fast wind has yet been detected for this nebula .
the feature at @xmath2 1.3 kev requires enhanced mg abundance ; we were unable to find alternative explanations for this feature .
we thus fit our x - ray spectrum starting with nebular abundances as determined by bernard salas ( 2001 ) but allowing o and mg to be free .
the resulting fit ( fig . 2 and tables 1,2 ) requires o enhanced to 9 times solar and mg enhanced to 3 times solar . in all cases we allowed the intervening column density n@xmath3 to be a free parameter .
best - fit values of @xmath4 ( table 2 ) are consistent with our previous results and with the typical value of @xmath5 toward ngc 7027 ( kastner 2001,2002 ) ; however , the temperature is a factor of @xmath6 larger , if the abundances of elements above mg in ngc 7027 are solar as assumed here .
a higher temperature implies that the fast wind material that was shocked and formed the presently x - ray emitting gas was blown at a somewhat higher speed , and/or the adiabatic cooling , studied by soker & kastner ( 2003 ) , was less efficient . " like ngc 7027 , past analysis of bd + 303639 suggests that this pn s x - ray emission arises primarily from wind and/or nebular material ( kastner 2000 ) . here ,
we perform several tests to determine whether the observed x - ray spectrum is consistent with reported nebular and/or stellar wind abundances .
the extracted spectrum shows distinct peaks near 0.5 kev and 0.95 kev .
no significant emission was detected below 0.3 kev and above 1.5 kev .
we restrict our fits to the energy range 0.3 - 1.5 kev .
we first assume nebular abundances reported from optical and ultraviolet observations by pwa , pottasch , & mo ( 1986 ) , aller & hyung ( 1995 ) , and grevesse & sauval ( 1998 ) .
the best - fit parameters are listed in table 2 and the model is shown in figure 3 .
we then use abundances reported for the stellar wind by arnaud ( 1996 ) .
the stellar wind abundances provide a significantly better fit than the nebular abundances but strong features near 0.95 and 0.6 kev remain ( figure 4 ) .
we then attempted to fit the x - ray spectrum starting with the stellar wind abundances but allowing o and ne to be free parameters .
this resulted in an acceptable fit ( @xmath7 ) with o enhanced to 3.6 times solar and ne enhanced to 21 times solar .
although this ne enhancement is not predicted by studies at longer wavelengths , it is consistent with earlier x - ray studies ( arnaud et al ( 1996 ) ; kastner et al .
( 2000 ) ) suggesting that the feature at 0.9 kev can only be fit with an enhanced ne abundance .
the model is shown overlaying the data in figure 5 , and the best - fit parameters are listed in tables 1 , 2 . our best - fit value for t is 0.7 times higher and for n@xmath3 two times higher than that obtained by kastner et al .
( 2001 ) , whose result for @xmath4 is consistent with the measured optical obscuration of the nebula ( @xmath5 ) given a conversion between @xmath4 and @xmath5 appropriate for the interstellar medium .
our result for @xmath4 would therefore suggest that the gas to dust ratio in this nebula is elevated over that of the interstellar medium , along lines of sight toward the x - ray - emitting gas .
ngc 7027 and bd + 303639 show strong emission features that require enhanced mg abundance relative to solar ( to fit lines at @xmath21.3 kev ) for ngc 7027 , and enhanced ne abundance relative to solar ( to fit lines at @xmath20.9 kev ) for bd + 303639 ( kastner 2000 ; kastner 2001 ) .
both of these results are surprising since abundance values for the nebular and stellar wind material obtained at other wavelengths indicate depleted mg and ne ( beintama 1996 ; keyes & aller 2001 ; middlemass 1990 ) .
kastner ( 2000 ) proposed that the high abundance of ne found in bd + 303639 results from the central star which itself is a carbon - rich wolf - rayet ( [ wc ] ) object ( grosdidier 2000 ) . here
we suggest that the source of the large abundance of ne in bd + 303639 is a nova eruption on an oxygen - neon - magnesium ( onemg ) wd companion .
novae in cataclysmic variables , which are thought to occur on the surface of onemg wds , are known to eject material with ne abundance up to @xmath8 times the solar abundance , and the total ejected mass is @xmath9 ( starrfield 1998 ) .
the estimated mass in the x - ray emitting gas in bd + 303639 is @xmath10 ( kastner 2000 ) .
therefore , a nova eruption can easily account for the @xmath11 times solar ne abundance .
the mg abundance in such nova eruptions is typically a factor of @xmath12 times lower than the ne abundance ( starrfield et al .
therefore , it is not surprising we do not detect the mg line in the spectrum of bd + 303639 .
the shaping of some bipolar pns by nova eruption on a wd companion has been proposed by soker ( 2002 ) , who reviews many routes for the formation of bipolar pns , several involving novae eruptions .
the accretion rate onto the wd can not be too high in that case , so the orbital separation can not be too small .
we do note that an accretion disk may mix ne in the ejected jets even when there is steady nuclear burning .
observations supporting this proposed scenario come from symbiotic systems .
some symbiotic systems are known to experience nova outbursts ( such systems are referred to as `` symbiotic novae '' ; e.g. , rx puppis , mikolajewska et al .
1999 ) , and some bipolar symbiotic nebulae appear to have gone through an eruptive mass loss event , possibly due to a nova eruption , as in the cases of v1016 cyg ( corradi et al .
1999 ) and he 2 - 104 ( corradi et al . 2001 ) .
for the symbiotic nova rx puppis , mikolajewska et al .
( 1999 ) argue for an orbital separation of @xmath13 .
although the bright ionized region of bd + 303639 has a general elliptical shape , high resolution radio molecular - line imaging has revealed the presence of two , oppositely - directed `` bullets '' of fast - moving , dense gas , suggestive of the action of collimated outflows or jets ( bachiller et al .
2000 ) . to conclude ,
we propose that the central star of bd + 303639 has a massive , @xmath14 onemg wd companion , at an orbital separation of @xmath15 . in principle
, peculiar abundance in the x - ray emitting gas of ngc 7027 could also result from material ejected from a wd companion ; indeed , kastner et al .
( 2001 ) attributed the emission peak near 0.9 kev to ne lines ( whereas here they are modeled as a complex of lines of highly ionized fe ) . however , we suggest a simpler explanation .
the nebular abundances derived from previous studies of ngc 7027 are for the gas in the optical nebulae , and do not include metals locked in dust particles .
ngc 7027 is known to have a significant amount of dust in the nebular shell ( sanchez contreras 1998 ) hence it is expected that many metals will be depleted in the gas phase . however , in the x - ray emitting gas we do nt see this depletion , which implies that the dust in this medium was either destroyed or was never formed . assuming that the mg abundance in the x - ray emitting gas is that of the wind lost by the progenitor , then the depletion in this nebula is by a factor of @xmath16 ( table 1 ) .
mg depletion in pns is well established .
for example , in the galactic pn ngc 3918 , harrington , monk , & clegg ( 1988 ) find depletion by a factor of 3 .
dopita et al .
( 1997 ) find mg depletion by a factor of @xmath17 in pns in the magellanic clouds and note that mg is not expected to be formed or destroyed during the evolution of these pns ( since an onemg wd is not formed ) , and that the magnesium resides in dust grains which are not destroyed by the uv radiation of the central star .
the depletion fraction we find , then , implies that the star was born with a mg abundance of @xmath18 times the solar value .
this is feasible , since the distance of ngc 7027 from the galactic center is similar to that of the sun , and it is close to the galactic plane .
mg abundance of @xmath18 times solar are found in the solar neighborhood and closer to the galactic center ( e.g. , smartt 2001 ) .
in addition , a high mg abundance due to destruction of dust is unlikely because the destruction time of the dust in this pn is likely to be longer than the time elapsed since the star left the agb , which we deduce in the following way .
the destruction time of dust particles in the temperature range @xmath19 depends weakly on the temperature ( draine & salpeter 1979 ; smith et al .
1996 ) . however , the destruction time depends on the grain size . for the pn ngc 3918 , for example ,
harrington ( 1988 ) find that @xmath2 50% of the dust mass is in dust particles of size @xmath20 . in ngc 7027
, the dust grains may be much larger , with sizes of @xmath20 , and up to @xmath21 ( sanchez contreras 1998 ) . scaling from draine & salpeter ( 1979 ) for values appropriate here , in particular grain sizes of @xmath22 ( jura 1996 )
, we find the destruction time of dust in ngc 7027 is longer than @xmath23 for an ism distribution of grain sizes the destruction time at a pressure of @xmath24 @xmath25 k , as in the x - ray emitting gas of ngc 7027 , is @xmath26 yr , according to the calculation of smith et al.(1996 ) .
the nebular dust can not be destroyed in the shock , since the shock moving through the optical - nebular gas is relatively slow .
the dynamical age of ngc 7027 is @xmath27 yr ( masson 1989 ; latter et al . 2000 ) .
we conclude that the dust ca nt be destroyed , even if the nebular gas is heated to @xmath28 k. because the dust can not be destroyed for this pn , we speculate that the x - ray emitting gas results from wind segments which never formed dust . these segments could result either from a post - agb wind or a cfw blown by an accreting companion ( for discussion of these possibilities see soker 2002 ) .
aller , l. h. & hyung , s. 1995 , , 276 , 1101 lcccc * ngc 7027 * ( 0.5 - 2.0 kev ) & & & & + nebular & 70@xmath34 & 7.9@xmath35 & 3.0e-14 & 2.0 + modified nebular & 41@xmath36 & 8.4@xmath35 & 3.6e-14 & 0.9 + + * bd + 30@xmath37 3639*(0.3 - 1.5 kev ) & & & & + nebular & 74@xmath36 & 1.2@xmath38 & 4.8e-13 & 9.3 + stellar wind & 24@xmath39 & 2.1@xmath38 & 6.1e-13 & 2.1 + modified stellar wind & 25@xmath39 & 2.1@xmath38 & 6.1e-13 & 1.2 + + | we revisit _ chandra _ observations of the planetary nebulae ngc 7027 and bd + 303639 in order to address the question of abundance anomalies in the x - ray emitting gas .
enhanced abundances relative to solar of magnesium ( mg ) for ngc 7027 and neon ( ne ) for bd + 30 3639 are required to fit their x - ray spectra , whereas observations at optical and infrared wavelengths show depleted mg and ne in these systems .
we attribute the enhancement of mg in ngc 7027 in the x - ray , relative to the optical , to the depletion of mg onto dust grains within the optical nebula . for bd + 30 3639 , we speculate that the highly enhanced ne comes from a wd companion , which accreted a fraction of the wind blown by the asymptotic giant branch progenitor , and went through a nova - like outburst which enriched the x - ray emitting gas with ne . |
the large - scale structure of the universe is one of the most powerful probes in cosmology .
the statistical nature of primordial density fluctuations can be investigated by large - scale distributions of galaxies .
geometrical effects on the observed clustering pattern of galaxies can constrain the nature of dark energy , through the alcock - paczinski effect @xcite or baryon acoustic oscillations ( bao ) @xcite . recently , it is found that the presence of primordial non - gaussianity introduces scale - dependent bias in the halo clustering @xcite . therefore , many possible models in cosmology , such as inflationary scenarios , dark energy models , modified gravity , and so on , should be constrained by precision measurements of the large - scale structure in near future . to compare the observations with theory , it is crucial to make precise predictions of observable quantities from a given cosmological model .
the linear theory applies on very large scales @xcite .
however , the linear theory is not sufficiently accurate for purposes in the precision cosmology .
accurate predictions of statistical measures of galaxy clustering beyond the linear theory are provided by nonlinear theories .
the method of numerical simulation is one of the most straightforward ways of investigating nonlinear dynamics .
however , they are not free from numerical artefacts and systematics , such as finite - volume effects , finite - resolution effects , and so forth .
fortunately , the analytical perturbation theory is applicable on large scales where density fluctuations are small .
thus the nonlinear perturbation theory of gravitational instability attracts renewed interests in recent years .
the nonlinear perturbation theory have been developed since decades ago @xcite .
the traditional perturbation theory is formulated in eulerian space , and such theory is referred to as the standard perturbation theory ( spt ) .
the perturbation theory in lagrangian space is also formulated @xcite , which is called the lagrangian perturbation theory ( lpt ) .
the first - order lpt corresponds to the classic zeldovich approximation @xcite . in these years ,
the renormalized perturbation theory ( rpt ) @xcite and other approaches @xcite have been developed to improve the accuracy of perturbation theory in nonlinear regime , partially taking into account higher - order effects of the spt .
some of those approaches are based on the reformulation of fluid equations using the propagator , the vertex , and a source @xcite , which provides a way to use standard tools of field theory .
nevertheless , various levels of approximations and ansatz should be employed in those approaches .
the rpt and its variants mentioned above significantly improve the perturbation theory of dark matter in real space . however , one of the most important applications of the perturbation theory is to interpret the large - scale clustering of galaxies or other astronomical objects , observed by redshift surveys .
the observable quantity in redshift surveys is the distribution of objects in redshift space .
even though the rpt and its variants could be powerful in predicting the nonlinear power spectrum of dark matter in real space , one could not directly compare the theoretical prediction with observations .
there are two obstacles to the comparison between the improved perturbation theories and observations .
the first one is the redshift - space distortions : the redshift as a measure of the radial distance is contaminated by peculiar velocities .
it is straightforward to take them into account in the spt framework @xcite .
however , the spt in redshift space breaks down at larger scales than in real space and the applicability range of scales is fairly narrow @xcite , since the spt does not sufficiently reproduce the nonlinear smearing effects , known as the fingers - of - god ( fog ) effect @xcite .
nonlinear modelings of the redshift - space distortions beyond the spt are proposed @xcite , in which the fog effects are phenomenologically put by hand .
it is found in those studies that the fog effects can be represented by putting a gaussian damping factor in front of the power spectrum .
it is shown that the gaussian damping factor in redshift space is naturally derived from the lpt @xcite , where the phenomenological gaussian factor should be modified and additional mode - coupling terms should be taken into account in nonlinear redshift space . the second obstacle to the comparison between perturbation theories and observations is the biasing .
any astronomical objects which can be observed are biased tracers of underlying mass distributions . in the galaxy redshift surveys , the tracers are galaxies .
the exact relationship between the distribution of mass and that of galaxies depends on the complex , nonlinear process of galaxy formation which is not clearly understood
. analytic models of biasing have been proposed , including the model of local eulerian bias @xcite , the halo model @xcite , peaks model @xcite , etc .
the first model is a purely phenomenological parameterization of the bias , assuming a local relation between the mass and galaxy distributions .
the last two models are relatively more physical than the first one , and are categorized as the lagrangian bias , i.e. , the locations of the galaxy formation are specified in initial density fields . the location of a peak , for example
, is displaced by the dynamical evolution of density fluctuations . in the framework of spt , the local eulerian bias is usually adopted @xcite .
however , as shown below in this paper , the halo model and the peaks model are not compatible with the local eulerian bias in nonlinear regime , because the gravitational evolutions are nonlocal process in general . in fig .
[ fig : biases ] , the relation between the eulerian and lagrangian biases is shown .
the relation between eulerian and lagrangian biases .
the eulerian bias is expressible by the lagrangian bias ( dotted arrow ) and vice versa ( dashed arrow ) , only when the biases are allowed to be nonlocal .
note that nonlinear evolutions and formation of objects are nonlocal processes in nonlinear regime . ]
the final mass density field is the result of nonlinear evolutions from the initial mass density field .
the final number density field of objects , which are observable such as galaxies , is the result of nonlinear evolutions and formations of those objects .
the sites of the formation in lagrangian space define the initial number density field .
thus the initial number density field also depends on the redshift of observation .
the lagrangian bias corresponds to the relation between the initial mass density field and the initial number density field , while the eulerian bias corresponds to the relation between the final mass density field and the final number density field .
the initial number density field is constructed only when the formation process is known .
for example , the halo model and the peaks model give prescriptions of constructing the initial number density field from the initial mass density field ; i.e. , these models specify the lagrangian bias . in the halo model with the peak - background split @xcite , the initial number density
is locally determined by initial mass density field .
thus the halo model corresponds to a local lagrangian bias . in the peaks model ,
the sites of the formation is not locally determined , since the peaks are defined not only by local values , but also by spatial derivatives of the field @xcite .
thus the peaks model corresponds to a ( semi-)nonlocal lagrangian bias .
such models of lagrangian bias are more physically motivated than the eulerian bias .
the bias depends on the properties of observed objects , such as the mass and luminosity .
the lagrangian bias schemes described above actually depend on the mass ( or the peak hight ) of collapsed objects . the local lagrangian bias and the local eulerian bias are not compatible to each other in nonlinear regime .
this fact is well illustrated in fig .
[ fig : biases ] .
the eulerian bias is expressible by the lagrangian bias ( dashed arrow ) and vice versa ( dotted arrow ) .
the relations involves nonlinear evolutions of mass , and nonlinear evolutions and formation of observed objects .
nonlinear evolutions and formations are nonlocal processes . therefore , the eulerian bias should be nonlocal even when the lagrangian bias is local .
the local biasing schemes are compatible only in the case that the linear theory and a local approximation of formation process are valid .
such conditions do not apply in generally nonlinear regime . in reality , the bias is definitely nonlocal .
for example , the galaxies are largely affected by their environment @xcite .
the nonlocality should be taken into account in a precise modeling of the bias .
the exact forms of the nonlocality in formation processes of observed objects have not been fully understood .
it requires a lot of future work with analytic , numerical , and observational studies of complicated astrophysical processes to understand the exact nonlocality of bias .
one of the main purposes of this paper is to formulate a nonlinear perturbation theory which can handle both the eulerian and lagrangian nonlocal biases in general .
there mainly two merits in this formulation .
first , one can distinguish general properties of nonlinear clustering which do not depend on details of formation processes from those which largely depends on models of bias .
second , it is straightforward to predict observable quantities in any models of bias .
effects of redshift - space distortions and primordial non - gaussianity are also included in the formalism .
resummation techniques in the presence of nonlocal bias are introduced as well .
the resummations of bias are shown to be essential to handle the nonlocal bias in a general way .
this paper is organized as follows . in sec .
ii , both eulerian and lagrangian perturbation theories in real space are extended to include both the nonlocal bias and the primordial non - gaussianity .
diagrammatic methods with graphical representations are introduced .
the relation between the perturbative kernels of ept and lpt with nonlocal bias is derived as well . in sec .
iii , the formalism of the previous section is extended to include the effect of redshift - space distortions .
techniques of the vertex resummations in our formalism are introduced in sec .
iv , and finally , some models of the lagrangian bias are considered , illustrating how higher - order bias factors are evaluated in our formalism .
in the following , the density contrast of observed objects x at a comoving , eulerian position @xmath0 is denoted by @xmath1 . the observed objects x can be any astronomical objects such as galaxies , quasars , absorption lines , 21 cm emissions , and so forth , which are selected and catalogued in a given redshift survey .
the density field of the objects is not generally a local nor linear function of the underlying mass density contrast @xmath2 .
instead , they are nonlocally and nonlinearly related to each other in general . in other words ,
the density contrast @xmath3 of objects is a functional of the mass density contrast @xmath4 .
the functional relation between density fields of mass and objects is deterministic on scales we are interested in .
one might think that the formation process of objects is determined not only by the density field of mass , but also by other physical factors such as the local radiation density and its spectrum , merger histories of galaxies , etc . however , the dynamical evolutions in the structure formation are deterministic , and the initial density field uniquely determines all the subsequent states of the universe , including the above complex factors .
consequently , when only the growing mode solutions are kept , density contrasts of mass and objects are nonlocal functionals of initial density contrast @xmath5 : we have @xmath6 $ ] and @xmath7 $ ] , where @xmath8 and @xmath9 represent nonlocal functionals . on scales where the perturbation theory is applicable
, the motion of dark matter is single streaming . in which case , the spatial distribution of dark matter uniquely inverted to give the initial density field @xmath10 $ ] .
thus we have a deterministic functional relation @xmath11 = { \cal f}_{\rm x}[{\cal f}_{\rm m}^{-1}[\delta_{\rm m}]]$ ] .
taking into account the translational invariance , the taylor expansion of the functional is generally given by @xmath12 the nonlocal bias functions , \{@xmath13 } , specify the relation between the number density field of objects and mass density field . the first term with @xmath14 in eq .
( [ eq:1 - 1 ] ) gives just a constant @xmath15 , which is determined by other functions @xmath13 to ensure a condition @xmath16 .
the constant term @xmath15 is irrelevant as we consider connected moments of the density contrast . in fourier space
, moreover , the constant term just disappears when nonzero modes @xmath17 are considered .
therefore , we do not retain the constant term @xmath15 in the below .
the taylor expansion of eq .
( [ eq:1 - 1 ] ) is applicable only when the functional dependence is smooth . however , as shown in sec .
[ sec : vresum ] below , the technique of vertex resummation relaxes this constraint , and even singular dependences of the number density field on the mass density field can be handled with this technique .
the local biasing ansatz is recovered when @xmath18 is replaced by @xmath19 , where @xmath13 is now a constant for each @xmath20 , and @xmath21 is a smoothing kernel with a smoothing radius of @xmath22 . however , the eulerian local biasing scheme is not a natural model when we consider nonlinear dynamics , as we explicitly show below in this paper . in fourier space , eq .
( [ eq:1 - 1 ] ) reduces to @xmath23 where we use a notation , @xmath24 throughout this paper .
some variables like @xmath4 , @xmath3 and @xmath13 in fourier space are denoted by the same symbols as those in real space , instead of properly using symbols like @xmath25 , @xmath26 , @xmath27 , etc .
we will work in fourier space in most of this paper .
the convention of fourier transform and its inverse in this paper is given by @xmath28 since the bias relations should not depend on the coordinates system , the bias functions @xmath13 should be rotationally invariant . for example , the first - order bias function @xmath29 is actually a function of the magnitude @xmath30 and can be denoted as @xmath31 .
similarly , the second - order bias function @xmath32 is actually a function of @xmath33 , @xmath34 , and @xmath35 which characterize relative configuration of @xmath36 and @xmath37 .
thus the function can be denoted as @xmath38 , where the first two arguments are symmetric under a permutation , but the last one .
similarly , the third - order bias function can be denoted as @xmath39 , and higher - order bias functions depend only on rotationally invariant set of variables . in the case of local biasing ansatz ,
@xmath40 is replaced by @xmath41 , where @xmath13 is now a constant for each @xmath20 , and @xmath42 is a smoothing window function .
when the smoothing radius @xmath22 is much smaller than the clustering scales we are interested in , the smoothing window function can be dropped and each bias function is simply considered as a constant @xmath13 in fourier space . the mass density contrast @xmath43 is also a nonlocal and nonlinear functional of a linear density field @xmath44 .
we have a taylor expansion which has a similar form with eq .
( [ eq:1 - 2 ] ) in fourier space : @xmath45 where @xmath46 s are perturbative kernels .
we consider the density fields at any given redshift , and time - dependences are suppressed in the above notations .
for the linear density field , @xmath47 , where @xmath48 is the linear growth factor at a redshift @xmath49 , and @xmath50 is the linear density contrast at the present time @xmath51 .
we adopt a normalization @xmath52 in this paper .
the evolution of the mass density field can be evaluated perturbatively in quasi - nonlinear regime .
the spt evaluates the perturbative kernel @xmath46 in eulerian space order by order @xcite . for @xmath53 , for example , we have @xmath54 expressions of @xmath55 and @xmath56 are explicitly given in ref . @xcite .
although above kernels @xmath46 @xmath57 are exact only in the einstein de sitter universe , @xmath58 , @xmath59 , those are good approximations in other cosmological models with @xmath60 , @xmath61 @xcite . combining eq .
( [ eq:1 - 2 ] ) and eq .
( [ eq:1 - 3 ] ) , we have a formal expansion of the form @xmath62 where @xmath63 \nonumber\\ & \quad + b_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) , \label{eq:1 - 6c}\end{aligned}\ ] ] and so forth , where @xmath64 in each expression of @xmath65 .
the @xmath66-point polyspectrum @xmath67 of the field @xmath68 is defined by @xmath69 where @xmath70 denotes the cumulant , which corresponds to the connected part of the @xmath66-point expectation value .
the 2-point polyspectrum is the power spectrum @xmath71 .
the 3- and 4-point polyspectra are the bispectrum @xmath72 and the trispectrum @xmath73 , respectively . substituting the eq .
( [ eq:1 - 5 ] ) into eq .
( [ eq:1 - 7 ] ) , one can perturbatively evaluate the polyspectra @xmath67 in terms of the polyspectra @xmath74 of the linear density contrast @xmath5 , which is similarly defined by @xmath75 the linear polyspectra @xmath74 are proportional to the primordial spectra . when the initial density field is random gaussian ,
only the primordial power spectrum is present and higher - order linear polyspectra all vanish . in calculating the polyspectra ,
diagrammatic methods are quite useful .
[ fig : eptreal ] shows the diagrammatic rules for the eulerian perturbation theory in real space .
diagrammatic rules for the eulerian perturbation theory in real space . in each vertex , @xmath76 should be satisfied .
the upper and lower rules correspond to the expansions in eq .
( [ eq:1 - 2 ] ) and ( [ eq:1 - 3 ] ) , respectively . a dashed line should be `` internal '' : one end of a dashed line should be connected to a vertex with double solid line , and the other end should be connected to a vertex with solid lines . ]
the first rule corresponds to each term in the expansion of eq .
( [ eq:1 - 2 ] ) , and the second rule corresponds to each term in eq .
( [ eq:1 - 3 ] ) . in both vertices ,
a momentum conservation @xmath77 should be satisfied , according to the dirac s delta function in each corresponding equation .
the double solid line , dashed line , and single solid line correspond to @xmath3 , @xmath4 , and @xmath5 , respectively . since the variable @xmath4 is expanded according to eq .
( [ eq:1 - 3 ] ) , dashed lines in the upper rule should always be connected to the vertices of the lower rule , i.e. , the dashed lines are `` internal . '' to evaluate the @xmath66-point polyspectra of eq .
( [ eq:1 - 7 ] ) with expansions of eqs .
( [ eq:1 - 2 ] ) and ( [ eq:1 - 3 ] ) , we need cumulants of the linear density contrast @xmath5 , which are given by eq .
( [ eq:1 - 7 - 1 ] ) .
this procedure is diagrammatically equivalent to applying rules in fig .
[ fig : linspec ] .
diagrammatic rules for contributions from the primordial polyspectra .
all the free ends of solid lines in fig .
[ fig : eptreal ] should be connected to each other by these rules . when the initial density field is random gaussian , the lower graph with @xmath78 does not exist . in the lower graph
, @xmath79 should be satisfied .
the case of @xmath80 in the lower graph is equivalent to the upper graph . ]
the open circles with cross represent the primordial polyspectra .
when the initial density field is random gaussian , higher - order polyspectra all vanish , @xmath81 for @xmath82 , and only the upper rule in fig .
[ fig : linspec ] is relevant . for the evaluations of @xmath66-point polyspectra @xmath83 in eq .
( [ eq:1 - 7 ] ) , we first consider @xmath66 vertices with double solid lines .
next we consider possible ways of connecting those vertices with dashed lines and solid lines according to the rules in figs .
[ fig : eptreal ] and [ fig : linspec ] .
the polyspectra @xmath84 is given by the sum of terms which correspond to all the possible diagrams , with appropriate statistical factors which are explained at the end of this subsection . since the @xmath66-point polyspectra are defined by the connected part in eq .
( [ eq:1 - 7 ] ) , only connected diagrams should be taken into account .
discarding unconnected diagrams is equivalent to taking the connected part . when there exist internal wavevectors @xmath85 , which are not uniquely determined from external wavevectors @xmath86 , those internal wavevectors should be integrated with a weight of @xmath87 , i.e. , @xmath88 .
the number of internal wavevectors to be integrated is the same as the number of loops in a given diagram .
every terms in the perturbative expansion of the polyspectra @xmath67 of eq .
( [ eq:1 - 7 ] ) corresponds to the diagrams constructed by above rules .
when the hierarchical orders for the linear polyspectra @xmath89 hold , the number of loops is equal to the order of @xmath90 in a given diagram . in any case
, the perturbative order in a given diagram is apparent from the number and kind of crossed circles .
when the mixed polyspectra of different types of objects , such as cross power spectra @xmath91 , @xmath92 , etc .
, are need to be evaluated , we just use different vertices with corresponding set of bias functions .
when the bias is not present , @xmath93 , @xmath94 , only one dashed line can be connected to each vertex of the lower rule in fig .
[ fig : eptreal ] . in this case
one does not need to consider the dashed line at all , and the diagrammatic rules of figs .
[ fig : eptreal ] and [ fig : linspec ] are equivalent to the ones which were previously introduced in ref .
@xcite , in which gaussian initial conditions are assumed .
therefore , our diagrammatic rules of figs .
[ fig : eptreal ] and [ fig : linspec ] are generalization of the previous rules to the case when the eulerian nonlocal bias and primordial non - gaussianity are present in general . the diagrammatic rule for the expansion of eq . ( [ eq:1 - 5 ] )
is shown in fig .
[ fig : shrunkkernel ]
. shrunk vertex .
the triangle represents all the possible tree graphs constructed by the rules in fig .
[ fig : eptreal ] of ept .
the shrunk vertex can also be expressed by lpt diagrams . ]
the triangle vertex corresponds to shrinking the vertex in terms of the diagrammatic rules of fig .
[ fig : eptreal ] .
in fact , for @xmath95 , the shrunk vertices are diagrammatically given by fig .
[ fig : eptshrinkreal ] . shrunk vertices in eulerian perturbation theory in real space . ]
these diagrammatic representations are equivalent to eqs .
( [ eq:1 - 6a])([eq:1 - 6c ] ) . as an example , fig .
[ fig : oneloopps ] shows diagrams for the power spectrum up to the one - loop order in terms of shrunk vertices .
diagrams up to one - loops for the power spectrum . ]
it is a good exercise for readers to explicitly calculate the power spectrum from eqs .
( [ eq:1 - 5 ] ) and ( [ eq:1 - 7 ] ) , without resorting diagrams , and compare the result with the diagrammatic representation of fig .
[ fig : oneloopps ] .
the result is given by @xmath96 ^ 2 p_{\rm l}(k ) \\
+ \frac12 \int \frac{d^3k'}{(2\pi)^3 } \left[k_2(\bm{k}',\bm{k}-\bm{k}')\right]^2 p_{\rm l}(k ' ) p_{\rm l}(|\bm{k}-\bm{k}'| ) \\
+ k_1(k ) p_{\rm l}(k ) \int \frac{d^3k'}{(2\pi)^3 } k_3(\bm{k},\bm{k}',-\bm{k } ' ) p_{\rm l}(k ' ) \\
+ k_1(k ) \int \frac{d^3k'}{(2\pi)^3 } k_2(\bm{k}',\bm{k}-\bm{k } ' ) b_{\rm l}(\bm{k},\bm{k}'-\bm{k},-\bm{k } ' ) .
\label{eq:1 - 8}\end{gathered}\ ] ] one should be careful to put a correct statistical factor in each diagram .
when there are equivalent pieces in a diagram , one should put a statistical factor @xmath97 , where @xmath98 is the number of that equivalent pieces .
for example , the second diagram in fig .
[ fig : oneloopps ] has a pair of equivalent lines connecting both vertices .
it has a two equivalent pieces , and the resulting statistical factor is @xmath99 . in another way of viewing the statistical factor of this diagram
, each vertex has a factor @xmath100 because of the prefactor @xmath101 in eq .
( [ eq:1 - 5 ] ) , and there is two ways of connecting the solid lines from the vertices , resulting in the final statistical factor of @xmath102 , which agrees with the previous consideration of equivalent pieces . in the third diagram ,
two solid lines in the loop connected to the left vertex are equivalent pieces , and the statistical factor is @xmath103 . in another way of viewing ,
left vertex has a factor @xmath104 , and there are three ways of choosing which solid lines are associated to the loop , resulting in the final statistical factor of @xmath105 , which agrees with the previous consideration of equivalent pieces .
the fourth diagram gives the same expression with the third , due to the parity symmetry .
similarly , the fifth and sixth diagrams have the statistical factor of @xmath99 in both ways of viewing . after some experience
, one can put a correct statistical factor in a given diagram .
two ways of counting as in the above examples are helpful for cross - checking . in the lagrangian view ,
the dynamical evolution of cosmological density fields is tracked by a set of trajectories of mass element , @xmath106 , where @xmath107 is the initial lagrangian coordinates of the trajectory .
a displacement field @xmath108 is defined by @xmath109 and is considered as a fundamental quantity in the lagrangian view of perturbations .
since the initial density field is sufficiently uniform , the eulerian mass density field @xmath110 satisfies the continuity relation , @xmath111 where @xmath112 is the comoving mean density of mass . on the other hand ,
the fluid elements in which observed objects reside are not uniformly distributed in lagrangian space .
the continuity relation is given by @xmath113 where @xmath114 is the density field of the observed objects in lagrangian space .
note that both @xmath115 and @xmath116 depend on the time of observation , because objects are identified by observers at a given time . the comoving mean density of the objects , @xmath117 is common to both density fields , and the eq .
( [ eq:1 - 103 ] ) is equivalent to the following equation : @xmath118 \delta_{\rm d}^3[\bm{x}-\bm{q}-\bm{\varpsi}(\bm{q } ) ] .
\label{eq:1 - 104}\ ] ] when the density field is not biased , @xmath119 everywhere .
the expansion of the number density field by the linear density field in lagrangian space is formally given by @xmath120 where @xmath121 is the @xmath20-th order nonlocal bias function in lagrangian space .
the bias functions are essentially infinite - dimensional taylor coefficients , and given by functional derivatives : @xmath122 applying the fourier transform to the eq .
( [ eq:1 - 104 ] ) , and expanding the exponent of the displacement field , we have @xmath123 e^{-i\bm{k}\cdot\bm{\varpsi}(\bm{q } ) } - ( 2\pi)^3\delta_{\rm d}^3(\bm{k } ) \\
= \sum_{n+m\geq 1}^\infty \frac{(-i)^m}{n!m ! } \int \frac{d^3k_1}{(2\pi)^3}\cdots\frac{d^3k_n}{(2\pi)^3 } \frac{d^3k_1'}{(2\pi)^3}\cdots\frac{d^3k_m'}{(2\pi)^3 } \\
\times ( 2\pi)^3\delta_{\rm d}^3(\bm{k}_{1\cdots n}+\bm{k}'_{1\cdots m}-\bm{k } ) b^{\rm l}_n(\bm{k}_1,\ldots,\bm{k}_n ) \\ \times \delta_{\rm
l}(\bm{k}_1)\cdots\delta_{\rm l}(\bm{k}_n ) [ \bm{k}\cdot{\tilde{\bm{\varpsi}}}(\bm{k}'_1)]\cdots [ \bm{k}\cdot{\tilde{\bm{\varpsi}}}(\bm{k}'_m ) ] , \label{eq:1 - 105}\end{gathered}\ ] ] where we define @xmath124 above just for @xmath14 , and @xmath125 is the fourier transform of the displacement field . in the following we use the displacement field both in configuration space and fourier space , and we notationally distinguish between @xmath126 and @xmath125 . the displacement field @xmath126 is similarly expanded in lagrangian space : @xmath127 the evolution of the displacement field can be perturbatively evaluated in quasi - nonlinear regime .
the lpt evaluates the perturbative kernel @xmath128 order by order @xcite . for @xmath129
, we have @xmath130 , \label{eq:1 - 108b}\end{aligned}\ ] ] where @xmath131 for @xmath132 . as in eq .
( [ eq:1 - 4b ] ) , the above kernel @xmath132 is exact only for einstein de - sitter universe , and the expression is a good approximation in other cosmological models . the explicit form of the third - order kernel @xmath133 is given in @xcite .
the kernel @xmath128 is not proportional to @xmath134 for @xmath78 , in general .
diagrammatic rules for eqs .
( [ eq:1 - 105 ] ) and ( [ eq:1 - 107 ] ) are given in fig .
[ fig : lptdiag ] .
diagrammatic rules for the lagrangian perturbation theory in real space .
the symbols @xmath135 and @xmath136 are @xmath137-components of @xmath134 and @xmath128 , respectively .
the wavy lines are internal , and also carry wavevectors
. the momentum conservation should be satisfied at each vertex .
the diagrammatic rules in redshift space are just given by replacements @xmath138 . the double solid line and single solid line correspond to @xmath68 and @xmath5 , respectively .
the wavy line correspond to the @xmath20-th order perturbation of the displacement vector @xmath126 , and has an index @xmath137 which corresponds to a component of the vector .
the wavy lines also carry wavevectors , and momentum conservations should be satisfied at every vertices . combining eqs .
( [ eq:1 - 105 ] ) and ( [ eq:1 - 107 ] ) , we have a formal expansion which should be identical to the eq .
( [ eq:1 - 5 ] ) in ept .
the first several kernels are given by @xmath139 [ \bm{k}\cdot\bm{l}_1(\bm{k}_2 ) ] \nonumber\\ & \quad + b^{\rm l}_1(\bm{k}_1)[\bm{k}\cdot\bm{l}_1(\bm{k}_2 ) ] + b^{\rm l}_1(\bm{k}_2)[\bm{k}\cdot\bm{l}_1(\bm{k}_1 ) ] \nonumber\\ & \quad + b^{\rm l}_2(\bm{k}_1,\bm{k}_2 ) , \label{eq:1 - 109b}\\ k_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) & = \bm{k}\cdot\bm{l}_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) \nonumber\\ & \quad + \left\ { [ \bm{k}\cdot\bm{l}_1(\bm{k}_1 ) ] [ \bm{k}\cdot\bm{l}_2(\bm{k}_2,\bm{k}_3 ) ] + \mbox{cyc . }
\right\ } \nonumber\\ & \quad + [ \bm{k}\cdot\bm{l}_1(\bm{k}_1 ) ] [ \bm{k}\cdot\bm{l}_1(\bm{k}_2 ) ] [ \bm{k}\cdot\bm{l}_1(\bm{k}_3 ) ] \nonumber\\ & \quad + \left\ { b^{\rm l}_1(\bm{k}_1)[\bm{k}\cdot\bm{l}_2(\bm{k}_2,\bm{k}_3 ) ] + \mbox{cyc . }
\right\ } \nonumber\\ & \quad + \left\ { b^{\rm l}_1(\bm{k}_1)[\bm{k}\cdot\bm{l}_1(\bm{k}_2 ) ] [ \bm{k}\cdot\bm{l}_1(\bm{k}_3 ) ] + \mbox{cyc . }
\right\ } \nonumber\\ & \quad + \left\ { b^{\rm l}_2(\bm{k}_1,\bm{k}_2)[\bm{k}\cdot\bm{l}_1(\bm{k}_3 ) ] + \mbox{cyc . }
\right\ } \nonumber\\ & \quad + b^{\rm l}_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) \label{eq:1 - 109c}\end{aligned}\ ] ] and so forth , where @xmath64 for @xmath65 .
those equations are diagrammatically represented in fig .
[ fig : lptshrink ] . when the bias is not present , @xmath140 , @xmath141 , and @xmath142 for @xmath143 . in this case , the equivalence of eqs .
( [ eq:1 - 6a])([eq:1 - 6c ] ) and eqs .
( [ eq:1 - 109a])([eq:1 - 109c ] ) indicates @xmath144 [ \bm{k}\cdot\bm{l}_1(\bm{k}_2 ) ] , \label{eq:1 - 110b}\\
f_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) & = \bm{k}\cdot\bm{l}_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) \nonumber\\ & \quad + \left\ { [ \bm{k}\cdot\bm{l}_1(\bm{k}_1 ) ] [ \bm{k}\cdot\bm{l}_2(\bm{k}_2,\bm{k}_3 ) ] + \mbox{cyc . } \right\ } \nonumber\\ & \quad + [ \bm{k}\cdot\bm{l}_1(\bm{k}_1 ) ] [ \bm{k}\cdot\bm{l}_1(\bm{k}_2 ) ] [ \bm{k}\cdot\bm{l}_1(\bm{k}_3 ) ] , \label{eq:1 - 110c}\end{aligned}\ ] ] and so forth , where @xmath64 for @xmath46 .
the above equations give the relations of kernels in lagrangian and eulerian perturbation theories for dark matter evolutions .
therefore , those equations are identities .
it is easily seen that eqs .
( [ eq:1 - 110a ] ) , ( [ eq:1 - 110b ] ) explicitly hold for eqs .
( [ eq:1 - 4a ] ) , ( [ eq:1 - 4b ] ) , ( [ eq:1 - 108a ] ) , ( [ eq:1 - 108b ] ) . in the presence of bias ,
the equivalence of eqs .
( [ eq:1 - 6a])([eq:1 - 6c ] ) and eqs .
( [ eq:1 - 109a])([eq:1 - 109c ] ) indicates the relation between the eulerian and lagrangian bias .
first several relations are given by @xmath145 b^{\rm l}_1(\bm{k}_2 ) + ( \bm{k}_1 \leftrightarrow \bm{k}_2 ) \right\ } , \label{eq:1 - 111b}\\ b_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) & = b^{\rm l}_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) \nonumber\\ & \quad - \left [ f_2(\bm{k}_1,\bm{k}_2 ) b_2^{\rm
l}(\bm{k}_{12},\bm{k}_3 ) + \mbox{cyc . } \right ] \nonumber\\ & \quad + \left [ \bm{k}\cdot\bm{l}_1(\bm{k}_3 ) b_2^{\rm l}(\bm{k}_1,\bm{k}_2 ) + \mbox{cyc . }
\right ] , \nonumber\\ & \quad + \bigl [ f_2(\bm{k}_1,\bm{k}_2 ) f_2(\bm{k}_{12},\bm{k}_3 ) + \mbox{cyc . } \nonumber\\ & \hspace{7pc } - f_3(\bm{k}_1,\bm{k}_2,\bm{k}_3 ) \bigr]b_1^{\rm l}(\bm{k } ) \nonumber\\ & \quad - \left [ \bm{k}\cdot\bm{l}_1(\bm{k}_3 ) f_2(\bm{k}_1,\bm{k}_2 ) b_1^{\rm l}(\bm{k}_{12 } )
\ + \mbox{cyc.}\right ] \nonumber\\ & \quad + \bigl\ { \bigl[\bm{k}\cdot\bm{l}_2(\bm{k}_1,\bm{k}_2 ) + \bm{k}\cdot\bm{l}_1(\bm{k}_1 ) \bm{k}\cdot\bm{l}_1(\bm{k}_2 ) \nonumber\\ & \qquad\quad - \bm{k}\cdot\bm{l}_1(\bm{k}_{12 } ) f_2(\bm{k}_1,\bm{k}_2 ) \bigr ] b_1^{\rm l}(\bm{k}_3 ) + \mbox{cyc.}\bigr\ } , \label{eq:1 - 111c}\end{aligned}\ ] ] and so forth , where @xmath64 for @xmath13 .
an immediate consequence of the above formulas is that the biasing can not be local simultaneously both in eulerian and in lagrangian space , since the bias parameters are all constants in local bias models .
for example , when the lagrangian bias is local and parameters @xmath146 are scale - independent constants , the higher - order eulerian parameters @xmath147 are inevitably scale - dependent according to eqs .
( [ eq:1 - 111a])([eq:1 - 111c ] ) .
the reason for the incompatibility of local biases is that nonlinear evolutions are nonlocal process in general , as we already described in introduction . in spite of that
, local relations between the lagrangian halo bias and the local eulerian bias parameters are known in the halo model @xcite .
these relations are derived by applying the spherical collapse model @xcite , in which the density evolutions are local process , and therefore the local biases are compatible .
this compatibility does not hold in generally non - spherical collapse @xcite .
the purpose of this subsection is to find a relation between the bias parameters of general perturbation theory and that of spherical collapse model .
as mentioned above , the spherical collapse model is used in the halo approach to find a local relation between the ( lagrangian ) halo bias parameters and eulerian bias parameters .
following the same manner of ref .
@xcite , one can derive general relations between local bias parameters in lagrangian space and in eulerian space .
such relations are derived in appendix [ app : sphericalmodel ] . up to the third order
, we have @xmath148 where both the eulerian bias parameters @xmath149 and the lagrangian bias parameters @xmath150 are local and scale - independent at the same time . remarkably , the local bias relations in eqs .
( [ eq:1 - 113a])([eq:1 - 113c ] ) for the spherical model can be derived from eqs .
( [ eq:1 - 111a])([eq:1 - 111b ] ) when the lagrangian bias is local and eulerian bias functions are averaged over directions of wavevectors .
each bias function @xmath151 is replaced by a constant @xmath152 in the local lagrangian bias .
angular averages of eqs .
( [ eq:1 - 111a])([eq:1 - 111c ] ) in this case are straightforwardly calculated by using the explicit forms of @xmath153 , @xmath154 and @xmath132 in eqs .
( [ eq:1 - 4b ] ) , ( [ eq:1 - 108a ] ) and ( [ eq:1 - 108b ] ) .
the angular average of @xmath55 is given by @xmath155 @xcite .
as a result , the angular averages of eqs .
( [ eq:1 - 111a])([eq:1 - 111c ] ) exactly reduce to the right - hand sides of eqs .
( [ eq:1 - 113a])([eq:1 - 113c ] ) .
therefore , the scale- and angular - dependences of bias are neglected in the bias relations of the spherical collapse model , which are widely used in the halo approach .
it is only when those dependences of bias are not important that the local bias relations of eqs .
( [ eq:1 - 113a])([eq:1 - 113b ] ) are useful . in general perturbations without spherical symmetry , one should use the nonlocal relations of bias in eqs .
( [ eq:1 - 111a])([eq:1 - 111c ] ) . in the framework of local bias models ,
the deterministic property of bias is not viable in reality .
the number density of observed objects is not solely determined by the local density of mass .
the stochastic biasing scheme @xcite is a phenomenological model to treat the biasing as a nondeterministic process , in the framework of local bias models .
one of the characteristic parameters of stochasticity is the correlation coefficient , defined by @xmath156 where @xmath157 , @xmath158 are square roots of the variances of density contrasts @xmath159 , @xmath160 which are smoothed by a radius @xmath22 at a point in eulerian space . if the deterministic relation @xmath161 exactly holds , the stochasticity parameter @xmath162 is identically unity .
deviations of the stochasticity parameter from unity characterize how the stochasticity is important .
it is also common to define the correlation coefficient in fourier space , @xmath163 where @xmath91 is the cross power spectrum of mass and objects , @xmath164 and @xmath165 are the power spectra of mass and objects , respectively .
this approach is purely phenomenological in the sense that the stochasticity itself does not correspond to any fundamental physics .
instead , the stochasticity represents our ignorance on the formation process of observed objects .
dynamical evolutions of density field and the formation process of observed objects are deterministic at the fundamental level .
if we would know all the detailed physics of the formation process , any stochastic character should not appear when we properly describe precise dependences of the number density of objects on physical quantities , not only a local density of mass .
for the above reason , the parameters of stochastic bias should be derived from the nonlocal bias functions , @xmath40 . at linear order , the correlation coefficient in fourier space @xmath166 is identically unity , since the bias is multiplicative , @xmath167 , which is a consequence of the translational invariance . even in this case , the correlation coefficient in configuration space , eq .
( [ eq:1 - 201 ] ) , is less than unity in general when the linear bias parameter @xmath31 is scale - dependent @xcite . at nonlinear orders ,
the correlation coefficient even in fourier space becomes less than unity .
it is straightforward to calculate the eq .
( [ eq:1 - 202 ] ) in our framework of nonlocal biasing .
the lowest - order contribution to @xmath168 is given by one - loop diagrams .
the relevant diagrams are similar to fig .
[ fig : oneloopps ] , and the final result simply reduces to @xmath169 ^ 2 p_{\rm l}(k ) } \\ \times \int\frac{d^3k'}{(2\pi)^3 } \left[b_2(\bm{k}',\bm{k}-\bm{k}')\right]^2 p_{\rm l}(k ' ) p_{\rm l}(|\bm{k}-\bm{k}'| ) .
\label{eq:1 - 203}\end{gathered}\ ] ] this result is valid even when the initial density field is non - gaussian .
the bispectrum contributions to one - loop power spectra cancel out in the combination of eq .
( [ eq:1 - 202 ] ) . in a case of local lagrangian bias , including the halo bias , the lagrangian functions @xmath121 are constants and the second - order eulerian bias function of eq .
( [ eq:1 - 111b ] ) reduces to @xmath170 b^{\rm l}_1 .
\label{eq:1 - 204}\ ] ] in the large - scale limit , @xmath171 , we have @xmath172 , and from eq .
( [ eq:1 - 202 ] ) , @xmath173 ^ 2 p_{\rm l}(k ) } \int\frac{d^3k'}{(2\pi)^3 } \left[p_{\rm l}(k')\right]^2 .
\label{eq:1 - 205}\ ] ] it is interesting to note that the stochasticity emerges even in the large - scale limit when the second - order lagrangian bias parameter @xmath174 is nonzero .
one can not find such kind of properties in an original approach of the stochastic biasing , since the correlation coefficient is just a free parameter in the latter .
the halo model actually predicts the nonzero value of the second - order bias parameter @xmath174 ( see sec .
[ subsec : halobias ] below ) . to summarize this subsection ,
the stochastic bias is a phenomenology which is conveniently introduced in the context of local bias models , and is not needed in nonlocal bias models .
stochastic properties of the local bias are derived from the deterministic nonlocal bias .
the comoving position @xmath0 in real space and @xmath175 in redshift space are related by @xcite @xmath176 in the plane - parallel limit of the distant - observer approximation , where @xmath177 is the unit vector along the line of sight , and @xmath178 is the velocity component along @xmath177 .
the number density field in real space @xmath179 and that in redshift space @xmath180 are related by a continuity relation : @xmath181 therefore , the density contrast in redshift space is given by @xmath182j^{-1 } - 1 , \label{eq:2 - 23}\ ] ] where @xmath183 is the jacobian of the mapping from real space to redshift space .
one can easily calculate the fourier transform of eq .
( [ eq:2 - 23 ] ) : @xmath184 e^{ik_z u_z(\bm{x } ) } , \label{eq:2 - 24}\ ] ] where @xmath185 and @xmath17 is assumed . a different expression of the above formula is seen in the eq .
( 4 ) of ref . @xcite , and it can be shown by partial integration that the two expressions are actually equivalent .
the above relation is applicable even in the fully nonlinear regime . in fourier space , @xmath186 where @xmath187 is the fourier transform of the normalized velocity convergence , @xmath188 .
expanding the peculiar velocity field in eq .
( [ eq:2 - 24 ] ) , we have @xmath189 where we define @xmath190 above just for @xmath191 , and @xmath192 , @xmath193 are direction cosines of wavevectors .
it can be shown that this equation is equivalent to the eq .
( 5 ) of ref .
@xcite , although the apparent expressions are somewhat different . in the spt , only growing - mode solutions in each order are retained , and the peculiar velocity field is consistently assumed to be irrotational .
thus , the velocity field is fully characterized by the velocity divergence @xcite , which is expanded by a linear density contrast as @xmath194 where @xmath195 is the linear growth rate which corresponds to the logarithmic derivative of the linear growth factor .
the perturbative kernels @xmath196 are given by spt @xcite . for @xmath129 , we have @xmath197 where the expression @xmath198 is exact only in einstein de - sitter universe , and only weakly depends on time in general cosmology .
[ fig : eptred ] shows the diagrammatic rules for the eulerian perturbation theory in redshift space .
diagrammatic rules for the eulerian perturbation theory in redshift space . in the first rule , @xmath199 should be satisfied , and @xmath200 should be satisfied in the last two rules .
the dashed lines and dotted lines are internal . ]
the first rule corresponds to the expansion of eq .
( [ eq:2 - 26 ] ) , and the second rule corresponds to eq .
( [ eq:2 - 27 ] ) .
the rules are used in a similar way of those in real space .
the third rule is common to the rule in real space , and corresponds to eq .
( [ eq:1 - 3 ] ) .
the momentum conservation should satisfied in each vertex .
the meanings of double solid line , dashed line and single solid line are the same as in real space .
the dotted line corresponds to the velocity convergence @xmath201 . substituting the perturbative expansions of eqs .
( [ eq:1 - 3 ] ) and ( [ eq:2 - 27 ] ) into eq .
( [ eq:2 - 26 ] ) , we have a formal series of the biased field in redshift space : @xmath202 where @xmath203 s are derived kernels . for @xmath129
, we have @xmath204 .
\label{eq:2 - 30b}\end{aligned}\ ] ] in the case of local biasing where @xmath205 and @xmath206 are constants , these kernels are equivalent to the eqs .
( 11 ) and ( 12 ) in ref
. @xcite .
those equations are diagrammatically represented in fig .
[ fig : eptshrinkred ]
. shrunk vertices in eulerian perturbation theory in redshift space . ]
redshift - space distortions are naturally derived in the lagrangian picture .
the velocity of a mass element with lagrangian coordinates @xmath107 is given by a simple derivative of eq .
( [ eq:1 - 101 ] ) : @xmath207 where the dot represents a partial derivative by the proper time @xmath208 .
note that we suppress the argument of the time @xmath208 in our variables .
from eq .
( [ eq:2 - 21 ] ) , the displacement field in redshift space is simply given by @xmath209 the redshift - space distortions are exactly linear mappings of the displacement field , even in the nonlinear regime @xcite .
a similar equation has been applied to the analysis of the zeldovich approximation @xcite . in our approximation
that the perturbative kernels @xmath128 are independent on time , we have @xmath210 , where @xmath211 is the configuration - space counterpart of @xmath20-th order term in eq.([eq:1 - 107 ] ) . therefore , the time derivative of the displacement field is simply given by @xmath212 and the displacement field of each order in redshift space is related to the real - space displacement via @xmath213 which is just a linear mapping of the displacement field in each order .
this linear transformation is characterized by a redshift - space distortion tensor @xmath214 for each @xmath20 @xcite , which is defined by @xmath215 the eq .
( [ eq:2 - 53 ] ) reduces to @xmath216 , or in a vector notation , @xmath217 . as a result
, each perturbative kernel of the redshift - space displacement is given by a linear transformation from the real - space kernel , @xmath218 for @xmath129 , we have @xmath219 , \label{eq:2 - 56b}\end{aligned}\ ] ] where @xmath131 for @xmath220 .
all the formalism of sec .
[ subsec : lagrangianbias ] in real space applies with substitutions of @xmath221 and @xmath222 in redshift space .
corresponding to eqs .
( [ eq:1 - 109a ] ) and ( [ eq:1 - 109b ] ) , we have @xmath223 [ \bm{k}\cdot\bm{l}^{\rm s}_1(\bm{k}_2 ) ] \nonumber\\ & \quad + b^{\rm l}_1(\bm{k}_1)[\bm{k}\cdot\bm{l}^{\rm s}_1(\bm{k}_2 ) ] + b^{\rm l}_1(\bm{k}_2)[\bm{k}\cdot\bm{l}^{\rm s}_1(\bm{k}_1 ) ] \nonumber\\ & \quad + b^{\rm l}_2(\bm{k}_1,\bm{k}_2 ) , \label{eq:2 - 57b}\end{aligned}\ ] ] and so forth , where @xmath64 for @xmath203 . substituting eqs .
( [ eq:2 - 56a ] ) , ( [ eq:2 - 56b ] ) , and eqs .
( [ eq:1 - 111a ] ) , ( [ eq:1 - 111b ] ) , the above equations are equivalent to eqs .
( [ eq:2 - 30a ] ) , ( [ eq:2 - 30b ] ) .
thus , the relation of the nonlocal biases of eqs . ( [ eq:1 - 111a ] ) , ( [ eq:1 - 111b ] ) , which are derived in real space , also consistently applies in redshift space .
such consistency should hold for any higher - order kernels .
the formalism presented so far is a natural extension of the spt , simultaneously including the nonlocal bias , redshift - space distortions and primordial non - gaussianity . in recent years
, various methods beyond the spt are developed as mentioned in introduction . in the rpt @xcite
, the `` propagator '' plays an important role .
the concept of the propagator is extended to the `` multi - point propagator '' @xcite , in which the original propagator is identified as the one - point propagator . in this section ,
we show that the multi - point propagators can be obtained by resumming the external vertices in our formalism .
most of the resummation methods known so far are only applied to the dark matter clustering in real space .
exceptions are the lagrangian resummation method @xcite , in which local lagrangian bias and redshift - space distortions are included , and the time - renormalization - group method @xcite ) , in which the halo bias is included .
therefore , the identification of the multi - point propagators in our formalism is an important step toward including all the effects of nonlocal biasing , redshift - space distortions and primordial non - gaussianity into the resummation methods .
we first consider how the multi - point propagator @xcite is related to our formalism .
we only consider one - component propagators with the density sector .
the @xmath20-th order propagator @xmath224 of the density sector is defined by an ensemble average of the functional derivative : @xmath225 the original multi - point propagator @xmath226 with density and velocity sectors , defined in refs .
@xcite , is related to our definition of density propagator by @xmath227 , where @xmath228 is the linear growth factor at the initial time @xmath229 .
the appearance of the dirac s delta function on the right - hand side of eq .
( [ eq:3 - 1 ] ) is due to the translational symmetry . when the initial density field is random gaussian , the multi - point propagator of eq .
( [ eq:3 - 1 ] ) corresponds to the coefficient of orthogonal expansion by a series of generalized wiener - hermite functionals @xcite . substituting eq .
( [ eq:1 - 3 ] ) into the left - hand side of eq .
( [ eq:3 - 1 ] ) , we have @xmath230 the last factor is proportional to @xmath231 for the translational symmetry , and the remaining factor is further decomposed into products of connected polyspectra , @xmath232 .
for example , @xmath233 and so forth , where @xmath234 , @xmath235 , @xmath236 , @xmath237 , etc .
thus we have @xmath238 this equation is equivalent to eq . (
23 ) of ref .
@xcite when the last term is decomposed into connected parts .
the graphical representation of eq . ( [ eq:3 - 4 ] ) is similar to fig . 2 of ref .
@xcite .
let us extend the above multi - point propagator of mass density field to include the effects of biasing and redshift - space distortions .
we define the @xmath20-th order propagator @xmath239 of the objects x by @xmath240 substituting eq .
( [ eq:1 - 5 ] ) into the above equation , and following the same way of obtaining the eq .
( [ eq:3 - 4 ] ) , we have @xmath241 to include the redshift - space distortions , one simply apply the substitution @xmath242 in the above equation . the diagrammatic representation of this equation is given in fig . [ fig : multip ] .
the usage of the multi - point propagator is parallel to the one in refs .
the multi - point propagator corresponds to the summation of all the loops which are attached to each external vertex .
therefore , the polyspectra @xmath67 of eq .
( [ eq:1 - 7 ] ) are represented by using the multi - point propagators and corresponding diagrams do not have any loop which is attached to a single external vertex of the multi - point propagator .
for example , fig .
[ fig : pspec ] represents the power spectrum of the object for the gaussian initial conditions .
diagrammatic representation of the power spectrum with multi - point propagators when the initial density field is gaussian .
effects of biasing and redshift - space distortions are included in the multi - point propagators in our formalism . ]
the resulting power spectrum is given by @xmath243 as described in ref .
@xcite , it is important to note that each term in the sum is positive and the subsequent contributions add constructively .
the multi - point propagators are still difficult to be exactly evaluated .
one of the remarkable results in the rpt is a derivation of the propagators in the high-@xmath244 limit @xcite .
it is not trivial whether or not the same limit can be calculated in the presence of bias .
investigations on this line will be interesting for future work , and we leave them as an open question . instead of evaluating the full expression of the multi - point propagators , we consider partial resummations of external vertex in the lagrangian representation of the perturbation theory in this section .
as obviously seen in fig .
[ fig : multip ] , the multi - point propagator is essentially a renormalized external vertex .
first we consider partial resummation of the external vertex given in fig .
[ fig : resuml ] .
vertex resummation with displacement legs . ] in this figure , the gray ellipse represents all the possible graphs which are attached to a single external vertex with @xmath245 wavy lines .
the graphs in the gray ellipse are not necessarily connected , and can be disconnected as illustrated in the second line of the figure . according to the diagrammatic rules of fig .
[ fig : lptdiag ] and appropriate statistical factors , the corresponding factor of fig .
[ fig : resuml ] reduces to @xmath246 \cdots [ \tilde{\varpsi}_{j_r}(\bm{k}'_r ) ] \right\rangle \nonumber\\ & \hspace{7pc } \times b^{\rm l}_n(\bm{k}_1,\ldots,\bm{k}_n ) k_{i_1}\cdots k_{i_m } k_{j_1}\cdots k_{j_r } \nonumber\\ & \quad = \sum_{r=0}^\infty \frac{(-i)^r}{r ! } \left\langle \left [ \int\frac{d^3k'}{(2\pi)^3 } \bm{k}\cdot\tilde{\bm{\varpsi}}(\bm{k } ' ) \right]^r \right\rangle b^{\rm l}_n(\bm{k}_1,\ldots,\bm{k}_n ) k_{i_1}\cdots k_{i_m } \nonumber\\ & \quad = \left\langle e^{-i\bm{k}\cdot\bm{\varpsi } } \right\rangle b^{\rm l}_n(\bm{k}_1,\ldots,\bm{k}_n ) k_{i_1}\cdots k_{i_m } , \label{eq:3 - 30}\end{aligned}\ ] ] where @xmath247 is the displacement vector at the origin .
the factor @xmath248 is the characteristic function of the one - point distribution of the displacement field , where @xmath249 is the one - point probability function of the displacement field .
this characteristic function is a generating function of moments of @xmath126 at a single point in configuration space : @xmath250 the characteristic function is represented by a connected moments by the cumulant expansion theorem @xcite : @xmath251 \nonumber\\ & = \exp \left [ \sum_{n=1}^\infty \frac{(-i)^n}{n ! } k_{j_1}\cdots k_{j_n } \left\langle \varpsi_{j_1 } \cdots \varpsi_{j_n } \right\rangle_{\rm c } \right ] .
\label{eq:3 - 34}\end{aligned}\ ] ] in real space , the cumulants of the last line of the above equation are nonzero only when @xmath20 is an even number for the rotational symmetry , and have the form @xmath252 where the factor @xmath253 is the number of all the possible pairings among indices @xmath254 , and equals to the number of terms in the parenthesis .
when all @xmath255 take the same component , e.g. , @xmath49-axis , we have @xmath256 where @xmath201 is the polar angle of @xmath126 , and we used the fact that the variables @xmath257 and @xmath201 are independent and the directional cosine is randomly distributed in real space . comparing eqs .
( [ eq:3 - 35 ] ) and ( [ eq:3 - 36 ] ) , we have @xmath258 this equation can also be directly confirmed by contracting eq .
( [ eq:3 - 35 ] ) for each @xmath20 .
thus eq .
( [ eq:3 - 34 ] ) reduces to @xmath259 = \exp\left[\sum_{n=1}^\infty \frac{(-1)^n \langle |\bm{\varpsi}|^{2n } \rangle_{\rm c}}{(2n+1 ) ( 2n ) ! } k^{2n } \right ] .
\label{eq:3 - 37}\ ] ] when the higher - order cumulants of displacement field are negligible on large scales , the above equation reduces to a gaussian damping factor , @xmath260 therefore the large - scale power spectrum is smeared by nonlinear effects , and such smearing is important in the analysis of bao @xcite . the above eq .
( [ eq:3 - 37 ] ) , however , is valid only in real space where the clustering is statistically isotropic . in redshift space ,
the clustering is not statistically isotropic and eq .
( [ eq:3 - 34 ] ) should be evaluated with the displacement field @xmath261 of eq .
( [ eq:2 - 51 ] ) .
when the lowest - order ( zeldovich ) approximation is valid on large scales , and higher - order cumulants of the displacement field is negligible ( i.e. , primordial non - gaussianity is weak ) , we have again a gaussian damping factor , @xmath262\langle|\bm{\varpsi}|^2\rangle \right\ } , \label{eq:3 - 39}\ ] ] where @xmath263 is the direction cosine of the wavevector with respect to the lines of sight .
this damping factor represents both effects of nonlinear smearing and fog in redshift space @xcite , which are present even on large scales .
the damping factor of fog is similar to , but somewhat different from that adopted in a phenomenological modeling @xcite , @xmath264 , where @xmath265 is equal to @xmath266 at the linear order .
next we consider partial resummation of the external vertex given in fig .
[ fig : resumb ] . as in fig .
[ fig : resuml ] , the gray ellipse represents all the possible graphs which are attached to a single external vertex with @xmath245 solid lines .
the graphs in the ellipse are not necessarily connected .
the corresponding factor of fig .
[ fig : resumb ] reduces to @xmath267 where we define the renormalized nonlocal bias function in lagrangian space : @xmath268 this expression is contrasted with eq .
( [ eq:1 - 106 - 1 ] ) . instead of evaluating the functional derivatives at @xmath269 , taking statistical averages of them
gives the renormalized bias functions .
vertex resummation with bias legs . ] while the original function @xmath121 can be determined solely by a functional relation between biased field and linear density field in lagrangian space , the renormalized function @xmath270 depends on the statistical properties of the linear ( initial ) density field . in configuration space ,
( [ eq:3 - 41 ] ) is equivalent to @xmath271 where the translational invariance is taken into account .
we use the same symbols for variables both in fourier space and configuration space as long as the notation is obvious .
when a model of the lagrangian bias is provided in configuration space , the renormalized bias functions are evaluated by the above equation .
taylor expansions are possible only when the number density field @xmath272 is a smooth functional of the linear density field @xmath5 .
however , the renormalized bias function @xmath270 can be evaluated even when the biased field @xmath273 is not a smooth functional of @xmath5 and does not have a taylor expansion , since statistical average in eq .
( [ eq:3 - 41 ] ) is possible even when the functional derivative in the bracket is a singular functional .
we will see some examples below in which the bias can not be expanded by a taylor series while the renormalized bias functions are still well - defined . putting the diagrams of fig .
[ fig : resuml ] and fig .
[ fig : resumb ] together , we define the partial resummation of the external vertex in fig .
[ fig : resumlb ] .
partial resummation of the vertex .
] the gray box in this figure represents the partial resummation with all the sub - graphs which are attached to an external vertex with only wavy lines and with only solid lines .
the partial resummation of fig .
[ fig : resumlb ] results in the factor @xmath274 connected graphs with both wavy and solid lines attached to an external vertex are missed in this resummation .
for example , graphs like fig .
[ fig : nonresumex ] are not included in the factor of eq .
( [ eq:3 - 43 ] ) .
an example of graphs that are not resummed in eq .
( [ eq:3 - 43 ] ) or fig .
[ fig : resumlb ] . ] to obtain the full multi - point propagators , all kinds of graphs like the one in fig .
[ fig : nonresumex ] should be added .
for example , the first - order propagator @xmath275 is diagrammatically given by fig .
[ fig : propone ] up to one - loop contributions .
first - order propagator with partially resumed vertex up to one - loop contributions . ]
in the case of the local lagrangian bias , the linear density field in lagrangian space , @xmath276 is given by a single function of the smoothed mass density field @xmath277 at the same position , where @xmath22 is a smoothing radius . in configuration space
, we have @xmath278 where @xmath279 is generally a nonlinear , univariate function , and the smoothed mass density field in lagrangian space is given by @xmath280 where the window function @xmath281 is spherically symmetric . applying the taylor expansion and the fourier
transform to eq .
( [ eq:4 - 1 ] ) , the lagrangian bias functions @xmath121 in eq .
( [ eq:1 - 106 - 1 ] ) reduces to @xmath282 where @xmath283 is the @xmath20-th derivative of the function @xmath279 , and @xmath42 is a ( 3-dimensional ) fourier transform of the smoothing window function @xmath284 . the renormalized bias function @xmath270 of eq .
( [ eq:3 - 41 ] ) reduces to @xmath285 by denoting @xmath286 as the probability distribution function of @xmath287 , the first factor of the above equation is given by @xmath288 where @xmath289 . on scales which are larger than the smoothing scale , @xmath290 ( @xmath291 )
, one can ignore the window function @xmath42 in eqs .
( [ eq:4 - 3 ] ) and ( [ eq:4 - 4 ] ) .
in this case the lagrangian bias functions are constants : @xmath292 one notices that eq .
( [ eq:4 - 3 ] ) is well - defined only when the bias function @xmath279 is a smooth function , because the factor @xmath293 corresponds to a coefficient of the taylor series .
however , eq . ( [ eq:4 - 4 ] )
is applicable even when the bias function is not a smooth function . as an illustration
, we consider a threshold bias given by @xmath294 where a constant @xmath295 is a threshold value , @xmath296^{-1 } , \label{eq:4 - 7}\ ] ] and @xmath297 is the step function .
the taylor expansion of this function is not well - defined since all the derivatives at the origin are zero , @xmath298 ( @xmath299 ) , and we have @xmath300 from eq .
( [ eq:4 - 3 ] ) .
however , the expectation value in eq .
( [ eq:4 - 4 ] ) is not zero in this case , and using eq .
( [ eq:4 - 5 ] ) we have @xmath301 for @xmath299 . therefore , the concept of renormalized bias functions @xmath270 extends the applicability to the case when the simple taylor expansion of the bias in eq .
( [ eq:1 - 106 ] ) does not work . in the limit @xmath302 , the threshold bias can be considered as an approximation to the peaks bias in the high - peak limit .
lagrangian statistics in this limit have been widely studied in 1980 s @xcite .
the methods developed in those studies are essentially equivalent to applying the renormalized bias in this work to the case of local lagrangian bias . the use of the partially resummed vertex of eq .
( [ eq:3 - 43 ] ) in calculating loop - corrections to the power spectrum is equivalent to applying a recent formalism developed in ref .
@xcite in a case of local lagrangian biasing and gaussian initial conditions .
the present formalism is applicable even in cases of nonlocal lagrangian biasing with primordial non - gaussianity . in a local bias model ,
the number density field @xmath303 is a function of the single variable @xmath287 of eq .
( [ eq:4 - 2 ] ) .
we next consider the case in which the number density field is a multivariate function of variables which are convolutions of the linear density field .
the multiple variables @xmath304 , ( @xmath305 ) in lagrangian space are given by @xmath306 where @xmath307 s are the convolution kernels .
for example , if one of the variables @xmath304 is the linear gravitational potential @xmath308 in lagrangian space , the corresponding convolution kernel is given by @xmath309 .
if one of the variables is a derivative of a smoothed density field , say @xmath310 , the kernel is @xmath311 , and so forth .
the peaks bias is described by a function of the smoothed density field and its derivatives up to second order , @xmath312 @xcite . a multivariate bias model with two variables , @xmath287 and @xmath308 , is recently considered in the context of scale - dependent halo bias with primordial non - gaussianity @xcite . a multivariate eulerian bias model is also proposed @xcite .
in general , we consider the biased field @xmath313 is a local function of a finite number of the variables @xmath314 at the same lagrangian position . in configuration space , we have @xmath315 in this case , the lagrangian bias function of eq . ( [ eq:1 - 106 ] ) reduces to @xmath316 where @xmath317 is the fourier transform of @xmath318 .
the renormalized bias function of eq . ( [ eq:3 - 41 ] ) reduces to @xmath319 in the peaks model , for example , the variables @xmath304 contain spatial derivatives of a smoothed density field , @xmath320 , @xmath321 , and corresponding kernel windows are @xmath322 , respectively .
when one of the variables @xmath304 is given by the linear gravitational potential @xmath308 , the kernel window @xmath307 in fourier space is given by @xmath323 .
one of the most popular models of biasing in nonlinear structure formation is provided by the halo approach @xcite , which is based on the extended press - schechter theory @xcite .
the peak - background split is applied in this approach , and the bias of halos are considered as a local lagrangian bias .
the bias functions can be calculated by eq .
( [ eq:4 - 3 ] ) or eq .
( [ eq:4 - 4 ] ) in this case . in this subsection , we derive explicit expressions of the bias functions of halos .
the derivation is similar to that of ref .
@xcite , in which the gaussian initial conditions are assumed , and unfortunately the powers of growth factor in final expressions are incorrect .
below we correct the expressions of the last reference and give a derivation which applies even when the initial density field is non - gaussian in general .
the mass of halo is related to the lagrangian radius @xmath22 of a spherical cell by @xmath324 , or @xmath325^{1/3 } \,h^{-1}{\rm mpc}$ ] , where @xmath326 is the solar mass , and @xmath327 is the density parameter at the present time .
henceforth , @xmath328 denotes the variance of density fluctuations smoothed on a mass scale @xmath329 which is linearly extrapolated to the present time . according to the press - schechter theory and its extensions , the comoving number density of halos with a mass range @xmath330 around @xmath329 , identified at redshift @xmath49 ,
is given by @xmath331 where @xmath332 is the typical amplitude of fluctuations that produce those halos , @xmath333 , and @xmath334 is the critical overdensity for spherical collapse at the redshift @xmath49 . in the einstein - de sitter model ,
the critical overdensity is independent on redshift , @xmath335 , and only weakly depends on cosmological parameters and redshift in general cosmology .
since the condition of collapse is always expressed by the linearly extrapolated overdensity at the present time , the growth factor is absorbed into the critical overdensity @xmath336 .
the multiplicity function @xmath337 is normalized by @xmath338 to ensure all the mass in the universe is contained in halos in the limit @xmath339 . in the original press - schechter ( ps )
theory , the multiplicity function @xmath337 is given by @xmath340 the original ps mass function is improved by sheth and tormen ( st ) @xcite to give a better fit in numerical simulations of cdm - type cosmologies with gaussian initial conditions . the corresponding multiplicity function is given by @xmath341\sqrt{q}\,\nu e^{-q\nu^2/2 } , \label{eq:4 - 104}\ ] ] where @xmath342 , @xmath343 are numerically fitted parameters , and @xmath344^{-1}$ ] is the normalization factor .
the st mass function is applicable only for gaussian initial conditions .
when the non - gaussianity is present in the initial density field , the multiplicity function should have the correction factor @xcite .
in the extended ps theory , the number density of halos of mass @xmath329 , identified at redshift @xmath49 , in a region of lagrangian radius @xmath345 in which the linear overdensity extrapolated to the present time is @xmath346 , is given by @xcite @xmath347 where @xmath348^{1/2}},\ \sigma_0 = \sigma(m_0),\ m_0 = 4\pi\bar{\rho}r_0 ^ 3/3 .
\label{eq:4 - 106b}\ ] ] the halo of mass @xmath329 is collapsed at @xmath49 , while @xmath349 is assumed uncollapsed at @xmath51 , and thus we always have @xmath350 .
the conditional number density of eq .
( [ eq:4 - 105 ] ) represents the biasing for the lagrangian number density of halos .
the smoothed density contrast @xmath346 of mass modulates the number of halos .
the density contrast of halos in lagrangian space is given by @xmath351 since @xmath352 $ ] , we have @xmath353 this relation gives the function @xmath354 of eq .
( [ eq:4 - 1 ] ) , where the smoothing radius in eq .
( [ eq:4 - 1 ] ) correspond to @xmath355 here .
we assume the redshift @xmath49 of halo identification is the same as the redshift of halo observation . the smoothed linear density field in eq .
( [ eq:4 - 1 ] ) corresponds to @xmath356 here , because @xmath346 is the value extrapolated to the present time .
to evaluate the bias functions , the derivatives @xmath357 in eqs .
( [ eq:4 - 3 ] ) , ( [ eq:4 - 4 ] ) need to be derived .
we consider a limit of the peak - background split for simplicity , and assume @xmath358 ( however , see ref .
@xcite for limitations of this commonly used method . ) .
in this limit , we have @xmath359 for @xmath299 , where @xmath360 is the @xmath20-th derivative of the multiplicity function @xmath361 . the substitution
@xmath362 is equivalent to @xmath363 in the present limit . in the same limit ,
taking the statistical average over the distribution of @xmath346 also equivalent to substituting @xmath364 in eq .
( [ eq:4 - 109 ] ) , because the distribution function of @xmath346 is highly peaked at @xmath365 and its variance is much smaller than @xmath366 .
therefore , we have @xmath367 in the limit of peak - background split , @xmath358
. the lagrangian bias functions of eqs .
( [ eq:4 - 3 ] ) , ( [ eq:4 - 4 ] ) reduces to @xmath368 where the window function to define the background field @xmath346 is dropped , assuming the large - scale limit @xmath369 .
the last approximation is consistent with that of the peak - background split . the above eq .
( [ eq:4 - 111 ] ) is applicable even in non - gaussian initial conditions , as long as effects of non - gaussianity are taken into account in the multiplicity function . in the case of st mass function of eq .
( [ eq:4 - 104 ] ) with gaussian initial conditions , the above bias functions are given by @xmath370 , \label{eq:4 - 112a}\\ b^{\rm l}_2 & = \frac{1}{{\vardelta_{\rm c}}^2 } \left[q^2\nu^4 - 3 q\nu^2 + \frac{2p(2q\nu^2 + 2p -1)}{1
+ ( q\nu^2)^p}\right ] , \label{eq:4 - 112b}\end{aligned}\ ] ] and so forth .
essentially the same expressions are derived in ref .
@xcite . unfortunately , the factor @xmath371 is incorrectly missing in eq .
( 55 ) of ref .
accordingly , the factor @xmath336 which appear in eqs .
( 57)(59 ) of that paper should be replaced by @xmath372 . in the peaks formalism ,
a density peak in lagrangian space is considered as a location of structure formation @xcite . in the smoothed density field @xmath277 ,
the number density of peaks above a height @xmath373 is given by @xmath374 where @xmath375 , and @xmath376 is the smallest eigenvalue of the matrix @xmath377_{ij } = -\partial_i \partial_j \delta_r$ ] .
thus the number density of peaks is a multivariate function of a scalar @xmath287 , a vector @xmath378 , and a tensor @xmath379 at each position .
the number density of eq .
( [ eq:4 - 150 ] ) is a singular function .
therefore the taylor expansion can not be applied .
the unrenormalized bias functions @xmath13 are not well - defined in this case , and it is crucial to consider the renormalized bias functions of eq .
( [ eq:3 - 41 ] ) .
the peaks bias is one of the multivariate lagrangian bias as described in sec .
[ subsec : mlbias ] , and the renormalized bias function is given by eq .
( [ eq:4 - 53 ] ) where @xmath380 is a 10-dimensional vector .
since @xmath379 is a symmetric tensor , only six components of @xmath381 are independent . the corresponding kernels in eq .
( [ eq:4 - 50 ] ) are given by @xmath382 $ ] where @xmath383 .
the calculation of the coefficient @xmath384 for general @xmath20 in the peaks model is quite involved . in ref .
@xcite , the correlation function and the power spectrum of peaks are calculated up to second order in the case of gaussian initial condition . in this paper
, we do not derive explicit forms of the coefficients .
instead , we consider formal properties of the bias functions derived from the rotational symmetry below . for @xmath385 , we have @xmath386 , \label{eq:4 - 151}\end{gathered}\ ] ] where @xmath387 and @xmath388 . in the above equation
, the function @xmath279 is symmetrized with respect to the off - diagonal derivatives @xmath381 , and partial derivatives are taken as if @xmath381 and @xmath389 were independent when @xmath390 . from the rotational symmetry , the second term in the square parenthesis in eq .
( [ eq:4 - 151 ] ) identically vanishes .
the last term is proportional to @xmath391 , since @xmath392 .
thus , the scale dependence of the first - order bias function should have a form , @xmath393 where @xmath394 this form is exact for peaks bias models @xcite , and higher - order powers of @xmath395 with @xmath78 do not appear .
it is easily understood that the first bias function of peaks should have the form of eq .
( [ eq:4 - 152 ] ) .
the peaks are defined by up to second derivatives of the smoothed field , and thus the first bias functions of eqs .
( [ eq:4 - 52 ] ) , ( [ eq:4 - 53 ] ) involve only polynomials of wavevector @xmath134 up to second order . since the bias function is rotationally invariant , only the form of eq .
( [ eq:4 - 152 ] ) is allowed . for @xmath80 ,
the same considerations show that the bias function should have a form , @xmath396 , \label{eq:4 - 154}\end{gathered}\ ] ] where @xmath397 it is again easily understood that the second bias function of peaks should have the form of eq .
( [ eq:4 - 154 ] ) , since the function is rotationally invariant and made from polynomials of wavevectors @xmath36 and @xmath37 up to second order for each .
the explicit evaluations of the above coefficients are tedious . in ref .
@xcite , second - order biased correlation function with a gaussian initial condition is analytically calculated .
similar techniques should be also useful in our formalism , which we leave for future work .
in this paper , the standard nonlinear perturbation theory of the gravitational instability is extended in several directions .
one of the main extensions is the inclusion of the nonlocal bias , which is a general framework of biasing .
nonlocal biases both in eulerian and lagrangian spaces are formulated and consistently included in ept and lpt , respectively .
the nonlinear eulerian and lagrangian biases are compatible only in the framework of nonlocal bias .
the relations among perturbation kernels of ept and lpt with nonlocal biases are derived .
effects of redshift - space distortions and primordial non - gaussianity are also included in our formalism .
therefore , our formalism provides a complete theory to predict the observable quantities in redshift surveys , once a model of bias and cosmology are given .
the concept of vertex resummations in the presence of nonlocal bias is introduced .
we show that the vertex resummation of the bias extends the applicability of the formalism to the case when the bias function(al ) can not be expanded into a taylor series .
this extension is essential for handling , e.g. , the threshold bias and the peaks model , in which the bias involves singular functions such as the heaviside s step function , dirac s delta function , etc .
calculation of perturbative bias functions in our formalism is exemplified by considering some models of local and nonlocal models of lagrangian bias , such as the threshold bias model , the multivariate bias model , the halo model , and the peaks model .
the scale dependence of bias functions are straightforwardly obtained in our formalism once a model of nonlocal bias is given .
the scale dependence of eulerian bias arises both from the nonlocal lagrangian bias and nonlinear evolutions .
the formalism of the present paper provides a basic methodology for future applications of the perturbation theory .
for example , the scale - dependent bias in the presence of primordial non - gaussianity has been derived by adopting either halo models @xcite or the local eulerian bias @xcite .
our formalism allows to calculate the scale - dependent bias in any models of bias in a consistent manner @xcite .
the power spectrum with bao in any given models of bias can be calculated with our formalism .
the result of applying our formalism to a local lagrangian bias is equivalent to the work in ref .
more precise modeling of the bias would be required in future analysis of the bao in the galaxy power spectrum to constrain the nature of dark energy .
our formalism provides a way to perturbatively calculate the nonlinear power spectrum , bispectrum , trispectrum , and other polyspectra .
these polyspectra are fundamental statistics and any other statistics in the large - scale structure , such as the correlation functions , counts - in - cells , genus statistics , etc
. are expressible by these polyspectra , in principle .
the formalism developed in this paper would have an essential importance in the era of precision cosmology with the large - scale structure of the universe .
i wish to thank r. sheth for helpful discussion .
i acknowledge support from the ministry of education , culture , sports , science , and technology , grant - in - aid for scientific research ( c ) , 21540263 , 2009 , and grant - in - aid for scientific research on priority areas no .
467 `` probing the dark energy through an extremely wide and deep survey with subaru telescope . ''
this work is supported in part by jsps ( japan society for promotion of science ) core - to - core program `` international research network for dark energy . ''
in this appendix , the relations between the eulerian and lagrangian bias parameters in the spherical collapse model are derived , following and generalizing the argument of ref .
the einstein
de sitter universe is assumed in the following equations for simplicity .
the dependences of the results on cosmological parameters are weak .
the time evolution of proper radius @xmath245 for a spherical mass shell as a function of the scale factor @xmath398 is given by a parametric form @xcite @xmath399 where @xmath400 and @xmath401 are initial values of @xmath245 and @xmath398 , respectively , and @xmath402 is the initial density contrast at @xmath401 . for @xmath403 ,
the parameter @xmath201 is a positive real number . for @xmath404
, the replacement @xmath405 should be applied to have the parameter real and positive .
the comoving radius is given by @xmath406 , and therefore an overdensity @xmath407 of any kind in the spherical volume is enhanced by a factor of @xmath408 .
thus , the density contrasts of mass @xmath4 and of biased object @xmath3 in the spherical volume are given by @xmath409 in the limit of @xmath410 , and @xmath411 , where @xmath272 is the density contrast in lagrangian space .
thus , we have @xmath412 the form of eq .
( [ eq : a-2a ] ) is well - known @xcite .
( [ eq : a-2b ] ) can also be derived from general eqs .
( [ eq:1 - 102 ] ) and ( [ eq:1 - 103 ] ) in the case of spherical perturbations .
the linear density contrast @xmath5 is proportional to the scale factor @xmath398 , and from eq .
( [ eq : a-1b ] ) , we have @xmath413 for @xmath414 , the relation between @xmath4 and @xmath5 is derived as a power series by taylor expansions of eqs .
( [ eq : a-2a ] ) and ( [ eq : a-3 ] ) with respect to the parameter @xmath201 .
the results are @xmath415 the eq .
( [ eq : a-4a ] ) is derived in ref .
although the eq .
( [ eq : a-4b ] ) is described in refs .
@xcite , they put incorrect numbers in the coefficients of third- and fourth - order terms . the coefficients @xmath416 and @xmath417 in their eq .
( a4 ) of ref .
@xcite should be replaced by @xmath418 and @xmath419 .
the dynamical evolutions are local in the spherical collapse model . in this special case
, both the eulerian and lagrangian biases can be simultaneously local .
we have expansions @xmath420 where @xmath13 and @xmath121 are constant bias parameters .
putting eqs .
( [ eq : a-2b ] ) , ( [ eq : a-4b ] ) , ( [ eq : a-5b ] ) together , we have a series expansion of @xmath3 in terms of @xmath421 . equating the resulting coefficients with eq .
( [ eq : a-5a ] ) gives @xmath422 since the relation of eq .
( [ eq : a-4b ] ) depends only very weakly on cosmological model @xcite , the relations of eqs .
( [ eq : a-6a])([eq : a-6d ] ) also do so . c. alcock and b. paczynski , , * 281 * , 358 ( 1979 ) .
w. e. ballinger , j. a. peacock , and a. f. heavens , mon . not .
r. astron .
* 282 * , 877 ( 1996 ) .
t. matsubara and y. suto , astrophys .
j. letters , * 470 * , l1 ( 1996 ) .
n. dalal , o. dor , d. huterer and a. shirokov , , 77 , 123514 ( 2008 ) .
s. matarrese and l. verde , astrophys .
j. letters , 677 , l77 ( 2008 ) .
a. slosar , c. hirata , u. seljak , s. ho , n. padmanabhan , j. cosmol .
8 , 31 ( 2008 ) .
a. taruya , k. koyama and t. matsubara , , * 78 * , 123534 ( 2008 ) . v. desjacques , u. seljak and i. t. iliev , mon . not .
r. astron .
soc . , * 396 * , 85 ( 2009 ) .
r. juszkiewicz , mon . not .
r. astron .
* 197 * , 931 ( 1981 ) .
e. t. vishniac , mon . not .
r. astron .
* 203 * , 345 ( 1983 ) .
j. n. fry , , * 279 * , 499 ( 1984 ) .
m. h. goroff , b. grinstein , s .-
rey , and m. b. wise , , * 311 * , 6 ( 1986 ) .
y. suto , m. sasaki , , * 66 * , 264 ( 1991 ) .
n. makino , m. sasaki , and y. suto , , * 46 * , 585 ( 1992 ) .
b. jain and e. bertschinger , , * 431 * , 495 ( 1994 ) .
r. scoccimarro and j. frieman , astrophys . j. suppl
. ser . , 105 , 37 ( 1996 ) .
f. bernardeau , s. colombi , e. gaztaaga , and r. scoccimarro , phys .
, * 367 * , 1 ( 2002 ) t. buchert , astron .
astrophys . , * 223 * , 9 ( 1989 ) .
f. moutarde , j .-
alimi , f. r. bouchet , r. pellat , and a. ramani , , * 382 * , 377 ( 1991 ) .
t. buchert , mon . not .
r. astron .
, * 254 * , 729 ( 1992 ) .
t. buchert and j. ehlers , mon . not .
r. astron .
, * 264 * , 375 ( 1993 ) .
t. buchert , mon . not .
r. astron .
soc . , * 267 * , 811 ( 1994 ) .
e. hivon , f. r. bouchet , s. colombi and r. juszkiewicz , astron .
, * 298 * , 643 ( 1995 ) .
p. catelan , mon . not .
r. astron .
* 276 * , 115 ( 1995 ) .
p. catelan and t. theuns , mon . not .
r. astron .
* 282 * , 455 ( 1996 ) .
j. ehlers and t. buchert , general relativity and gravitation , * 29 * , 733 ( 1997 ) . m. crocce and r. scoccimarro , , * 73 * , 063519 ( 2006 ) . m. crocce and r. scoccimarro , , * 73 * , 063520 ( 2006 ) .
p. mcdonald , , * 75 * , 043514 ( 2007 ) .
p. valageas , astron . astrophys . ,
* 465 * , 725 ( 2007 ) .
s. matarrese and m. pietroni , j. cosmol .
, 0706 ( 2007 ) 026 . s. matarrese and m. pietroni , mod .
a , * 23 * , 25 ( 2008 ) a. taruya and t. hiramatsu , , * 674 * , 617 ( 2008 ) . m. pietroni , j. cosmol .
* 10 * , 36 ( 2008 ) t. matsubara , , * 77 * , 063530 ( 2008 ) r. scoccimarro , ann .
* 927 * , 13 ( 2001 ) .
n. kaiser , mon . not .
r. astron .
soc . , * 227 * , 1 ( 1987 ) . a. j. s. hamilton , astrophys .
j. letters , * 385 * , l5 ( 1992 ) .
a. f. heavens , s. matarrese , and l. verde , mon . not .
r. astron .
* 301 * , 797 ( 1998 ) .
r. scoccimarro , h. m. p. couchman , and j. a. frieman , , * 517 * , 531 ( 1999 ) h. j. mo and s. d. m. white , mon . not .
r. astron .
* 282 * , 347 ( 1996 ) .
h. j. mo , y. p. jing , and s. d. m. white , mon . not
r. astron .
* 284 * , 189 ( 1997 ) .
r. k. sheth and g. tormen , mon . not .
r. astron .
, * 308 * , 119 ( 1999 ) .
r. scoccimarro , r. k. sheth , l. hui , and b. jain , , * 546 * , 20 ( 2001 ) .
a. cooray and r. sheth , phys .
, * 372 * , 1 ( 2002 ) .
n. kaiser , astrophys .
j. letters , * 284 * , l9 ( 1984 ) .
m. davis , g. efstathiou , c. s. frenk , and s. d. m. white , , * 292 * , 371 ( 1985 ) .
j. bardeen , j. r. bond , n. kaiser , and a. s. szalay , , * 304 * , 15 ( 1986 ) .
e. regos and a. s. szalay , mon . not .
r. astron .
* 272 * , 447 ( 1995 ) .
v. desjacques , , * 78 * , 103503 ( 2008 ) .
v. desjacques and r. k. sheth , , * 81 * , 023526 ( 2010 ) .
v. desjacques , m. crocce , r. scoccimarro and r. k. sheth , , * 82 * , 103529 ( 2010 ) .
a. dekel and o. lahav , , * 520 * , 24 ( 1999 ) .
t. matsubara , , * 525 * , 543 ( 1999 ) .
t. matsubara , , * 535 * , 1 ( 2000 ) .
a. n. taylor and a. j. s. hamilton , mon . not .
r. astron .
soc . , * 282 * , 767 ( 1996 ) f. bernardeau , m. crocce and r. scoccimarro , , * 78 * , 103521 ( 2008 ) f. bernardeau , m. crocce and e. sefusatti , , * 82 * , 083507 ( 2010 ) t. matsubara , , * 78 * , 083519 ( 2008 ) ; * 78 * , 109901(e ) ( 2008 ) a. elia , s. kulkarni , c. porciani , m. pietroni and s. matarrese , arxiv:1012.4833 . t. matsubara , astrophys . j. suppl .
ser . , * 101 * , 1 ( 1995 ) s .- k .
ma , _ statistical mechanics
_ ( world scientific , singapore , 1985 ) , sec .
12 m. crocce and r. scoccimarro , , * 77 * , 023533 ( 2008 )
. h. d. politzer , and m. b. wise , astrophys .
j. letters , * 285 * , l1 ( 1984 )
. l. g. jensen , and a. s. szalay , astrophys .
j. letters , * 305 * , l5 ( 1986 ) .
s. matarrese , f. lucchin , and s. a. bonometto , astrophys .
j. letters , * 310 * , l21 ( 1986 ) .
p. mcdonald , , * 78 * , 123519 ( 2008 ) .
m t. giannantonio , and c. porciani , , * 81 * , 063530 ( 2010 ) .
t. baldauf , u. seljak , and l. senatore , arxiv:1011.1513 ( 2010 ) .
p. mcdonald and a. roy , j. cosmol .
* 8 * , 20 ( 2009 ) .
w. h. press and p. schechter , , * 187 * , 425 ( 1974 ) . r. i. epstein , mon . not .
r. astron .
* 205 * , 207 ( 1983 ) .
j. a. peacock and a. f. heavens , mon . not .
r. astron .
* 243 * , 133 ( 1990 ) .
r. g. bower , mon . not .
r. astron .
soc . , * 248 * , 332 ( 1991 ) .
j. r. bond , s. cole , g. efstathiou , and n. kaiser , , * 379 * , 440 ( 1991 ) .
s. matarrese , l. verde and r. jimenez , , 541 , 10 ( 2000 ) .
m. lo verde , a. miller , s. shandera and l. verde , j. cosmol .
phys . , 4 , 14 ( 2008 ) . | the standard nonlinear perturbation theory of the gravitational instability is extended to incorporate the nonlocal bias , redshift - space distortions , and primordial non - gaussianity .
we show that local eulerian bias is not generally compatible to local lagrangian bias in nonlinear regime .
the eulerian and lagrangian biases are nonlocally related order by order in the general perturbation theory .
the relation between eulerian and lagrangian kernels of density perturbations with biasing are derived .
the effects of primordial non - gaussianity and redshift - space distortions are also incorporated in our general formalism , and diagrammatic methods are introduced .
vertex resummations of higher - order perturbations in the presence of bias are considered .
resummations of lagrangian bias are shown to be essential to handle biasing schemes in a general framework . |
we consider a thin circular washer with inner radius @xmath10 and outer radius @xmath11 .
the washer is broken into a set of concentric loops with width and turn - to - turn spacing @xmath12 .
the vector potential @xmath13 at radius @xmath14 and produced by a current @xmath15 through a loop at radius @xmath16 is given by @xmath17 the differential term @xmath18 has magnitude @xmath19 in the direction parallel to the radius vector of @xmath14 and @xmath20 in the tangential direction . upon integration
, only the tangential direction is non - zero .
the flux enclosed in a loop at radius @xmath14 is given by @xmath21,\end{aligned}\ ] ] where we have defined @xmath22 and where @xmath23 and @xmath24 are the complete elliptic integrals of the first and second kind , respectively .
the functional dependence of @xmath25 is plotted in fig . 1 . versus normalized distance @xmath26 .
] the self and mutual inductances between loops @xmath27 at radius @xmath14 and @xmath28 at radius @xmath16 are given by @xmath29/r)+a_c([x - w/2]/r)}{2 } \ , \end{aligned}\ ] ] where we have approximately averaged over the width of the loop by evaluating @xmath25 at @xmath30 .
the currents in loops 1 to @xmath31 are described by a vector @xmath32 ; the induced flux @xmath33 in the loops is then given by @xmath34 in the case of a superconducting washer with vanishingly small london penetration depth @xmath35 , the flux quantization condition @xmath36 requires a constant distribution of fluxes through the washer .
the current distribution is determined by setting @xmath37 for all loops and multiplying by the inverse of the mutual inductance matrix : @xmath38
a single vortex trapped in the squid washer induces an effective flux offset in the squid .
the magnitude of this effect can be understood by considering a few simple examples .
first , for a vortex trapped in an infinite superconducting sheet , the phase change between points above and below the vortex is @xmath39 , corresponding to a @xmath40 flux shift . by symmetry , this phase change is unchanged if the vortex is centered in a superconducting strip of finite width .
a vortex in a squid with a narrow washer @xmath41 will similarly produce a flux offset of @xmath40 in the squid if the vortex is placed in the center of the wire at @xmath42 . for a vortex placed near the inner radius @xmath10
, the flux offset will approach @xmath43 , whereas a vortex placed near the outer radius @xmath11 will give a flux offset approaching @xmath44 . for vortices
trapped uniformly throughout the narrow washer , the average flux offset will be @xmath40 . to calculate the coupling between a squid and a vortex trapped in the squid washer at radius @xmath16 , we consider a radially symmetric geometry that is amenable to the calculation technique described above . from symmetry , the flux coupling is independent of angular orientation of the vortex around the washer . thus , the flux coupling from one vortex is equivalent to @xmath45 vortices of magnitude @xmath46 placed around the washer with uniform angular spacing and identical radius . as @xmath47 ,
the vortices are equivalent to a cut in the washer with a net flux through the cut of magnitude @xmath5 .
this problem can now be solved using the matrix inversion technique of eq .
[ eq : invert ] . for the solution
, we need to calculate the effective flux @xmath48 that must be applied to the squid so that the total circulating current is unchanged when a flux quantum is applied to the cut .
the flux profile in the loops is thus @xmath49 the offset flux @xmath48 is solved using eq .
[ eq : invert ] with the constraint that the total circulating current @xmath50 is zero .
the solution of this problem @xmath51 is shown in fig . 3 for both a narrow and wide washer . by solving for @xmath51 over the range @xmath52 and integrating over the area of the squid , we compute the average flux @xmath53 coupled to the squid from a vortex with a uniform distribution across the washer : @xmath54 numerical calculations for a circular washer show that @xmath55 for a narrow washer , as expected ; for our device dimensions ( inner radius 100 @xmath2 m and outer radius 500 @xmath2 m ) , we find a coupling factor @xmath56 . of a vortex to a squid : ( blue )
@xmath57 = 1.02 ; ( red ) @xmath57 = 5 .
the average flux offset for the thin washer is @xmath58 , whereas for the large washer we find @xmath59 . ]
what is the physical source of the offset flux ?
vortex currents , although localized mostly around the core , fall off over a long distance .
part of the current must then circulate around the center hole of the squid washer , giving an effective flux offset to the squid loop .
the total magnetic flux through a vortex is quantized , and has contributions from magnetic fields generated by the currents circulating around the vortex along with any externally applied magnetic field .
this external magnetic field is typically considered to be zero because of screening from the superconductor . however ,
the surface spins located within the vortex are strongly coupled to the vortex , so their magnetic flux can not be neglected .
as these spins become polarized , they couple a flux @xmath60 to the vortex ; the quantization condition requires a redistribution of currents in the vortex .
the change in average flux coupled to the squid is thus @xmath61 per vortex for the thick washer geometry consider previously . to calculate @xmath62 we rely on reciprocity
: if we know the current distribution in the vortex , we can calculate the magnetic fields at all points in space .
this allows us to calculate the coupling to a spin at an arbitrary location ; the total flux is obtained by integrating over the area of the vortex .
we first discuss calculation of the current distribution in the vortex , then we describe the coupling of the vortex to surface spins .
we model the vortex as a superconducting washer with inner radius @xmath63 and outer radius @xmath64 , where @xmath65 is the coherence length .
the current distribution is calculated using the techniques described in section i. in this case , however , we have a penetration depth @xmath35 that is larger than the inner dimensions of the washer , and flux is not constant across the superconductor .
instead , single - valuedness of the superconducting wavefunction imposes the following quantization condition : @xmath66 where the integral on the right - hand side is london s fluxoid .
the new term adds additional elements on the diagonal of the inductance matrix : @xmath67/r\right)+a_c\left(\left[x - w/2\right]/r\right)}{2 } \nonumber\\ & + \delta_{ij}\ , 2\pi \mu_0 x \left(\frac{\lambda_l^2}{dw}\right ) , \label{eq : mlon}\end{aligned}\ ] ] where @xmath68 is the kronecker @xmath69-function and @xmath70 is the film thickness .
inversion of the matrix @xmath71 allows determination of the current profile in the vortex .
a two - dimensional solution of the vortex shows that , for the materials parameters we are interested in here , the currents are uniform along the @xmath72 direction to within about 10@xmath73 .
therefore we may simplify the matrix inversion part of the calculation by using 1-d discretization and replacing @xmath74 in eq .
[ eq : mlon ] by @xmath75 , where @xmath76 is the thin - film penetration depth . in fig .
4 we plot the current and flux distributions in the vortex for @xmath65 = 30 nm and @xmath77 = 100 nm .
= 30 nm and @xmath77 = 100 nm . ] the vortex self - inductance @xmath78 is easily determined from the total current @xmath79 in the washer : @xmath80 for @xmath65 = 30 nm and @xmath77 = 100 nm , we find a vortex self - inductance @xmath78 = 0.24 ph .
once the currents in the vortex are known , it is straightforward to calculate the in - plane @xmath81 and out - of - plane @xmath82 magnetic fields at all points in space by summing over the contributions @xmath83 of the individual loops of current @xmath27 . for a current loop with radius @xmath84 centered at @xmath85 in the @xmath86 plane
, we have @xmath87 where now we have @xmath88 and where again @xmath23 and @xmath24 are the complete elliptic integrals of the first and second kind , respectively . and out - of - plane @xmath89 magnetic fields for a vortex with @xmath65 = 30 nm and @xmath77 = 100 nm . here
we consider a film with thickness @xmath70 = 80 nm and evaluate the fields on the surface of the film , assuming a uniform current density throughout the thickness of the film . ]
we now consider magnetic coupling between a vortex and a surface spin located at radius @xmath16 .
we first think of the spin as an infinitesimal current loop with effective sensing area @xmath90 . the flux coupled to the spin by the vortex
is thus @xmath91 where @xmath92^{1/2}$ ] is the magnitude of the field at @xmath16 .
note that the geometrical coupling of the spin to the vortex is unity , since the spin is polarized in the direction of the local field .
the mutual inductance of the spin to the vortex is then @xmath93 the flux coupled to the vortex by a spin at @xmath16 with moment @xmath94 is thus @xmath95 the total flux coupled to the vortex by spins with surface density @xmath8 is therefore @xmath96 where we have introduced the effective spin polarization
@xmath97 , defined as @xmath98 from the known current distribution in the vortex we can calculate @xmath97 .
note that in eq .
[ eq : peff ] we are also integrating the current over the thickness of the washer .
because we have a uniform current distribution in @xmath72 , we average the fields over the thickness of the washer .
as the squid is cooled to millikelvin temperatures , a change in temperature yields a change in effective spin polarization in the vortex , which in turn induces a change in flux coupled to each vortex . the total flux change @xmath99 at the squid is obtained by summing over all vortices , taking into account the coupling factor 0.14 .
we find @xmath100 where @xmath101 is the area of the squid washer , and where @xmath102 is the density of vortices in the squid washer . for our field - cool experiments , we have @xmath103 , where @xmath1 is the magnitude of the cooling field . in these experiments ,
we see a linear dependence of flux drift @xmath104 on vortex density .
a linear fit to the data of fig .
1b yields a slope @xmath105/mt .
this slope is related to spin density as follows : @xmath106 where @xmath107 . for our materials parameters , we find @xmath108 = 0.037 .
using @xmath101 = 0.96 mm@xmath109 and the slope @xmath7 = 1.3 @xmath5/mt from fig .
1b , we extract a spin density @xmath8 = @xmath110 m@xmath111 .
this density of surface spins is compatible with densities considered in recent theoretical models of @xmath9 flux noise from surface spins @xcite .
we acknowledge useful discussions with l. faoro , l.b .
ioffe , b.l.t .
plourde , and c.c .
some devices were fabricated at the ucsb nanofabrication facility , part of the nsf - funded nnin .
this work was supported in part by the u.s .
the views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies , either expressly or implied , of the u.s .
government .
the al - alox - al squids were fabricated at uc santa barbara on oxidized silicon substrates , and comprised sputter - deposited al films , thermally - grown alox , pecvd - grown sio@xmath112 , and evaporated cuau shunt resistors .
the nb - alox - nb squids were fabricated at the university of gieen on oxidized silicon or sapphire substrates , and comprised sputtered nb , sputtered al , thermally grown alox , and sputtered pd shunt resistors ; in addition , some of the nb - alox - nb devices included a sputtered sio wiring dielectric . | here we present details of a calculation that allows us to extract a surface density of unpaired spins from flux _ vs. _ temperature experiments performed on field - cooled dc superconducting quantum interference devices ( dc squids ) .
our recent measurements show an unexpected dependence of squid flux on bath temperature in the millikelvin temperature range @xcite .
the flux change scales as @xmath0 as temperature is lowered .
this behavior has been observed in both al and nb devices , made both with and without a wiring dielectric , and prepared in different facilities according to different fabrication recipes @xcite .
paramagnetic impurities in the materials of the squid would naturally give rise to such a signature , and we interpret the @xmath0 dependence of the flux through the squids as strong evidence for unpaired spins , most likely in the native oxides of the superconductors . in order to clarify the source of the temperature - dependent flux , we have performed a series of field - cool experiments in which a magnetic field @xmath1 is applied to a 350 ph square - washer nb squid ( with inner dimension 200 @xmath2 m and outer dimension 1 mm ) as it is cooled through @xmath3 ; the field cool freezes magnetic flux vortices into the nb film , with density @xmath4 @xcite . when the device is well below @xmath3 , the magnetic field is removed , and the squid is maintained in a flux - locked loop as it is cooled to millikelvin temperatures . in fig .
1a we plot the flux threading the squid as a function of temperature for eight different values of the cooling field .
the cooling field strongly affects the temperature - dependent flux , enhancing or even reversing the polarity of the observed signal . in fig .
1b we plot the flux change on cooling from 500 mk to 100 mk as a function of the cooling field ; a linear fit to the data yields a slope of 1.3 @xmath5/mt .
clearly , vortices contribute significantly to the measured temperature - induced flux shift . applied as the device was cooled through the superconducting transition .
( b ) temperature - induced flux change @xmath6 on cooling from 500 mk to 100 mk , as a function of cooling field @xmath1 .
a linear fit to the data yields a slope @xmath7 = 1.3 @xmath5/mt . ]
the linear dependence of flux drift on vortex density suggests the following interpretation of the data .
as the superconducting films are cooled through @xmath3 in an applied field , vortices nucleate and relatively large magnetic fields are frozen into the films . the observed signal is due to the magnetization of unpaired electron spins on the surface of the superconductor in the strong fields produced by the trapped vortices . for our nb thin films we expect magnetic fields in the vortex to be of the order of 10 mt . this field strength yields characteristic temperatures of order 10 mk for single electron spins , compatible with the low energy scale seen in the experiments .
careful analysis of our field - cool data allows us to extract the surface density @xmath8 of spins , a key parameter in models of @xmath9 flux noise from surface magnetism @xcite .
we note that there is negligible _ direct _ coupling to the squid from spins polarized out of the plane of the superconducting films ; this is easily understood from reciprocity , as surface magnetic fields due to currents in the squid have vanishing perpendicular component .
however , the topology of our device is quite different from that of a continuous superconducting washer , due to the presence of the vortices .
indeed , the thermal polarization of unpaired surface spins in the vortex forces a redistribution of the circulating supercurrents in the vortex , due to the requirement to conserve magnetic flux .
these currents , in turn , couple strongly to the squid loop .
calculation of the spin density from the data of fig .
1 therefore proceeds in two stages : ( 1 ) calculation of the coupling between a vortex and the squid loop , and ( 2 ) calculation of the coupling between surface spins and a vortex . for both parts of this calculation
we need to solve for the currents in a superconducting washer with radial symmetry ( in the case of the squid this is a simplifying assumption ) ; therefore , we first describe the numerical solution to this problem .
this note is organized as follows . in section
i , we describe calculation of the current distribution in a thin superconducting washer . in section
ii , we discuss the coupling of a vortex to a squid . in section iii , we describe the coupling of surface spins to a vortex . finally ,
in section iv , we combine the results of sections ii and iii to extract a surface density of spins from the data of fig . 1 . |
the gyromagnetic factor of a particle is defined as @xmath21 where @xmath22 , @xmath23 , @xmath24 , and @xmath25 are the magnetic moment , the spin , the mass , and the charge of that particle . for a pointlike lepton dirac theory
gives @xmath26 .
virtual radiative corretions cause the @xmath27 value for leptons to differ from 2 .
the anomalous @xmath27-value is defined as @xmath28 .
the anomalous @xmath27-value of the electron and muon , @xmath29 and @xmath3 , played an important role in the development of particle physics and continue to serve as fundamental quantities for testing the validity of the standard model and putting stringent constraints on speculative theories beyond the standard model .
the value of @xmath29 is measured to the precision of 4 parts per billion ( ppb ) .
this value is currently the most precise measurement of fine structure constant @xmath30 . for the muon , because the muon mass @xmath31 is larger than the electron mass @xmath32 , the contribution to @xmath3 from heavier particles such as hadrons or weak vector bosons through virtual processes is greater than that to @xmath29 typically by a factor of @xmath33 .
this sensitivity enhancement of @xmath3 applies generally also for postulated new particles such as supersymmetric particles .
hence , the measurement of @xmath7 provides us more information than that of @xmath29 though @xmath7 is less well known than @xmath29 .
the muon anomalous magnetic moment was measured three times in cern @xcite , and the precision of the last cern measurement is @xmath34 ppm .
the muon @xmath35 experiment at brookhaven national laboratory ( e821 ) measures the @xmath7 of both the positive and the negative muon to @xmath36 ppm .
two values agree with each other and the precision of the combined value of @xmath7 is @xmath37 ppm@xcite .
for a muon moving in a uniform magnetic field @xmath38 , which is perpendicular to the muon spin direction and to the plane of the orbit , and with an electric field @xmath39 , both the spin and the momentum precess . the angular frequency difference , @xmath40 , between the spin precession frequency @xmath41 and the cyclotron frequency @xmath42 is given by @xmath43 \label{eq : nomagic}\ ] ] for @xmath44 and @xmath45 , where @xmath46 and @xmath25 are the mass and charge of the muon , @xmath47 is the speed of light , @xmath48 is the ratio of speed of the muon and the speed of light , and @xmath49 . in muon
@xmath35 experiment , the dependence of @xmath40 on the electric field is eliminated by storing muons with the `` magic '' @xmath50 such that @xmath51 this corresponds to a muon momentum @xmath52 gev / c . hence , the measurement of @xmath40 and @xmath53 determines @xmath3 , @xmath54 in our experiment , the magnetic field is measured by an nmr system which is calibrated with respect to the free proton nmr angular frequency @xmath55 as @xmath56 where @xmath57 is the magnetic moment of the free proton . by combination with [ eq : amdef0 ] ,
[ eq : bfreep ] and using muon spin @xmath58 and the definition of @xmath59 , @xmath60 so that @xmath61 the ratio @xmath62@xcite , as determined from muonium spectrocopy @xcite .
the frequency @xmath40 is measured by counting high energy decay positrons . in the muon rest frame ,
parity violation in the decay @xmath63 causes @xmath64 to be emitted preferentially along the muon spin direction .
positrons emitted along the muon momentum direction get the largest lorentz boost and have highest energy in the laboratory frame .
hence , the angle between the muon momentum and muon spin determines the number of high energy decay positrons . while the muon spin precesses with angular frequency @xmath41 and the muon momentum precesses with angular @xmath42 in the magnetic field , the number of decay positrons with energy greater than some threshold energy @xmath65 oscillates with the angular frequency @xmath66 and is ideally described by @xmath67\}\ ] ] in which @xmath68 is the dilated muon life time , @xmath69 is the asymmetry , @xmath70 is the phase of the oscillation .
both @xmath69 and @xmath70 depend on the energy threshold of the decay positrons .
however , the frequency @xmath40 does not depend on @xmath65 .
the general arrangement of the muon @xmath35 experiment is shown in fig .
[ fig : ring ] .
the proton beam from the alternating gradient synchrotron(ags ) strikes a nickel target to produce pions .
the pions are captured into a secondary beamline where a fraction of about @xmath71 decays to muons .
we select the high energy muons which come from the forward decay .
they are polarized due to the parity violation in the decay @xmath72 .
the muons are injected through a superconducting inflector@xcite into the muon storage ring with 1.45 t uniform magnetic field@xcite which is measured by a nmr system@xcite relative to the free proton nmr frequency .
a pulsed magnetic kicker@xcite at about @xmath73 from the injection point gives muons a 10 mrad deflection which places the muons onto stored orbits .
the electrostatic quadrupole@xcite fields provide vertical focusing .
the decay positrons are detected by 24 lead / scintillating fiber electromagnetic calorimenters@xcite read out by waveform digitizers .
both the waveform digitizer and nmr clocks were phase locked to the loran c frequency signal .
the data analysis in the experiment is divided into two independent parts , @xmath55 and @xmath40 analyses .
the absolute values of @xmath55 and @xmath40 from analyses were modified by arbitrary offsets .
no one knew both absolute values of @xmath55 and @xmath40 .
the values of @xmath74 and then @xmath7 were evalueated only after both analyses had been finalized . the measurements of the magnetic field at sub - ppm lever is done with nmr probes and is expressed relative to the free proton nmr frequency @xmath55 .
the nmr frequencies are measured with respect to a standard probe , which is a spherical water sample .
the calibration of the standard probe with respect to the free proton is obtained from other experiments@xcite .
a trolley with 17 nmr probes was used to measure the field around the storage ring .
the 17 trolley probes are calibrated with respect to the standard probe through a secondary probe called plunging probe , which can be mounted on a stand and moved in radial and vertical direction in vacuum chamber at one azimuthal location .
the calibrations were done by comparing the nmr readings of the plunging probe and the trolley probes taken at the same location .
the errors come from the position uncertainties of the trollye probes and of the plunging probe , and from the b field inhomogeneity , large magnetic sextupole moments and field gradient in azimuth direction were introduced to measure the relative position of active volume of trolley probes and the plunging probe . the effect on the trolley probe calibration from temperature ,
power supply and magnetic field were also carefully studied .
we measured the magnetic field with trolley about twice per week .
one of the measurements of the magnetic field with the trolley center probe is shown vs. azimuth in figure [ fig : multipole ] . for most of the ring
, the magnetic field varies about @xmath75 ppm .
the field measurements are averaged over azimuth for all 17 probes individually .
since the field component in radial and azimuthal direction are much smaller than that in vertical direction , the field can be expanded in 2-dimension @xmath76 where @xmath77 is the radial direction in our ring , and @xmath78 are polar coordinates .
the center of the storage region is at @xmath79 and the direction pointing outward is @xmath80 .
the value of @xmath81 is set to be 4.5 cm . a half ppm contour plot and the multipoles of the azimuthal average of the field measured in one trolley run are shown in figure [ fig : multipole ] .
the field between trolley runs was measured by a fixed probe system of 378 nmr probes .
more than 120 probes are used in the analysis because of their good qualities .
the average measurements from those fixed probes during the trolley run is calibrated with respect to the trolley measurements .
the muon distribution is taken into account by using the field value at the center of the muon distribution .
the radial muon distribution is measured by fast rotation with uncertainty of 1 mm .
the vertical muon distribution is symetric around the central plane determined by electric quadruples , and the uncertainty is 2 mm .
the contributions to the total systematic error from different sources are given in table [ tb : berror ] for 2001 run . and
the result for @xmath55 weighted by the muon distribution in 2001 is found to be @xmath82 .systematic uncertainties for @xmath55 analysis of the 2001 data [ cols="<,<",options="header " , ] @xmath83 ags background , timing shifts , e field and vertical oscillations , beam debunching / randomization , binning and fitting procedure .
[ tb : omegaa ] the value of @xmath7 was determined after the analyses of @xmath55 and @xmath40 had been finalized , @xmath84 in which the correlations between the systematic uncertainties have been taken into account . in eq .
[ eq : amdef ] , @xmath85 is measured by @xcite to high precision for @xmath86 only .
hence @xmath87 and @xmath88 can not be done directly .
instead , we compared @xmath74 values from @xmath86 ad @xmath89 .
the difference is @xmath90 and is consistent with 0 . assuming cpt , we combine the results from @xmath86 and @xmath89 and obtain @xmath91 fig . [
fig : finalresult ] shows the measurements and the theoretical prediction of @xmath7 .
in the standard model , the contributions to @xmath7 are divided into three types , qed , hadronic and weak .
@xmath92 the qed contribution has the largest contribution but least uncertainty .
the calculation up to @xmath93 order of @xmath94 gives @xmath95 a recent recalculation of the terms of @xmath96 order of @xmath94 increases the qed contribution to @xmath97 the weak contribution has been calculated to the second order and is evaluated as @xmath98 the hadronic contribution can not be calculated from pertubative qcd alone because it involves low energy scales near the muon mass .
however , the leading order of hadronic contribution @xmath99@xcite can be determined directly from the annihilation of @xmath5 to hadrons through a dispersion integal @xmath100 in which @xmath101 is the energy in the center of mass frame , @xmath102 is a kinematic factor and @xmath103 the value @xmath104 can be indirectly determined using data from hadronic @xmath6 decays , the conserved vector current hypothesis , plus the appropriate isospin correction . in principle , the large statistics of @xmath6 decay data could improve the precision of @xmath7(had ) .
however , discrepancies between the @xmath6 and @xmath5 data exist as shown in fig .
[ fig : discrepancy ] .
the @xmath5 data were dominated by the data from cmd2 experiment .
the recent result from radiative return data@xcite with comparable uncertainty agree with the cmd2 experiment .
the next leading order hadronic contribution@xcite and the hadronic light - by - light scattering@xcite are calculated by serveral groups and the combined the results are @xmath105 the large discrepancies between @xmath6 and @xmath5 data prevent us from combining two to obtain a single standard model prediction .
the differences between experimental measurement and the standard model prediction are @xmath106 @xmath107 the question concern the discrepancies betweent @xmath5 and @xmath6 data are under further theoretical scrutiny and we expected more radiative return measurements and efforts from latice calculation .
99 g. charpak _
_ , nuovo cim . * 37 * 1241 ( 1965 ) .
j. bailey _
et al . _ , nuovo cim . * a9 * 369 ( 1972 ) . j. bailey _
et al . _ ,
b150 * , 1 ( 1979 ) .
muon @xmath108 collaboration , r.m .
et al . _ ,
lett . * 82 * , 1632 ( 1999 ) .
muon @xmath108 collaboration , h.n .
et al . _ ,
phys . rev . * d62 * , 091101 ( 2000 ) .
muon @xmath108 collaboration , h.n .
et al . _ ,
. lett . * 86 * , 2227 ( 2001 ) .
muon @xmath108 collaboration , g.w .
et al . _ ,
lett . * 89 * , 101804 ( 2002 ) .
muon @xmath108 collaboration , g.w .
lett . * 92 * , 161802 ( 2004 ) .
d.e . groom _
et al . _ ,
j. * c15 * , 1 ( 2000 ) .
lett * 82 * , 711 ( 1999 ) .
a. yamamoto _
et al . _ ,
instrum meth .
* a491 * , 23 ( 2002 ) . g. danby _
et al . _ ,
a457 * , 151 ( 2001 ) .
r. prigl _ et al .
a374 * 118 ( 1996 ) .
e. efstathiadis _
et al . _ ,
instum meth .
* a496 * , 8 ( 2003 ) . y. semertzidis _
et al . _ ,
. meth . * a 503 * 458 , ( 2003 ) s. sedykh _
et al . _ ,
a455 * ( 2000 ) x. fei , v. hughes , and r. prigl , nucl .
instrum , meth . *
a394 * , 349 ( 1997 ) .
t. kinoshita and w.j .
marciano in _ quantum electrodynamics _ ed .
t. kinoshita ( world scientific , singapore , 1990 ) , p. 419 .
t. kinoshita and m. nio , hep - ph/0402206 , ( 2004 ) .
a. czarnecki , w.j .
marciano , and a. vainshtein , phys . rev . * d67 * 073006 ( 2003 )
. m. davier , s. eidelman , a. hcker , z. zhang , eur .
j. * c31 * , 503 ( 2003 ) .
s. ghozzi and f. jegerlehner , hep - ph/0310181 , ( 2004 ) .
k. hagiwara , a. martin , d nomura , and t. teubner , phys .
d69 * , 093003 , ( 2004 ) .
akhmetshin , _ et al .
_ , ( cmd-2 ) , phys . lett . *
b527 * , 161 ( 2002 ) .
kloe collaboration , hep - ex/0407048 , ( 2004 ) .
b. krause , phys .
b390 * , 392 ( 1997 ) .
r. alemany , m. davier , a. hcker , eur .
j. * c2 * , 123 , ( 1998 ) .
m. hayakawa and t. kinoshita , hep - ph/0112102 ( 2001 ) .
j. bijnens , e. pallante , and j. prades , nucl .
* b 626 * , 410 ( 2002 ) .
m. knecht and a. nyffeler , phys . rev . *
d65 * 073034 ( 200 ) m. knecht , a. nyffeler , m. perrottet , and e. de rafael , phys .
* 88 * 071802 ( 2002 ) .
i. blokland , a. czarnecki , and k. melnikov , phys .
lett . * 88 * , 071803 ( 2002 ) . | the anomalous magnetic moments of both positive and negative muon have been meausred to a precision of 0.7 parts per million ( ppm ) at the brookhaven alternating gradient synchrotron .
the results are over an order of magnitude more precise than the previous measurement of the muon .
the results @xmath0 and @xmath1 , where the first uncertainty is statistical and the second is systematic , are consistent with each other .
the average for the muon anomaly is @xmath2 .
the standard model value of @xmath3 is calculated to a precision of @xmath4ppm , which is dominated by uncertainty of hadronic contributions .
there is a significant disagreement between the hadronic contribution calculated from the hadronic cross section of @xmath5 collision and that from hadronic @xmath6 decay .
the difference between experimental and theoretical values of @xmath7 is @xmath8 for @xmath5 data or @xmath9 for @xmath6 decay data .
_ @xmath10department of physics , boston university , boston , massachusetts 02215 + @xmath11brookhaven national laboratory , upton , new york 11973 + @xmath12budker institute of nuclear physics , novosibirsk , russia + @xmath13newman laboratory , cornell university , ithaca , new york 14853 + @xmath14 kernfysisch versneller instituut , rijksuniversiteit groningen , nl 9747aa groningen , the netherlands + @xmath15 physikalisches
institut der universitt heidelberg , 69120 heidelberg , germany + @xmath16 department of physics , university of illinois at urbana - champaign , illinois 61801 + @xmath17 kek , high energy accelerator research organization , tsukuba , ibaraki 305 - 0801 , japan + @xmath18department of physics , university of minnesota , minneapolis , minnesota 55455 + @xmath19 tokyo institute of technology , tokyo , japan + @xmath20department of physics , yale university , new haven , connecticut 06520 _
= 14.5pt |
* experiment * + _ trapping : _ we start with 2@xmath31 bose condensed @xmath1rb atoms in the @xmath32 , @xmath33 state in a crossed dipole trap@xcite with 1.8 w per beam and 160 beam waists . gravity is canceled by a magnetic field gradient . a blue - detuned 773.5 nm wavelength 3d optical lattice , made from two retroreflected horizontal 450 waist beams and a retroreflected vertical 700 waist beam ,
is turned on in 14 ms to a depth of @xmath4 , after which the two horizontal lattice beam pairs are increased to their full depth of 40@xmath8 in 44 ms .
we find that the results of the experiment ( including spatial distributions and occupancy fractions ) do not significantly change as long as the lattice turn - on times are 35 ms or longer .
that remains true if we wait for tens of ms in all the traps before starting the evolution .
+ _ flat lattice : _ since for technical reasons the lattice waists are much larger than the dipole trap waists , we can only create a flat lattice near the center of the trap . to fine tune
the cancelation of the two potentials , we start with trapped 1d gasses with no axial lattice , and choose the highest crossed dipole beam intensity at which no atoms remain trapped in the central region .
this gives a central potential that is flat to within 0.08@xmath8 over a length of 160 @xmath34 m .
+ _ small bump : _ the small bump on the left of the initial atom distributions in the experimental results of fig .
2 is due to spatial imperfections on the yag beams that make up the confining crossed dipole trap , where the atoms are trapped 3.4 rayleigh lengths from the beam focus .
after @xmath7=0 the atoms evolve in a much smoother potential , so only the initial distribution is affected by this issue , not the evolution .
the bump can serve as a feature , since it provides confirmatory evidence of the fraction of singlons that remain confined ( see , e.g. , the long time curves in fig . 2**f * * ) .
+ _ determination of the trapped singlon fraction : _ we analyze the singlon distributions ( see figures 2**d**2**f * * ) , by first determining the fwhm of the doublons ( see figures 2**g**2**i * * and figure 4**b * * ) at each time .
we then measure the difference between the peak of the singlon distribution and its value at the doublon fwhm positions .
we assume that the confined peak height is twice that difference , and has the doublon width .
we do not use this procedure at very early times , before there is a discernible shoulder in the singlon distribution , although the curves in figure 4**a * * change little if we add a few earlier points .
the assumption that the fwhms of doublons and trapped singlons are the same is not exactly true for a bundle of tubes .
this procedure also assumes that the unconfined singlon distribution is approximately flat in the central region , which is also not quite true .
our confidence in the reliability of this procedure is buttressed by the universality of the curves in figure 4**a * * , which holds despite the systematic differences in the doublon fwhms at different @xmath4 ( see figure 4**b * * ) .
* theoretical calculations * + for the theoretical analysis of the expansion , we model a collection of independent 1d tubes . in order to account for possible higher band effects during the dynamics , which may be important for the lowest lattice depths , we do not use the one - band approximation that results in the standard bose - hubbard model .
instead , we discretize the space along the tubes by introducing an artificial grid of spacing @xmath35 ( we take @xmath36 , where @xmath37 is the wavelength of the optical lattice ) to obtain a representation of the continuum in terms of an artificial one - dimensional bose - hubbard model.@xcite the particles are then subject to the periodic potential generated by the optical lattice and to the confining potentials generated by the crossed optical dipole trap . we simulate arrays of up to @xmath38 tubes , with each tube s length being @xmath39 .
we choose the theory initial conditions so that the fraction of singlons in the simulation matches the experiment .
this is done by arbitrarily tuning the strength of the @xmath40 axial and transverse trapping potentials , keeping their ratio fixed to the experimental value .
the parameters for @xmath41 are those in the experimental setup , except that the overall confining potential is set exactly to zero .
the calculations are carried out within the gutzwiller mean - field approximation,@xcite where the initial state is selected to be the ground state in the absence of tunneling between the 1d tubes .
our results are robust to further reduction of the value of @xmath42 . for further details ,
see supplementary information .
we are indebted to andrew daley and michael fleischauer .
this work was supported by nsf grant no .
phys-167830 , the army research office , and the office of naval research ( j.c . and m.r . ) .
j.c . acknowledges support from the john templeton foundation .
research at perimeter institute is supported through industry canada and by the province of ontario through the ministry of research & innovation . | ultracold atomic gases have revolutionized the study of non - equilibrium dynamics in quantum many - body systems .
many counterintuitive non - equilibrium effects have been observed , such as suppressed thermalization in a one - dimensional ( 1d ) gas,@xcite the formation of repulsive self - bound dimers,@xcite and identical behaviors for attractive and repulsive interactions.@xcite here , we observe the expansion of a bundle of ultracold 1d bose gases in a flat - bottomed optical lattice potential . by combining in situ measurements with photoassociation,@xcite
we follow the spatial dynamics of singly , doubly , and triply occupied lattice sites .
the system sheds interaction energy by dissolving some doublons and triplons .
some singlons quantum distill out of the doublon center,@xcite while others remain confined.@xcite our gutzwiller mean - field model captures these experimental features in a physically clear way .
these experiments might be used to study thermalization in systems with particle losses@xcite or the evolution of quantum entanglement,@xcite or if applied to fermions , to prepare very low entropy states.@xcite department of physics , the pennsylvania state university , university park , pa 16802 , usa perimeter institute for theoretical physics , waterloo , ontario , canada n2l 2y5 quantum distillation is a previously unobserved phenomenon in which atoms at singly occupied lattice sites ( singlons ) escape the central region of an untrapped lattice gas , leaving doubly occupied lattice sites ( doublons ) behind ( see figure [ fig:1]*a*).@xcite it depends on one readily achievable condition , that there be an energy mismatch that prevents isolated doublons from disintegrating into two singlons .
singlons in a sea of doublons can be understood as vacancies .
the tunneling rate of a singlon in an empty lattice is @xmath0 , while the bosonic vacancy tunneling rate is 2@xmath0 , since the vacancy moves when either of the adjacent doublon s atoms tunnels .
when bosonic vacancies reach a doublon sea edge only those with intermediate energies can transmit into the empty lattice while conserving energy,@xcite as illustrated in figure [ fig:1]*b*. vacancies with quasimomenta outside that limited range reflect from the edge , confining them in the doublon sea .
when a singlon does exit , it purifies and shrinks the doublon sea by one lattice site .
it has been hypothesized that collisions of vacancies with triplons can thermalize the vacancies , possibly transferring them into transmissible quasimomentum states.@xcite in contrast , fermionic vacancies have the same tunneling energy as singlons in the empty lattice , so they always pass through the edge of the doublon sea .
a previous non - equilibrium experiment with bosons in flat 1d lattices mostly focused on initial single - atom number states in a deep lattice , and studied the expansion dynamics after a quench of the on - site interaction energy.@xcite such a quench leaves the many - body wavefunction out of local equilibrium .
in contrast , we start our experiments with trapped , superfluid , 1d bose gases in lattices with an average of between one and two atoms per site , and we quench by suddenly removing the trap . this is a fundamentally different quench , a geometric quench , after which the many - body wavefunction is still locally in equilibrium .
geometric quenches have been theoretically shown to lead to remarkable universal phenomena , such as quasi - condensation at finite momenta,@xcite dynamical fermionization of the tonks - girardeau gas,@xcite and identical expansions of bosons and fermions.@xcite our experimentally observed expansion dynamics are qualitatively reproduced by a gutzwiller mean - field calculation ( see methods ) . by theoretically and experimentally studying the spatial evolution of site occupancy , we obtain a straightforward physical interpretation of the dynamics .
it exhibits clear signatures of quantum distillation@xcite and confinement@xcite of vacancies in the doublon sea .
in our experiment , bose condensed @xmath1rb atoms in a crossed dipole trap are slowly loaded into an array of 1d tubes formed by a blue - detuned 2d optical lattice ( wavevector @xmath2=2@xmath3/773 nm ) , with a superposed axial optical lattice of variable depth @xmath4 and a red - detuned crossed dipole trap for overall confinement ( see methods ) . for the density and @xmath4 used in this work ,
the initial ground states are predominantly superfluid.@xcite since each lattice site starts with a superposition of number states , pictures like those in figure [ fig:1]*a * represent one of many distributions whose coherent sum is the state of the system . as we will see , most of the qualitative behavior of what are thus delocalized singlons and doublons can be understood using localized pictures , leavened by the understanding that each picture represents only a small piece of the overall wavefunction . at @xmath5 , we suddenly lower the depth of the crossed - dipole trap , leaving enough power to cancel the residual anti - trap due to the 2d lattice beams over a range of @xmath6 ( see methods ) .
we observe the subsequent spatial evolution in three ways ( referred to as m1 , m2 , and m3 ) , which together allow us to separately determine the spatial evolution of the probability distributions of singlons , doublons , and atoms at more highly occupied sites . in m1 , we measure all the atoms after a given @xmath7 by switching to a 27@xmath8 deep 1d axial lattice ( where the recoil energy @xmath9 , and @xmath10 is the rb mass ) and allowing the atoms to expand radially so that the density is low enough for absorption imaging .
information about the transverse distribution among tubes is lost , but the axial distribution is preserved , with a resolution of @xmath6 . in m2 , at @xmath7 we suddenly switch to a 27@xmath8 lattice in each of three directions and turn on a photoassociation pulse for 1.5ms,@xcite which is long enough to eliminate all doublons , 2/3 of the triplon atoms , and most of the atoms from sites with higher occupancies .
the axial distribution measurement is then made as in m1 . in m3 , at @xmath7 we suddenly switch to the 27@xmath8 3d lattice and wait for 50ms , which is long enough for three - body inelastic collisions@xcite to empty the triplon sites and most atoms from more highly occupied sites .
the axial distribution is then measured as in m1 . .
a singlon in the empty lattice tunnels at @xmath0 .
a vacancy at the edge of the doublon sea is the same as a singlon at the edge of the empty lattice .
it is important to note that , at our low temperature and our lattice depths , atoms are delocalized .
each 1d tube is occupied by a superposition of many distributions like the one pictured , generally including sites with higher occupancies and empty sites , as well as different doublon edge positions . *
b*. ground energy bands within the doublon sea ( on the right ) and in the empty lattice ( on the left ) . only atoms with quasimomentum between 1/3 and 2/3 of the band edge can tunnel from the doublon sea into the empty lattice . outside that range transmission into the empty lattice does not conserve energy .
this limitation on quantum distillation is absent for fermions.[fig:1 ] ] figures [ fig:2]*a*[fig:2]*c * show the evolution of the total atom distribution ( m1 ) for @xmath4=3@xmath8 , 4@xmath8 , and 5@xmath8 .
these depths correspond to @xmath11 of 4.7 , 6.8 and 9.6,@xcite respectively , in the one - band hubbard model,@xcite where @xmath12 is the onsite repulsion energy .
all are characterized by a central core of atoms that steadily releases atoms that tunnel away from the center .
figures [ fig:2]*d*[fig:2]*f * show the distributions of single atoms , which are derived from m1 , m2 and m3 ( see supplementary information figs . s1**a**s1**f * * ) .
these curves show two dominant features .
first , the cores contain many single atoms .
second , the broader pedestals of the m1 distributions are composed nearly exclusively of single atom sites .
the velocities of the leading edges of the pedestals equal , to within @xmath610% systematic uncertainties , the calculated maximum possible velocity ( @xmath13 , where @xmath14 is the lattice spacing ) for single atoms tunneling in the lowest band ( see insets in figures [ fig:2]*a*[fig:2]*c * ) .
these velocities start to decrease at the end of the compensation range of the crossed dipole trap , ultimately bragg scattering backwards ; we do not display data after atoms return to the core .
figures [ fig:2]*g*[fig:2]*i * show the distribution of doublons , derived from all three measurement types .
the number of doublons steadily decreases after the quench , but the widths of the doublon distributions barely change .
the triplon distributions ( see supplementary information figs
. s1**g** s1**i * * ) ) have the same width as the doublon distributions to within a 10@xmath15 uncertainty . figures [ fig:2]*j*[fig:2]*l * ( and supplementary information figs .
s1**j**s1**l * * ) show the results of a gutzwiller mean - field calculation@xcite ( see methods ) , which simulates an array of identical tubes with different atom numbers as in the experiment , and discretizes the direction along the tubes@xcite so that the standard single - band approximation is not used .
the theory assumes an initial zero temperature bec .
this means that finite temperature and quantum fluctuations due to the one - dimensional character of the system are not taken into account .
the initial distribution is thus an imperfect match to the experiment ( see methods ) .
the doublons initially expand farther in the theory than in the experiment , presumably because of the long range initial phase coherence in the theory .
the early decrease in the theory s doublon density no doubt affects the details of the ensuing dynamics , but qualitatively , the theory behaves like the experiment in all respects other than the shape and size of the doublon distributions . for additional comparison to theory ,
see supplementary information . at @xmath4=3@xmath8 , 4@xmath8 , and 5@xmath8 , respectively .
insets : the location of the distribution edges vs. time .
the measured velocity values for @xmath16 , @xmath17 , and @xmath18 are @xmath19 , @xmath20 , and @xmath21 mm / s , respectively , slightly slower than @xmath22 of 2.1 , 1.6 , and 1.2 mm / s . *
d - f*. the distributions of singlons at successive @xmath7 at @xmath4=3@xmath8 , 4@xmath8 , and 5@xmath8 , respectively .
these are derived from combining the raw measurements according to the formula m2-(m1-m3)/3 .
* g - i*. the distributions of doublons at successive @xmath7 at @xmath4=3@xmath8 , 4@xmath8 , and 5@xmath8 , respectively .
these are derived from combining the raw measurements according to the formula m1-m2 - 2(m1-m3)/3 .
note that all the experimental plots ( * a - i * ) have the same vertical scale . for a discussion of the small asymmetry in these figures , see methods .
* j - l . *
gutzwiller mean - field theory results for @xmath23 for the distribution of all atoms , singlons , and doublons , respectively .
we choose the theory initial conditions so that the fraction of singlons matches the experiment ( see methods ) .
the initial fraction of triplons and higher exceed what is seen in the experiment , and the initial cloud length is smaller .
the qualitative behavior seen in the theory ( and in the experiment ) is robust to the initial conditions .
all but the least occupied @xmath24 of tubes behave qualitatively like the average over all tubes , but with less doublon dissolution and quantum distillation in the most occupied tubes .
the inset of * j * shows the motion of the edges at @xmath22 , as expected when the single particle quasimomentum distribution includes the midpoint of the band .
[ fig:2 ] ] figures [ fig:3]*a*[fig:3]*c * show the number of singlons , doublons and triplons as a function of time , derived from the appropriate combinations of m1 , m2 , and m3 .
the numbers of doublons and triplons drop steadily , with corresponding increases in the numbers of singlons .
the theory shows similar behavior ( see fig .
[ fig:3]*d * ) .
although conservation of energy dictates that isolated doublons can not dissociate for @xmath25,@xcite in a predominantly doublon sea the aforementioned tunneling enhancement doubles this limit .
similarly , one can show that a sea of singlons increases the limit by 50@xmath15 . that our doublons live in a bath intermediate to these two seas explains why they dissociate , at least for @xmath26 , and why the dissociation rate decreases at long times ( see especially fig .
[ fig:3]*a * after 50 ms ) when the number of empty sites in the center increases .
dissociation for @xmath27 ( @xmath28 ) naively requires that energy be shared among more singlons,@xcite but it might be that the one - band hubbard model calculation@xcite overestimates the effective value of @xmath11 .
the latter interpretation is supported by the fact that , at short times , the evolution of all the curves in figs .
[ fig:2 ] and [ fig:3 ] are approximately self - similar when the time axes are multiplied by @xmath0 ( see supplementary information figure s2 ) , to within the small differences in the doublon distribution widths discussed below . though the physics is dominated by interacting particle effects , marginal changes in site occupancy scale with @xmath0 .
our mean field calculation allows us to explicitly track the conversion of potential energy ( interaction + lattice potential ) into kinetic energy ( see inset ) that results primarily from doublon dissolution
. experiment
. * b*. @xmath23 experiment .
* c*. @xmath28 experiment .
the error bars derive from the standard deviations determined from 8 separate measurements . when the curves in * a - c * are plotted together with a time axis rescaled by @xmath0 ( see supplementary figure s2 ) , they overlap well until the doublons reach their asymptotic number . * d*. @xmath23 theory . in all these subfigures the singlons are the blue circles , the doublons are the red diamonds , and the triplons are the black triangles .
inset to * d * : the interaction ( blue pluses ) , kinetic ( orange asterisks ) , and lattice potential ( green crosses ) energies vs. @xmath7 calculated for @xmath23 .
[ fig:3 ] ] quantum distillation is difficult to isolate at early times , since it occurs while initially unconfined singlons are also leaving the central region and singlons are being created by dissolution .
but quantum distillation dominates at 5@xmath8 ( fig .
[ fig:2]*f * ) after 20 ms , by which time the doublon number is stable ( see also figs .
[ fig:2]*i * and [ fig:3]*c * ) and the unconfined singlons present a locally flat background . the number of singlons confined in the doublon sea as a function of time is plotted in fig .
[ fig:4]*a * ( see methods ) .
its steady decrease is a clear signature of quantum distillation , further supported by the fact that the rate scales with @xmath0 ( see also supplementary information ) . at late times , the fraction of confined singlons levels off at @xmath29 , showing long term vacancy confinement in the doublon sea .
the mean - field calculations also show some singlons initially leaving and others remaining indefinitely ( see fig . [ fig:3]*d * ) , but fewer singlons distill out in the calculations .
this is expected because the real 1d gas has a broader initial quasimomentum distribution .
thus in the calculation there are more singlons with the lower energies that do not transmit out of the doublon sea . in @xmath0-rescaled time , after doublons stop dissolving at 5@xmath8 they are still dissolving at lower lattice depths . that the three sets of data points in fig .
[ fig:4]*a * overlap means that extra doublon dissolution does not affect the central singlon number .
this could be because the vast majority of singlons created when doublons dissolve have the right quasimomentum for immediate quantum distillation , and leave the center rapidly .
further evidence of quantum distillation is given in fig .
[ fig:4]*b * , which shows the evolution of the full width at half maximum ( fwhm ) of the doublon distribution .
there are three size changing processes , each dominating for a time .
the fwhms increase during the first 10 to 20 ms because doublons initially can expand into singly occupied sites and perhaps there is more doublon dissolution in the middle ( see figures [ fig:2]*g*[fig:2]*i * ) .
the fwhms then decrease due to the mechanics of distillation , where escaping singlons move the last doublon one site inward . when the rate of quantum distillation decreases ( near 4 ms@xmath30@xmath8 in fig .
[ fig:4]*a * ) then as long as the lattice depth is small = 3@xmath8 , 4@xmath8 and 5@xmath8 respectively in the experiment .
the squares are the theory results for @xmath4=4@xmath8 .
see methods for a description of how the data is derived from the singlon distributions .
* b*. the fwhm of the doublon distributions vs. tunneling - rescaled evolution time .
( same labels as in * a * ) .
the error bars in this figure derive from the standard deviation of 8 measurements .
[ fig:4 ] ] enough , shrinking is overtaken by expansion .
we suspect this to be the result of higher order processes involving confined singlons that compromise the stability of the edges of the doublon sea.@xcite this unanticipated higher order effect , undoubtedly absent in fermions and not present in the mean field theory results , further limits the effectiveness of bosonic quantum distillation in producing low entropy blocks of doublons .
the approximate stability or slight increase in the width of the doublon distribution implies that as time evolves , the number of empty sites among the doublons increases .
our work has concentrated on spatial distributions , which are local properties , but it should also be possible to use related techniques like time - of - flight measurements to study non - local properties , like quasimomentum distributions and correlations .
this simple lattice system , in which the doublon sea is open but nonetheless settles to a stable steady state , can help address major open questions in quantum dynamics such as how systems thermalize in the presence of particle losses.@xcite studying how quantum correlations grow after the quench should give more general insight into how entanglement spreads in quantum systems.@xcite the fact that our mean - field treatment ( only exact in infinite dimensions ) qualitatively captures the 1d dynamics , suggest that similar dynamics occur in higher dimensions .
finally , an experimental implementation with fermions holds the promise of producing superlatively low entropy doublon cores,@xcite which might allow the study of hitherto inaccessible models of quantum magnetism @xcite and high temperature superconductivity.@xcite |
it is widely believed that the origin of life required the formation of sets of molecules able to collectively self - replicate , as well as of compartments able to undergo fission and proliferate , i.e._protocells _ @xcite .
in particular , in order to observe a lifelike behavior it was necessary that some of the chemical reactions were coupled to the rate of proliferation of the compartments .
several protocell architectures have been proposed @xcite , most of them identifying the compartment with a lipid vesicle that may spontaneously fission under suitable circumstances . on the other hand , many distinct models were proposed to describe sets of reactions involving randomly generated molecules @xcite . in many cases , although this is not in principle required , it is assumed that only _ catalyzed _
reactions take place at a significant rate , therefore these sets are also termed _ catalytic reaction sets _ ( briefly , crss ) .
it is worth noting that the appearance of new molecules implies the appearance of new reactions involving those new molecules , so that both the set of molecular types and the set of reactions change in time .
hence , it is possible that at a certain time a set of molecules able to catalyze each other s formation emerges @xcite , and we will refer to it as an _ autocatalytic set _ ( acs ) .
it can be noticed that a crs can contain one or more acss , or none . + even though some models of protocell actually describe the coupling between reaction networks and the dynamics of a lipid container , they consider only a fixed set of molecular species and reactions @xcite , hence providing an incomplete representation of this complex interplay .
conversely , while there are several studies on collectively self replicating sets of molecules in a _ continuously stirred open - flow tank reactor _
, cstr @xcite including our own @xcite , they provide only limited information about the behavior of a protocell .
therefore , in order to develop a framework that may unify the crss and the protocell modeling approaches , it is necessary i ) to analyze the behavior of crss in a vesicle , and ii ) to investigate the coupling of the evolving chemical population with the growth of the lipid container and its fission . in this paper we propose a step towards the first goal , while deferring the second one to a further work .
in particular , we here analyze the behavior of a dynamical model of crss in a simplified model of a non - growing vesicle . to the best of our knowledge
this is a novel approach .
+ a few important remarks .
let us first observe that the cstr is not an _ a - priori _ good model of a protocell for at least two reasons : @xmath2 in general , in protocells there is no constant inflow and @xmath3 protocells have semipermeable membranes , which allow the inflow / outflow of some , but not all , molecular types .
on the contrary , in open flow reactors all that is contained in the inflow enters the reactor and all that is dissolved in the reactor can be washed out in the outflow .
another important limit of the cstr concerns its evolvability .
it has been argued @xcite that the presence of different asymptotic dynamical states and the ability to shift between them may be essential to achieve the viable evolution of the first forms of life .
recent works @xcite have found that , in models of catalytic reaction networks in cstrs , generally only one of these states is found , apart from fluctuations @xcite .
+ furthermore , in order to accomplish the goal of this work , we need to better specify both the model of catalytic reactions sets and that of the protocell .
as far as the former is concerned , we have studied the dynamics of random sets of molecules by revisiting a model by kauffman , who proposed an interesting way to build new molecular species from the existing ones ( see section [ sec : model ] for a description ) . the original version of the model relied on purely graph - theoretical arguments @xcite , which are important , but fail to appreciate the effects of the dynamics , including noise , fluctuations and small - number effects .
the dynamics has been later introduced by farmer et al .
@xcite , who described the kinetics by using ordinary differential equations .
however , this formalism does not account for
the chance of a species to become extinct in a finite amount of time , as it may instead well happen ( so the reaction graph may grow but never shrinks ) . in order to overcome these limitations ,
bagley proposed an empirical correction by setting to zero the concentration values that happen to fall below a certain threshold @xcite . in our works
we rather use from the very beginning a stochastic approach to analyze the dynamics , the well - known gillespie algorithm @xcite , in order to deal in a rigorous way with low concentrations and with their fluctuations .
+ note that the kauffman model largely relies upon randomness .
in particular , every polymer in the system has a fixed probability ( that may vanish ) to catalyze any possible reaction .
therefore , in different simulations the same species can catalyze different reactions leading to the formation of different _
chemistries_. thus , this is exactly the language we choose : a set of tuples @xmath4 , where the species catalyzes the reaction , will be called chemistry , because it describes a possible artificial world .
we can then simulate different chemistries and look for generic properties of the set of chemistries ; but in a different series of experiments we can also keep the chemistry fixed , and simulate various time histories . in principle , these may differ , since the discovery of a given catalyst at an early phase in a finite system might channel the following evolution in a way or another . since the number of molecules of some species may be very small ,
it is not in principle legitimate to ignore this aspect , and our stochastic model is particularly well suited to analyze it , as it will be shown in section [ sec : model ] .
of course , there can be conditions where all the simulations of a given chemistry converge asymptotically to the same chemical mixture . + moving now to the protocell model , note that they are usually based on lipid vesicles , i.e. approximately spherical structures with an aqueous interior and a membrane composed by a lipid bilayer , which spontaneously form when lipids are mixed with water under certain conditions @xcite .
even though different protocell architectures have been proposed , we will here consider the simplest model , namely that in which all the key reactions take place in the aqueous phase _ inside _ the protocell .
it would be indeed straightforward to model the coupling between some of these molecules and the growth of the protocell following an approach similar to that of our previous studies @xcite .
yet , the main objective of the present work is that of studying the dynamics of crss embedded in a vesicle , so we will simplify our treatment by ignoring the growth dynamics of the protocell , and keeping its volume fixed .
this implies that our study will be limited to time intervals that are short with respect to those describing the growth of the whole protocell .
+ the selective character of membranes is a key ingredient of our model : we will suppose for simplicity that all ( and only ) the molecules that are shorter than a certain length can cross the membrane .
the transmembrane motion of the permeable species is here supposed to be driven by the difference of their concentrations in the internal aqueous volume of the protocell and in the external aqueous environment .
we will assume that transmembrane diffusion is extremely fast , so that there is always equilibrium between the concentrations of the species that can cross the membrane ; this adiabatic hypothesis could be easily relaxed in the future .
furthermore , we assume that protocells are turgid , so that the constant - volume approximation implies that we will also neglect issues related to osmotic pressure .
another related aspect of the model is that , since it is assumed that the permeable species are at equilibrium , while the non - permeable ones never cross the barriers , infinite concentration growth is possible ; this is obviously a nonphysical behavior , so the model validity is limited in time .
all these simplifications , which will be removed in subsequent studies , are also justified by the fact that our main goal is that of studying how the dynamics of crss are affected by being embedded in a vesicle .
+ this model can be used in order to investigate the behavior of the system in different conditions and to address some important questions .
the first and perhaps most important one is the reason why compartments seem to be necessary for life .
indeed , the very first studies on self - replicating molecules @xcite were not interested in this aspect , so the crss were supposed to exist , e.g. , in a pond or in a beaker . yet life seems to require compartments , that are ubiquitous .
it is then important to understand whether there are major differences between what may happen in a protocell and what happens in the bulk phase
. it would be unconvincing to postulate _ a priori _ that the internal and external environments are different .
it is indeed more likely to assume that the vesicles form in a pre - existing aqueous environment , so the average internal milieu is essentially the same as the external one .
then , if a membrane surrounds a portion of the fluid , what can happen that makes a difference ?
+ let us first observe that protocells are small ( their typical linear dimensions ranging from @xmath5 to @xmath6 ) .
if we imagine that a population of protocells exists , and they are not overcrowed , their total internal volume will typically be much smaller than the total external volume ( this is _ a fortiori _
true for an isolated one ) . moreover ,
every point in the interior of a protocell is not allowed to be far away from the surface of the protocell that contains it .
these observations imply that the effect of surfaces will be much larger within protocells then outside them .
suppose for example that the membrane hosts some catalytic activities , so that important molecules are synthesized close to its boundaries , both inside and outside , and diffuse freely .
if the membrane width is much smaller than the protocell radius , then the internal and external surface areas are very close to each other , but the external volume is much larger than the external one : therefore the internal concentrations will be much higher than those in the external environment . in this case , the system behavior in the interior can be significantly different from the external one @xcite .
note also that this effect may be different for different molecules : the formation of some of them might be catalyzed by the membrane , while others might be unaffected : so even the relative concentrations of different chemicals may differ in the two cases .
+ indeed , there are important protocell models that are based on such an active catalytic role of the membrane @xcite . in these cases
it is easy to understand what the role of the protocell is , since it provides essential catalysts and a way to keep their products closer . but protocells might be able to give rise to an internal environment different from the bulk even if the catalytic activity is absent .
the reason for this seemingly counterintuitive behavior is , once again , the smallness of the protocells .
note that we are considering a case where new molecules are formed ( from those that are already in the interior of the protocell plus those that can cross the semipermeable membrane ) .
if the concentrations are not too high , it is likely that the total numbers of newly formed molecules are quite low , so that different protocells might host different groups of molecules .
it might even happen that a molecular type is present in some protocells and not in others .
+ in order to get a feeling for this possibility , let us provide some realistic estimates of the number of molecules of different types that can be present in a protocell .
let us consider typical vesicles ( linear dimension about @xmath7 ) and small ones ( @xmath8 ) .
typical concentrations of macromolecules may be in the millimolar to nanomolar range ; the excepted numbers of molecules in a single protocell are therefore given in table [ tab : expnumofmols ] .
.excepted number of molecules of a given species in a given protocell ; rows refer to protocell volumes , columns to concentrations .
[ cols="^,^,^,^,^,^",options="header " , ]
in this paper we introduced a simplified model of a non - growing protocell and we investigated the behavior of a stochastic model of catalytic reaction networks in such an environment . to the best of our knowledge
this is a novel approach
. the crucial importance of the small size of the protocell has been stressed , and the effects of the fact that some chemicals can be present in low numbers have been investigated . while a broader analysis is ongoing , we have here shown that it is possible to reach different compositions of the chemical species , in the particular case in which some species are present in the bulk at low concentrations .
we have also shown that there are two different , possibly overlapping reasons for this diversity : @xmath0 the random sequence of molecular events involving those species and @xmath1 the random differences in their initial concentrations .
we have also stressed the importance of raf sets in influencing the overall dynamics .
there are several ways in which this work might seed further research .
the most obvious is that of relaxing the physical limitations that have been considered , e.g. infinitely fast diffusion , yet we do not except that this may change the major conclusions summarized above . obviously , a very interesting direction is that of considering a protocell that is able to grow and divide .
the processes involved in protocell growth and replication are indeed complex and , in particular , a necessary condition for its existence and replication is the coupling between the rates of molecules replication and cell growth .
we have shown elsewhere @xcite that the very existence of this coupling suffices to guarantee ( under very general conditions ) that , in the long run , the rate of cell division and that of duplication of the replicating molecules converge to the same value , thereby allowing sustainable growth of a population of protocells . however
, those results were achieved by supposing a fixed set of genetic memory molecules , with some possible extinction .
it could be sound to extend this approach to the case where there are evolving chemical reaction sets and to verify whether synchronization occurs .
an important aspect to be addressed in the case of growing vesicles is also the effect of volume growth on the concentrations of the various chemicals : a preliminary investigation can be found in @xcite .
besides , we have not here explicitly considered the possibly catalytic role of membranes that , as it has been discussed in section [ sec : intro ] , might be a major cause of the difference between the intracellular environment(s ) and the bulk . in a fixed volume model this effect can be lumped in the effective reaction rates , but if we consider a growing protocell we have to take into account the differences between surfaces and volumes .
this might also lead to interesting phenomena that will be analyzed in future developments . to conclude
, different protocells may host different mixtures of molecular species , even if they share the same chemistry ( i.e. , they inhabit the same world ) .
it might be extremely interesting to model the behavior of populations of different protocells of this kind , which may show different growth rates , but may also undergo phenomena like coalescence , exchange of material , etc . thus , further investigations will be indeed necessary to assess different generations of protocell populations and their possible evolution pathways .
+ last but not least , it will be interesting to extend these studies to other protocell architectures like e.g. the los alamos bug , gard models , and others .
stuart kauffman , norman packard and wim hordijk kindly shared with us their deep understanding of autocatalytic sets in several useful discussions .
useful discussions with ruedi fchslin , davide de lucrezia , timoteo carletti , andrea roli and giulio caravagna are also gratefully acknowledged .
the authors are also grateful to giulia begal for kindly drawing the image of fig .
[ fig : protocell ] .
c.d . wishes to acknowledge the project sysbionet ( 12 - 4 - 5148000 - 15 ; imp .
611/12 ; cup : h41j12000060001 ; u.a .
53 ) for the financial support of the work .
+ the final publication is available at springer via http://dx.doi.org/10.1007/s11047-014-9445-6 + +
simulations were performed with the carness simulator developed by the research group . + in the following ,
the baseline setting of the system used in the simulations is reported ( for the parameters that were variated in the different experiments please refer to the text ) : @xmath9 alphabet : a , b , @xmath9 volume = @xmath10 , @xmath9 average catalysis probability = 1 catalyzed reaction for species , @xmath9 maximum length of the species , @xmath11 , @xmath9 @xmath12 , @xmath9 monomers and dimers do not catalyze , @xmath9 @xmath13 , @xmath9 @xmath14 , @xmath9 @xmath15 , @xmath9 @xmath16 .
a filisetti , a graudenzi , r serra , m villani , d de lucrezia , and i poli . . in lenaerts t , giacobini m ,
h bersini , p bourgine , m dorigo , and r doursat , editors , _ advances in artificial life , ecal 2011 proceedings of the eleventh european conference on the synthesis and simulation of living systems _ , pages 227234 .
mit press , cambridge , ma , 2011 .
roberto serra , timoteo carletti , irene poli , marco villani , and alessandro filisetti . .
in _ in j. jost and d. helbing ( eds ) : proceedings of eccs07 : european conference on complex systems .
cd - rom , paper n.68 _ , 2007 . | protocells are supposed to have played a key role in the self - organizing processes leading to the emergence of life .
existing models either @xmath0 describe protocell architecture and dynamics , given the existence of sets of collectively self - replicating molecules for granted , or @xmath1 describe the emergence of the aforementioned sets from an ensemble of random molecules in a simple experimental setting ( e.g. a closed system or a steady - state flow reactor ) that does not properly describe a protocell . in this paper
we present a model that goes beyond these limitations by describing the dynamics of sets of replicating molecules within a lipid vesicle .
we adopt the simplest possible protocell architecture , by considering a semi - permeable membrane that selects the molecular types that are allowed to enter or exit the protocell and by assuming that the reactions take place in the aqueous phase in the internal compartment . as a first approximation
, we ignore the protocell growth and division dynamics .
the behavior of catalytic reaction networks is then simulated by means of a stochastic model that accounts for the creation and the extinction of species and reactions . while this is not yet an exhaustive protocell model , it already provides clues regarding some processes that are relevant for understanding the conditions that can enable a population of protocells to undergo evolution and selection .
* keywords : * autocatalytic sets of molecules , catalytic reaction sets , origin of life , stochastic simulations , protocell |
in a recent paper@xcite , we have developed an analytic continuation method to determine nucleon resonances within a dynamical coupled - channel model of meson - baryon reactions@xcite ( msl ) .
the method has been applied@xcite to extract 14 nucleon resonances from the @xmath6 model developed in ref.@xcite ( jlms ) which has been extended to investigate @xmath7@xcite , @xmath8@xcite and @xmath9@xcite reactions .
the purpose of this paper is to explain how the residues of the extracted resonance poles are determined from the predicted @xmath10 and @xmath11 amplitudes . in section
ii , we briefly review the analytic continuation method developed in ref.@xcite .
section iii is devoted to explaining the determination of the residues of the nucleon resonance poles .
illustrative results for @xmath3 , @xmath4 , and @xmath5 nucleon resonances are presented in section iv and compared with the results from other analysis .
a summary is given in section v.
within the msl formulation@xcite , the partial wave amplitudes of two - body meson - baryon reactions can be written as @xmath12 where @xmath13 represent the meson - baryon ( mb ) states @xmath14 , @xmath15 , and @xmath16_{i , j } \bar{\gamma}_{\alpha , j}(p;e ) \label{eq : tr}\end{aligned}\ ] ] with @xmath17_{i , j}(e)&= & ( e - m_{n^*_i})\delta_{i , j } - \sigma_{i , j } ( e)\ , .
\label{eq : tr - g}\end{aligned}\ ] ] here @xmath18 denote the bare @xmath1 states defined in the hamiltonian .
@xmath19 are their masses . the first term ( called meson - exchange amplitude from nowon ) in eq.([eq : fullt ] )
are defined by the following equation @xmath20 where @xmath21 is defined by the meson - exchange mechanisms , and @xmath22 is the propagator for channel @xmath23 .
the dressed vertices and the energy shifts of the second term in eqs.([eq : tr])-([eq : tr - g ] ) are defined by @xmath24 where @xmath25 defines the coupling of the @xmath26-th bare @xmath1 state to channel @xmath27 . to search for nucleon resonances ,
we have developed@xcite an analytic continuation method to find poles of the scattering amplitude @xmath28 on the unphysical sheets of the complex energy @xmath29-plane . for multi - channel reactions considered here
, there can be many poles associated with a single resonance on different unphysical sheets .
the pole nearest to the physical sheet is supposed to play a dominant effect on physical observables , and is called the resonance pole traditionally .
the other poles are called shadow poles . since @xmath30 and the bare vertex @xmath31
are energy independent within the msl formulation , the analytic structure of the scattering amplitude defined above as a function of @xmath29 is mainly determined by the the green functions @xmath22 . thus the key for selecting the amplitude on physical sheet or unphysical sheet is to take an appropriate path of momentum integration @xmath32 in eqs.([eq : fullt])-([eq : selfe ] ) according to the locations of the singularities of the meson - baryon green functions @xmath33 as @xmath29 move to complex plane .
this can be done independently for each meson - baryon channel . for channel with stable particles such as @xmath6 and @xmath34
, the meson - baryon green function is @xmath35 which has a pole at the on - shell momentum @xmath36 defined by @xmath37 as an example , let us consider the analytic continuation of the amplitude to the unphysical sheet of the @xmath38 channel when the energy @xmath29 is above the threshold @xmath39 and @xmath40 .
the on - shell momentum @xmath36 for such a @xmath29 is on the second and the fourth quadrant of the complex momentum plane . as @xmath41 becomes more negative as illustrated in fig .
[ fig : path ] , the on - shell momentum ( open circle ) moves into the fourth quadrant .
the amplitude on the unphysical sheet can be obtained by deforming the path @xmath32 into @xmath42 or equivalently @xmath43 so that the on - shell momentum does not cross the integration contour .
we note here that for the energy below threshold ( @xmath44 ) the path @xmath43 will give amplitudes on the physical sheet . ) as energy e moves from a real value above the threshold energy to a complex value with negative imaginary part .
@xmath45 in ( a ) or @xmath43 in ( b ) is the integration path for calculating eqs.([eq : mext])-([eq : selfe ] ) amplitude for e on the unphysical plane.,width=453 ] for the channels with unstable particle such as the @xmath46 as an example , the green function is of the following form @xmath47 where @xmath48^{1/2}}. \nonumber \\ \label{eq - sigma - pid}\end{aligned}\ ] ] the @xmath49 green function eq .
( [ eq - g - unst ] ) has a singularity at momentum @xmath50 , which satisfies @xmath51 physically , this singularity corresponds to the @xmath49 two - body scattering state. there is also discontinuity of the @xmath49 green function associated with the @xmath52 cut in @xmath53 , as shown in the dashed line in fig .
[ fig : c2 ] , where @xmath36 is defined by @xmath54^{1/2}.\end{aligned}\ ] ] therefore , for @xmath55 , the integration contour @xmath32 must be chosen to be below the @xmath52 cut ( dashed line ) and the singularity @xmath56 , such as the contour @xmath57 shown in fig.[fig : c2 ] , for calculating amplitudes on the unphysical sheet .
the singularity @xmath58 of the integrand of eq .
( [ eq - sigma - pid ] ) depends on the spectator momentum @xmath59 @xmath60^{1/2 } \ , .
\label{eq : sing - q0}\end{aligned}\ ] ] thus @xmath58 moves along the dashed curve , illustrated in fig.[fig : c3 ] , when the momentum @xmath59 varies along the path @xmath57 of fig.[fig : c2 ] . to analytically continue @xmath61 to the unphysical sheet , the contour @xmath62 of eq .
( [ eq - sigma - pid ] ) must be below @xmath58 .
a possible contour @xmath62 is the solid curve in fig.[fig : c3 ] . for for calculating eqs.([eq : mext])-([eq : selfe ] ) for e on the unphysical plane with the unstable particle propagators , such as eq.([eq - g - unst ] ) for @xmath46 channel .
see the text for the explanations of the dashed line and the singularity @xmath56.,width=226 ] for calculating the @xmath63 self energy eq.([eq - sigma - pid ] ) on the unphysical sheet .
dashed curve is the singularity @xmath58 of the propagator in eq .
( [ eq : sing - q0 ] ) , which depends on the spectator momentum @xmath59 on the contour @xmath57 of fig.[fig : c2 ] . , width=226 ]
we emphasize here that we can deform the contour @xmath32 only in the region where the potential @xmath64 and the bare @xmath1 vertex @xmath65 are analytic .
the contours described above only from considering the singularities of @xmath38 and @xmath52 green functions . thus they must be modified according to the analytic structure of the considered @xmath64 and @xmath65 . within the msl formulation ,
the t - channel meson exchange potential @xmath66 has singularities at @xmath67 with @xmath68 or @xmath69 .
the form of @xmath65 is chosen such that its singularity is at the pure imaginary momentum .
thus the contours have to be chosen to also avoid these singularities . as an example we show in fig.[fig : actual - path ] the singularities associated with the @xmath49 channel at @xmath70mev . the dotted line for @xmath52 cut and
the circle shows @xmath71 are the singularities from the green s function , as discussed above .
the most relevant singularity of the meson - exchange potential in our investigation of electromagnetic pion production amplitude is due to the t - channel pion exchange of @xmath72 , which is shown as the dashed - dot curve .
thus the integration contour has to be modified to the solid curve in fig.[fig : actual - path ] . and @xmath56 are the singularities shown in fig.[fig : c2 ] .
the dashed - dot curve is the singularity of the pion - exchange @xmath8 matrix element at @xmath73 mev .
[ fig : actual - path ] , width=377 ]
the resonance energy ( @xmath74 ) is the position of the pole of the scattering amplitude which is on the sheet nearest to the physical sheet . in principle
the resonance pole can be found in the meson - exchange amplitude @xmath75 and/or resonance amplitude @xmath76 of eq .
( [ eq : fullt ] ) . within the @xmath6 model developed in ref.@xcite ( jlms ) , we find@xcite that resonance poles are only from @xmath76 .
we therefore will only explain how the residues of resonance poles are extracted from this term .
the poles of @xmath76 are found from the zeros of the determinant of @xmath1 propagator defined by eq.([eq : tr - g ] ) @xmath77&= & 0 \ , .
\label{eq : dete}\end{aligned}\ ] ] near the resonance energy @xmath78 , @xmath1 green function can be expressed as @xmath79 where @xmath18 denote the bare @xmath1 state in the hamiltonian and @xmath80 represents @xmath26-th component of the dressed @xmath1 and satisfies @xmath81\chi_j = 0.\end{aligned}\ ] ] if there is only one bare @xmath1 state , it is easy to see that @xmath82 where @xmath83_{e = m_r}$ ] .
if we have two bare @xmath1 states , we find that @xmath84 where @xmath85_{e = m_r}$ ] can be evaluated using eq.([eq : dete ] ) .
we now examine how the residues @xmath80 can be used to see the effects of resonance poles on the full amplitude @xmath86 defined by eq.([eq : fullt ] ) . near the resonance pole
we can perform laurent expansion of the on - shell amplitude @xmath87 for the cases that the resonance poles are from @xmath76 term of eq.([eq : fullt ] ) , one can see from the definitions eqs.([eq : tr])-([eq : tr - g ] ) and eq.([eq : pole - res ] ) that the vertex functions in the residue of eq.([eq : fullt - exp ] ) is determined by the dressed vertex @xmath88 defined by eq.([eq : dressf ] ) @xmath89 the terms @xmath90 and @xmath91 in eq.([eq : fullt - exp ] ) also depend on the matrix elements of meson - exchange amplitude @xmath75 of eq.([eq : fullt ] ) , but will not be discussed here .
the residue of the @xmath6 elastic scattering amplitude @xmath92 characterizes the strength of the coupling of the resonance with @xmath6 channel . using the standard notation
, we have at near resonance position @xmath78 @xmath93_{e\rightarrow m_r } \,,\end{aligned}\ ] ] where @xmath94 is the partial - wave s - matrix . in terms of the normalization of jlms model
, we have @xmath95 we thus have @xmath96 the elasticity of a resonance is then defined as @xmath97 the electromagnetic @xmath0-@xmath1 transition form factor is defined by matrix element of the electromagnetic currents between nucleon and @xmath1 @xmath98 where @xmath99 , @xmath100 with @xmath101 .
the above definition was originally introduced for the constituent quark model@xcite .
if @xmath102 is a resonance state , then the above expression at resonance position @xmath78 must be evaluated by using eq.([eq : gamma - r ] ) .
we thus have @xmath103 where @xmath104 here the additional factor @xmath105 is due to our normalization of the vertex function @xmath106 .
the above helicity amplitudes are in general complex number . @xmath107 and @xmath108 have similar expressions .
in this section , we illustrate our procedures by presenting the results for the pronounced resonances in @xmath4 , @xmath5 and in the most complex @xmath3 partial waves .
their pole positions were determined in ref.@xcite and are listed in table [ tab : poles ] .
our results for @xmath109 and @xmath110 agree well with the values listed by particle data group@xcite ( pdg ) . for @xmath3 channel we found three poles below 2 gev .
two of them near 1360 mev are close to the @xmath49 threshold .
this finding is consistent with the earlier analysis of vpi@xcite and cutkosky and wang@xcite , and the recent analysis by the gwu / vpi@xcite and juelich@xcite groups .
we have shown in ref .
@xcite that all of the three @xmath3 resonances listed in table [ tab : poles ] correspond to a single bare @xmath1 state at 1736 mev .
this is a dynamical verification of the resonance pole - shadow pole relation in coupled - channels reactions , as discussed by eden and taylor@xcite , kato@xcite , and morgan and pennington@xcite .
the extracted residues @xmath111 , defined in eq.([eq : pin - resid ] ) , for @xmath6 amplitude are compared with some of the previous works in table [ tab : residues - pin ] .
we see that the agreement in @xmath4 and @xmath5 are excellent . for @xmath3
, there are significant differences between four analysis .
as discussed in ref.@xcite , it could be mainly due to the differences in the employed reaction models .
on the other hand , the difference between the predicted @xmath3 amplitudes at @xmath112 about 1.6 gev could be the reason why the third @xmath3 pole is not found in juelich analysis .
from the values of @xmath78 of table [ tab : poles ] and @xmath113 of table [ tab : residues - pin ] , we can evaluate the elasticities @xmath114 using eq.([eq : elasticity ] ) .
the results are also listed in table [ tab : poles ] .
we see that our results agree well with pdg values , while some investigations are needed to understand better the comparisons for the two @xmath3 poles near 1360 mev which are close to @xmath49 threshold . to extract helicity amplitudes using eq.([eq : emff ] )
, we use the multipole amplitudes calculated from using the parameters determined in ref.@xcite .
our results at photon point are listed in table [ tab : residues - gn ] .
we observe that the real parts of our results for @xmath4 and @xmath5 are in good agreement with several previous results@xcite .
the large differences in @xmath3 indicate that more investigations are needed to understand the differences between our resonance extraction method within a coupled - channel model and other methods which are mainly based on the briet - wigner parametrization of single channel k - matrix amplitudes . for @xmath4 we can use the standard relation@xcite to evaluate the @xmath0-@xmath63 magnetic transition form factor @xmath115 in terms of helicity amplitudes .
the real parts of our results are the solid circles in fig.[fig : p33-ff ] , which are in good agreement with the previous analysis . in the same figure
, we also show that the imaginary parts of our results are much weaker .
this result and the results of table [ tab : residues - gn ] suggest that we can only make meaningful comparisons with the results from analysis based on the briet - wigner parametrization of single channel k - matrix amplitudes only for the cases that the imaginary parts are small .
this turns out to be also the case of the @xmath116 resonance . in fig.[fig : d13-ff ] , we see that the real parts of our @xmath117 and @xmath107 are in good agreement with the results from clas collaboration@xcite .
the large differences in @xmath108 perhaps are mainly from the fact that the longitudinal parts of the amplitudes can not be well determined with the available data . for @xmath3 ,
the imaginary parts of the calculated helicity amplitudes for the three poles listed in table [ tab : poles ] are very large .
thus it is not clear how to compare our results with previous results .
we thus show both the real parts ( solid circles ) and imaginary parts ( solid triangles ) in fig .
[ fig : p11-ff ] .
it seems that the structure of @xmath118 and @xmath119 are similar . in particular their real parts of @xmath107 change sign at low @xmath120 ,
similar to what haven been seen in the results from clas collaboration@xcite .
however , because of the double pole structure and the large imaginary parts , more detailed investigations are need to make meaningful comparison with previous results .
in this paper , we have briefly reviewed the analytic continuation method developed in ref.@xcite and explained how it is used to determine the residues of nucleon resonance poles . to illustrate our method ,
we have presented the results for resonances in @xmath4 , @xmath5 , and @xmath3 partial waves . for residues associated with @xmath6 channel , we agree with most of the previous results@xcite for @xmath109 , @xmath116 and two @xmath3 poles near 1360 mev . for @xmath121 ,
the calculated elasticity @xmath122 agree well with the value @xmath123 of pdg and ref.@xcite , while this resonance is not reported in the analysis of refs.@xcite . for residues associated with @xmath14 channel ,
the corresponding helicity amplitudes for @xmath4 ( 1232 ) and @xmath5 ( 1521 ) are dominated by their real parts which are in good agreement with other analysis based on the briet - wigner parametrization of k - matrix amplitudes . for @xmath3 resonances ,
the extracted helicities amplitudes have large imaginary parts and more investigations are needed to compare our results with previous analysis .
our next necessary task is to examine how to define the residues associated with unstable @xmath49 , @xmath124 , and @xmath125 channels .
our effort in this direction along with our complete results for the 14 nucleon resonances extracted in ref.@xcite will be reported elsewhere .
-@xmath63 ( 1232 ) transition form factor @xmath127 defined in ref.@xcite .
@xmath128 with @xmath129 ( gev / c)@xmath130 .
the solid circles ( solid triangles ) are the real ( imaginary ) parts of our results .
the other data points are from previous analysis@xcite . ,
width=302 ] form factors .
the data are from clas collaboration@xcite .
transition form factors , title="fig:",width=188 ] form factors .
the data are from clas collaboration@xcite .
transition form factors , title="fig:",width=188 ] form factors .
the data are from clas collaboration@xcite .
transition form factors , title="fig:",width=188 ] transition form factors of @xmath3 . the solid circles ( solid triangles ) are their real ( imaginary ) parts .
, title="fig:",width=188 ] transition form factors of @xmath3 .
the solid circles ( solid triangles ) are their real ( imaginary ) parts .
, title="fig:",width=188 ] transition form factors of @xmath3 .
the solid circles ( solid triangles ) are their real ( imaginary ) parts .
, title="fig:",width=188 ] | we explain how an analytic continuation we have developed recently is applied to determine the residues of the nucleon resonance poles within a dynamical coupled - channel model of meson - baron reactions .
a procedure for evaluating the electromagnetic @xmath0-@xmath1 transition form factors at resonance poles is developed .
illustrative results of the obtained @xmath2 transition form factors for @xmath3 , @xmath4 , and @xmath5 nucleon resonances are presented and compared with previous results . |
atlas and cms collaborations recently announced a small enhancement over smooth background of two photon events with invariant mass 750 gev @xcite .
though statistical significance of this enhancement is not large ( within 3 standard deviations ) , it induced a whole bunch of theoretical papers devoted to its interpretation .
the reason for this explosive activity is clear : maybe the standard model of particle physics is changed at one tev scale , and we are witnessing the first sign of this change . let us suppose that the observed enhancement is due to the @xmath2 decay of a new particle
. then it should be a boson with spin different from one ; the simplest possibility is a scalar particle @xmath0 with @xmath3 gev .
since it decays to two photons , it should be an @xmath4 singlet , and in @xmath5-collisions at the lhc it can be produced in gluon - gluon fusion through the loop of colored particles and in photon - photon fusion through the loop of charged particles .
let us suppose that particles propagating in the loops are heavy , and @xmath0 decays to them are kinematically forbidden .
reduces significantly which makes @xmath6 decays unobservable at the lhc . ]
production cross section is evidently larger in the case of gluon fusion , however @xmath6 branching ratio is suppressed in this case since @xmath7 decay dominates .
we suppose that the particles propagating in the loop are dirac fermions , so they have tree level masses , and that they are @xmath8 singlets .
nonzero hypercharges provide couplings of these particles with photon and @xmath9-boson .
these particles can be quark(s ) ( color triplets ) @xmath10 or lepton(s ) ( color singlets ) @xmath11 .
they couple with @xmath0 by yukawa interactions with coupling constants @xmath12 and @xmath13 correspondingly . in section
[ s : quarkophilic ] we will consider @xmath0 production and decay in the model with extra heavy quark(s ) , in which gluon fusion dominates @xmath0 production ; in section [ s : leptophilic ] we will consider the model with extra heavy lepton(s ) , where @xmath0 production occur in photon fusion , and @xmath14 decay dominates .
in the case of one heavy quark @xmath15 the following terms should be added to the standard model lagrangian : @xmath16 where @xmath17 and @xmath18 are gluon and @xmath19 gauge fields respectively , and @xmath20 are gell - mann matrices . @xmath0 coupling with gluons is generated by the @xmath15-quark loop : @xmath21 where @xmath22 , @xmath23,\ ] ] and @xmath24 for @xmath25 .
inclusive cross section of @xmath0 production in @xmath5 collision at the lhc through gluon fusion is given by : @xmath26 where the so - called gluon - gluon luminosity is given by the integral over gluon distributions : @xmath27 @xmath28 , @xmath29 , and we use @xmath30 . in fig .
[ fig : s - production ] the corresponding feynman diagram is shown .
integrating gluon distributions from @xcite for @xmath31 , @xmath32 , we get @xmath33 nb , @xmath34 . at @xmath35
tev for @xmath36 gev the luminosity @xmath37 , and therefore cross section , is @xmath38 times smaller . in order to take into account gluon loop corrections ,
should be multiplied by the so - called @xmath39-factor which is close to 2 for @xmath40 tev , according to @xcite ( see also fig . 2 in @xcite ) . production . ] in this way for @xmath41 and @xmath42 , substituting @xmath43 , we obtain : @xmath44 which should be multiplied by @xmath45 in order to be compared with experimental observations @xcite .
total width of @xmath0 is dominated by the @xmath46 decay , and from we get : @xmath47 four orders of magnitude smaller than the 45 gev width which ( maybe ) follows from the preliminary atlas data .
thus we conclude that for the models we consider , @xmath0 width should be much smaller than 45 gev .
let us note that cms data prefer narrow @xmath0 ; see also @xcite . decay . ]
@xmath15-quark loop contributes to @xmath6 decay as well ( see fig . [
fig : s - decay ] ) .
the corresponding matrix element equals @xmath48 where the factor @xmath49 corresponds to the three colors , and @xmath50 is the @xmath15-quark electric charge . for @xmath2
width we get : @xmath51 and @xmath52 where we substituted @xmath53 and @xmath54 . , @xmath55 .
] finally , from and we obtain : @xmath56 experimental data provides a value approximately 36 times larger : @xmath57_\text{exp } \approx 10\text { fb } , \label{pp->sx - experimental - value}\ ] ] since with @xmath58 luminosity collected by each collaboration at 13 tev and effectivity of @xmath2 registration @xmath59 @xcite they see about 15 events each . in order to reproduce experimental result we should suppose that six @xmath15-quarks exist .
in this case @xmath60 , @xmath45 remains the same , while the cross section of @xmath0 production should be multiplied by the same factor 36 , and is reproduced .
however , unappealing multiplication of the number of @xmath15-quarks can be avoided . for @xmath61
gev we have @xmath62 and @xmath63 is @xmath64 times larger than what is given in .
thus for @xmath65 we reproduce the experimental number .
is a parameter of perturbation theory , this value of @xmath66 is close to the maximum allowed value in order for the perturbation theory to make sense . ] in figure [ fig : pp->sx ] isolines of the product @xmath67 are shown on @xmath68 plot . . ] in the following we consider the model with one additional quark @xmath15 and @xmath69 @xmath0 can mix with the standard model higgs boson due to renormalizable interaction term @xmath70 , where @xmath71 is the higgs isodoublet .
such an extension of the standard model was studied in our recent paper @xcite .
doublet admixture in the 750 gev boson wave function results in tree level decays @xmath72 , @xmath73 , @xmath74 and @xmath75 , where @xmath76 is the 125 gev higgs boson .
according to eqs .
( 16)(20 ) from @xcite , the sum of these widths equals approximately @xmath77 gev , where @xmath78 is the mixing angle , and @xmath79 gev is the higgs boson vacuum expectation value
. ratio of partial widths at small @xmath78 is @xmath80 as a result , @xmath0 width grows and @xmath45 diminishes correspondingly .
thus , experimental result will not be reproduced . to reduce this effect
we should make the mixing angle @xmath78 small enough .
for example , for @xmath81 we obtain at most 12 mev ( or 11% ) increase of the width of @xmath0 , which is acceptable .
according to eq .
( 7 ) from @xcite , @xmath82 and it is less than @xmath83 for @xmath84 below 15 gev . let us check if @xmath85 decays do not exceed experimental bounds on their relative probability obtained at 13 and 8 tev at the lhc . since @xmath86 is below @xmath87 , we obtain @xmath88_{13\text { tev } } < 33\text { fb},\ ] ] well below experimental upper bound which , according to fig .
11 from @xcite , equals @xmath89 at @xmath90 ( see also @xcite ) .
gluon - gluon luminosity is @xmath38 times smaller at 8 tev , so we get @xmath91_{8\text { tev } } < 9.0 \text { fb},\ ] ] which should be compared with 60 fb experimental upper bound ( fig . 12 from @xcite ) .
more stringent upper bound comes from the search of @xmath92 decays @xcite and equals 40 fb , while in our case the cross section equals 10 fb . since as it has just been written above , at @xmath35 tev the gluon - gluon luminosity is @xmath38 times smaller that at @xmath40 tev , the cms bound from run 1 @xcite @xmath93_{8\text { tev } } < 1.5\text { fb}\ ] ] is ( almost ) not violated in the model considered .
it is natural to suppose that @xmath15-quark mixes with @xmath94- , @xmath95- , and @xmath96-quark which makes it unstable . to avoid lhc run 1 bounds on @xmath97 following from the search of the decays @xmath98 , @xmath99 and @xmath100 @xcite which exclude @xmath15-quark with mass below 700 gev
, we suppose that @xmath101 mixture is small , and @xmath15-quark mixing with @xmath94- and @xmath95-quarks dominates . in this
case bounds @xcite are avoided @xcite .
concerning @xmath0 decays , let us note that the dominant @xmath46 decay is hidden by the two jets background produced by strong interactions .
at 8 tev lhc energy the following upper bound was obtained @xcite : @xmath102_{8\text { tev}}^\text{exp } < 30 \text { pb}. \label{quark-8tev - bound}\ ] ] in our model @xmath103 . from eq . , using gluon - gluon luminosity at @xmath35 tev , parameters from eq . , and @xmath39-factor @xmath104 @xcite , @xcite
, we get @xmath105^\text{theor } \approx 0.39\text { pb } ,
\ { \mathrm{br}}(s \to gg ) \approx 1,\ ] ] two orders of magnitude smaller than the upper bound .
three modes of @xmath0 decays to neutral vector bosons do exist and have the following hierarchy : @xmath106 where @xmath107 @xmath108 is the sine ( cosine ) of electroweak mixing angle . with higgs doublet
is negligible ; in the opposite case @xmath109 can exceed @xmath110 . ]
thus if @xmath6 decays will be observed in future run 2 data , @xmath111 and @xmath85 decays should be also looked for .
if the existence of @xmath0 will be confirmed with larger statistics at the lhc , then it can be studied at @xmath112-colliders as well . for the cross section of two - photon @xmath0 production in the reaction @xmath113 , according to ( * ? ? ?
* eq . ( 48.47 ) ) , @xcite , we have : @xmath114 , \label{ee->ees - xsection}\ ] ] where @xmath115 and @xmath116 is given in eq . .
for @xmath112 collider clic with @xmath117 , substituting in eqs . ,
@xmath65 , @xmath61 gev , @xmath55 , and @xmath62 we obtain : @xmath118 with projected clic luminosity @xmath119 @xcite , during one accelerator year ( @xmath120 sec ) about 300 @xmath0 resonances should be produced .
let us suppose that heavy leptons @xmath11 which couple to @xmath0 have electric charges @xmath121 , and there are @xmath122 such degenerate leptons .
the lagrangian is similar to that of the heavy quarks case : @xmath123 where we assume equal lepton masses and @xmath0 couplings . for @xmath14 width
we obtain : @xmath124 production of @xmath0 at the lhc occurs through fusion of two virtual photons emitted by quarks which reside in the colliding protons .
let us estimate the production cross section .
for the partonic cross section we get : @xmath125 , \label{qq->qqs - xsection}\ ] ] where @xmath126 and @xmath127 are charges of the colliding quarks , @xmath128 is the invariant mass of the colliding quarks , and @xmath129 is given by
. we should multiply by quark distribution functions and integrate over @xmath130 and @xmath131 : @xmath132 where the sum should be performed over valence @xmath133 , @xmath134 , @xmath135 , and @xmath136 quark collisions , and sea quarks should be taken into account as well .
contribution constitutes 50% of the cross section at @xmath40 tev with another 24% coming from @xmath134 and @xmath137 . ]
quark luminosity equals : @xmath138 @xmath139 , @xmath140 .
we take @xmath30 and use quark distributions from @xcite .
quark and gluon luminosity functions for @xmath141 tev and @xmath142 tev are shown in fig .
[ fig : luminosity ] .
+ cross sections in the case of one heavy lepton with charge @xmath143 , yukawa coupling constant @xmath144 and mass @xmath145 gev are shown in table [ lepton - table ] . for @xmath146 mev and @xmath147
tev we get @xmath148 ab , while the experimental result is three orders of magnitude larger .
we come to the conclusion that @xmath149 is needed : we need either 30 leptons with unit charges , or one lepton with charge 6 , or several multicharged leptons .
ab , then 30 should be replaced with 20 . ]
it is natural to suppose that leptons with charge one mix with the standard model leptons and become unstable .
search for such particles was performed at the lhc , and the lower bound @xmath150 gev was obtained @xcite .
see also @xcite , where bounds on masses and mixings of @xmath151 are discussed . for masses above 200 gev the existence of @xmath151 is still relatively unconstrained .
cross section for quasielastic @xmath0 production can be estimated with the help of the following equation : @xmath152.\ ] ] for @xmath40 tev , @xmath144 and @xmath145 gev it equals 4.1 ab .
.cross sections ( in ab ) for double photon production in the leptophilic model for different values of @xmath153 and proton collision energies . [ cols="^,^,^,^,^ " , ]
we analyze the possibility that the enhancement at 750 gev diphoton invariant mass observed by atlas and cms is due to decays of a new scalar @xmath0 .
we found that production of @xmath0 in gluon fusion in a minimal model with one additional heavy dirac quark @xmath15 can have value of @xmath154 compatible with data .
an upper bound on the mixing of @xmath0 with @xmath155 is obtained .
if heavy leptons @xmath151 are introduced instead of @xmath15 , then @xmath0 can be produced at lhc in photon fusion .
however , in order to reproduce experimental data many leptons @xmath11 are needed and/or they should be multicharged .
if the existence of @xmath0 will be confirmed by future data then production of heavy vector - like quarks and/or leptons at the lhc should be looked for .
the search for @xmath156 and @xmath75 would be also of great importance .
s. g. , m. v. and e. zh .
are partially supported under the grants rfbr no .
14 - 02 - 00995 and 16 - 02 - 00342 , and by the russian federation government under the grant nsh-6792.2016.2
. s. g. and e. zh .
are also supported by mk-4234.2015.2 and 16 - 32 - 00241 .
in addition , s. g. is supported by rfbr under grants 16 - 32 - 60115 , by dynasty foundation and by the russian federation government under grant no .
11.g34.31.0047 .
9 the atlas collaboration , atlas - conf-2015 - 081 ( 2015 ) .
the cms collaboration , cms - pas - exo-15 - 004 ( 2015 ) .
l. a. harland - lang , a. d. martin , p. motylinkski , r. s. thorne , eur .
j. * c*75 ( 2015 ) 204 ; arxiv:1412.3989 .
j. baglio , a. djouadi , jhep 1103 ( 2011 ) 055 ; arxiv:1012.0530 .
r. v. harlander , w. kilgol .
88 ( 2002 ) 201801 .
m. r. buckley , arxiv:1601.04751 .
s. i. godunov , a. n. rozanov , m. i. vysotsky , e. v. zhemchugov , eur .
j. c 76 ( 2016 ) 1 ; arxiv:1503.01618 .
the atlas collaboration , atlas - conf-2015 - 59 .
r. franceschini _ et .
al . _ , arxiv:1512.04933 .
the atlas collaboration , eur .
j. c76 ( 2016 ) 1 , 45 ; arxiv:1507.05930 .
the atlas collaboration , phys .
d92 ( 2015 ) 092004 ; arxiv:1509.04670 .
the cms collaboration , phys .
lett . * b*750 ( 2015 ) 494 ; arxiv:1506.02301 .
the atlas collaboration , jhep 1510 ( 2015 ) 150 ; arxiv:1504.04605 .
the atlas collaboration , jhep 1508 ( 2015 ) 105 ; arxiv:1505.04306 .
the cms collaboration , phys .
rev . * d*93 ( 2016 ) 012003 ; arxiv:1509.04177
. m. buchkremer , proc . of 49th rencontres
de moriond on electroweak interactions and unified theories ( 2014 ) , p. 519 ; arxiv:1405.2586 .
the atlas collaboration , phys . rev . * d*91 ( 2015 ) 052007 ; arxiv:1407.1376 .
k. a. olive _ et al . _
( particle data group ) .
c , 2014 , * 38*(9 ) : 090001 .
v. m. budnev , i. f. ginzburg , g. v. meledin , v. g. serbo , phys .
reports 15 ( 1975 ) 181 . l. a. harland - lang , v. a. khoze , m. g. ryskin , arxiv:1601.07187 .
the atlas collaboration , jhep 09 ( 2015 ) 108 , arxiv:1506.01291 .
a. djouadi , j. ellis , r. godbole , j. quevillon , arxiv:1601.03696 . | if decays of a heavy particle @xmath0 are responsible for the diphoton excess with invariant mass 750 gev observed at the 13 tev lhc run , it can be easily accomodated in the standard model .
two scenarios are considered : production in gluon fusion through a loop of heavy isosinglet quark(s ) and production in photon fusion through a loop of heavy isosinglet leptons . in the second case
many heavy leptons are needed or / and they should have large electric charges in order to reproduce experimental data on @xmath1 . |
dynamics of multiple stars contains information on their origin . from this perspective , triple systems where elements of both outer and inner orbits are known and reasonably accurate are particularly interesting , as relative orbit orientation , period ratio , and interaction between subsystems can be studied ( e.g. * ? ? ?
however , such well - studied multiples are still rare , requiring a substantial observational effort .
this work deals with one of those , @xmath1 aquarii .
papers bearing nearly the same title as the present one have been published by @xcite and @xcite . the naked - eye ( @xmath2 mag )
visual binary @xmath1 aquarii was measured in 1779 by w. hershel , but an earlier observation is listed in the first double - star catalog by @xcite , so we do not know the name of its real discoverer . coordinates and basic identifiers of the components are given in table [ tab : id ] .
the proper motion is @xmath3mas yr@xmath4 , the trigonometric parallax is [email protected] @xcite , the spectral type is f3iii - iv .
the two pairs are designated in the wds @xcite as stf 2909 ab and ebe 1 aa , ab , although they have not been discovered by w. struve and j. ebersberger .
l l [ ht ] ab & ads 15971 , hip 110960 , 55 aqr + ab & wds j2228@xmath50001 , stf 2909 ab + a & 22:28:49.91 @xmath500:01:11.8 ( j2000 ) + a & @xmath6 aqr , hr 8559 , hd 213052 + b & @xmath7 aqr , hr 8558 , hd 213051 the wds database contains about 1160 measures of the pair a , b made with different techniques and with a varying precision ( figure [ fig : hist ] ) .
the binary traveled only a short arc during the 19th century , but in the 20th century it became closer and moved faster .
photographic measurements available since 1900 are more precise than visual micrometer estimates .
they revealed deviations from the pure keplerian motion with a 25-year period , first discovered by @xcite . since then
, several authors have undertaken reanalysis of the orbital motions in this triple system ( table [ tab : hist ] ) .
@xcite pointed out that dynamical interaction between components causes measurable deviation from the purely keplerian motion ( see also recent work by * ? ? ? * ) .
yet , @xcite argued that the variation of the osculating orbital elements is too slow to matter and fitted the data by two keplerian orbits .
as shown below , even this first - order approximation of the observed motion was , so far , quite inaccurate ; it is improved here to set the stage for a future dynamical analysis .
the motion of visual binaries is traditionally represented by an orbit of the secondary component b around the primary a. the `` wave '' in the motion was implicitly attributed to the `` dark '' ( unseen ) companion around b , until @xcite demonstrated that the perturbing body actually revolves around the component a. he categorically dismissed the detections of this dark companion to b by infrared interferometry @xcite and the claimed resolution of aa , ab by optical speckle interferometry @xcite , as they did not match the orbit . indeed , the large magnitude difference between aa and ab ( @xmath8 mag ) and the modest dynamic range of early speckle instruments made such a resolution impossible . despite the separation exceeding 05 , the putative subsystem aa ,
ab has not been resolved in subsequent speckle observations until 2009 .
note that @xcite calculated visibility from one - dimensional scans in the north - south direction that can not distinguish which of the binary components is resolved . their attribution of the tertiary to b was based entirely on the literature . at the time of their observation , 1981.94 , the position of aa , ab was at ( 1686 , 023 ) according to the new orbit , so it does not contradict their measured projected separation of 017@xmath0004 .
however , the relative photometry in @xcite disagrees with our estimates , throwing doubt on this claimed resolution .
the first real detection of the third body in the @xmath1 aqr system has been made by adaptive optics in the @xmath9 band at the southern astrophysical research telescope , soar @xcite .
the two stars a and b were observed separately , and the tertiary has been erroneously attributed to b. this has been corrected in the following observations , securely associating the low - mass tertiary with the main component a. direct images of the subsystem aa , ab are given by @xcite .
the tertiary ab , 5.4 mag fainter than aa in the @xmath9 band , is just detectable by modern speckle interferometry , which furnishes better spatial resolution and more accurate positions compared to the long - exposure imaging with partial ao correction ( figure [ fig : acf ] ) .
the resolved measures of aa , ab do not match the latest astrometric orbit of aa , ab by @xcite .
arbitrary adjustment of the orbit was suggested by @xcite to reach a better agreement . since the work by @xcite , the inner subsystem aa , ab has made one full revolution ,
so the next iteration on the orbits , now accounting for the resolved measures of aa , ab , can be made .
this is the purpose of the present work .
l ccc ccc @xcite & 400 & 3.403 & 0.60 & 25 & 0.080 & 0 + @xcite & 600 & 4.013 & 0.45 & 25.5 & 0.097 & 0.20 + @xcite & 856 & 5.055 & 0.50 & 25.5 & 0.072 & 0.26 + @xcite & 760 & 4.507 & 0.50 & 25.7 & 0.076 & 0.59 + scardia+ ( 2010 ) & 487 & 3.380 & 0.43 & 25.8 & 0.062 & 0.13 + this work & 540 & 3.496 & 0.42 & 26.0 & 0.110 & 0.87
in my first attempt to update the orbits , i used an obvious but wrong strategy : fit the motion of a , b in the outer orbit and then represent the residuals from this motion by the inner astrometric orbit of aa , ab .
the inner orbit is eccentric , therefore the star aa , practically coincident with the photo - center a , spends most of the time near apastron and is displaced from the center of mass of aa , ab .
the outer orbit must describe the motion of b around the center of mass , not around the photo - center ( component a ) . when the outer orbit is computed from the positions of a , b , the average residuals are minimized , and the perturbation ( deviations of a from the outer orbit ) becomes centrally symmetric , leading to the false inner orbit with a small eccentricity .
this explains why astrometric orbits tend to have small eccentricities in general .
whenever an astrometric binary is resolved , it often turns out that its true visual orbit has a larger eccentricity than the astrometric orbit .
@xcite discussed this effect and insisted on fitting the two orbits simultaneously .
yet , @xcite ignored his warning and derived the inner orbit of @xmath1 aqr with a small eccentricity after fitting and subtracting the outer orbit .
i include the resolved measures of aa , ab together with the positions of a , b in the data set and fit 15 free parameters : seven elements of a , b , seven elements of aa , ab , and @xmath10 , the ratio of the true and astrometric semimajor axes in the inner orbit . if the subsystem belonged to the b , the parameter @xmath10 would be negative ( the reflex displacement of b is oposite to the vector ba , bb ) , but here it is positive .
i modified the idl code orbit.pro to fit both orbits simultaneously .
the least - squares fitting implicitly assumes that the measurement errors are known and normally distributed . in this ideal case ,
the weights are inversely proportional to the squares of the measurement errors @xmath11 .
it is well known , however , that visual micrometer measures of binary stars are not normally distributed , showing both large occasional deviations and systematic errors ( see figure [ fig : hist ] ) .
there is no general rule allowing to describe these data by normally distributed random variables .
photographic measurements of binary stars can also have larger - than - usual errors caused by poor seeing , systematic errors proper to photography ( interaction of closely spaced images in the emulsion ) , and subjective errors of persons who measured the plates .
the ccd - based measurements should be better than the photography , although this is not the case for some low - quality ccd measures made by amateurs . in the speckle interferometry ,
the uncertainty of the pixel scale and detector orientation , difficult to quantify , is often a major contributor to the position errors of wide pairs such as @xmath1 aqr .
published errors of speckle measures do not always include the calibration errors .
the only viable approach to this problem is to censor the data by rejecting obvious outliers and to treat the errors of the remaining measurements as unknowns , estimating them from the orbital solution itself .
the large number of measurements of @xmath1 aqr favors this strategy .
i used averaged micrometer measures made prior to 1900 , taking them from the paper by @xcite , and assumed their errors of 01 .
all micrometer measures made after 1900 are ignored , as more accurate and objective photographic observations became available .
i adopted the errors of photographic measures prior to 1940 as 50 mas , then as 30 mas , same as for the accepted ccd measures .
the speckle measures are assumed to have errors of 3 to 9 mas .
obvious outliers were deleted .
then , in the process of fitting the orbits , those measures that deviated by more than @xmath12 in either coordinate were given an increased @xmath11 to bring their deviation to the @xmath13 level .
after several rounds of manual error adjustment and automatic down - weighting of outliers , the resulting normalized goodness of fit @xmath14 ( @xmath15 being the number of degrees of freedom ) became close to one in both coordinates .
the 322 retained observations of a , b , their adopted errors , and residuals to the orbits are listed in table [ tab : obsab ] . given the large number of measures , it is not practical to cite references to all of them .
a similar table [ tab : obsaa ] gives the resolved measures of aa , ab , all coming from the speckle interferometry at soar .
the latest measure in 2016.39 is still unpublished . in the following ,
i denote the orbital elements of the outer and inner pairs by the indices 1 and 2 , respectively , and use common notation ( @xmath16 period , @xmath17 epoch of periastron , @xmath18 eccentricity , @xmath19 semimajor axis , @xmath20 position angle of the node , @xmath21 argument of periastron , @xmath22 orbital inclination ) .
the orbital elements found by fitting 15 parameters are listed in table [ tab : orb ] . for the inner orbit , the astrometric semimajor axis @xmath23 is given ; it should be multiplied by @xmath10 to get the true axis .
the weighted rms residuals to the measures of a , b are 15mas in both coordinates .
the resolved measures of aa , ab have rms residuals of 10mas .
they are less accurate than most speckle data from soar because of the large magnitude difference between aa and ab .
the periods @xmath24 and @xmath25 are similar to their values found in the previous studies , but the inner eccentricity @xmath26 found here is substantially larger ( see table [ tab : hist ] ) .
although the overall character of the orbital motion is well defined , the range of inner orbits compatible with the data is quite large , more than suggested by the formal errors of the elements .
additional consideration helps to select the most plausible inner orbit .
owing to the faintness of ab , it is safe to assume that the photo - center of the composite component a is located at the primary aa , so the measurements of a , b and aa , b are equivalent . in such case , the mass ratio in the inner pair @xmath27 .
the mass sum of the inner pair is proportional to @xmath28 , the mass sum of the outer pair is proportional to @xmath29 . as the stars aa and b have very similar brightness , we expect equal masses @xmath30 .
so , the ratio of the mass sums in the inner and outer orbits should be close to @xmath31 .
this leads to @xmath32 both periods and the outer semimajor axis @xmath33 are rather well defined , hence the inner semimajor axis should be @xmath34 according to this equation .
the unconstrained least - squares fit produces an inner orbit with @xmath35 and @xmath36 , too large to match the expected inner mass sum .
convergence of astrometric orbits derived from noisy data to nearly parabolic solutions has been studied by @xcite , and here we have a similar situation . as the line of apsides of aa , ab is almost perpendicular to the line of nodes ( @xmath37 ) , increasing the inclination @xmath38 , the eccentricity @xmath39 , and the semimajor axis @xmath23 simultaneously has little effect on the apparent orbital ellipse .
the measures alone do not constrain well enough the semimajor axis of the inner orbit and allow solutions with unrealistically large @xmath23 .
the final orbits ( figure [ fig : ab ] ) were obtained by fixing @xmath23 to the value given by ( [ eq : masses ] ) and fitting all other elements plus @xmath10 .
then the new value of @xmath23 was computed and the constrained fit was repeated , until the iterations converged to @xmath40 .
there is no doubt that the inner orbit has a high eccentricity ; for example , a fixed @xmath41 leads to substantially larger residuals .
visual measures in the 19th century seem to deviate systematically from the orbit of a , b .
this could be caused by an over - estimation of the angular separation .
however , such residuals might also result from dynamical interaction between the orbits , according to @xcite .
one can not help noting that the outer and inner pairs move in the opposite directions , retrograde and prograde , hence their orbital angular momenta can not be aligned .
there are no radial velocity ( rv ) data of adequate precision to establish the correct orbital nodes .
examination of the disparate published rvs indicates that rv(a ) was less than rv(b ) in the first half of the 20th century , meaning that the node @xmath42 corresponds to the component a. the amplitude of the rv difference between a and b is about 4 km s@xmath4 , presently it is + 3 km s@xmath4 according to the new orbit .
however , the orbit of aa , ab is seen almost face - on , so the rvs of a might be of little help for defining the inner node .
fortunately , the two possible angles between the orbital angular momentum vectors computed without knowledge of the true nodes are close to each other : @xmath43 and @xmath44 .
the period ratio is @xmath45 .
relative orientation of the orbits and the high inner eccentricity are strongly suggestive of the kozai - lidov cycles in this triple system .
if the inner orbit were originally almost perpendicular to the outer orbit , it would evolve towards high eccentricity and relative inclination of 39or 141 . in a system of three point masses
the inclination and inner eccentricity oscillate with a period on the order of @xmath46kyr , but real stars can tidally interact at periastron , locking the inner orbit in the high-@xmath18 state , with subsequent circularization @xcite .
however , the distance between aa and ab at periastron is about 1.4 a.u .
, too large for tidal interaction .
r rrr rr 1826.4000 & 358.8 & 3.6000 & 0.100 & 1.5 & @xmath50.0793 + 1832.6000 & 354.7 & 3.6000 & 0.100 & 1.5 & 0.0295 + 1838.8900 & 350.4 & 3.7500 & 0.100 & 0.8 & 0.1918 + & & & & & + 2014.7631 & 164.5 & 2.2928 & 0.009 & 0.0 & @xmath50.0020 + 2015.7379 & 163.7 & 2.3261 & 0.003 & 0.1 & 0.0094 + 2016.3901 & 162.7 & 2.3212 & 0.003 & @xmath50.2 & @xmath50.0091 r rrr rr 2009.7552 & 192.0 & 0.4057 & 0.009 & 4.9 & 0.0015 + 2009.7552 & 187.9 & 0.4115 & 0.009 & 0.8 & 0.0073 + 2010.8914 & 193.4 & 0.4760 & 0.009 & 0.5 & @xmath50.0031 + 2010.9681 & 193.2 & 0.4976 & 0.009 & 0.0 & 0.0140 + 2012.9229 & 200.2 & 0.5835 & 0.009 & 0.2 & 0.0042 + 2013.7364 & 201.8 & 0.6211 & 0.009 & @xmath50.4 & 0.0113 + 2014.7631 & 204.7 & 0.6218 & 0.009 & @xmath50.1 & @xmath50.0198 + 2015.7379 & 207.5 & 0.6700 & 0.009 & 0.4 & 0.0046 + 2016.3901 & 208.0 & 0.6665 & 0.009 & @xmath50.5 & @xmath50.0116
l l rrr rrr r ccc a , b & 540 & 1981.50 & 0.419 & 3.496 & 131.3 & 269.3 & 142.0 & + & @xmath015 & @xmath00.58 & @xmath00.011 & @xmath00.046 & @xmath00.8 & @xmath01.7 & @xmath00.4 & + aa , ab & 25.95 & 2006.52 & 0.872 & 0.110 & 293.7 & 100.9 & 11.8 & 3.50 + & @xmath00.048 & @xmath00.13 & @xmath00.006 & fixed & @xmath074 & @xmath073 & @xmath06.7 & @xmath00.09
l ccc [ ht ] a+b & 3.65 & 3.04 & 2.72 + b@xmath5a & 0.21 & 0.27 & 0.33 + ab@xmath5aa & ( 6.3 ) & 5.44 & ( 4.0 ) + aa & 4.30 & 3.67 & 3.21 + b & 4.51 & 3.94 & 3.93 + ab & ( 10.60 ) & 9.11 & ( 7.3 ) the _ hipparcos _ parallax of [email protected] ( distance modulus 2.25 mag ) relates orbital elements to masses .
the mass sum of ab is [email protected] @xmath47 and the mass sum of a is [email protected] @xmath48 . as noted above , the inner mass sum is not tightly constrained , for this reason the near - equality of the masses of aa and b was imposed in the orbit fit .
the masses of aa , b , and ab deduced from the orbits and the mass ratio @xmath49 are 1.4 , 1.4 and 0.6 @xmath48 , with an uncertainty of 10% .
available combined and differential photometry is collected in table [ tab : ptm ] .
the component a is brighter and redder than b , as the magnitude difference between a and b increases with wavelength : @xmath50 mag , @xmath51 mag @xcite , @xmath52 mag @xcite , @xmath53 mag @xcite
. six measurements of the magnitude difference between aa and ab in the @xmath9 band made at soar average at 5.44 mag with the rms scatter of 0.13 mag .
considering that the absolute @xmath54 magnitude of ab @xmath55 mag matches a main - sequence star of 0.6 @xmath47 , its magnitudes in @xmath56 and @xmath57 are estimated from the isochrone , not measured directly ( numbers in brackets ) .
individual magnitudes of the components are computed from the combined and differential photometry .
figure [ fig : iso ] places the components of @xmath1 aqr on the @xmath58 color - magnitude diagram .
the most massive stars aa and b are located above the main sequence and match the 3-gyr dartmouth isochrone for solar metallicity @xcite .
the corresponding masses are 1.42 @xmath59 , in excellent agreement with the orbit .
this means that this system is unlikely to contain additional close ( spectroscopic ) stellar companions to any of the stars .
the near - equality of the masses of a and b , adopted above as a hypothesis , is supported by the isochrones because at this evolutionary stage the dependence of color on mass is strong .
the age of this system is close to 3gyr .
the stars aa and b have a fast axial rotation ( about 50 km s@xmath4 ) , not uncommon for their spectral type and evolutionary status as subgiants .
fast rotation is related to the high chromoshheric activity and x - ray luminosity @xcite . @xcite
measured a high lithium abundance in the photosphere of these stars .
most spectroscopic studies treat the object as a single star , given that the light of two closely spaced and nearly equal components aa and b is usually mixed in the slit .
when the spectra of both components were taken separately , the unusual slit illumination possiblly caused discordant rvs and false claims of rv variability .
@xcite found a constant rv of 25.90 km s@xmath4 over a 3.8-yr time span in the combined light of both components . using this rv and the _ hipparcos _ astrometry , the galactic velocity is @xmath60 km s@xmath4 ( @xmath61 is directed away from the galactic center ) .
the spatial motion does not match any known kinematic group , supporting the view that this multiple system is not very young .
comparable separations in the inner and outer orbits of @xmath1 aqr raise concern of its dynamical stability .
there are no exact formulations for a criterion of dynamical stability of triple systems , while the existing approximate criteria do not differ substantially between themselves .
one of the popular criteria by @xcite for coplanar orbits can be written as @xmath62 in present notation .
this translates to @xmath63 for @xmath64 and @xmath65 . according to this criterion ,
@xmath1 aqr with @xmath66 is within the stability boundary . on the other hand ,
the `` empirical stability criterion '' by @xcite , @xmath67 , calls for @xmath68 , placing this triple system in the instability zone .
given the age of this system , we know that it is in fact stable . in any case , the dynamical interaction between inner and outer orbits should be strong , and the description of the motion by two keplerian orbits can certainly be improved by taking it into account .
the stability criteria can be checked against multiple systems with known outer and inner orbits .
a selection of such systems from the current ( 2016 ) version of the sixth catalog of visual binary stars @xcite is shown in figure [ fig : stab ] .
points above both curves are due to the wrong orbital elements ( visual orbits with long periods and small coverage lack credibility ) .
even such a famous binary as @xmath1 aqr had outer orbits with periods ranging within a factor of two ( table [ tab : hist ] ) .
there is a group of several other triples with similar outer eccentricities and period ratios on the order of 20 .
three triple systems in this group have inner secondary components of very low mass discovered by @xcite . properties of multiple stellar systems are related to their origin .
the orbital architecture of @xmath1 aqr is suggestive of chaotic dynamical interactions , rather than dissipative evolution in accretion disks associated with coplanar multiple systems like hd 91962 @xcite .
however , the masses of the main components aa and b and their axial rotation are remarkably similar ( a twin binary ) .
it is likely that the axes of stars aa and b are aligned with each other and , possibly , with the outer orbit , rather than with the inner orbit .
the outer binary aa , b could have been formed by a collapse of a rotating core with subsequent accretion that made the components masses nearly equal . in this case , the third star ab is an intruder that met the binary aa , b at a later time and was captured on a nearly perpendicular orbit that is now undergoing kozai - lidov oscillations .
the dynamical capture of ab could happen in the nascent cluster from which this system formed , or in an unstable quadruple system where a pair of low - mass stars was disrupted , leaving one of its components bound to a , b and ejecting another .
some other multiple systems with a high mutual inclination of their orbits could be produced by such dynamical interplay .
@xcite noted the similarity between @xmath1 aqr and the nearby quadruple system @xmath69 uma ( ads 8119 ) .
the latter consists of two solar - mass stars on a 60-year outer orbit @xcite .
each star has a low - mass close companion .
the orbit of aa , ab is inclined to the outer orbit a , b by 130 and has a substantial eccentricity , just like aa , ab in @xmath1 aqr .
the subsystem ba , bb with a period of 4 days and a minimum bb mass of only 0.04 @xmath59 could be a product of kozai - lidov cycles with tidal dissipation if its initial orbit were nearly perpendicular to the orbit of a , b .
however , the orbital periods in @xmath69 uma are @xmath7010 times shorter than in @xmath1 aqr .
the next periastron passage in the inner subsystem of @xmath1 aqr will occur in 2032.6 .
knowing that the orbit is very eccentric , this event should be observed by spectroscopy and imaging .
rv variations near the periastron , if detected , will inform us on the exact inner eccentricity and inclination .
meanwhile , the motion of aa , ab should be monitored by speckle interferometry in the visible , while imaging in the near - infrared is needed to secure differential photometry and to measure the colors of ab .
the pair aa , ab is now approaching its maximum separation of almost 07 , facilitating its resolved photometry and , possibly , spectroscopy .
i thank b. mason for extracting all measures and references from the wds database .
this work used the simbad service operated by centre des donnes stellaires ( strasbourg , france ) , bibliographic references from the astrophysics data system maintained by sao / nasa , and the washington double star catalog maintained at usno .
mccarthy , d. w. , low , f. j. , kleinmann , s. g. & arganbright , d. v. 1982 , apj , 259 , l75 muterspaugh , m. w. , lane , b. f. , kulkarni , s. r. et al . 2010 , aj , 140 , 1657 nordstrm , b. , mayor , m. , andersen , j. et al . 2004 , a&a , 418 , 989 | zeta aquarii is a bright and nearby ( 28pc ) triple star with a 26-year astrometric subsystem .
almost a half of the outer 540-year visual orbit has been covered in 238 years of its observations .
both inner and outer orbits are revised here taking into account recent direct resolution of the inner pair aa , ab .
the inner orbit has a high eccentricity of 0.87 and is inclined to the outer orbit by 140@xmath010 , suggesting that kozai - lidov cycles take place .
the masses of the stars aa , b , and ab are 1.4 , 1.4 , and 0.6 solar .
the age of the system is about 3gyr , and the two main components have just left the main sequence .
hypothetically , this system could have formed by a dynamical capture of the small star ab in the twin binary aa , b . |
the linear stability analysis with the parallel base flow assumption is termed as local stability analysis where the base flow is varying only in wall normal direction .
the linear stability of shear flows is governed by classical orr - sommerfeld equation .
the numerical solutions to such problems deals with solutions of orr - sommerfeld equation obtained for disturbance amplitude functions are of wave - like nature .
it is assumed that streamwise wavenumber is constant at each streamwise location and disturbance amplitude functions are a function of wall normal coordinate only .
the numerical solution of orr - sommerfeld equation parallel base flow assumption can be achieved with very limited computational efforts @xcite .
the numerical solutions of orr - sommerfeld equation obtained for disturbance amplitude functions are of wave - like nature .
it is assumed that streamwise wavenumber is constant at each streamwise location and disturbance amplitude functions are a function of wall normal coordinate only . in spatially growing boundary layers ,
the base flow velocity is varying in streamwise direction which results in a considerable wall normal velocity . in these cases ,
non - parallel effects are very strong and can not be neglected .
this is true at low and moderate reynolds numbers .
the parallel flow assumption is not a good approximation for real flows where the flow is varying in streamwise direction . under such conditions
the flow becomes two - dimensional , and only wave - like nature of the solution is not expected .
however , it may be one of the solution .
therefore , stability analysis with the parallel flow assumption may not represent the complete flow physics , and hence , it is having very limited scope for parallel and weakly non - parallel ( wnp ) flows only .
+ in the last decade , the stability analysis of non - parallel flow have received much attention through global stability analysis . here
the base flow is varying in both , radial and axial directions .
the term global instability is used to represent instability analysis of the such flows @xcite .
the base flow considered here is non - parallel due to continuous growth of boundary layer in spatial directions and not self - similar due to transverse curvature effect @xcite .
+ global temporal modes are time harmonic solutions of the homogeneous linearized disturbance equations with homogeneous boundary conditions in space . such solution obtained numerically for a base flow with strong non parallel effects @xcite .
if the base flow is varying slowly on the scale of a typical instability wavelength , global modes can be recovered by wkbj type analysis @xcite .
a boundary layer at high reynolds numbers and flow through diverging channels at small angles of divergence are common examples of weakly non - parallel flows .
the boundary layer at low and moderate reynolds number have strong non - parallel effects and hence wkbj approach may not be the right choice .
the prediction by non - parallel theory also agrees better with the experiments at moderate reynolds numbers .
+ joseph is the first among all to introduce the term global stability analysis , where he studied perturbation energy and established lower bounds for flow stability @xcite .
bert reported the discovery of short - wavelength elliptic instability in inviscid vortex flows , a problem related to both transition and turbulence research .
bert discovered short - wavelength instability for inviscid vortex flow related to transition and turbulence research @xcite .
jackson reports first computations on viscous flow where he studied flow past various shaped bodies @xcite .
onset of periodic behavior in two - dimensional laminar flow past bodies of various shapes is examined by means of finite - element simulations .
zebib studied growth of symmetric and asymmetric perturbations for flow past a circular cylinder @xcite .
they found that the symmetric disturbances are stable , the asymmetric perturbations become unstable at a reynolds number about 40 with a strouhal number about 0.12 .
the global approach is used by christodoulou for free surface flow @xcite , tatsumi for rectangular duct flow @xcite and lin for boundary layer flows @xcite .
the matrices generated by the discretization of global stability equations are very large , non - symmetric and sparse , which requires significant computational resources .
all above said work have implemented qr algorithm for computations of eigenvalues and eigenvectors of resulting matrices .
this algorithm computes all the eigenvalues .
however , flow becomes unstable in case of a shear flow due to very few dangerous modes only , it is more economical to calculate that modes only .
iterative techniques like arnoldi s algorithm with shift - invert strategy computes few selected eigenvalues near the shift value .
theofilis studied 2d steady laminar separation bubble using wnp and dns and got an excellent agreement of the stability results for both 2d and 3d @xcite . by performing a global stability analysis
, they proved the existence of new instability modes that are not explored by either of the approaches i.e. wnp & dns .
theofilis studied the global stability of separated profiles in three different flow configurations , and they show that the amplitude of the global mode is less in the separated region than in the wake region or shear layer region @xcite .
ehrenstein studied the global instability of a flat plate boundary layer @xcite . by an optimal superposition of the global temporal modes
, they were able to simulate the convective nature of instability of the boundary layer .
alizard performed a global stability analysis on a flat plate boundary layer and obtained an excellent agreement with the local stability results and the wnp results @xcite .
monokrousos performed a global stability analysis of the flat plate boundary layer and calculated the maximum energy growth with a reduced order model and found that optimal energy growth can not be obtained with least stable global modes only but need to consider few globally stable modes also @xcite .
this is essential for flow control .
tezuka has first time solved initial value problem based on the same approach for three - dimensional disturbances @xcite .
+ the study of stability analysis of incompressible axisymmetric boundary layer is found very sparse in literature . the available literature on axisymmetric boundary layers
are limited to local stability analysis .
rao first studied the stability of axisymmetric boundary layer @xcite .
he found that non - axisymmetric disturbances are less stable than that of two - dimensional disturbances .
the estimated critical reynolds number based on free stream velocity and body radius of the cylinder was 11,000 .
later work of tutty investigated that for non - axisymmetric mode critical reynolds number increases with azimuthal wave number n @xcite .
the critical reynolds number found to be 1060 for n=1 mode and 12439 for n=0 mode .
the axisymmetric mode is found least stable fourth mode .
recently vinod investigated that higher non - axisymmetric mode @xmath0 is linearly stable for a small range of curvature only @xcite .
the helical mode @xmath1 is unstable over a significant length of the cylinder , but never unstable for curvature above @xmath2 .
thus , transverse curvature has overall stabilizing effect over mean flow and perturbations .
malik studied the effect of transverse curvature on the stability of incompressible boundary layer @xcite .
they investigated that the body curvature and streamline curvature are having significant damping effects on disturbances .
they also found that traveling waves are the most amplified waves in three - dimensional boundary layers .
the secondary instability of an incompressible axisymmetric boundary layer is also studied by vinod @xcite .
they found that laminar flow is always stable at high transverse curvature to secondary disturbances .
parallel base flow assumption is considered in all above investigations .
+ the main aim of this paper is to study the global stability characteristics of the axisymmetric boundary layer and the effect of transverse curvature on it .
the characteristics of the global eigenmodes of axisymmetric boundary layer are compared with the flat plate boundary layer(with zero transverse curvature ) at different reynolds numbers.the comparison shows that the transverse curvature has overall stabilizing effect on the disturbances .
we follow the standard procedure to derive linearized navier - stokes equations for disturbance flow quantities .
the navier - stokes equations for mean flow and instantaneous flow are written in polar cylindrical coordinates ( @xmath3 ) .
the equations are non - dimensionalised using free stream velocity ( @xmath4 ) and displacement thickness ( @xmath5 ) at the inlet of the domain . following the standard procedure of stability analysis
, we get the set of equations for disturbance flow quantities .
the base flow is axisymmetric and disturbances are three - dimensional in nature .
the reynolds number is defined as , @xmath6 the flow quantities are split into base flow and perturbations , @xmath7 we assume normal mode form for disturbances with amplitude varying in radial(@xmath8 ) and streamwise(@xmath9 ) directions .
it is correct to assume periodicity in azimuthal(@xmath10 ) direction as the flow is axisymmetric .
@xmath11}\ ] ] where , + q = [ u , v , w , p ] + q = [ u , v , w , p ] + @xmath12 $ ] + @xmath13 = frequency of the waves + n= azimuthal wave number + @xmath4=free stream velocity + @xmath14 = displacement thickness + @xmath15 = body radius of the cylinder + where , q , q and @xmath16 are the perturbation , mean and instantaneous flow quantities respectively .
u , v and w are the disturbance velocity components in the axial(@xmath9 ) , radial(@xmath8 ) and azimuthal(@xmath10 ) directions respectively .
the linearized navier - stokes equations for instability analysis are as follows , @xmath17=0\ ] ] @xmath18=0\ ] ] @xmath19=0\ ] ] @xmath20 where , @xmath21 at wall , on the solid surface of cylinder no slip and no penetration boundary conditions are applied .
all disturbance velocity components have zero magnitude at the solid surface of the cylinder .
+ @xmath22 at free stream far away from the surface of cylinder exponential decay of disturbances is expected .
homogeneous dirichlet boundary conditions are applied for velocity and pressure disturbances at free - stream .
@xmath23 the boundary conditions in streamwise direction are not straight forward .
homogeneous dirichlet boundary conditions are applied on disturbance velocity components at the inlet .
this boundary condition is appropriate as suggested by theofilis ( 2003 ) because we are interested in the disturbances evolved within the basic flow field only . at outlet
one may apply boundary conditions based on the incoming / outgoing wave information @xcite .
such conditions impose wave - like nature of the disturbances and it is more restrictive in nature .
this is not appropriate from the physical point of view in instability analysis .
even the wave number @xmath24 is not known initially in case of a global stability analysis .
an alternative way is to impose numerical boundary condition which extrapolate information from the interior of the computational domain .
linear extrapolated conditions are applied by several investigators .
review of literature on global stability analysis suggests linearly extrapolated boundary conditions are the good approximations @xcite .
we also tried second order extrapolated boundary conditions , however the difference is only marginal . thus , we considered linear extrapolated conditions at outlet .
@xmath25 @xmath26-u_{n-1}[x_n - x_{n-2}]+ u_n[x_{n-1}-x_{n-2}]=0 % \ ] ] similarly , one can write extrapolated boundary conditions for wall normal and azimuthal disturbance components v and w respectively .
the boundary conditions for pressure do not exist physically at the wall . however , compatibility conditions derived from the linearized navier - stokes equations are collocated at the wall of cylinder @xcite . @xmath27 -u\frac{\partial u } { \partial x}- v\frac { \partial u } { \partial r}\ ] ] @xmath28 -u\frac{\partial v}{\partial x}- v\frac { \partial v}{\partial r}\ ] ]
the linearized navier - stokes equations are discretized using chebyshev spectral collocation method . the chebyshev polynomial generates non - uniform grids and generates more collocation points towards the ends .
it is a favourable arrangement for the boundary value problems .
@xmath29 @xmath30 where n and m are number of collocation points in streamwise and wall normal direction .
the gradient of disturbance amplitude functions is very large near the wall region within the thin boundary layer , which requires a large number of grid points to increase the spatial resolution .
grid stretching is applied via the following algebraic equation @xcite .
@xmath31 in the above grid stretching method , half number of the collocation points are concentrated within the @xmath32 distance from the lower boundary only .
the non - uniform nature of distribution for collocation points in the streamwise direction is undesirable .
the maximum and minimum distance between the grid points are at center and end respectively .
thus , it makes a poor resolution at the center of the domain and a very small distance between the grids at the end gives rise to gibbs phenomenon . to improve the resolution and to minimize the gibbs oscillation in the solution , grid mapping
is implemented in streamwise direction using following algebraic equation @xcite .
@xmath33 the value of @xmath34 is selected carefully to improve spatial resolution in the streamwise direction .
a very small value of @xmath34 keeps the grid distribution like chebyshev and near to unity almost uniform grid . for the detail description of the grid mapping readers
are suggested to refer @xcite .
to incorporate the effect of physical dimensions of the domain [ @xmath35 , @xmath36 along with grid stretching and mapping it is required to multiply the chebyshev differentiation matrices by proper jacobean matrix .
once all the partial derivatives of the lns are discretized by spectral collocation method using chebyshev polynomials , the operator of the differential equations formulates the matrices a and b. these matrices are square , real and sparse in nature , formulates general eigenvalues problem . @xmath37
\left [ \begin{array}{c } u \\ v \\ w \\
p \end{array } \right]= i\omega \left [ \begin{array}{cccc } b_{11 } & b_{12 } & b_{13 } & b_{14 } \\
b_{21 } & b_{22 } & b_{23 } & b_{24 } \\
b_{31 } & b_{32 } & b_{33 } & b_{34 } \\
b_{41 } & b_{42 } & b_{43 } & b_{44 } \end{array } \right ] \left [ \begin{array}{c } u \\ v \\ w \\
p \end{array } \right]\ ] ] @xmath38[\phi ] = i\omega [ b][\phi]\ ] ] where a and b are square matrix of size 4 nm , @xmath39 is an eigenvalues and @xmath40 is a vector of unknown amplitude of disturbance flow quantities u , v , w and p. the above mentioned all the boundary conditions are properly incorporated in the matrix a and b. the matrix a and b are sparse in nature and of very large size . for @xmath41 and @xmath42 the number of eigenvalues are @xmath43 .
i.e. in the order of @xmath44 . the eigenvalues problem being very large to solve for all the eigenvalues .
however for instability analysis , the few eigenvalues with its largest imaginary parts , which makes the flow temporally unstable , are important .
hence , we are interested in the few eigenvalues and corresponding eigenvectors only .
the qz algorithm computes the full spectrum of eigenvalues and hence obviously it is not economical .
the iterative method based on arnoldi s algorithm is a proper choice .
the krylov subspace provides the possibility of extracting major part of the spectrum using shift and invert strategy .
the computations of krylov subspace along with arnoldi s algorithm applied to eigenvalues problem becomes easy .
@xmath45 where , @xmath46 being the shift parameter and @xmath47 is the eigenvalues of the converted problem
. sometimes it is also called spectral transformation , which converts generalized eigenvalues problem to standard eigenvalues problem .
the krylov subspace may be computed by successive resolution of the linear system with matrix @xmath48 , using lu decomposition .
full spectrum method is employed for this small subspace to get a good approximate solution to the original general eigenvalues problem @xcite .
the large size of the krylov subspace extracts major part of the spectrum . generally , the computed spectrum is always nearby the shift parameter .
the good approximation of shift parameter reduces the number of iterations to converge the solution to required accuracy level .
however , larger size of subspace makes the solution almost independent from the shift parameter @xmath49 .
we tested the code for several values of shift parameter @xmath49 .
the convergence of the solution depends on the value of shift parameter .
the good approximation of the shift value needs less number of iterations .
however we have taken the maximum number of iterations equal to 300 ; hence convergence of the solution may not be affected by shift parameter .
given the large subspace size of @xmath50 , the part of the spectrum for our instability analysis could be recovered in the one computation that took about 2.36 hours on intel xenon(r ) cpu e5 26500@@xmath51 .
the base flow velocity profile is obtained by solving the steady navier - stokes equations ( 21)-(23)with the finite - volume code ansys fluent .
the axisymmetric domain is considered with 20 and 200 radius in radial and axial directions as shown in figure [ axi_model ] .
@xmath52 @xmath53 @xmath54 appropriate boundary conditions are applied to close the formulation of above problem .
uniform streamwise velocity @xmath55 was imposed at the inlet ( ab ) and no - slip and no penetration boundary conditions at the surface of the cylinder ( fed ) .
slip boundary condition(@xmath56 ) was imposed at free stream far away from the solid surface ( bc ) and outflow boundary conditions at outlet(cd ) , that consists @xmath57 , @xmath58 and @xmath59 .
line afg is the axis of the cylinder .
the steady navier - stokes equations was solved using simple algorithm with under - relaxation , to get a stable solution .
quick , a weighted average of second order upwind and second order central scheme was used for spatial discretization of momentum equations .
thus obtained base flow velocity profile is interpolated to spectral grids using cubic spline interpolation to perform global stability analysis .
the solution for the base flow has converged with the total 251251 number of grid points with 1001 and 251 grids in streamwise and radial directions respectively .
the distribution of the grids is geometric in both the directions .
we started simulation initially with coarse grid size of 501 and 125 grids in axial and radial direction and refined it with a factor of 1.4142 in each directions .
the discretization error was computed through grid convergence index study on three consecutive refined grids@xcite .
the monotonic convergence is found for all this refined grids . in our problem
we computed error and gci for two field values u(x=75,r=1.2 ) and v(x=75,r=1.2 ) near the surface of the cylinder , where the gradient of the velocity field is maximum .
the gci and error was computed for three different grids as shown table [ mean_gci ] the error between mesh 1 and mesh 2 is too small .
the gci has also reduced with the increased spatial resolution .
it shows that solution is converging monotonically towards the grid - independent one .
further refinement in the grid will hardly improve the accuracy of the solution while increases the time for the computations .
the grid is thoroughly refined near the wall where velocity gradient is very high .
the convergence order n for u and v is 2.00 , which is in agreement with the second order discretization scheme used in the finite volume code ansys fluent .
a mesh 1 was checked with domain height of 25 times cylinder radius to study the effect of far - field condition .
the percentage error with this two domain height is within 0.1 % .
thus , mesh 1 was used in all the results presented here in to compute velocity field for the base flow . .
the reynolds number is calculated based on body radius of the cylinder .
the velocity profile is interpolated to collocation points for instability analysis .
the actual domain height in wall normal direction is 20 radius of the cylinder .
, title="fig:",width=211,height=120 ] .
the reynolds number is calculated based on body radius of the cylinder .
the velocity profile is interpolated to collocation points for instability analysis .
the actual domain height in wall normal direction is 20 radius of the cylinder .
, title="fig:",width=211,height=120 ] /a)(left ) and shape factor(h)(right ) in streamwise direction for different reynolds number .
reynolds number is based on body radius(a ) of the cylinder .
axial and radial length is normalized with cylinder radius ( a ) .
, title="fig:",width=192,height=120 ] /a)(left ) and shape factor(h)(right ) in streamwise direction for different reynolds number .
reynolds number is based on body radius(a ) of the cylinder .
axial and radial length is normalized with cylinder radius ( a )
. , title="fig:",width=192,height=120 ] llllllll mesh & grid size & u & @xmath60 & gci(% ) & v & @xmath61&gci(% ) + # 1 & 1001 @xmath62 251 & 0.034138 & 0.00879 & 0.0262 & 1.655@xmath62 @xmath63 & 0.060&0.1812 + # 2 & 708 @xmath62 177 & 0.034135 & 0.0175 & 0.05274 & 1.656@xmath62 @xmath63 & 0.12&0.3622 + # 3 & 501 @xmath62 125 & 0.034129 & & & 1.658@xmath62 @xmath63 & & +
for the purpose of preliminary comparisons of global stability results with local stability results , we reduced the global nature of the problem to an equivalent local stability problem .
the domain length equal to one wavelength ( lx=2@xmath64/@xmath24 ) is taken in streamwise direction for a least stable axisymmetric mode(n=0 ) .
non - parallel effects from the mean flow velocity profile are removed to apply parallel flow assumption i. e. @xmath65 & @xmath66 . in wall
normal direction , the boundary conditions are same as that of local stability analysis @xcite . to impose wave - like behaviour of disturbances , robin and periodic boundary conditions are applied in the streamwise direction at inlet and outlet @xcite .
the robin boundary conditions with constant streamwise wavenumber @xmath24 is prescribed at inlet and outlet .
the robin boundary conditions are derived from the @xmath67}$ ] .
we applied robin boundary conditions along with periodic boundary conditions at inlet and outlet with constant streamwise wavenumber @xmath24 .
@xmath68 @xmath69 . here
@xmath70 is a critical reynolds number for local stability analysis based on body radius of the cylinder .
, width=336,height=192 ] based on body radius of the cylinder .
the reynolds number based on the displacement thickness ( @xmath5 ) is 1264 .
, title="fig:",width=240,height=240 ] based on body radius of the cylinder .
the reynolds number based on the displacement thickness ( @xmath5 ) is 1264 .
, title="fig:",width=240,height=240 ] above boundary conditions are applied to disturbance velocity components in a streamwise direction at inlet and outlet .
the second derivatives are considered to avoid complex quantities in the boundary conditions and hence fast computations can take place .
here we considered axisymmetric mode(@xmath71 ) for the validation of global stability results .
the critical reynolds number , streamwise wavenumber and streamwise location are 12439 , 2.73 and 47 respectively @xcite .
the reynold number is based on the body radius of the cylinder .
streamwise domain length is taken equal to one wavelength(@xmath72 ) .
the domain in streamwise direction is very limited and hence the solution is converged with the number of collocation points n=41 in streamwise direction and m=101 in the radial direction .
the eigenspectrum and eigenfunctions are in good agreement with the results of local stability analysis as shown in figure [ spectrum ] and figure [ eig_fun ] .
the streamwise wavenumber @xmath73 is computed from the eigenfunctions obtained through global stability analysis and it is equal to 2.73 .
hence , the applied value of @xmath24 is recovered in the global stability analysis .
the perturbation for the disturbances u and v can be written as @xmath74}$ ] and , @xmath75}$ ] , where @xmath76 and @xmath77 are phase angles of wave with respect to x and r. the phase angles of waves can be defined by the following relationship @xcite .
@xmath78 where r and i denotes real and imaginary parts of the eigenfunctions u and v. the phase is dependent on the normal coordinate only .
thus the position of maximum absolute magnitude of each disturbances remains same at all the streamwise locations .
the streamwise wave number ( @xmath24 ) is calculated by @xcite . @xmath79
in the present analysis reynolds number is varying from 261 to 693 with azimuthal wave numbers @xmath80 , @xmath2 , @xmath81 , @xmath82 , @xmath83 and @xmath84 .
reynolds number based on the displacement thickness ( @xmath85 ) at the inlet of the domain is considered .
streamwise and wall normal dimensions are normalized with the displacement thickness at the inlet of the domain .
the streamwise domain length @xmath86 and wall normal height @xmath87 is selected for the instability analysis .
the number of collocation points considered are @xmath88 in streamwise direction and m=121 in wall normal direction .
the general eigenvalues problem is solved using arpack , which uses arnoldi s algorithm .
the computed eigenvalues are found to be accurate up to three decimal points .
the additional lower resolution cases were also run to confirm the proper convergence of the eigenspectrum .
heavy sponging is applied to avoid the spurious reflections at the outlet .
the eigenvalues with the largest imaginary part that is least stable ones is selected .
the two - dimensional mode structure of the selected modes is also checked for the spurious mode .
+ a grid convergence study was performed to check the accuracy of the solution and appropriate grid size .
table 2 shows the values of two leading eigenvalues computed for @xmath89 for axisymmetric mode using different three grid size .
the grid resolution was successively improved by a factor of 1.14 in axial and radial direction respectively .
the real and imaginary parts of the eigenvalues shows monotonic convergence of the solution with the increase in resolution . in this table
n and m indicates the number of grids in axial and radial directions respectively .
the relative errors are calculated between consecutive grids for real and imaginary parts of the eigenvalues .
the largest associated error among both the eigenmode is considered .
the relative error for mesh 1 is well within the limit and it is used for all the results reported here .
lllllllll mesh & lx & lr & n@xmath62 m & n & m & @xmath90 & @xmath91 & error(% ) + # 1 & 605&20 & 14641 & 121 & 121 & 0.03654 - 0.01763i & 0.02965 - 0.01770i & 3.1598 + # 2 & 605&20 & 11449 & 107 & 107 & 0.03644 - 0.01709i & 0.02951 - 0.01751i & 6.3791 + # 3 & 605&20 & 8649 & 93 & 93 & 0.03636 - 0.01642i & 0.02938 - 0.01646i & + to check the influence of domain height(lr ) , computations were performed with three different domain height of lr=15,20 and 25 .
the difference of results obtained with lr=20 and lr=25 is too small as shown in figure [ sp_ly ] .
thus lr=20 is selected for all global stability computations .
+ , as marked by circle in figure [ sp_0_383 ] .,width=432,height=96 ] , as marked by circle in figure [ sp_0_383 ] .,width=432,height=96 ] , as marked by circle in figure [ sp_0_383 ] .,title="fig:",width=192,height=115 ] , as marked by circle in figure [ sp_0_383 ] .,title="fig:",width=192,height=120 ] figure [ sp_0_383 ] shows the eigenspectrum for the axisymmetric mode ( n=0 ) with reynolds number , @xmath89 . in figure [ sp_0_383 ] ,
the eigenmodes marked by circle and square are stationary ( @xmath92 ) and oscillatory(@xmath93 ) modes respectively .
the stationary mode has a complex frequency @xmath94 .
this global mode is temporally stable because @xmath95 and hence disturbances decay in time .
figure [ cont_0_383_s ] shows the two dimensional spatial structure of the streamwise(u ) and wall normal ( v ) disturbance amplitudes for stationary mode @xmath94 .
the magnitudes of the velocity disturbances is zero at the inlet , as it is the inlet boundary condition .
as the fluid particles move towards the downstream , the disturbances evolve monotonically in time within the domain and progresses towards the downstream .
the magnitude of the u disturbance amplitudes is one order higher than that of v disturbance amplitudes .
the amplitude structure has opposite sign for u and v. the spatial structure of the amplitudes grow in size and magnitude towards the downstream .
figure [ plot_0_383_s ] shows the variation of u and v disturbance amplitudes in streamwise direction at different radial locations .
the magnitude of the disturbance amplitudes are very small near the cylinder surface due to viscous effect , increases in radial direction and vanishes at far - field . + , as marked by square in figure [ sp_0_383 ] .,width=432,height=96 ] , as marked by square in figure [ sp_0_383 ] .,width=432,height=96 ] , as marked by square in figure [ sp_0_383 ] .,title="fig:",width=201,height=120 ] , as marked by square in figure [ sp_0_383 ] .,title="fig:",width=201,height=120 ] the most unstable oscillatory mode has an eigenvalue @xmath96 .
the flow is stable for this global mode because @xmath95 .
the spatial structure of the above global eigenmode for u and v disturbance amplitudes are not monotonic in nature.the distribution in spatial directions is show in figure [ cont_0_383_o ] .
the amplitudes of the velocity disturbances is zero at the inlet as it is the imposed boundary condition at the inlet .
however , in the streamwise direction the disturbances grow in magnitudes and it contaminates the flow field towards the downstream .
the magnitudes of the disturbance amplitudes grows exponentially as they move further towards the downstream and hence flow is convectively unstable @xcite .
however , its variation in the normal direction is different from that of a stationary mode , as seen in the respective figures .
the magnitude of the normal disturbance amplitudes is one order less than that of the streamwise component .
+ figure [ plot_0_383_o ] presents variation of disturbance amplitudes in streamwise direction at various radial locations .
it shows that the most of the disturbances grow in magnitude as they move towards downstream .
the magnitude of the disturbance amplitudes are very small near the wall due to viscous effect , gradually increases in radial direction and finally vanishes at the far - field .
it shows that disturbances evolve within the flow - field in time and grow in magnitude and size while moving towards the downstream .
+ figure [ cont_0_383_omegar ] shows contour plot of the real part of streamwise disturbance velocity of two different eigen modes with frequency @xmath97=0.003689 & @xmath97=0.1234 for n=0 and re=383 .
the streamwise domain length is 345 .
the disturbances is seen to evolve in the vicinity of the wall and increasing the amplitudes when moving towards downstream .
the typical length scale of the the wavelet structure decreases with the increases in frequency(@xmath97 ) .
+ = 0.003689 ( b)@xmath97=0.1234 .
the corresponding reynolds number and azimuthal wave - number are re=383 and n=0 respectively.,width=432,height=96 ] = 0.003689 ( b)@xmath97=0.1234 .
the corresponding reynolds number and azimuthal wave - number are re=383 and n=0 respectively.,width=432,height=96 ] the body radius of the cylinder is another important length scale in case of the axisymmetric boundary layer in addition to the boundary layer thickness ( @xmath85 ) .
the inverse of the body radius is called the transverse surface curvature(s ) .
the flat plate boundary layer is a special case of a zero transverse curvature .
it is normalised with the displacement thickness ( @xmath85 ) at the inlet . )
boundary layer for @xmath98 for three different reynolds number ( a ) re=383 ( b ) re=557 and ( c ) re=909.,title="fig:",width=216,height=144 ] ) boundary layer for @xmath98 for three different reynolds number ( a ) re=383 ( b ) re=557 and ( c ) re=909.,title="fig:",width=216,height=144 ] ) boundary layer for @xmath98 for three different reynolds number ( a ) re=383 ( b ) re=557 and ( c ) re=909.,width=216,height=144 ] in case of the axisymmetric boundary layer it has significant effect on the characteristic of base flow as well as disturbance flow quantities . in order to understand the transverse curvature effect on the stability characteristic of the axisymmetric boundary layer ,
the stability characteristics are compared with the flat plate boundary layer(zero transverse curvature ) at the same reynolds number and domain size .
the reynolds number is computed based on the displacement thickness and free - stream velocity at the inlet .
the eigen spectrum , spatial amplification rate , effect of the domain length on the distribution of the spectrum and eigen functions are compared at the same reynolds number . )
boundary layer for three different reynolds number ( a ) re=383 ( b ) re=557 and ( c ) re=909.,title="fig:",width=168,height=120 ] ) boundary layer for three different reynolds number ( a ) re=383 ( b ) re=557 and ( c ) re=909.,title="fig:",width=168,height=120 ] ) boundary layer for three different reynolds number ( a ) re=383 ( b ) re=557 and ( c ) re=909.,width=168,height=120 ] figure [ sp_plate_cyl ] shows the comparison of eigen spectrum for three different reynolds numbers 383 , 557 and 909 ( based on the body radius of the cylinder these are @xmath99 , @xmath100 and @xmath101 ) with @xmath102=345 for axisymmetric and flat plate boundary layer .
the comparison is limited for the discrete part of the spectrum only .
it has been observed that at small reynolds number the damping rates of eigen modes are higher ( @xmath103 is smaller ) for axisymmetric boundary layer than that of flat plate boundary layer .
the difference of damping rate in eigen modes for axisymmetric and flat plate boundary layer reduces with the increase in reynolds number .
the eigen modes of axisymmetric boundary layer approaches to eigen modes of flat plate boundary layer at higher reynolds number .
this is primarily due to the effect of transverse curvature only .
the effect of transverse curvature is significant at low reynolds number .
the effect of transverse curvature reduces with the increase in reynolds number .
thus , it proves that the global modes are more stable at low reynolds number for axisymmetric boundary layer .
+ .,title="fig:",width=168,height=120 ] .,title="fig:",width=168,height=120 ] .,width=168,height=120 ] figure [ ax_plate_cyl ] shows the comparison of spatial amplification rates ( @xmath104)at three different reynolds number for both the boundary layers .
the least stable mode is selected to compute the spatial amplification rate .
the comparison shows that at low reynolds number , ( @xmath104)is higher for flat plate boundary layer .
the ( @xmath104 ) for axisymmetric boundary layer also increases with the increase in reynolds number and approaches to flat plate boundary layer at higher reynolds number .
+ .,title="fig:",width=144,height=144 ] .,title="fig:",width=144,height=144 ] three different families of eigenspectrum are shown in figure [ sp_lx_383 ] for axisymmetric and flat plate boundary layer .
the discrete part of the spectrum is only shown in the figures for the comparison .
the difference between the spectrum depends on the domain length .
the distance between two consecutive frequency reduces with the increase in domain length . to quantify this discretization of the eigen modes for re=323 ,
different domain lengths were considered with @xmath105,@xmath106 and @xmath107 .
figure [ mod_u ] shows the comparison of modulus of streamwise and normal disturbance velocity u and v respectively for re=383 at x=445 .
the magnitude is modulus of disturbances is higher for flat plate plat boundary layer .
the eigenspectrum for azimuthal wave number n=1 and @xmath108 is shown in figure [ sp_1_383 ] .
stationary eigen mode is also found corresponding to @xmath92 in the non - axisymmetric case too .
the corresponding growth rate @xmath109 is -0.02661 .
figure [ cont_1_383_o ] presents the global structure of disturbance amplitude of the oscillatory mode . the oscillatory mode with largest imaginary part is @xmath110 .
this global mode is also temporally stable because @xmath95 .
the eigenmode corresponding to n=1 is non - axisymmetric and therefore , the disturbance velocity have azimuthal component ( @xmath111 ) also .
the spatial structure for streamwise(u ) , radial ( v ) and azimuthal(w ) disturbance velocity amplitudes is shown in figure [ cont_1_383_o ] .
the disturbance amplitudes grows in streamwise direction when moving towards the downstream .
the magnitudes of the amplitudes also increases towards the downstream .
the wave - like nature of the disturbance amplitudes is also found in figure [ cont_1_383_o ] .
, as marked by square in figure [ sp_1_383].,title="fig:",width=432,height=96 ] , as marked by square in figure [ sp_1_383].,title="fig:",width=432,height=96 ] , as marked by square in figure [ sp_1_383].,title="fig:",width=432,height=96 ] , as marked by square in figure [ sp_1_383].,title="fig:",width=192,height=120 ] , as marked by square in figure [ sp_1_383].,title="fig:",width=192,height=120 ] , as marked by square in figure [ sp_1_383].,width=192,height=120 ] [ plot_1_383_o ] ) with reynolds number for different azimuthal wave numbers .,width=211,height=120 ] figure [ temp_growth ] shows the temporal growth rate of the eigenmodes for different reynolds number .
the growth rate increases with the increase in reynolds number for all the azimuthal wave numbers . as the largest imaginary part is negative for different reynolds
number the flow is temporally stable .
the transverse curvature varies inversely with the reynolds number . at low reynolds number growth rate
is small and with the increase in reynolds number growth rate increases .
which proves that transverse curvature has significant damping effect on the global temporal mode also . a global modes with higher wave numbers , n=3,4 and 5 are more stable than that of axisymmetric and helical mode n=1 and 2 .
at low reynolds number axisymmetric mode is having higher growth rate then that of n=1 and 2 .
as reynolds number increases , the growth rate of helical mode n=1 and 2 increases then that of axisymmetric mode .
the global temporal modes exhibit spatial growth / decay in streamwise direction when moving towards the downstream .
the disturbances in the streamwise direction at a particular radial location may decay or amplify .
however the overall effect of all disturbances together at each streamwise location can be quantified by computing the spatial amplitude growth , a(x ) @xcite .
@xmath112 where @xmath113 denotes the complex conjugate .
figure [ sp_amp ] shows the spatial amplitude growth of the disturbance waves for different azimuthal wave numbers at various reynolds number .
most unstable oscillatory ( @xmath114 ) temporal modes are considered to compute the spatial growth of the amplitudes .
the amplitude growth increases in the streamwise direction as the flow is convectively unstable . at low reynolds number
the the growth of disturbances start at early stage .
in streamwise direction x for ( a ) re=261 ( b ) re=477 and ( c ) re=628 for different azimuthal wave numbers . the most unstable temporal mode is considered to compute the spatial amplitude growth @xmath115.,title="fig:",width=201,height=120 ] in streamwise direction x for ( a ) re=261 ( b ) re=477 and ( c ) re=628 for different azimuthal wave numbers .
the most unstable temporal mode is considered to compute the spatial amplitude growth @xmath115.,title="fig:",width=201,height=120 ] in streamwise direction x for ( a ) re=261 ( b ) re=477 and ( c ) re=628 for different azimuthal wave numbers . the most unstable temporal mode is considered to compute the spatial amplitude growth @xmath115.,width=201,height=120 ] [ sp_amp ]
the global temporal modes are computed using linear stability theory for axisymmetric boundary layer .
the numerical eigenvalue problem is solved using arpack , which employs arnoldi s iterative algorithm .
the stability analysis conducted for various reynolds number and azimuthal wave numbers .
the largest imaginary part of the computed eigenvalues are negative for the range of reynolds number and azimuthal wave numbers considered here .
thus the flow is temporally stable .
the global modes in axisymmetric boundary layer , showed wavelike behaviour for the range of reynolds numbers considered .
the spatial properties of the global mode shows that the disturbances grow in size and magnitude within the flow domain in the streamwise direction when move towards the downstream , which hints that the flow is convectively unstable . at low reynolds number axisymmetric mode ( n=0 )
is least stable one , while at higher reynolds number(@xmath116 ) helical mode n=2 is observed least stable one .
the global modes with higher wave numbers n=3 , 4 and 5 are more stable .
the comparison of global modes of axisymmetric and flat plate boundary layer shows that at low reynolds number the effect of transverse curvature is significant , however with the increase in reynolds number this effect reduces and global modes of the axisymmetric boundary layer become less stable .
thus , the transverse curvature shows overall damping effect on the global modes .
the amplitude functions corresponding to the stationary mode is found to be monotonically growing in the streamwise direction .
rodriguez , d. , theofilis , v. : on instability and structural sensitivity of incompressible laminar separation bubbles in a flat - plate boundary layer . aiaa 38th fluid dynamics conference and exhibit , seattle , * 4148 * ( 2008 ) | this paper presents the linear global stability analysis of the incompressible axisymmetric boundary layer on a circular cylinder .
the base flow is parallel to the axis of the cylinder at inlet .
the pressure gradient is zero in the streamwise direction .
the base flow velocity profile is fully non - parallel and non - similar in nature .
the boundary layer grows continuously in the spatial directions .
linearized navier - stokes(lns ) equations are derived for the disturbance flow quantities in the cylindrical polar coordinates .
the lns equations along with homogeneous boundary conditions forms a generalized eigenvalues problem .
since the base flow is axisymmetric , the disturbances are periodic in azimuthal direction .
chebyshev spectral collocation method and arnoldi s iterative algorithm is used for the solution of the general eigenvalues problem .
the global temporal modes are computed for the range of reynolds numbers and different azimuthal wave numbers .
the largest imaginary part of the computed eigenmodes are negative and hence the flow is temporally stable .
the spatial structure of the eigenmodes shows that the flow is convectively unstable because the disturbance amplitudes grow in size and magnitude when moving towards downstream .
the global modes of axisymmetric boundary layer are more stable than that of 2d flat plate boundary layer at low reynolds number . however , at high reynolds number they approaches to 2d flat plate boundary layer .
thus , the damping effect of transverse curvature is significant at low reynolds number .
the wave - like nature of the disturbances is found for most unstable eigenmodes . |
the spectroscopic binary system hr 6902 ( g9 iib + b8 - 9 v ) belongs to the category of @xmath8 aur stars , where a hot companion ( generally a main sequence b star ) moves through the extended envelope of an evolved late - type star .
the specific binary geometry provides an exceptional opportunity to probe the spatial dependence of the physical properties of the outer atmosphere and wind of the primary .
observations of @xmath8 aur systems and related objects with the _ international ultraviolet explorer _ ( _ iue _ ) and more recently with the _ hubble space telescope _
( _ hst _ ) take advantage of the ultraviolet spectral region . at wavelengths below @xmath9 the hot
secondary dominates the observed flux allowing a separation of the spectral information . in the uv
the observed spectrum is purely that of the companion with superimposed circumstellar and interstellar lines .
further information on the binary technique can be found in several reviews ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . ground
based spectroscopic as well as uv and x - ray observations from space of stars in the cool half of the hr diagram have shown that the extended atmospheres of giants and supergiants undergo a dramatic physical change across a region in the hr diagram . to the right of this transition we observe cool and massive winds plus extended cool chromospheres , while to its left there is much hotter circumstellar matter .
this transition is accompanied by different spectral indicators defining distinct dividing lines .
the leftmost is the `` x - ray dividing line '' . all stars to its left show x - ray emission , probably arising from extended corona - like plasma ( e.g. , haisch , schmitt , & rosso 1991 , 1992 ) . in the ultraviolet
the same stars also show and and often emission ( @xcite ) .
additional dividing lines are defined by asymmetries in the and emission lines ( e.g. , @xcite , 1980b ) .
the strongest and most reliable evidence for the existence of a cool , massive wind is the appearance of circumstellar absorption features in the line ( @xcite ) .
a theoretical discussion of the dividing line concept based on hydromagnetic waves can be found in rosner ( 1993 ) and rosner et al .
( 1995 ) .
however , this simple picture is complicated by the existence of the so - called hybrid stars . hartmann , dupree & raymond ( 1980 ) and reimers ( 1982 ) found that some bright k giants and g supergiants show blue - shifted ( and partly ) absorption and emission lines like and signifying temperatures of at least @xmath10 .
hybrid stars are also x - ray emitters ( @xcite )
. the outflow properties of these stars are still very uncertain and it is controversial whether hybrids show cold ( @xmath11 ) or possibly warm ( @xmath12 ) winds .
previous _ iue _ observations of hr 6902 suggest the existence of warm circumstellar material that may represent the transitional physics of hybrid star atmospheres .
reimers , baade , & schrder ( 1990a ) and reimers et al.(1990b ) have presented a series of _ iue _ high resolution spectra ( @xmath13 ) taken a few days before and after eclipse . and column densities indicate a height - independent electron temperature of @xmath14 . as discussed by reimers et al .
( 1990b ) the density distribution is compatible with a low mass - loss wind . in the present paper
we report the results of a reanalysis of _ iue _ spectra of hr 6902 . compared to the first study ( @xcite , 1990b ) we utilize a substantially expanded observations database .
the paper is organized as follows : in sect .
[ sec2 ] we present the observations and basic stellar parameters used in this paper . sect .
[ sec3 ] summarizes the methodical principles of our absorption analysis technique and the application to the hr 6902 spectra . in sect .
[ sec4 ] we outline the multi - dimensional radiative transfer procedure used to infer global properties of the circumstellar shell .
while nearly all observations have been conducted by ourselves in several _ iue _ programs , the data used here are taken from the _ iue _ final archive .
we have compiled a sequence of large aperture spectra obtained between 1987 and 1990 .
full details of the _ newsips _ data reduction have been given by garhart et al .
( 1997 ) , and will not be repeated here .
it is , however , worth reiterating that the new processing techniques generally greatly enhance the quality of _ iue _ spectra .
the increase in signal - to - noise improves the detection limit of faint absorption features significantly .
furthermore , the new noise handling allows a reliable error estimate for each point in the flux spectrum .
the archival _ iue _ spectra exploited for this paper are listed in table [ tbl-1 ] . the analysis presented by reimers et al .
( 1990b ) is based on _ iue _ observations of the 1988 and 1989 eclipses , covering a range of impact parameters @xmath15 ( projected distance between the stellar components ) from @xmath16 giant radii .
our new study also includes the high resolution spectra of the 1990 eclipse showing and absorption up to @xmath17 .
these additional impact parameters allow us to directly examine the wind acceleration region .
one additional observation obtained 1987 yields an upper limit to the column densities at @xmath18 .
lcccc date & swp & exp .
time & @xmath15 & phase + & no . &
( min ) & @xmath19 & + 1987 & & & & + jul .
11 & 31325 & 140 & 5.26 & 0.9227 + 1988 & & & & + aug .
25 & 34131 & 140 & 1.15 & 0.9887 + aug .
25 & 34135 & 110 & 1.10 & 0.9898 + sep .
03 & 34177 & 130 & 1.19 & 0.0122 + sep .
04 & 34182 & 125 & 1.33 & 0.0147 + sep .
05 & 34187 & 100 & 1.48 & 0.0175 + 1989 & & & & + jun .
30 & 36589 & 140 & 9.12 & 0.7926 + sep .
26 & 37191 & 180 & 1.68 & 0.0209 + sep .
27 & 37195 & 190 & 1.84 & 0.0235 + sep .
28 & 37200 & 194 & 2.00 & 0.0261 + 1990 & & & & + sep .
20 & 39668 & 195 & 3.32 & 0.9534 + sep .
20 & 39669 & 195 & 3.29 & 0.9538 + sep .
23 & 39699 & 195 & 2.81 & 0.9612 + sep .
23 & 39700 & 180 & 2.79 & 0.9616 + sep .
26 & 39713 & 177 & 2.28 & 0.9694 + sep .
28 & 39721 & 170 & 1.98 & 0.9742 + oct .
08 & 39795@xmath20 & 410 & 0.84 & 0.9997 + @xmath21 low - dispersion spectrum [ tbl-1 ] a direct analysis of the circumstellar and lines is not feasible owing to the considerable photospheric background absorption . we extract the circumstellar contribution by dividing all high resolution spectra by the out - of - eclipse exposure swp 36589 .
this observation serves as suitable reference spectrum since the large impact parameter of @xmath22 implies a pure b star spectrum .
indeed , the final wind model presented in sect . [ sec3 ] confirms that there should be no detectable circumstellar absorption . as an example we present in fig .
[ fig1 ] the phase variation of the absorption line .
the observed line strength depends strongly on the relative position of the b star and decreases with increasing impact parameter .
though circumstellar absorption at 1335 is clearly detectable , the poor s / n ratio of the _ iue _ spectra does not allow an adequate absorption line reconstruction . in the case of
nearly saturated photospheric absorption lines the dividing procedure leads to spurious line profiles with artificial emission features .
unfortunately , the and doublets are the only circumstellar absorption lines that can be used for a quantitative analysis .
lc parameter & value + period ( days ) & 385.0 + eccentricity&0.311 + inclination ( deg)&87 + longitude of periastron ( deg)&146 + semimajor axis ( cm ) & @xmath23 + periastron passage ( jd)&2447306.33 + conjunction ( jd)&2447788.18 + spectral type ( primary)&g9 iib + spectral type ( secondary)&b8 - 9 v + @xmath24&33.0 + @xmath25&3.0 + @xmath26&3.86 + @xmath27&2.95 + @xmath28 ( k)&4900 + @xmath29 ( k)&11600 + @xmath30 ( mag ) & @xmath31 + @xmath32(mag)&@xmath33 + @xmath34(mag)&@xmath35 + distance ( pc)&@xmath36 + [ tbl-2 ] great significance must be attached to the emission line spectrum visible during total eclipse of the b star . the 1988 total eclipse observation ( swp 34143 , @xmath37 ) was part of the first hr 6902 program and has been reported by reimers et al . ( 1990a ) .
the deep @xmath38 exposure swp 39795 ( fig .
[ fig2 ] ) obtained on 1990 october 8 provides a significantly improved signal - to - noise ratio and will be subject of this paper .
we note that the large difference in quality does not allow a comparative analysis in order to search for long - term variations .
while in `` normal '' @xmath8 aur systems hundreds of scattering lines ( e.g. , and ) are seen , only a few are observed in the spectrum of hr 6902 .
a comparison with stars of similar spectral type suggests that the emission line spectrum can not be due to the g star alone ( @xcite ) . instead , the observed and doublets are formed primarily by resonance scattering of b star photons in the circumstellar envelope .
griffin et al . (
1995 ) have redetermined the orbital and stellar parameters of hr 6902 by means of photometric observations during the eclipses of 1987 and 1989 ( table [ tbl-2 ] ) .
the relative orbit and the eclipse geometry are shown in fig .
it should be noted that the stellar parameters of giants belonging to @xmath8 aur systems can be much better determined than for single stars .
a recent evolutionary analysis suggests that the bright giant primary of hr 6902 is already evolved into its blue loop phase ( @xcite ) .
a comparison of the total eclipse emission line spectrum with the out - of - eclipse observations shows immediately that the continuum flux is orders of magnitude larger than the scattered line flux .
hence , the circumstellar absorption profiles can be interpreted as pure absorption lines formed in the extended envelope of hr 6902 . in the subsequent sections we present details of the analysis and interpretation of the spectra of hr 6902 . the line formation in hr 6902
is dominated by velocity fields on different scales . to simplify the concept
we introduce a line broadening velocity @xmath39 containing both the thermal velocity @xmath40 and a stochastic component @xmath41 associated with turbulence or other short - scale nonthermal motions . assuming a gaussian distribution of the stochastic velocity the spectroscopic line broadening parameter becomes @xmath42 .
the macroscopic velocity field is characterized by a spherically symmetric expansion described by an a priori unknown velocity law @xmath43 . throughout this paper
we assume the line broadening velocity to be constant in the relevant parts of the envelope .
it is convenient to define a dimensionless frequency variable @xmath44 , where @xmath45 is the line - center frequency and @xmath46 the corresponding doppler width .
macroscopic velocities are measured in the same units , i.e. @xmath47 . regarding the b star as a point light - source of intensity @xmath48 the emergent flux for a given impact parameter @xmath15
can be written @xmath49.\ ] ] the optical depth along a ray from the observer to the b star is given by @xmath50\ , \rm { d}z,\ ] ] where @xmath51 denotes the dimensionless profile function .
the geometry is defined by the distance from the primary @xmath52 and the integration path @xmath53 .
the boundaries @xmath54 and @xmath55 depend on the position of the b star and the adopted outer radius of the envelope ( chosen as large that the results are unaffected by its actual value ) , respectively .
the line opacity per absorber reads @xmath56 where @xmath57 is the oscillator strength and @xmath58 the rest wavelength of the transition under consideration . in the static case ( i.e. without considering the expansion explicitly )
[ eq-2 ] degenerates to a simple relation between the optical depth and the column density @xmath59 , viz @xmath60 our analysis is carried out using a non - linear least square fit technique based on the levenberg - marquard method ( e.g. , @xcite ) to derive the column density @xmath59 and the line broadening parameter @xmath41 .
the _ newsips _ error estimate provides statistical errors of the deduced parameters and allows to assess the goodness - of - fit . as an additional constraint
the best - fit parameters are forced to be equal for all lines of the particular multiplet under consideration , i.e. , we favor the strategy of resonance doublet modeling instead of analyzing individual absorption lines .
our fitting procedure allows to vary the continuum flux level in order to avoid the difficulties associated with a continuum definition by eye .
finally , the theoretical profile has to be convoluted with the _ iue _ line - spread function .
the instrumental profile is assumed to be well represented by a gaussian with fwhm in accordance with the actual determination of the _ newsips _ resolution ( @xcite ) .
as a starting point of our analysis we use a quasi - static approximation neglecting macroscopic velocity fields .
the objective of this approach is to provide a provisional set of column densities and stochastic velocities .
however , line broadening effects due to the wind outflow calls the static model into question . in sect .
[ sec33 ] we will examine possible consequences for the derived parameters . in the spectrum swp 31325 ( impact parameter @xmath61 ) we can not identify any signature of or absorption lines .
instead , we use the error information of the _ iue _ flux to determine upper limits of the column densities .
the results of the static analysis are summarized in table [ tbl-3 ] .
[ fig4 ] shows the run of column densities as a function of the impact parameter . as already noticed by reimers et al .
( 1990b ) , the egress observations indicate significant differences from epoch to epoch .
the obtained stochastic velocities show a considerable scatter with a weighted mean of @xmath62 .
cccccc phase & @xmath63 $ ] & @xmath64 $ ] & @xmath65 & @xmath66 & @xmath67 $ ] + @xmath68 & @xmath69 & @xmath70 & & &
+ @xmath71 & @xmath72 ( @xmath73 ) & @xmath74 ( @xmath75 ) & @xmath76 ( @xmath77 23.1 ) & @xmath78 ( @xmath775.0 ) & @xmath79 ( @xmath80 ) + & @xmath81 ( @xmath82 ) & @xmath83 ( @xmath82 ) & @xmath84 ( @xmath77 29.4 ) & @xmath78 ( @xmath775.3 ) & @xmath85 ( @xmath86 ) + @xmath87 & @xmath88 ( @xmath75 ) & @xmath89 ( @xmath82 ) & @xmath90 ( @xmath77 23.0 ) & @xmath78 ( @xmath774.5 ) & @xmath91 ( @xmath92 ) + & @xmath93 ( @xmath94 ) & @xmath83 ( @xmath82 ) & @xmath95 ( @xmath77 31.7 ) & @xmath96 ( @xmath775.6 ) & @xmath91 ( @xmath97 ) + @xmath98 & @xmath99 ( @xmath94 ) & @xmath100 ( @xmath101 ) & @xmath102 ( @xmath77 16.2 ) & @xmath103 ( @xmath773.2 ) & @xmath104 ( @xmath105 ) + & @xmath106 ( @xmath82 ) & @xmath107 ( @xmath108 ) & @xmath109 ( @xmath77 20.3 ) & @xmath103 ( @xmath773.7 ) & @xmath110 ( @xmath111 ) + @xmath112 & @xmath113 ( @xmath82 ) & @xmath114 ( @xmath115 ) & @xmath116 ( @xmath77 12.5 ) & @xmath117 ( @xmath773.2 ) & @xmath118 ( @xmath119 ) + & @xmath120 ( @xmath121 ) & @xmath122 ( @xmath123 ) & @xmath124 ( @xmath77 22.3 ) & @xmath125 ( @xmath774.1 ) & @xmath126 ( @xmath111 ) + @xmath127 & @xmath128 ( @xmath129 ) & @xmath130 ( @xmath129 ) & @xmath131 ( @xmath77 5.9 ) & @xmath125 ( @xmath771.9 ) & @xmath132 ( @xmath133 ) + & @xmath83 ( @xmath129 ) & @xmath134 ( @xmath129 ) & @xmath135 ( @xmath77 11.8)&@xmath117 ( @xmath772.4 ) & @xmath136 ( @xmath137 ) + @xmath138 & @xmath139 ( @xmath140 ) & @xmath141 ( @xmath129 ) & @xmath142 ( @xmath77 6.9 ) & @xmath143 ( @xmath771.8 ) & @xmath144 ( @xmath145 ) + & @xmath146 ( @xmath140 ) & @xmath147 ( @xmath148 ) & @xmath149 ( @xmath77 10.3)&@xmath150 ( @xmath772.3 ) & @xmath151 ( @xmath145 ) + @xmath152 & @xmath153 ( @xmath154 ) & @xmath155 ( @xmath101 ) & @xmath156 ( @xmath77 4.9 ) & @xmath157 ( @xmath772.4 ) & @xmath158 ( @xmath159 ) + & @xmath160 ( @xmath161 ) & @xmath162 ( @xmath161 ) & @xmath163 ( @xmath77 9.4 ) & @xmath117 ( @xmath773.0)&@xmath164 ( @xmath165 ) + @xmath166 & @xmath167 ( @xmath168 ) & @xmath169 ( @xmath115 ) & @xmath170 ( @xmath77 4.9 ) & @xmath171 ( @xmath773.2 ) & @xmath172 ( @xmath173 ) + & @xmath174 ( @xmath115 ) & @xmath155 ( @xmath168 ) & @xmath175 ( @xmath77 11.4 ) & @xmath96 ( @xmath774.3)&@xmath176 ( @xmath173 ) + @xmath177 & @xmath178 ( @xmath123 ) & @xmath179 ( @xmath115 ) & @xmath180 ( @xmath77 5.0 ) & @xmath181 ( @xmath773.0)&@xmath182 ( @xmath183 ) + & @xmath184 ( @xmath161 ) & @xmath185 ( @xmath161 ) & @xmath186 ( @xmath77 12.9)&@xmath171 ( @xmath773.6 ) & @xmath187 ( @xmath188 ) + @xmath189 & @xmath184 ( @xmath108 ) & @xmath190 ( @xmath123 ) & @xmath191 ( @xmath77 5.3 ) & @xmath192 ( @xmath772.8)&@xmath193 ( @xmath194 ) + & @xmath195 ( @xmath108 ) & @xmath130 ( @xmath161 ) & @xmath196 ( @xmath77 13.6 ) & @xmath197 ( @xmath773.4)&@xmath198 ( @xmath199 ) + @xmath200 & @xmath114 ( @xmath201 ) & @xmath202 ( @xmath75 ) & @xmath203 ( @xmath77 12.4)&@xmath204 ( @xmath774.8 ) & @xmath205 ( @xmath206 ) + & @xmath207 ( @xmath168 ) & @xmath208 ( @xmath209 ) & @xmath210 ( @xmath77 24.1 ) & @xmath211 ( @xmath776.2 ) & @xmath212 ( @xmath213 ) + @xmath214 & @xmath215 ( @xmath216 ) & @xmath217 ( @xmath148 ) & @xmath218 ( @xmath77 6.7 ) & @xmath219 ( @xmath771.5 ) & @xmath220 ( @xmath148 ) + & @xmath215 ( @xmath216 ) & @xmath179 ( @xmath140 ) & @xmath221 ( @xmath77 9.5 ) & @xmath222 ( @xmath771.8 ) & @xmath220 ( @xmath145 ) + @xmath223 & @xmath114 ( @xmath129 ) & @xmath174 ( @xmath129 ) & @xmath218 ( @xmath77 15.0 ) & @xmath125 ( @xmath771.7)&@xmath205 ( @xmath224 ) + & @xmath225 ( @xmath140 ) & @xmath226 ( @xmath148 ) & @xmath227 ( @xmath77 17.5 ) & @xmath228 ( @xmath772.0 ) & @xmath229 ( @xmath224 ) + @xmath230 & @xmath231 ( @xmath129 ) & @xmath147 ( @xmath108 ) & @xmath232 ( @xmath77 9.0 ) & @xmath233 ( @xmath772.0 ) & @xmath234 ( @xmath235 ) + & @xmath236 ( @xmath148 ) & @xmath147 ( @xmath237 ) & @xmath238 ( @xmath77 15.4 ) & @xmath222 ( @xmath772.6 ) & @xmath239 ( @xmath224 ) + [ tbl-3 ] assuming the simplifying conditions of a low density plasma the and density ratio can be used to derive an ionization temperature @xmath240 .
we solve the ionization balance using the radiative - collisional equilibrium program cloudy ( version 9003 ; @xcite ) .
the metal abundances are supposed to be solar ( @xmath241 [ @xcite ] ) . as an approximation
we identify the ionization temperature with the kinetic temperature @xmath242 knowing that the ionization state may freeze in , so that we have @xmath243 in the outer wind .
it turns out that the electron temperature appears to be constant within the observed part of the envelope ( @xmath244 ) with a weighted mean of @xmath245 .
however , the ion ratio may be affected by different cno abundances due to the first dredge up which can deplete the carbon abundance significantly ( e.g. , @xcite ) .
for example , a carbon abundance reduced by a factor of 4 would lead to a temperature of @xmath246 . with known electron temperature
we are able to transform the ion density into total hydrogen densities . in the next step
we try to match the derived hydrogen column densities employing different types of density laws .
each density law implies a physical model defined by a specific set of parameters .
again we implement a least square fit procedure using the levenberg - marquard method . as a first attempt we use the hydrostatic equation for an isothermal plasma @xmath247 where @xmath248 defines the bottom of the envelope with the corresponding hydrogen density @xmath249 .
@xmath250 denotes the density scale height .
this approach is motivated by the fact that at least the subsonic region will be well represented by the hydrostatic density law . in order to fit the empirical column densities we adjust the density parameter @xmath249 and the scale height @xmath250 .
identifying @xmath248 with the giant radius the least square procedure yields @xmath251 and @xmath252 .
the pressure may be composed of two parts , the thermal gas component and an effective turbulent pressure leading to the total dynamical pressure @xmath253 where @xmath254 is the mean molecular weight measured in units of the hydrogen mass .
the velocity @xmath255 denotes an isotropic turbulent motion with a spatial scale smaller than the pressure scale height . assuming the ionization equilibrium to be described by the physics of a low density plasma with @xmath256 we obtain @xmath257 [ for solar composition according to anders & grevesse ( 1989 ) ] .
if we identify the stochastic velocity of @xmath258 with the turbulent velocity we obtain a scale height of @xmath259 cm , in agreement with the best - fit parameter of the hydrostatic model .
however , at the impact parameter of @xmath61 the theoretical column density exceeds the observed upper limit by a factor of 3 ( see fig . [ fig5 ] ) .
as expected the hydrostatic equation provides an adequate description of the inner shell ( @xmath260 ) , but overestimates the densities in the outer envelope considerably . a more realistic model of the large - scale density distribution requires an appropriate kinematic approach .
we follow two different strategies to introduce a velocity law : an analytical @xmath4 power - law approach traditionally used in @xmath8 aurigae studies , and a hydrodynamic description of an isothermal , isoturbulent wind ( where @xmath261 ) . for the sake of simplicity
we adopt a steady , spherically symmetric outflow .
we are aware of the fact that these assumptions may be a point of criticism , since the observations suggest a considerable variability . however , the amplitude of these variations are small and justify a kind of `` average '' description .
the first kinematic model uses a velocity law proven to be adequate for different classes of stellar winds including the envelopes of evolved late - type stars .
the analyses of several @xmath8 aur binaries suggest a velocity field of the form @xmath262 where @xmath263 is the terminal wind velocity and @xmath4 the acceleration parameter which defines the steepness of the spatial velocity gradient associated with the expansion .
the wind velocity @xmath264 increases monotonically outward and approaches the terminal velocity at large distances from the giant .
the number density is assumed to vary in accordance to the equation of continuity .
the hydrogen density becomes @xmath265 where the density parameter @xmath266 is related to the mass - loss rate and terminal wind velocity : @xmath267 the hydrogen abundance by weight @xmath268 is fixed at @xmath269 assuming solar chemical composition .
the least square procedure yields a set of best - fit parameters @xmath270 @xmath271 and @xmath272 .
the small value of @xmath4 indicates a rapid acceleration to @xmath273 within @xmath274 .
we note that the wind models of `` classical '' @xmath8 aurigae systems ( @xmath8 aur , 31 cyg , 32 cyg ) indicate a gradual acceleration with @xmath4 in the range between 2.5 and 3.5 ( e.g. , @xcite ) . in the case of hr 6902 the terminal velocity eludes a spectroscopic determination , since we can only observe the tangential velocity component .
the poor quality of the spectra does not allow the use of alternative velocity indicators like absorption . as a consequence
the mass - loss information is restricted to the ratio @xmath275 .
the winds from hybrid stars show high terminal velocities ( typically 50 to @xmath276 ) as measured by the circumstellar absorption ( e.g. , @xcite ) .
furthermore , the flow velocities of luminous cool stars are always less than the escape speed from the photosphere ( for hr 6902 we find @xmath277 ) .
so we expect @xmath263 to lie in the range @xmath278 .
the corresponding mass - loss rate is @xmath279 with a statistical error of @xmath280 .
finally , we try to attack the mass motion using the equations of gas dynamics .
conservation of momentum gives the one - dimensional euler equation @xmath281 where @xmath282 denotes a potential extra force per unit - mass . in the present study we set @xmath283 and introduce an effective pressure term containing an unspecified nonthermal contribution assuming a proportionality of the form @xmath284
this simplification leads to a parker - type wind .
the euler equation can be integrated immediately and yields a solution with a positive velocity gradient at all distances ( e.g. , @xcite ) : @xmath285 = \left(\frac{r_{\rm c}}{r}\right)^2 \exp \left(-\frac{2r_{\rm c}}{r } + \frac{3}{2}\right).\ ] ] the transcendental equation can be solved numerically to give @xmath286 .
the critical radius is defined by the singularity of the euler equation : @xmath287 the hydrogen density may be expressed by an equation of the form @xmath288 where the density parameter @xmath289 is related to the mass - loss rate by @xmath290 this formalism suggests varying the density parameter and the critical velocity to model @xmath291 .
the critical velocity specifies the singularity of the momentum equation and thus determines the slope of the resulting density distribution .
we note that there is no need to take explicitly care of specific pressure constituents .
our least square fit procedure yields @xmath292 and @xmath293 .
the density parameter leads to a mass - loss rate of @xmath294 . using eq .
[ eq-12 ] the critical radius becomes @xmath295 , indicating a high velocity close to the surface of the giant .
[ fig5 ] demonstrates that both wind models lead to similar density laws .
indeed , the approximate relationship between @xmath296 and @xmath4 derived by harper et al . [ 1995 , eq .
( 10 ) ] yields an acceleration parameter of 0.54 quite similar to the result of our beta - law analysis .
this finding may justify the @xmath4 power - law approach in other red giant winds , where we have an unknown hydrodynamic situation . assuming the effective pressure to be supplied by both gas and turbulence pressure the critical velocity is given by @xmath297 the best - fit model with @xmath298 yields a turbulence of @xmath299 which exceeds the observed stochastic velocity by a factor of 1.8 .
this discrepancy may become even worse when we allow for macroscopic line broadening .
either the identification of turbulent and stochastic velocity is completely inappropriate , or the momentum eq .
[ eq-10 ] contains unknown external forces or pressure terms .
furthermore , the relation between the observed line broadening velocity and the turbulent velocity is not clear and depends on the driving mechanism(s ) , the geometry , and the statistics of the absorbing structures . in the preceding analysis
we have ignored line broadening due to the wind velocity . in order to examine the effect on the derived parameters we repeat the absorption line analysis with the complete optical depth quadrature according to eq .
the wind outflow is assumed to be well described by the @xmath4 power - law .
again we consider terminal velocities ranging from 50 to @xmath300 .
it turns out that high velocity models with @xmath301 can not reproduce the observed profile shapes adequately .
the deduced stochastic component of the line broadening lies in the interval @xmath302 showing partly a large standard error .
the range of stochastic velocities reflects both noisy _ iue _ spectra and true irregularities of the local velocity field . as an example we present in table [ tbl-3 ] the results for @xmath303 which represents the most probable fit model .
though this model yields considerably reduced stochastic velocities with a weighted mean ( i.e. , averaged over all heights ) of @xmath304 , the derived column densities are fairly close to the quasi - static model .
this finding even holds for larger wind speeds and we conclude that the quasi - static analysis has produced column densities with a reasonable degree of confidence .
reimers et al . ( 1990a ) have shown that the and emission lines are formed by resonance scattering of b star photons in the envelope of the primary .
profile modeling requires the solution of a two - dimensional radiative transfer problem .
the difficulty arises due to the fact that the light source ( i.e. the b star ) is displaced from the center of wind symmetry .
it is well - known that the scattering lines are formed in a large volume compared to the scale of the binary orbit ( @xcite ) .
hence , the emission line diagnostics allows to study the large - scale properties of the circumstellar envelope . for the 2-d radiative transfer calculations we employ the sei ( sobolev with exact integration ) method , i.e.
, the source function is approximated using the escape probability formalism followed by an exact formal solution using short characteristics .
this procedure is simple in concept and can be applied to very extended geometrical grids . in normal @xmath8 aur systems
the sobolev approach leads to erroneous results , since the wind velocities are typically only a few times the stochastic velocities .
however , a comparison with a more refined 2-d radiative transfer code ( @xcite , 1990 ) demonstrates the reliability of the sobolev solution in the case of hr 6902 .
the reason is the low mass - loss rate ( @xmath305 ) compared to the `` classical '' @xmath8 aur systems ( @xmath306 ) . as a consequence the line opacities are
small and multiple line photon scattering is of minor importance . in the framework of
the first order escape probability theory the scattering integral is given by ( e.g. , @xcite ) @xmath307s({\bf r})+\beta_{\rm c}({\bf r})i_{\rm c},\ ] ] where @xmath308 denotes a constant core intensity representing the b star .
the escape probabilities are given by @xmath309 and @xmath310 with the sobolev optical thickness @xmath311 .
the so - called sobolev length in the direction @xmath312 may be defined in tensor notation : @xmath313 the quantity @xmath314 denotes the probability that a photon emitted at the point @xmath315 escapes from the outer boundary and @xmath316 symbolizes the probability that a photon emitted by the core ( i.e. the b star ) penetrates to the point @xmath317 . in the case of pure resonance scattering the line source function is simply given by @xmath318 detailed information can be found in hempe ( 1982 ) who applied the sobolev formalism to the specific conditions of the binary configuration . with known source function on a two - dimensional grid , we can calculate the emergent radiation field .
the intensities are supplied by the formal solution of the transfer equation and can be integrated along a set of parallel rays through the envelope . in a subsequent step the integration over all solid angles and
normalization yields the residual flux @xmath319 . for a detailed description
we refer to baade ( 1990 ) and kirsch & baade ( 1994 ) . to calculate synthetic emission lines we extrapolate the wind models derived in sect .
[ sec3 ] to the whole envelope assuming the temperature and thus the ionization state to remain fixed .
though this simplistic approach implies spurious energy sources to maintain the temperature it may be sufficient to explain qualitatively the observed flux . in the first step we use the and line ratio observed during total eclipse to derive the ionization temperature .
unfortunately , the clearly visible line eludes the diagnostic procedure , since the photospheric background profile is not known precisely .
furthermore , the loci of line formation may be quite different .
we calculate a grid of theoretical flux ratios in the temperature range between 60000 and @xmath320 .
the ionization problem is solved using the cloudy code as described in sect .
[ sec32 ] . to consider the influence of different wind velocities we have a closer look at four different models defined either by the @xmath4 power - law or the @xmath5 wind according to sect .
[ sec33 ] . as shown in fig .
[ fig6 ] the theoretical flux ratios depend only slightly on the specific wind model .
the observed line ratio suggests an electron temperature of @xmath321 consistent with the mean temperature obtained on the basis of the absorption line analysis .
we use the derived electron temperature to calculate integrated line fluxes of the and emission lines .
the observed values ( table [ tbl-4 ] ) can be reconstructed using the parameters @xmath322 @xmath271 for the @xmath4 power - law and @xmath323 @xmath271 for the @xmath5 wind model . in both cases
the density parameter differs only by a factor of @xmath324 from the results of the absorption line analysis .
the discrepancy may be a result of the inadmissible extrapolation of the isothermal density distribution .
indeed , we would expect a decreasing temperature at large distances from the giant causing a reduced absorber density .
lcc ion & @xmath325 & f + & ( ) & @xmath326 + @xmath327 & 1395/1408 & @xmath328 + @xmath329 & 1548/1550 & @xmath330 + [ tbl-4 ] the presence of a macroscopic velocity field will influence the profile shapes and cause wavelength shifts of the emission lines .
however , the low resolution _ iue _ spectra ( @xmath331 ) does not allow to examine these effects more closely . to demonstrate the expected situation we present in fig .
[ fig7 ] a selection of theoretical flux profiles .
it is obvious that the asymptotic flow velocity determines the character of the line profiles . in combination with a semi - empirical density model the line positions and profile shapes give reliable information about the global wind structure
this allows us to predict that high - resolution uv observations will unravel the kinematic nature of the circumstellar envelope of hr 6902 .
in the present paper we have analyzed _ iue _ observations of the @xmath8 aurigae type binary system hr 6902 . the hot secondary serves as a convenient probe for the extended envelope of the bright giant . by means of a least square profile - fit procedure
we have deduced and column densities and stochastic velocities .
the density distribution of the inner envelope ( @xmath332 ) can be described by a quasi hydrostatic model .
we have shown that pressure support by turbulence is sufficient to explain the empirical scale height . however , in order to match the empirical column densities adequately we had to introduce an appropriate outflow model .
we have demonstrated that both the common @xmath4 power - law and a @xmath5 wind yield appropriate density distributions and explain the absence of and absorption at larger distances from the giant .
[ fig8 ] demonstrates that the hydrostatic solution as well as the different outflow models lead to similar density laws . despite this similarity
the hydrodynamic implications are completely different .
the required driving force per unit mass @xmath282 should be compared to theoretical models .
we have derived mass - loss rates in the range @xmath333 .
this result indicates that the mass - loss rate of hr 6902 scales roughly with the solar mass flux , i.e. , @xmath334 .
finally , we have demonstrated that the emission line spectrum obtained during total eclipse confirms essentially the wind models derived from the absorption line analysis .
this finding supports the picture of a steady and spherically symmetric flow without excessive deviations from the mean wind model .
it is shown that hr 6902 possesses warm ( @xmath335 ) circumstellar material , observed in a range of impact parameters between 1.1 and @xmath336 .
recently , the innermost region immediately above the photosphere ( @xmath337 ) has been examined by schrder , marshall , & griffin ( 1996 ) using optical spectra .
they derived a temperature of @xmath338 and a turbulent velocity of @xmath339 .
however , in order to fit the k - line wing they postulated a second absorption component with a considerably larger turbulent velocity of @xmath340 quite similar to the stochastic velocities derived in the present work . combining these results we suspect a dramatic change of the physical conditions in the giant s atmosphere at @xmath341 associated with a steep rise of the temperature from about 5000 to @xmath342 . the large nonthermal line broadening parameter and its relative constancy ( apart from local fluctuations ) may be a characteristic property of the wind of hr 6902 .
the simplistic identification of turbulence and the observed stochastic velocity can not explain the large pressure required to explain the density distribution by a @xmath5 wind .
however , the true relation between the observed line width and turbulent motions associated with wave - driven winds depends on the specific driving mechanism , geometrical details , and the randomness of the absorbing structures ( e.g. , @xcite ) .
the wind models suggest a rapid wind acceleration within @xmath343 above the photosphere .
this result is similar to the outflow characteristics of hybrid stars [ @xmath344 aqr ( g2 ib ) , @xmath4 aqr ( g0 ib ) , @xmath345 her ( k1 iia ) , @xmath346 aur ( k3 ii ) , @xmath347 aql ( k 3ii ) , and @xmath344 tra ( k3 ii ) ; @xcite ; @xcite ; @xcite ] .
furthermore , a recent analysis of _ orfeus ii _ spectra of the hybrid stars @xmath344 tra and @xmath344 aqr ( @xcite ) have revealed that the stellar winds are not confined to cool circumstellar material .
they pointed out that ions like through indicate expanding material up to a temperature of @xmath348 with wind velocities from 100 to @xmath349 .
we argue that the wind properties of hybrid stars are comparable to the extended atmosphere of hr 6902 . with its location in the hr diagram
hr 6902 represents the physically important connection between solar - like stars with hot coronae and late type supergiants with cool outer atmospheres .
the binary technique provides spatially resolved information about the physical structure of the extended envelope and wind that are unobtainable by other means . with _
iue _ we have gained a first insight , but additional high resolution uv observations are indispensable to construct a detailed model . with this unique opportunity hr 6902 may play a crucial role in the elucidation of atmospheric heating processes and mechanisms to drive winds of cool stars . + this work is supported by the verbundforschung of the bundesministerium fr bildung , wissenschaft , forschung und technologie under grant no .
50 or 96016 .
this research has made use of iuefa data obtained through the esa _ iue _ data server .
the authors would like to thank graham harper for discussions and useful suggestions .
gary ferland is kindly acknowledged for providing his ionization code cloudy .
we made use of the vizir search engine , a joint effort of cds ( strasbourg , france ) and esa - esrin .
ahmad i.a . , 1993 , in the realm of interacting binary stars , sahade a. et al .
( dordrecht : kluwer ) , 305 anders e. , grevesse n. , 1989 , geochim .
cosmochim .
acta 53 , 197 baade r. , 1989 , in rev .
astron . 2 , 324 baade r. , 1990 , a&a 233 , 486 baade r. , kirsch t. , reimers d. , toussaint f. , bennett p.d .
, brown a. , harper g.m .
, 1996 , apj 466 , 979 baade r. , 1998 , in ultraviolet astrophysics , beyond the _ iue _ final archive , harris b. ( ed . ) ( esa sp-413 ) , 325 brosius j.w .
, mullan d. j. , 1986 , apj 301 , 650 che - bohnenstengel a. , reimers d. , 1986 , a&a 156 , 172 dupree a.k . ,
brickhouse n.s . , 1998 ,
apj 500 , l33 dupree a.k . , reimers d. , 1987 , in scientific accomplishments of the iue , kondo y. ( ed . ) ( dordrecht : reidel ) , 321 dupree a.k . , whitney b.a .
, avrett e. h. , 1992 , in proc .
7th cambridge workshop on cool stars , stellar systems , and the sun , giampapa m.s .
, bookbinder j.a .
( asp conf . ser . ) , vol .
26 , 525 ferland g.j .
, 1997 , univ .
kentucky dept .
physics and astron .
internal rept .
garhart m.p . ,
smith m.a .
, levay k.l . ,
thompson r.w . , 1997 , international ultraviolet explorer new spectral image processing system information manual , version 2.0 griffin r.e.m . ,
marshall k.p . ,
griffin r.f . , schrder k .-
, 1995 , a&a 301 , 217 guinan , e.f .
, 1990 , in evolution in astrophysics , rolfe e.j .
( esa sp-310 ) , 74 haisch b.m . ,
schmitt j.h.m.m . , rosso c. , 1991 , apj 383 , l15 haisch b.m . ,
schmitt j.h.m.m . ,
rosso c. , 1992 , apj 388 , l61 harper g.m . ,
, linsky j.l . , bennett p.d . , ayres t.r . , brown a. , 1995 , apj 452 , 407 harper g. , 1996 , in proc .
9th cambridge workshop on cool stars , stellar systems , and the sun , pallavicini r. , dupree a.k .
. ser . ) , vol .
109 , 481 hartmann l. , dupree a.k .
, raymond j.c . , 1980 ,
apj 236 , l143 hempe k. , 1982 , a&a , 115 , 133 judge p.g . , 1992 , in proc .
7th cambridge workshop on cool stars , stellar systems , and the sun , giampapa m.s .
, bookbinder j.a.(eds . )
( asp conf .
ser . ) , vol .
26 , 403 kirsch t. , baade r. , 1994 , a&a 291 , 535 lambert d. c. , 1981 , in physical processes in red giants , iben i. , renzini a. ( eds . ) ( dordrecht : reidel ) , 115 linsky j.l . , haisch , b.m . , 1979 , apj 229 , l27 mihalas d. , 1978 , in stellar atmospheres ( 2d .
ed . ; san francisco : freeman ) press w.h .
, teukolsky s.a .
vetterling w.t . ,
flannery b.p . , 1992 , numerical recipes in fortran , 2nd ed .
( cambridges : cambridge university press ) reimers d. , 1977 , a&a 57 , 395 reimers d. , 1982 , a&a 107 , 259 reimers d. , 1989 , in fgk stars and t tauri stars , cram l.e . , kuhi l.v .
( nasa sp-502 ) , 53 reimers d. , baade r. , schrder k .-
p . , 1990a , a&a 227 , 133 reimers d. , baade r. , schrder k .-
, ising j. , 1990b , a&a 236 , l25 reimers d. , hnsch m. , schmitt j.h.m.m . , toussaint f. , 1996 , a&a 310 , 813 rosner r. , 1993 , in physics of solar and stellar coronae , linsky j.l .
, serio s. ( eds . ) ( dordrecht : kluwer ) , 549 rosner r. , musielak z.e .
, cattaneo f. , moore r.l . ,
suess s.t . , 1995 , apj 442 , l25 schrder k .-
, 1985 , a&a 147 , 103 schrder k .- p . , marshall k. p. , griffin r.e.m .
, 1996 , a&a 311 , 613 schrder k .-
, pols o.r . ,
eggleton p.p . , 1997 ,
mnras 285 , 696 stencel r.e .
, mullan d.j .
, 1980a , apj 238 , 221 stencel r.e .
, mullan d.j .
, 1980b , apj 240 , 718 | _ iue _ observations of the eclipsing binary system hr 6902 obtained at various epochs spread over four years indicate the presence of warm circumstellar material enveloping the g9 iib primary .
the spectra show and absorption up to a distance of 3.3 giant radii ( @xmath0 ) .
line ratio diagnostics yields an electron temperature of @xmath1 which appears to be constant over the observed height range . applying a least square fit absorption line analysis we derive column densities as a function of height .
we find that the inner envelope ( @xmath2 ) of the bright giant is consistent with a hydrostatic density distribution .
the derived line broadening velocity of @xmath3 is sufficient to provide turbulent pressure support for the required scale height .
however , an improved agreement with observations over the whole height regime including the emission line region is obtained with an outflow model .
we demonstrate that the common @xmath4 power - law as well as a @xmath5 wind yield appropriate fit models . adopting a continuous mass outflow
we obtain a mass - loss rate of @xmath6 @xmath7 depending on the particular wind model .
the emission lines observed during total eclipse are attributed mostly to resonance scattering of b star photons in the extended envelope of the giant . by means of a multi - dimensional line formation study
we show that the global envelope properties are consistent with the wind models derived from the absorption line analysis .
we argue that future high resolution uv spectroscopy will resolve the large - scale velocity structure of the circumstellar shell . as an illustration we present theoretical and emission profiles showing model - dependent line shifts and asymmetries . |
qcd cosmology is important as the universe exits the radiation dominated era and enters the matter dominated era , eventually evolving into the dark energy dominated era @xcite .
the effect of the strong interactions on cosmology was considered early on @xcite but the nonperturbative nature of the strong interactions at low energy limited the progress of the subject . today with nonperturbative approaches like
lattice qcd @xcite and ads / qcd @xcite it is important to revisit the subject to see what can be said quantitatively about qcd cosmology . also besides the time dependence of the scale factor in qcd cosmology , one can also look at the effect of recent advances on the qcd equation of state for exotic astrophysical objects like quark or strange stars @xcite .
in this paper we consider a spatially flat friedmann - robertson - walker universe where the radius of the universe @xmath0 , energy density @xmath1 , and pressure @xmath2 all depend on time .
conservation of the local energy - momentum tensor leads to : @xmath3 the friedmann equation relates the time evolution of the universe to energy density and is is given by : @xmath4 where @xmath5 is the reduced planck mass .
the equation of state refers to the dependence of energy density and pressure on temperature @xmath6 . for a given equation of state one can rewrite the above two equations in a form which determines the expansion factor @xmath0 .
first we can determine the time dependence of the temperature from : @xmath7 this can be integrated to give : @xmath8 then one can invert the function @xmath9 to yield @xmath10 .
next one can determine the radius @xmath0 time dependence from equation ( [ eq : emtensor ] ) which is integrated to yield : @xmath11 for a simple example consider a radiation dominated universe with equation of state : @xmath12 where @xmath13 is a constant .
solving equation ( [ eq : dtdt ] ) we find @xmath14 and @xmath15 . performing the integration ( [ eq : at ] )
we obtain : @xmath16 this @xmath17 time dependence also gives the very high temperature and early time behavior of qcd cosmology . for nonzero cosmological constant @xmath18 the friedmann equation
is modified to : @xmath19 for all the models of low energy qcd that we use in this paper the cosmological constant provides only a small modification .
this is because the energy density scale of low energy qcd is typically @xmath20 whereas the cosmological constant is @xmath21 @xcite .
modifications are expected by the cosmological constant as can been seen , e.g. , by the scale factor for matter domination @xmath22 which reduces to @xmath23 at small times and @xmath24 at large times .
we will not investigate corrections with respect to the cosmological constant in this paper and leave this for further study at a later stage .
another possible modification to the cosmological equations for @xmath0 involves the bulk viscosity .
the inclusion of bulk viscosity modifies the conservation equation by : @xmath25 where @xmath26 is the bulk viscosity .
qcd bulk viscosity is currently being studied in heavy ion collisions and qcd , see e.g. @xcite and @xcite , respectively , thus it is of interest to see what cosmological effect the bulk viscosity may have .
we leave the study of the cosmological effects of the bulk viscosity for future work . for any realistic equation of state related to qcd
the above steps which determine @xmath9 , @xmath10 , @xmath27 and @xmath0 all have to be done numerically .
we shall find a simple radiation dominated picture for the scale factor @xmath0 at early times corresponding to deconfinement . at late times
the time dependence of the scale factor is quite complex .
one can use an effective description using masses of various resonances form the particle data table as in the hadronic resonance gas model ( hrg)@xcite , one can use an effective hagedorn string picture @xcite as weinberg and huang did in @xcite , one can use a ads / cft approach to the equation of state @xcite , a symmetry which relates small radius to large radius qcd - like theories if one thinks of the time direction as compactified as in an imaginary time formalism @xcite , or one can use lattice simulations to determine the equation of state @xcite . in this paper
we compare the results of some of these approaches applied to the calculation of the scale factor @xmath0 . this paper is organized as follows . in section 2
we discuss the numerical determination of the scale factor of the universe as a function of time , @xmath0 , from the lattice qcd equation of state .
we work with zero chemical potential and bulk viscosity .
we plan to include these in future work . in section 3
we determine the scale factor from the hadronic resonance gas model model . in section 4
we discuss the scale factor time dependence for an early universe model which has a hagedorn ultimate temperature . in section 5
we discuss the scale factor determined from the ads / cft approach to the equation of state following a similar approach to gubser and nellore @xcite . in section 6 we state the main conclusions of the paper .
lattice qcd is a modern tool which allows one to systematically study the non - perturbative regime of the qcd equation of state . utilizing supercomputers the qcd equation of state
was calculated on the lattice in @xcite with two light quarks and a heavier strange quark on a @xmath28 size lattice .
the quark masses have been chosen to be close to their physical value , i.e. the pion mass is about @xmath29 . for further details
we refer the reader to @xcite , however , we like to remark that the equation of state was calculated at a temporal extent of the lattice @xmath30 and for @xmath30 sizable lattice cut - off effects are still present @xcite . the data for energy density @xmath31 , pressure @xmath32 and trace anomaly @xmath33 and entropy @xmath34 of ref .
@xcite are given in table [ tab1 ] .
all the analysis in this section is derived from this data . besides the strange quark
one can also include the effect of the charm quark as well as photons and leptons on the equation of state .
these have important cosmological contributions as was shown in @xcite .
recent references on lattice qcd at high temperature are @xcite@xcite@xcite .
.[tab1 ] data from lattice qcd study with @xmath35 @xcite .
temperature in units of @xmath36 , energy density , pressure and trace anomaly in units of @xmath37 , entropy density @xmath34 in units of @xmath38 . [ cols="^,^,^,^,^",options="header " , ] rewriting the equation ( [ eq : dtdt ] ) that determines the time dependence of the temperature in terms of entropy and energy density we have : @xmath39 then using the correspondence from table [ tab2 ] one can determine the radius of the 4d universe @xmath0 from ( [ eq : dtdts ] ) .
this is plotted in figure [ fig10 ] .
one can see from the plot on right hand side of figure [ fig10 ] that over the temperature range covered by the lattice qcd data one never leaves the radiation regime and the scale factor is well approximated by @xmath40
. in this ads approach it is is important not to confuse the scale factor @xmath0 with the dual radius @xmath41 .
one uses the black hole solution including @xmath41 to determine the entropy , temperature and energy density and then uses ( [ eq : dtdts ] ) to determine the scale factor @xmath0 through the correspondence in table [ tab2 ] . because the vacuum solution to @xmath42 gravity does not match with the lattice qcd data at low temperature and stays in the radiation regime over the temperature range of lattice qcd , one looks for non vacuum solutions that can mimic the lattice qcd equation of state . in @xcite
the potential @xmath43 was used to describe a @xmath42 black hole with nonzero dilaton field . for nonzero @xmath44 and @xmath45
one can solve the equations of motion numerically .
in @xcite it was shown that the speed of sound associated with the potential ( 5.2 ) closely approximates the speed of sound from lattice qcd . for nonzero dilaton
it is convenient to replace the coordinate @xmath46 by @xmath47 .
then the asymptotic region @xmath48 corresponds to @xmath49 . in the coordinate @xmath50 , the five dimensional metric ansatz is taken to be : @xmath51 it is convenient to choose the gauge : @xmath52 in this gauge the equations of motion become : @xmath53 where the prime refers to the derivative with respect to @xmath50 and as before where we have defined @xmath54 and @xmath55 .
we found that the asymptotic form of the dilaton at small @xmath50 had an important effect on the values of the entropy and mass of the black hole thus one needs to carefully study the effects of various dilaton boundary conditions to mimic the lattice qcd equation of state . in the asymptotic regime near for small @xmath50
we seek a black hole solution to to the equations of motion with a small asymptotic value for the dilaton field @xmath56 . for small @xmath50
we set : @xmath57 where @xmath58 is a deformation parameter . in terms of @xmath59 and @xmath60
this becomes : @xmath61 one can then use the second equation in ( 5.3 ) to solve for the asymptotic form of the dilaton by integrating : @xmath62 as in @xcite .
this leads to the asymptotic form of the dilaton : @xmath63 expanding this expression for small @xmath50 we find : @xmath64 one can use this expression to define a boundary condition on the dilaton field for small @xmath50
. then one can numerically solve for the black hole solution with potential ( 5.2 ) and a given deformation parameter @xmath58 .
as @xmath58 goes to zero one has the vacuum black hole solution discussed in the previous subsection . for @xmath65 the entropy , mass ( energy density ) , and temperature
are all modified by the dilaton . in figure [ fig11 ]
we plot the entropy and mass as a function of temperature for @xmath65 .
one can then use equation ( [ eq : dtdts ] ) to calculate the time dependence of the physical radius @xmath0 .
this is shown in figure [ fig12 ] .
the inclusion of the dilaton creates nontrivial deviation from the radiation type expansion at late times and low temperatures .
in particular , we observe in figure [ fig12 ] the upward bending hockey stick behavior that is qualitatively similar to that of lattice qcd . without the dilaton
the scale factor @xmath0 can be described by the radiation expression @xmath17 over the temperature range covered by the lattice qcd calculation . with the dilaton it is possible to describe a late time behavior similar to lattice qcd .
we have computed the time dependence of the scale factor from the lattice qcd equation of state with @xmath30 .
we find that the scale factor is described by a radiation dominated universe @xmath17 at early times with a complicated time dependence at late times which seems to be closer to a matter dominated universe .
we compared our results from lattice qcd with other approaches to the low temperature equation of state including the hadronic resonance gas model , the hagedorn model and the ads / cft equation of state .
we found that on a log log plot the scale factor displayed an upward pointing hockey stick behavior for the lattice qcd data .
for the hadronic resonance gas model over the temperature range of lattice qcd data , the hrg model lies above the lattice qcd data @xcite .
the slope of the log log plot for the scale factor is between the radiation value .5 and the matter dominated value .666 for the hrg model in this regime .
the hagedorn model leads to a limiting temperature as a function of time and works better a low temperature and late times .
the equations are simpler than the hrg model because less experimental input is required but lead to a diverging energy density at the hagedorn temperature unlike the lattice qcd data .
the hagedorn cosmological model is still of interest for theories where the limiting temperature is at the tev scale or the planck scale . for the ads
/ qcd model we found that without the dilaton the entropy and energy density led to a @xmath17 expansion over the temperature range of lattice qcd .
we introduced a parameter @xmath58 that determined the boundary condition at infinity for the dilaton . turning the dilaton on and
tuning the parameter @xmath58 allowed us to derive a time dependence for the scale factor that was qualitatively similar to that of lattice qcd through the ads / qcd correspondence . with future improvements in lattice qcd calculations and deeper understanding of the role of ads
/ qcd it will be interesting to revisit the subject of qcd cosmology by adding the effects of bulk viscosity , chemical potentials and interaction with leptons . eventually one would like to make contact with astrophysical measurements of the early universe as envisioned in early references such as @xcite .
we wish to thank frithjof karsch , michael creutz , dmitri kharzeev and mikko laine for useful discussions and suggestions .
this manuscript has been authored in part by brookhaven science associates , llc , under contract no . de - ac02 - 98ch10886 with the u.s .
department of energy .
e. witten , `` cosmic separation of phases , '' phys .
d * 30 * , 272 ( 1984 ) .
d. j. schwarz , `` the first second of the universe , '' annalen phys .
* 12 * , 220 ( 2003 ) [ arxiv : astro - ph/0303574 ] .
l. m. mller , master s thesis , `` the quark - hadron phase transition in the early universe '' , ( 2001 ) .
n. borghini , w. n. cottingham and r. vinh mau , `` possible cosmological implications of the quark - hadron phase transition , '' j. phys .
g * 26 * , 771 ( 2000 ) [ arxiv : hep - ph/0001284 ] .
j. i. kapusta , `` quark - gluon plasma in the early universe , '' arxiv : astro - ph/0101516 .
d. chandra and a. goyal , `` dynamical evolution of the universe in the quark - hadron phase transition and possible nugget formation , '' phys .
d * 62 * , 063505 ( 2000 ) [ arxiv : hep - ph/9903466 ] .
a. a. coley and t. trappenberg , `` the quark - hadron phase transition , qcd lattice calculations and inhomogeneous big bang nucleosynthesis , '' phys .
d * 50 * , 4881 ( 1994 ) [ arxiv : astro - ph/9307031 ] .
h. suganuma , h. ichie , h. monden , s. sasaki , m. orito , t. yamamoto and t. kajino , `` qcd phase transition at high temperature in cosmology , '' arxiv : hep - ph/9608333 .
k. a. olive , `` the quark - hadron transition in cosmology and astrophysics , '' science * 251 * , 1194 ( 1991 ) .
r. hagedorn , `` statistical thermodynamics of strong interactions at high - energies , '' nuovo cim .
suppl . * 3 * , 147 ( 1965 ) .
k. huang and s. weinberg , `` ultimate temperature and the early universe , '' phys .
lett . * 25 * , 895 ( 1970 ) .
m. cheng _ et al .
_ , `` the qcd equation of state with almost physical quark masses , '' phys . rev .
d * 77 * , 014511 ( 2008 ) [ arxiv:0710.0354 [ hep - lat ] ] .
s. s. gubser and a. nellore , `` mimicking the qcd equation of state with a dual black hole , '' arxiv:0804.0434 [ hep - th ] .
s. s. gubser , a. nellore , s. s. pufu and f. d. rocha , `` thermodynamics and bulk viscosity of approximate black hole duals to finite temperature quantum chromodynamics , '' arxiv:0804.1950 [ hep - th ] .
s. s. gubser , s. s. pufu and f. d. rocha , `` bulk viscosity of strongly coupled plasmas with holographic duals , '' arxiv:0806.0407 [ hep - th ] . b. freedman and l. d. mclerran , `` quark star phenomenology , '' phys . rev .
d * 17 * , 1109 ( 1978 ) .
s. chakrabarty , `` equation of state of strange quark matter and strange star , '' phys . rev .
d * 43 * , 627 ( 1991 ) .
w. m. yao _ et al .
_ [ particle data group ] , `` review of particle physics , '' j. phys .
g * 33 * , 1 ( 2006 ) .
m. luzum and p. romatschke , `` conformal relativistic viscous hydrodynamics : applications to rhic , '' arxiv:0804.4015 [ nucl - th ] .
d. kharzeev and k. tuchin , `` bulk viscosity of qcd matter near the critical temperature , '' arxiv:0705.4280 [ hep - ph ] .
f. karsch , k. redlich and a. tawfik , `` hadron resonance mass spectrum and lattice qcd thermodynamics , '' eur .
j. c * 29 * , 549 ( 2003 ) [ arxiv : hep - ph/0303108 ]
. f. karsch , k. redlich and a. tawfik , `` thermodynamics at non - zero baryon number density : a comparison of lattice and hadron resonance gas model calculations , '' phys .
b * 571 * , 67 ( 2003 ) [ arxiv : hep - ph/0306208 ] .
k. sakthi murugesan , g. janhavi and p. r. subramanian , `` can the phase transition from quark - gluon plasma to hadron resonance gas affect primordial nucleosynthesis ?
, '' phys .
d * 41 * , 2384 ( 1990 ) .
a. tawfik , `` the qcd phase diagram : a comparison of lattice and hadron resonance gas model calculations , '' phys .
d * 71 * , 054502 ( 2005 ) [ arxiv : hep - ph/0412336 ] .
m. shifman and m. unsal , `` qcd - like theories on @xmath66 : a smooth journey from small to large @xmath67 with double - trace deformations , '' arxiv:0802.1232 [ hep - th ] .
r. gupta , the eos from simulations on bluegene l supercomputer at llnl and nyblue , pos * lat2008 * 170 ( 2008 ) m. laine and y. schroder , `` quark mass thresholds in qcd thermodynamics , '' phys .
d * 73 * , 085009 ( 2006 ) [ arxiv : hep - ph/0603048 ] .
m. cheng [ rbc - bielefeld collaboration ] , `` charm quarks and the qcd equation of state , '' pos * lat2007 * , 173 ( 2007 ) [ arxiv:0710.4357 [ hep - lat ] ] . g. endrodi , z. fodor , s. d. katz and k. k. szabo , `` the equation of state at high temperatures from lattice qcd , '' pos * lat2007 * ,
228 ( 2007 ) [ arxiv:0710.4197 [ hep - lat ] ] .
d. e. miller , `` lattice qcd calculation for the physical equation of state , '' phys .
* 443 * , 55 ( 2007 ) [ arxiv : hep - ph/0608234 ] .
p. braun - munzinger , k. redlich and j. stachel , `` particle production in heavy ion collisions , '' arxiv : nucl - th/0304013 ; a. andronic , p. braun - munzinger and j. stachel , `` hadron production in central nucleus nucleus collisions at chemical freeze - out , '' nucl .
a * 772 * , 167 ( 2006 ) [ arxiv : nucl - th/0511071 ] .
r. d. carlitz , `` hadronic matter at high density , '' phys .
d * 5 * , 3231 ( 1972 ) .
p. castorina , d. kharzeev and h. satz , `` thermal hadronization and hawking - unruh radiation in qcd , '' eur .
j. c * 52 * , 187 ( 2007 ) [ arxiv:0704.1426 [ hep - ph ] ] .
a. nayeri , r. h. brandenberger and c. vafa , `` producing a scale - invariant spectrum of perturbations in a hagedorn phase of string cosmology , '' phys .
lett . * 97 * , 021302 ( 2006 ) [ arxiv : hep - th/0511140 ] .
p. tyson , `` absolute hot : is there an opposite to absolute zero ? ''
+ http://www.pbs.org/wgbh/nova/zero/hot.html
e. poisson and w. israel , `` internal structure of black holes , '' phys .
d * 41 * , 1796 ( 1990 ) .
w. fischler , d. morgan and j. polchinski , `` quantization of false vacuum bubbles : a hamiltonian treatment of gravitational tunneling , '' phys .
d * 42 * , 4042 ( 1990 ) .
n. pidokrajt , `` black hole thermodynamics '' , master s thesis ( 2003 ) .
j. louko , j. z. simon and s. n. winters - hilt , `` hamiltonian thermodynamics of a lovelock black hole , '' phys .
d * 55 * , 3525 ( 1997 ) [ arxiv : gr - qc/9610071 ] .
h. quevedo and a. sanchez , `` geometrothermodynamics of asymptotically de sitter black holes , '' arxiv:0805.3003 [ hep - th ] .
w. de paula , t. frederico , h. forkel and m. beyer , `` dynamical ads / qcd with area - law confinement and linear regge trajectories , '' arxiv:0806.3830 [ hep - ph ] . | we numerically determine the time dependence of the scale factor from the lattice qcd equation of state , which can be used to define a qcd driven cosmology .
we compare a lattice approach to qcd cosmology at late times with other models of the low temperature equation of state including the hadronic resonance gas model , hagedorn model and ads / cft . 0.5 in 0.0 in 5.5 in 0.0 in .25 in = 9.0 in addtoresetequationsection |
the development of linear scaling computational methods for electronic structure calculations in molecules and solids with a very large number of atoms , ( i.e. methods whose computational demands grow as the first power of the size of the system , ) has been a very active and successful field in the last decade.@xcite linear scaling and low order scaling techniques exist for the computation of the one - electron effective hamiltonian matrix in first - principles calculations ( with density functional theory and wave function based methods ) , @xcite as well as for the energy minimization step , @xcite which is a common step to first - principles and semiempirical calculations . in low order scaling energy minimization methods ,
the traditional cubic scaling diagonalization of the matrix representation of the effective one - electron hamiltonian in a finite basis set , @xcite which leads to the canonical orbitals , is substituted by different algorithms which , either solve directly for the unique optimal density matrix , @xcite or for some sort of arbitrary optimal localized orbitals .
@xcite in parallel to linear scaling electronic structure methods , a significant development has also been made in embedding methods , which focus the computational effort on local properties of a system @xcite ( see refs . and for recent reviews ) . in this paper
, we present a new method for the energy minimization step which can be used in semiempirical and first - principles hartree - fock and kohn - sham calculations .
it is expected to be useful as a complement of linear - scaling methods for the computation of the hamiltonian matrix , which are nowadays available for local exchange - correlation potentials and coulomb potentials , @xcite as well as for exact exchange fields .
@xcite the energy minimization method is based on exploiting the self - consistent calculation of the optimum localized orbitals of any localization method of choice and the use of orbital - specific basis sets .
the @xmath0 occupied localized orbitals of a very large molecule which correspond to a chosen localization method ( e.g. the popular methods of boys , @xcite edmiston - ruedenberg , @xcite and pipek - mezey , @xcite or any other available or newly developed localization method ) can be regarded as a mosaic of orbitals , where each of its component contains only a few orbitals . in the present method
we define an orbital as a subset of the occupied localized orbitals ( of the method of choice ) which are localized in some region of real space and compute the localized orbitals of each out of one specific pseudoeigenvalue equation to be solved in a basis set expansion approximation , using a basis set specific to the , or orbital - specific basis set . in other words , a set of building - block embedded cluster pseudoeigenvalue coupled equations , one for each , is solved in a self - consistent manner . doing so , the method becomes parallel ( the are computed independently of each other in the self - consistent procedure ) and exhibits a linear - scaling dependence with the size of the molecule .
we call the present method . early works on localized orbitals proposed the ideas of computing them directly by a self - consistent procedure @xcite and computing one or several localized orbitals out of separate eigenvalue equations starting with a set of qualitatively localized orbitals .
@xcite these ideas have been used by several authors to propose methods leading to some particular sets of localized orbitals or to localized orbitals dependent on the initial guess .
@xcite we apply them to the computation of the occupied localized orbitals of any localization method of choice . here
, we do not pay attention to the computation of virtual localized orbitals ; methods for the determination of virtual localized orbitals useful in wave function based correlation methods have been used already in early works @xcite .
the idea of using different basis sets for different regions of the system is also present in early work , @xcite it has been used and discussed by several authors , @xcite and is a common procedure in embedded cluster and effective core potential calculations .
@xcite besides its use as a linear scaling method in large molecules and solids , the method can be used as an embedded cluster method , where only a few localized orbitals of a large system are treated variationally while the rest is taken from a calculation on a similar molecule and frozen , with obvious computational advantages .
this can be done because the solutions of the building - block embedded cluster coupled equations are the localized orbitals corresponding to a given localization method of choice , which enables their transferability between similar molecules . in sec .
[ sec : method ] we present the details of the method .
we performed monitoring calculations on large poly(ethylen oxide ) molecules and three dimensional carbon monoxide clusters using an extended hckel semiempirical hamiltonian , which are presented in sec .
[ sec : results ] .
they are aimed at showing the convergence of the parallel calculation to the right solution , the convergence of the total energy with the size of the orbital - specific basis sets towards the exact value , the linear - scaling characteristics of the method , and the performance of embedded cluster approximations .
let us consider the hartree - fock equations in wave function based _ ab initio _ methods , @xcite the kohn - sham equations in density functional theory , @xcite or the effective hartree - fock equations in semiempirical methods .
@xcite the canonical form of the spin - restricted closed - shell version of these equations can be written as @xmath1 with an appropriate choice of the one - electron hamiltonian @xmath2 for each case , where @xmath3 is a row vector of @xmath0 occupied molecular orbitals , @xmath4 and @xmath5 is an @xmath6 diagonal matrix of orbital energies .
this and what follows can be generalized to the spin - unrestricted cases if the orbitals and the hamiltonian in eqs .
[ eq : caneq ] and [ eq : canmos ] are adequately substituted by the @xmath7 and @xmath8 choices ; @xcite here we will continue with the detailed description of the spin - restricted closed - shell case for the sake of clarity .
the virtual orbitals , @xmath9 , are also solutions of eq .
[ eq : caneq ] ; in this paper , however , we will focus our attention on the occupied spectrum and , unless specified , we will refer only to different sets of occupied orbitals from now on .
the fock - dirac one - electron density operator is defined as @xmath10 it is invariant under arbitrary unitary transformations of the occupied orbitals , @xmath11 where we use the notation @xmath12 for the unitary matrix that transforms canonical occupied orbitals onto localized orbitals of a given localization method , which we label @xmath13 . also , @xmath14 is the projection operator of the occupied space , @xmath15 the fact that wave function , one - electron density , total energy , and fock operator are invariant under unitary transformations within the occupied space ( eq . [ eq : canul ] ) has been exploited to define localized orbitals , which are useful to facilitate large scale calculations and to bridge the gap between extensive numerical calculations and qualitative chemical thinking .
a localization method , say @xmath13 , can be defined by its choice of @xmath12 .
very sound and popular localization methods are the methods of boys , @xcite edmiston - ruedenberg , @xcite and pipek - mezey , @xcite although others have been proposed .
all of the above can be used in first - principles methods ; pipek - mezey s can also be applied in semiempirical methods .
the common procedure to compute localized orbitals is to complete firstly a canonical calculation and , later , use the canonical orbitals in an iterative optimization process converging to @xmath12 .
gilbert @xcite has pointed out that any set of occupied localized orbitals formally fulfills the eigenvalue equation of an effective fock ( or kohn - sham ) hamiltonian defined as @xmath16 , @xmath17 \,\underline{\varphi}^{l } = \underline{\varphi}^{l } \ , \underline{\lambda } \,,\end{aligned}\ ] ] where @xmath18 is a hermitean localization operator and @xmath19 is a diagonal matrix whose diagonal elements are the eigenvalues of @xmath20 and of @xmath18 , @xmath21 localization operators @xmath18 corresponding to the above mentioned localization methods are , in general , not known and eq .
[ eq : loceq ] has not been exploited to compute the localized orbitals via diagonalization of the matrix @xmath22 , to our knowledge .
[ eq : loceq ] has been discussed by several authors @xcite and it has been used as a basis for a building - block technique based on a self - consistent series of embedded cluster calculations . @xcite the iterative solution of the building - block equations of ref . , which contain arbitrary localization operators , leads to localized orbitals dependent on the initial guess and on the iteration procedure .
although in principle this is not detrimental for total energy calculations , it is an undesirable property because it creates problems of transferability of localized orbitals between similar molecules . here , on the basis of the same ideas than in ref .
, we present a new building - block and embedding method that , starting from an arbitrary guess and a choice of a particular localization method ( boys , edmiston - ruedenberg s , pipek - mezey s , or any other ) , leads , in a controlled manner , to the corresponding localized orbitals
. we will call this method .
let us suppose we set the goal of computing the @xmath0 occupied localized orbitals corresponding to a localization method @xmath13 for the ground state of a molecule , @xmath23 and related properties such as electron density and total energy .
we see the whole set of localized orbitals as a mosaic of orbitals , and define subsystem , fragment , cluster , or _
tessera _ as a subset of these orbitals which are localized in some region of real space .
( for example , in an organic acid r - cooh we may define one of the subsystems or _
tesserae _ as that made of nine orbitals localized in the spatial region close to the cooh nuclei . )
the terms subsystem , fragment , and cluster have been used by many authors with different meanings and we prefer to use the term _ tessera _ all over the paper for the present definition of subsystem . in this way , the whole mosaic of localized orbitals is made of @xmath24 _ tesserae _ , and the vector of localized orbitals can be rewritten as @xmath25 where @xmath26 is a row vector with the @xmath27 occupied orbitals of a , @xmath28 @xmath29 a row vector with the @xmath30 orbitals of b , @xmath31 and so on , with @xmath32 .
the orbitals of a define a subspace of the occupied space whose density and projection operator is @xmath33 the density operators of the fulfil @xmath34 where the sum in eq .
[ eq : rhocond - b ] extends over all the @xmath24 of the system . in the method we seek to compute the localized orbitals of each out of its own eigenvalue equation : @xmath35 \,\underline{\varphi}^{l}_a = \underline{\varphi}^{l}_a\,\underline{\lambda}_a \ , , \\
\label{eq : bb - b } \mbox{\it tessera}\,\,b & : & \nonumber \\ \hat{f}^{l}_b\,\underline{\varphi}^{l}_b & = & \left [ \hat{f } - \hat{\rho}\,\hat{f}\,\hat{\rho } + \hat{\rho}\,\hat{l}_b\,\hat{\rho } \right ] \,\underline{\varphi}^{l}_b = \underline{\varphi}^{l}_b\,\underline{\lambda}_b \ , , \\
\nonumber & \ldots & \hfill \hspace{1mm}\end{aligned}\ ] ] ( @xmath36 , @xmath37 , , being diagonal matrices of size @xmath38 , @xmath39 , , ) under the conditions of eqs .
[ eq : rhoa]-[eq : rhocond - c ] .
these conditions are fulfilled if all the orbitals are eigenfunctions of the same hermitean operator . in other words ,
all the orbitals @xmath40 for @xmath41 must be eigenfunctions of @xmath42 , all the orbitals @xmath43 for @xmath44 must be eigenfunctions of @xmath45 , and so on .
this means that the subsystem localization operators @xmath46 , @xmath47 , of eqs .
[ eq : bb - a ] , [ eq : bb - b ] ,
must be such that the effective subsystem hartree - fock or kohn - sham hamiltonians @xmath42
, @xmath45 , , have the same eigenfunctions as @xmath20 in eq . [ eq : loceq ] but different eigenvalues .
( note that the virtual orbitals @xmath9 are also eigenfunctions of eqs .
[ eq : bb - a ] , [ eq : bb - b ] , since @xmath48 ; their calculation has not been stated as a goal here and they will not be referred to in the rest of the section . ) as commented above , the localization operator @xmath18 in eq .
[ eq : loceq ] corresponding to a localization method @xmath13 is usually not known .
however , every localization method @xmath13 has a well defined procedure for the computation of the unitary transformation matrix @xmath12 ( eq .
[ eq : canul ] ) in a given molecule .
@xcite this procedure can also be applied to any orthogonal basis of the occupied space @xmath49 other than the canonical orbital basis @xmath3 ; the result is then the unitary matrix @xmath50 that transforms the initial non - canonical set of occupied orbitals onto the @xmath13-method localized orbitals , @xmath51 at this point , we can express the operator @xmath18 of eq .
[ eq : loceq ] as its spectral representation in any basis of the occupied space , e.g. @xmath49 , @xmath52 now , consistently with the discussion following eqs .
[ eq : bb - a ] , [ eq : bb - b ] , , we can define the following hermitean subsystem or localization operators : @xmath53 where @xmath54 , @xmath55 ,
are diagonal matrices of size @xmath6 with arbitrary diagonal real data . a choice consisting on low values for the @xmath27 diagonal elements of @xmath54 corresponding to the localized orbitals of @xmath56 , @xmath43 , and sufficiently higher values for the @xmath57 remaining diagonal elements , guarantees that the localized orbitals of @xmath56 are computed as the lowest @xmath27 eigenvectors of the effective hartree - fock or kohn - sham hamiltonian @xmath42 ( eq . [ eq : bb - a ] ) , which is a convenient choice for safe and efficient orbital selection in the iterations of self - consistent procedures .
obviously , the same comments stand for all .
we may remark that a choice consisting on negative values for the @xmath27 , @xmath30 , cited diagonal elements and zero values for the @xmath58 , @xmath59 ,
remaining elements , although not necessary , is efficient and simplifies the expression of the subsystem localization operators , @xmath60 where @xmath61 , of size @xmath38 , is the diagonal eigenvalue matrix of eq .
[ eq : bb - a ] , and @xmath62 is a rectangular matrix made of @xmath27 columns of @xmath50 , and similarly for all .
using eqs .
[ eq : la ] , [ eq : lb ] , in [ eq : bb - a ] , [ eq : bb - b ] , , plus the fact that @xmath14 is the projection operator of the occupied space ( eq . [ eq : rhoisproj ] ) , results in the working equations of the method for a localization method of choice @xmath13 : @xmath63 \,\underline{\varphi}^{l}_a \nonumber \\ & = & \underline{\varphi}^{l}_a\,\underline{\lambda}_a \ , , \\ \label{eq : mosaico - b } \mbox{\it tessera}\,\,b & : & \nonumber \\ \hat{f}^{l}_b\,\underline{\varphi}^{l}_b & = & \left [ \hat{f } - \hat{\rho}\,\hat{f}\,\hat{\rho } + \underline{\varphi}^{(0 ) } \,_{0}\underline{u}^{l}\ , \,\underline{\lambda}_{b(n)}\ , \,\,_{0}\underline{u}^{l \dagger}\ , \underline{\varphi}^{(0 ) \dagger } \right ] \,\underline{\varphi}^{l}_b \nonumber \\ & = & \underline{\varphi}^{l}_b\,\underline{\lambda}_b \ , , \\
\nonumber & \ldots & \end{aligned}\ ] ] eq .
[ eq : mosaico - a ] is the mosaico equation for the embedded @xmath56 , eq .
[ eq : mosaico - b ] for the embedded @xmath64 , and so on .
they are pseudoeigenvalue equations ( even in the case of semiempirical methods with density - independent @xmath2 hamiltonians . )
they can be solved with standard scf iterative procedures . starting with an initial guess , @xmath49 , the procedure of the localization method of choice , @xmath13 ,
is applied in order to compute @xmath50 , which , together with @xmath14 and @xmath2 , give the embedded hamiltonians @xmath42 , @xmath45 , for the current iteration .
( note that the elements of the diagonal matrices @xmath54 , @xmath55 , , are input real numbers . ) solving the eigenvalue equations leads to new orbitals which are used to update @xmath49 and iterate . at convergence ,
@xmath65 and @xmath50 is the unit matrix . also
, the eigenvalues of the are identical to the input values of @xmath54 , @xmath55 , .
for instance , the @xmath27 non - zero values of @xmath61 coincide with the non - zero values of @xmath54 associated with the localized orbitals of a. several options are open for iterative procedures leading to the solutions of the mosaico equations [ eq : mosaico - a ] , [ eq : mosaico - b ] ,
they all have to face two basic types of iterations : ( 1 ) the microiterations , or iterations , which are standard scf iterations addressed to solve one of the embedded equations , e.g. eq . [ eq : mosaico - a ] , for fixed orbitals of the other .
all the methods available for speeding the solution of the canonical hartree - fock or kohn - sham equations can be used here .
( 2 ) the macroiterations , or mosaic iterations , which involve repeated solutions of all the embedded equations .
the macroiterations can be performed sequentially on a given list of , e.g. @xmath56 , @xmath64 , , @xmath56 , @xmath64 , , meaning that the orbitals computed for @xmath56 are used in the calculation of @xmath64 , and so on .
most interestingly , they can be performed in parallel , meaning that the calculations of all @xmath56 , @xmath64 ,
are done using the whole mosaic of orbitals from the previous macroiteration .
this alternative is very important , because it allows to take full advantage of parallelism at a high level of the calculation . in other words ,
the full calculation is made of a sequence of macroiterations , each of them consisting of a parallel set of hartree - fock , kohn - sham , or semiempirical calculations on the embedded @xmath56 , @xmath64 , .
each of these calculations , which can be time consuming , can be performed in a separate processor or computer .
alternataively , the macroiterations can be done _
la carte _ , that is restricted to only one or several selected .
this is to say that only some of the localized orbitals are optimized whereas the rest of them are frozen .
this is the embedded cluster approximation .
its reliability rests upon the transferability of the localized orbitals .
embedded cluster calculations are significantly cheaper than the full calculation of a complete system .
they can be performed on a molecule or solid if a calculation on a similar molecule or solid has been carried out before in order to provide the orbitals of the embedding frozen .
they are specially useful to study defects in solids , chemisorption , series of molecules with different substituents , or reactions taking place in local regions of large molecules . the equations ( eqs .
[ eq : mosaico - a ] , [ eq : mosaico - b ] , ) , which give the localized orbitals of a given localization method @xmath13 and , in consequence , the same total energy and electron density than the canonical calculation , are the basis for approximations that make them useful in practice .
these approximations , which resort to truncations supported by the localized nature of the orbitals , are systematic and converge to the exact result .
they are basically twofold : ( 1 ) an orbital - specific basis set approximation can be adopted , where the localized orbitals of a are expanded in a different basis set than the localized orbitals of another .
this can be done because the @xmath27 occupied localized orbitals of @xmath56 are the only orbitals to be computed by solving equation [ eq : mosaico - a ] , whereas the @xmath58 orbitals of the other are discarded ; consequently , a local basis set can be used as long as it is sufficient for a good representation of the localized orbitals @xmath43 .
obviously , this is true for any .
( 2 ) a localization algorithm based on local rotations can be used in each interation instead of the standard algorithm of the localization method of choice , @xmath13 . in the local rotations algorithm ,
the localized orbitals of @xmath56 are computed in each iteration using a subset of the @xmath49 orbitals which is localized in not too distant from @xmath56 .
this can be done because the target localized orbitals are computed out of input localized orbitals rather than out of canonical orbitals . in a standard molecular calculation ,
all the orbitals of a molecule are expanded in a common basis set , which is the basis set of the molecule and consists of @xmath66 basis set functions , @xmath67 which could be contracted gaussian functions or other sort of local functions . in the orbital - specific basis
set approximation , osbs , the localized orbitals of @xmath56 are expanded in a local basis set made of @xmath68 functions , the localized orbitals of @xmath64 in a local basis set made of @xmath69 functions , and so on .
although not necessary , it is very convenient that the local basis sets are subsets of the global basis set , and that several share a number of basis set functions .
the local basis sets can be represented by the following row vectors , @xmath70 accordingly , eqs .
[ eq : mosaico - a ] , [ eq : mosaico - b ] , , take the usual matrix form , @xcite @xmath71 in eq .
[ eq : a ] , for instance , @xmath72 and @xmath73 are the @xmath74 matrix representations of the @xmath75 and the unit operators in the @xmath76 basis set , @xmath77 and @xmath78 is the @xmath79 matrix of localized orbital coefficients , @xmath80 the solution of eq .
[ eq : a ] can be attained using standard ( @xmath81 ) diagonalization procedures . except for low gap materials , where the degree of localization attainable
is limited and the localized orbitals may decay slowly with distance , @xcite the local basis set size @xmath68 is expected to remain within reasonable limits for a diagonalization or , more generally , for an embedded cluster calculation .
obviously , the growth of @xmath68 will impose more practical limitations in 3d systems , like bulk solids and very big clusters , than in 2d and 1d systems , like many molecules .
in general , all the methods useful to speed up a standard molecular calculation , like convergence acceleration methods , can be used with eq .
[ eq : a ] .
but , in this case , additional advantage can be taken from the fact that only a small number of eigenvalues / eigenfunctions are computed for each , and efficient diagonalization algorithms used in ci calculations , like the multiroot davidson - liu method , @xcite can be applied to significantly reduce the prefactor of the ( @xmath81 ) dependence .
let us comment on the calculation of the effective hamiltonian matrix of a , @xmath72 , where the localized nature of the orbitals is also profitable . for simplicity , we will call @xmath82 the row vector of the @xmath0 current localized orbitals ( @xmath83 , eq .
[ eq:0ul ] ) at a given iteration , which can be written as the union of the row vectors @xmath84 , @xmath85 , , that contain the @xmath27 , @xmath30 , current localized orbitals of @xmath56 , @xmath64 , respectively , @xmath86 the @xmath74 effective hamiltonian matrix of @xmath56 is @xmath87 the first term in the right hand side of eq .
[ eq : feffa ] is a diagonal block of the hamiltonian matrix of the full system .
the second term can be expanded in inter- terms by inserting eq .
[ eq : rhocond - b ] : @xmath88 \,.\end{aligned}\ ] ] in the evaluation of eq .
[ eq : rhohrho ] , we can take advantage of the locality of basis sets and molecular orbitals by evaluating a @xmath89 interaction table and a @xmath90 interaction table .
@xmath89 is set to @xmath91 if all the elements of the @xmath92 matrix @xmath93 and all the elements of the @xmath94 matrix @xmath95 have an absolute value lower than a given threshold , and is set to @xmath96 otherwise .
similarly , @xmath90 is set to @xmath91 if all the elements of the @xmath97 matrix @xmath98 have an absolute value lower than a given threshold , and is set to @xmath96 otherwise . rearranging eq .
[ eq : rhohrho ] and using these interaction tables , we can write @xmath99 \,.\end{aligned}\ ] ] the use of the interaction tables @xmath89 and @xmath90 guarantees that the calculation of the present term does not scale as @xmath100 because the number of in the neighborhood of ( or interacting with ) @xmath56 does not increase indefinitely with the size of the molecule .
also , note that the interaction tables do not need to be updated every macroiteration , because the localized orbitals do not experience big changes in size after a few iterations .
it should be noticed that we assume that a linear - scaling method is used for the computation of the hamiltonian matrix of the full system , @xcite of which diagonal and nondiagonal blocks are needed in eqs .
[ eq : feffa ] and [ eq : rhohrho - ok ] .
the last term in eq .
[ eq : feffa ] can be written as @xmath101 where the @xmath89 interaction table is used and @xmath102 is the @xmath39 diagonal matrix resulting from the elements of @xmath54 corresponding to the localized orbitals of @xmath64 . for the particular choice of the arbitrary diagonal matrix @xmath54 leading to eq .
[ eq : la2 ] , eq .
[ eq : thirdterm ] is further simplified : @xmath103 where all @xmath104 must be negative . the computation of the @xmath72 matrices ( eqs .
[ eq : feffa ] , [ eq : rhohrho - ok ] , and [ eq : thirdterm ] ) and their diagonalization ( eq . [ eq : a ] ) can be performed in parallel .
this is graphically indicated in the upper part of fig .
[ fig : loops ] . the global scaling of this diagonalization step
is @xmath105 ; in the case of all having the same local basis set size , @xmath106 , and same number of interactions with other , the scaling is @xmath107 . + usually , a given set of localized orbitals of a molecule , @xmath83 , is computed out of the canonical orbitals ( eq . [ eq : canul ] ) .
well defined algorithms are routinely used to calculate the @xmath6 unitary transformation matrix @xmath12 , which are normally of order @xmath108 or higher .
@xcite in the present method , however , they are computed in each macroiteration out of other set of localized orbitals ( eq . [ eq:0ul ] ) .
we can expect the contributions of the initial localized orbitals @xmath49 to the target localized orbitals @xmath83 to decay with distance .
accordingly , a @xmath109 local rotation table can be computed , where @xmath110 if the initial localized orbitals of @xmath64 are used to compute the target localized orbitals of @xmath56 ( and viceversa ) and @xmath111 otherwise .
several options to compute the @xmath109 local rotation table are possible .
reasonable choices are to use the same criterium as for the @xmath112 interaction table except for the use of a different threshold or , simply , make the @xmath109 local rotation table identical to the @xmath112 interaction table . arranging the initial localized orbitals in , @xmath113 the approximation of local rotations for the localization step can be written as @xmath114 in eq .
[ eq : locrot - a ] , @xmath115 is the @xmath116 block of @xmath117 with the contributions of the initial localized orbitals of @xmath118 to the target localized orbitals of @xmath56 . the column block made of all @xmath115 with @xmath119 is represented by the fill rectangular box shown in fig .
[ fig : lrot ] . this column block can be calculated approximately as the column block of the submatrix of @xmath120 indicated in fig .
[ fig : lrot ] by the open square box .
we can call this submatrix @xmath121 , so that if @xmath122 is a row vector with the initial localized orbitals of all @xmath118 having @xmath123 ( of length ) , then @xmath124 @xmath121 can be computed by simple application of the localization method of choice to the set of @xmath125 initial localized orbitals @xmath122 .
note that only the @xmath27 orbitals of @xmath126 corresponding to @xmath56 ( @xmath127 ) are taken from eq .
[ eq : ulra ] . lowering the threshold of the local rotation table
improves the precision of this approximation systematically .
+ + the computation of the @xmath121 matrices ( eqs .
[ eq : ulra ] ) can be performed in parallel .
this is graphically indicated in the lower part of fig .
[ fig : loops ] .
the global scaling of this localization step is @xmath128 , where @xmath129 is the order of the localization method of choice @xmath13 ( e.g. 3 in pipek - mezey and 5 in edmiston - ruedenberg methods ) ; if the same number of orbitals are used in the local rotations of all , @xmath130 , the scaling is @xmath131 .
since the localization by local rotations among initial localized orbitals goes together with the use of orbital - specific basis sets , the resulting localized orbitals of @xmath56 must be represented with the basis set @xmath76 alone .
this can be achieved by truncation of @xmath26 after the transformation ( eq .
[ eq : locrot - a ] ) or , alternatively , by its projection on the @xmath76 space , @xmath132 .
we have not observed any practical advantage in projecting instead of truncating , neither in the precision attainable in the total energy nor in the convergence , in all the tests performed .
the orbitals resulting from the present method do not belong to irreducible representations of the molecular point symmetry group because they are not eigenfunctions of the totally symmetric one - electron hamiltonian @xmath2 . however , except for symmetry breaking localization methods , they are related by the symmetry operations of the molecule and this fact can be used to reduce computing time .
@xcite in effect , all orbitals can be obtained by applying molecular symmetry operations , @xmath133 , to a list of symmetry independent , so that if @xmath56 is a symmetry independent and @xmath64 is obtained from @xmath56 by symmetry operation @xmath133 , then @xmath134 . in other words , the diagonalizations and localizations
described in sections [ sec : osbs ] and [ sec : lr ] can be performed only on the list of symmetry independent .
the computation of @xmath72 ( eq . [ eq : feffa ] ) can also profit from this property .
also , the site symmetry of the can be used in exactly the same way that molecular symmetry is used in standard molecular calculations , because the matrix of the effective hamiltonian of any embedded is blocked according to the irreducible representations of the local point symmetry group of that particular . besides the use of symmetry
, one can also take advantage of quasisymmetry for speeding up purposes .
so , if two @xmath56 and @xmath64 are quasisymmetry related by an operation @xmath135 ( @xmath136 ) , they can be treated as symmetry related for a number of macroiterations , in which the list of where the diagonalization and localization steps are performed can be shortened , finally releasing all quasisymmetry restrictions up to a full convergence of the calculation . in summary , the algorithm for a calculation of a molecule is the following : 1 .
[ mosaico - step-1 ] take the current guess of localized orbitals of the molecule and do , in parallel , for each symmetry independent : 1 . compute the embedded effective hamiltonian matrix ( eqs .
[ eq : feffa ] , [ eq : rhohrho - ok ] , and [ eq : thirdterm ] ) , 2 .
diagonalize it ( eq .
[ eq : a ] ) and compute new orbitals ( eq . [ eq : neworb ] ) , 2 .
[ mosaico - step-2 ] take all the orbitals produced in step [ mosaico - step-1 ] and do , in parallel , for each symmetry independent : 1 .
compute the @xmath121 unitary matrix ( eq . [ eq : ulra ] ) corresponding to the localization method of choice @xmath13 by applying the corresponding localization algorithm , and take the columns that correspond to the current , 2 .
compute the target localized orbitals ( eq . [ eq : locrot - a ] ) , 3 .
check for convergence and iterate on step [ mosaico - step-1 ] if necessary .
compute properties upon convergence .
[ eq : project ] the parallel steps [ mosaico - step-1 ] and [ mosaico - step-2 ] are schematically represented in fig .
[ fig : loops ] .
in this section , we present the results of monitoring calculations on poly(ethylene oxide ) molecules , , ( sec .
[ sec : peo ] ) and three dimensional co clusters , ( co)@xmath137 , ( sec .
[ sec : co ] ) aimed at showing the convergence of the parallel calculations to the right solutions , the convergence of the total energies with the size of the orbital - specific basis sets towards the exact values , and the linear - scaling of the method .
we also include embedded cluster calculations on defective systems resulting from a chemical substitution of one o atom by a s atom in poly(ethylene oxide ) , ( sec .
[ sec : ec ] ) aimed at showing the performance of embedded cluster calculations vs. full system calculations .
all the calculations are single point energy calculations with an extended hckel ( eh ) hamiltonian .
@xcite this is a convenient choice to monitor the procedure because , on the one hand , the hamiltonian and overlap matrices of the eh method are isomorphous with their counterparts , and on the other , the hamiltonian is not self - consistent and its computation is straightforward . in this way , the analysis of timing and scaling is focused on the orbital optimization or diagonalization part and free from contaminations due to the computation of the hamiltonian matrix .
the calculations have been performed with a 2 gb ram personal computer with the program mosaico ; @xcite the extended hckel hamiltonian and overlap matrices have been calculated with the program eht .
@xcite although the loops performed mimic parallel loops ( fig .
[ fig : loops ] ) , the total elapsed times shown hereafter correspond to sequential loops performed in a single processor . in the present version of the program
we have paid special attention to the scaling features , whereas the prefactors are highly improvable .
this fact , together with the average performance of the personal computer used , makes the absolute values of the elapsed times shown in this section of little value ; instead , it is the scaling of the method what is relevant .
poly(ethylene oxide ) ( peo ) @xcite is a polymer widely used in the field of polymer electrolytes of molecular formula .
the molecular calculations we present for different numbers of monomers , @xmath138 , use the localization method of projected localized molecular orbitals ( plmo ) @xcite ( see appendix ) using a very simple set of reference orbitals : @xmath139 for all a - b pairs of bonded atoms , plus @xmath140 and @xmath141 for all oxygen atoms , where @xmath142 and @xmath143 are local cartesian axis on the oxygen assuming the c - o - c atoms define a @xmath144 plane with the @xmath142 axis bisecting the c - o - c angle . in all these calculation we used the following definition for the or subsystems : one made of 10 orbitals localized in the sigma bonds and lone pairs of ch@xmath145och@xmath146 [ which , in the localization method used for these particular calculations , are the 10 orbitals with one - to - one maximum overlap with the reference orbitals @xmath147 , @xmath148 , @xmath149 , @xmath150 , @xmath151 , @xmath152 , @xmath153 , @xmath154 , @xmath155 , @xmath156 , plus @xmath157 made of 9 orbitals localized in the sigma bonds and lone pairs of the next ch@xmath146och@xmath146 groups , and a final of 9 orbitals localized in the sigma bonds and lone pairs of the terminal ch@xmath146och@xmath145 .
we performed the calculations using three different orbital - specific basis sets : in the calculations labeled 1n , the orbitals of each have been represented with a subset of the global basis set consisting of all the basis set functions of the atoms involved in the bonds and lone pairs of the plus those of the atoms involved in bonds and lone pairs of the first neighbor or monomers . in the 2n and 3n calculations , the orbital - specific basis sets were extended to second and third neighbor monomers .
all the calculations converge to the same results regardless of the initial guess and the iteration procedure ( parallel or any kind of sequential choice ) .
the 9 localized orbitals which constitute one of the the bulk of h(ch@xmath146och@xmath146)@xmath158h are shown in fig .
[ fig : peo - tess9 ] . + fig .
[ fig : peo - etot ] shows the total energy of h(ch@xmath146och@xmath146)@xmath158h as a function of the orbital - specific basis sets used , which converges to the exact energy in the limit of a standard calculation where all orbitals are spanned in a common basis set . as the size of the osbs increases , the lack of full orthogonality between orbitals of differente originated by the basis set truncation becomes negligible and , accordingly , the difference between the total energy properly computed and the total energy computed under the assumption that the orbitals are fully orthogonal vanishes .
the energy loss per monomer due to the use of orbital - specific basis sets instead of a common basis set for all orbitals , is presented in table [ tab : one ] for several orbital - specific basis sets as a function of the molecular size .
the predictability of the energy losses is apparent .
[ fig : peo - evsiter ] shows the convergence with macroiterations of the total energy of the h(ch@xmath146och@xmath146)@xmath158h polymer at the 2n level of orbital - specific basis set . + ccrrr + & osbs & + + & & & + + & 1n & @xmath159 & @xmath160 & @xmath161 + & 2n & @xmath162 & @xmath163 & @xmath164 + & 3n & @xmath165 & @xmath165 & @xmath166 + + & & & + + & 1n & @xmath167 & @xmath168 + & 2n & @xmath169 & @xmath170 + & 3n & @xmath171 & @xmath172 + fig .
[ fig : peo - time ] shows the wall clock elapsed time per macroiteration in the calculation of the h(ch@xmath146och@xmath146)@xmath173h molecules as a function of the number of monomers @xmath138 ( which in this case coincides with the number of , @xmath24 ) , of atoms , and of basis set functions .
the times per macroiteration and spent in the diagonalization step and in the localization step are shown in fig .
[ fig : peo - tesseratime ] .
the horizontal lines reflect the ( @xmath24 ) scalings of both steps .
the ( @xmath174 ) scaling of the diagonalization step is clearly reflected in the separation between the 1n , 2n , and 3n horizontal lines .
the drop of the lines at low number of monomers is due to surface effects : the ratio between edge and bulk is significant in small polymers and , since edge are less demanding in terms of orbital - specific basis set and in terms of number of inter- interactions , they reduce the computing time with respect to what it would be if all of them were bulk . as we will see below , this fortunate surface effect , which lowers time with respect to a set of @xmath24 bulk ,
is much more pronounced in 3d systems .
the localization times shown in the bottom part of fig .
[ fig : peo - tesseratime ] are very similar in the 1n , 2n and 3n calculations .
this is so because they depend basically on the number of orbitals used in the local rotations , which is the same in the three calculations . for the particular localization method we used for these calculation , plmo ,
the elapsed times scale as @xmath175(@xmath176 ) @xmath177 ( @xmath178)(@xmath24 ) .
the small dependence with the size of the orbital - specific basis sets is related with the lengths of the matrix transformations in eq .
[ eq : plolr ] .
+ + + we performed calculations on three dimensional ( co)@xmath173 clusters of several sizes , extracted from the crystalline structure of the @xmath7 phase of solid carbon monoxide ( @xmath179 spatial group , co bond length @xmath180(c - o ) = 1.128 , cell constant @xmath181 = 5.64 ) , @xcite which is an interesting material known to experience irreversible photopolymerization under pressure .
@xcite in this crystal , a bulk co molecule has a first - neigbhor coordination number 12 ( co molecules at a distance between centers of gravity of 3.99 ) , and second- and third - neigbhor coordination numbers 18 and 42 ( co molecules at 5.64 and 6.91 , respectively ) . the ( co)@xmath182 cluster is represented in fig .
[ fig : co63 ] as an example .
+ + in these calculations we also used the plmo localization method , with the reference orbitals defined as the @xmath183 valence occupied canonical orbitals of the @xmath138 isolated co molecules .
we defined each to be made of 5 orbitals localized in the spatial region of a co molecule , which , for the chosen localization method , means the 5 localized orbitals with maximum overlap with the canonical orbitals of the co molecule .
the orbital - specific basis sets used have been labeled 1n , 2n , and 3n when the basis set of a consists of the basis set functions of its c and o atoms plus the basis set functions of the atoms of the first , second , and third - neighbor co molecules , respectively ( see above ) .
the energy losses per co molecule ( table [ tab : one ] ) are small and diminish as the size of the osbs increases .
times per macroiteration and spent in the diagonalization step and in the localization step are shown in fig .
[ fig : co - tesseratime ] .
the diagonalization times spent in a inner or bulk , which is the most demaning , are included as a reference .
this time is constant for all clusters where the inner is a truly bulk , that is , where its osbs and its interaction tables die off before the limits of the cluster ; this situation is reached in a smaller cluster for the 1n osbs and in a bigger cluster for the 3n osbs .
naturally , the times spent per in the full system calculations are lower than the times spent in the bulk , the differences showing the importance of surface effects : smaller clusters and bigger orbital - specific basis sets have a larger ratio of surface / bulk and , accordingly , a larger time reduction with respect to the bulk . in a large @xmath138 regime ,
the bulk are dominant and set the assymptotic limit of the full system times , which scale linearly with the cluster size .
the ( @xmath184 ) dependence is shown by the asymptotic values of the 1n , 2n , and 3n lines , as well as by the values of @xmath185 printed along the 3n line in fig .
[ fig : co - tesseratime ] .
the localization times ( bottom of fig .
[ fig : co - tesseratime ] ) are very much independent of the size of the osbs because they are directly dependent in the number of occupied orbitals included in the local rotations , which is the same in 1n , 2n , and 3n calculations ; the small dependence shown in the figure is due to the fact that the transformations down to the basis set level depend on the osbs size .
+ the method can be used for embedded cluster calculations , where the computational effort is focused on an active site of a molecule , comprising only a number of relevant , while the rest of it is taken from a previous calculation on a similar molecule and frozen . in this section
we show the results of embedded cluster calculations on h(ch@xmath146och@xmath146)@xmath186-ch@xmath146sch@xmath146-(ch@xmath146och@xmath146)@xmath186h .
h(ch@xmath146och@xmath146)@xmath186-ch@xmath146sch@xmath146-(ch@xmath146och@xmath146)@xmath186h can be regarded as the result of creating a chemical defect in h(ch@xmath146och@xmath146)@xmath173h , with @xmath187 , by substitution of the central oxygen by a sulfur atom .
we may expect the localized orbitals distant from the s atom in the defective " molecule to be very similar to the orbitals of the perfect " molecule localized in the same spatial regions and , accordingly , we may take them from a previous calculation on h(ch@xmath146och@xmath146)@xmath173h and use them in a calculation of h(ch@xmath146och@xmath146)@xmath186-ch@xmath146sch@xmath146-(ch@xmath146och@xmath146)@xmath186h where they are kept frozen ; this defines an embedded cluster calculation .
canonical and calculations on the @xmath188 polymer ( using the same nuclear configurations before and after the creation of the s defect ) reveal that the precision reached with a 3n orbital - specific basis set on the perfect polymer requires a better osbs after creating this chemical defect : a 5n basis set for the central defective and its 5 neighbor together with a 3n basis set for the remaining , gives an energy error of @xmath189 hartree / monomer . taking this into account , we performed embedded cluster calculations using a 5n osbs for the variational and taking the orbitals that remain frozen for the rest of in the polymer from the 3n osbs calculation on the perfect " polymer molecule h(ch@xmath146och@xmath146)@xmath173h .
the total energy errors of these embedded cluster calculations ( with respect to the full molecular calculations ) on the polymers of 21 and 201 monomers are shown in fig .
[ fig : peo - s - ec ] as a function of the active cluster size ( number of active ) .
+ + it is shown that the errors are too large to be acceptable if only the central defective is active , whereas they drop to an acceptable value of @xmath190 hartree when the orbitals of the first - neigbhor are optimized as well , and to less than @xmath191 hartree when second - neigbhor are part of the variational embedded cluster .
these errors are the same in the two @xmath188 and @xmath192 polymers , as corresponds to the local nature of the chemical defect .
the times spent in the embedded cluster calculations , as a fraction of the times spent in the respective full system calculations , are shwon at the bottom of fig .
[ fig : peo - s - ec ] .
overall , this figure illustrates the potentiallity of embedded cluster calculations where the transferability of the localized orbitals of a localization method of choice is exploited .
we presented a new linear - scaling method for the energy minimization step of semiempirical and first - principles hartree - fock and kohn - sham calculations , which we abbreviated under the name . in this method , a set of embedded pseudoeigenvalue coupled equations is solved in a building - block self - consistent fashion , which results in optimum occupied localized orbitals of any localization method of choice , represented with orbital - specific basis sets .
the method is parallel at a high level of the calculation .
it can be used in full system calculations as well as in embedded cluster calculations , where only an active fraction of the localized molecular orbitals of the whole system are variational while the rest are taken from a similar molecule and kept frozen .
we presented the results of monitoring , single point energy calculations with the extended hckel hamiltonian on poly(ethylene oxide ) molecules and three dimensional carbon monoxide clusters with very large number of basis set functions .
total energy losses due to the use of orbital - specific basis sets are small for reasonably small sizes of these and total energies converge to the canonical values when the orbital - specific basis sets are increased towards the limit of a common basis set for all the localized orbitals
. convergence of total energy with self - consistent macroiterations is good and elapsed times per macroiterations have been shown to scale linearly with the molecular size . besides the full system calculations , the good performance of the much less demanding embedded cluster approach has been shown in total energy calculations on defective systems resulting from chemical substitution of an oxygen atom by a sulfur atom in poly(ethylene oxide ) molecules .
the transferability of the localized orbitals of a given localization method between similar molecules has been shown to lead to the same total energy precision than full molecular calculations at a fraction of the computational cost .
this research was supported in part by direccin general de investigacin , ministerio de ciencia y tecnologa , spain , under contract bqu2002 - 01316 , and ministerio de educacin , cultura y deportes , spain , accciones de movilidad pr2003 - 0024 and pr2003 - 0027 .
we are very grateful to professor martin head - gordon , university of california , berkeley , and to his research group for their hospitality .
a simple localization method has been proposed by ruedenberg , @xcite which has not reached the popularity of other methods like boys , @xcite edmiston - ruedenberg , @xcite and pipek - mezey @xcite methods . although it has been formulated in a context of atoms - in - molecules , it is a general method . here ,
we reformulate it in the general case and in the case of localization by local rotations ( sec . [ sec : lr ] ) , using the present notation .
given a set of @xmath0 input occupied ( canonical or localized ) orthogonal orbitals of a molecule , @xmath193 , and a set of @xmath0 reference arbitrary orbitals , @xmath194 , the @xmath0 projected localized orthogonal orbitals , @xmath195 , are defined as those resulting from a unitary transformation of @xmath49 which are most similar to @xmath196 .
the plmos correspond to maximizing the one - to - one overlaps with the reference orbitals @xmath196 , or the functional @xmath197 , under orthogonality constraints @xcite and can be computed as @xcite @xmath198^{-\frac{1}{2 } } = \underline{\varphi}^{(0)}\ , _ 0\underline{u}^{plmo } \,,\end{aligned}\ ] ] with @xmath199^{-\frac{1}{2 } } \,.\end{aligned}\ ] ] in the approximation of local rotations ( section [ sec : lr ] ) , the projected localized orbitals of a @xmath56 are computed out of @xmath125 ( @xmath200 ) input localized orbitals @xmath122 and @xmath125 reference orbitals @xmath201 , both of them including the localized / reference orbitals belonging to the included in the local rotations , with the expression : @xmath202 being @xmath203^{-\frac{1}{2 } } \big\}_{n_a-\mbox{col } } \,,\end{aligned}\ ] ] where it has been indicated that only the @xmath27 columns that correspond to @xmath56 are computed and used . among the advantages of the plmo localization method are its speed and its simplicity , because the usual iterative optimization procedures involved in localization @xcite are substituted by a one - step calculation of the reciprocal square root of a symmetrical matrix , which is an ( @xmath108 ) process ( or ( @xmath204 ) in local rotations ) .
its main disadvantage is the requirement of an external , arbitrary set of reference orbitals , @xmath196 .
although this is a limitation in calculations of reactivity , it is not a practical problem in molecular structure calculations where the nature of the bonds is known in advance and finding good reference bond and lone pair orbitals is not difficult .
we may remark that the application of the plmo method with a reference consisting of a given set of localized orbitals , e.g. edmiston - ruedenberg s , leads exactly to that set of orbitals .
this property can be exploited in many ways and , in particular , in order to remove the arbitrariness inherent to the plmo method .
for instance , it can be used to produce reference orbitals for the plmo method ( to be used in large molecules ) out of edmiston - ruedenberg s or other non - arbitrary localized sets computed in selected sets of relatively small molecules .
also , a calculation addressed to produce a given set of localized orbitals like edmiston - ruedenberg s can be safely performed using the chosen localization method in some macroiterations and the faster plmo method , with the current er orbitals as a reference set , in the rest of them . | we present a new linear scaling method for the energy minimization step of semiempirical and first - principles hartree - fock and kohn - sham calculations .
it is based on the self - consistent calculation of the optimum localized orbitals of any localization method of choice and on the use of orbital - specific basis sets .
the full set of localized orbitals of a large molecule is seen as an orbital mosaic where each is made of only a few of them .
the orbital are computed out of a set of embedded cluster pseudoeigenvalue coupled equations which are solved in a building - block self - consistent fashion . in each iteration ,
the embedded cluster equations are solved independently of each other and , as a result , the method is parallel at a high level of the calculation .
in addition to full system calculations , the method enables to perform simpler , much less demanding embedded cluster calculations , where only a fraction of the localized molecular orbitals are variational while the rest are frozen , taking advantage of the transferability of the localized orbitals of a given localization method between similar molecules . monitoring single point energy calculations of large poly(ethylene oxide ) molecules and three dimensional carbon monoxide clusters using an extended hckel hamiltonian
are presented . |
observing from mcdonald observatory at a maximum altitude of 14@xmath2 , @xcite reported the b2 star hd124448 to `` show no hydrogen lines , either in absorption or in emission , although the helium lines are sharp and strong ... the abundance of hydrogen appears to be very low in the atmosphere of this star '' .
indeed , hd124448 turned out to be the first of around 20 b- and early a- supergiants , with apparent magnitude @xmath3 , in which hydrogen comprises less than 1 part per thousand of the atmosphere . in terms of kinematics , metallicity and galactic distribution
, they have the properties of the galactic bulge @xcite . with luminosities
approximately ten thousand times solar , the number count is complete for the observable parts of the galaxy ( fainter stars would lie beyond ) .
in addition to their low hydrogen abundances , the true `` extreme helium stars '' ( ehes ) show atmospheres which are enriched in nitrogen by a factor ten , carbon by between one and three parts per hundred , and sometimes oxygen by a similar amount .
the combination of extreme surface composition and extreme rarity makes these stars interesting , and presents a challenge for the theory of stellar evolution .
the general properties of ehes have prompted suggestions of a connection to the cooler rcoronae borealis stars @xcite , which are better known for their spectacular and unpredictable light variability @xcite .
figure [ f : hdef ] shows the distribution by surface gravity and effective temperature of several classes of hydrogen - deficient star ; stars falling on an imaginary line parallel to the eddington limit would have the same luminosity - to - mass ratio , corresponding to the evolutionary path of a giant contracting to become a white dwarf .
links between non - variable hydrogen - deficient carbon ( hdc ) , rcoronae borealis ( rcb ) , extreme helium ( ehe ) , luminous helium - rich subdwarf o ( hesdo@xmath4 ) and o(he ) stars have all been suggested at one time or another .
diagram for several classes of low - mass hydrogen - deficient stars described in detail by @xcite .
ehe stars are denoted by squares ( light grey , green online ) , as are rcb ( dark grey , red online ) , hdc ( open squares ) , hesdo@xmath4 ( back , blue online ) and o(he ) ( mid - grey , violet online ) stars.,width=453 ] in order to establish the origin of ehes and related objects , temperatures and gravities are required .
additional observables are provided by the surface composition and by the fact that many ehes are pulsating .
pulsations provide an opportunity to make direct measurements of mass and radius .
this paper reviews the spectroscopic data on surface composition , summarizes the principal theories for the evolutionary origin , and discusses the major pulsation properties .
it concludes with a synopsis of the extraordinary pulsating ehe v652herculis .
the depleted hydrogen surfaces of ehes indicate that a mixture of nuclear - processed helium and carbon has been exposed .
abundance analyses have been carried out on the basis of optical and ultraviolet spectroscopy using primarily lte methods @xcite ; some partial non - lte analyses have been executed @xcite .
the majority of these results can be summarised as follows @xcite : * h is a relic of the original star and only appears as a trace element . *
ca , ti , cr , mn and ni scale with fe and are essentially unprocessed .
* is proportional to [ ( c@xmath5+n@xmath5+o@xmath5)/fe ] where primes refer to a primordial value , indicating that the original carbon and oxygen has been converted to nitrogen via the hydrogen - burning cno - cycle .
* @xmath6 , indicating the presence of 3@xmath7 product . * @xmath6 , indicating @xmath8c+@xmath7 and/or @xmath9n+@xmath7 product . * @xmath6 , indicating @xmath9n+@xmath10 product . *
mg , si , s , ... are unremarkable and scale with fe .
* @xmath6 @xcite ; fluorine is a by - product of the @xmath9n+@xmath7 reaction .
* y , zr ( s - process ) overabundances are measured in a few cases @xcite , indicating material from a thermal - pulsing phase on the asymptotic giant branch ( agb ) .
* @xmath6 ; possible phosphorus overabundances are not understood , but may also have an agb origin .
together , these results point to the co - existence of material produced by cno - cycled hydrogen burning , and by 3@xmath7 , @xmath8c+@xmath7 and @xmath9n+@xmath7 helium burning .
there is evidence from rcbs that the large oxygen excesses are due to a short - lived phase of @xmath9n+@xmath11o burning @xcite .
there are also relics of primordial hydrogen and , in some stars , s - process elements from a previous agb phase . a theory for the origin of ehes must be able to explain the co - existence of all of these products on their surfaces .
@xcite reviewed extant proposals for the origin of rcbs and ehes .
several authors have investigated the evolution of single helium stars from the helium main - sequence to become helium - shell burning giants @xcite .
such models mimic the general properties of rcbs and ehes , but generally fail to explain the detailed surface mixtures even if it could be established how the helium main - sequence stars would form in the first place .
@xcite had proposed a model involving `` deep envelope mixing '' on the asymptotic giant branch , which fell short of establishing a mechanism to drive the necessary mixing .
@xcite introduced the notion of `` hot bottom burning '' in which a vigorous nuclear - burning shell is sufficiently luminous to drive convection to the stellar surface .
meanwhile , @xcite and @xcite concluded that ehes `` can be understood as the remnants of red giant stars that have lost their hydrogen - rich envelope during their ascent along the asymptotic giant branch . ''
all three proposals struggle to fully account for the observed surface chemistries .
@xcite proposed that stars which experience a late thermal pulse ( ltp , or helium - shell flash ) after leaving the agb would return to the cool giant regime and there be identified as rcb stars .
there would be mixing between the hydrogen - rich surface layers , the helium - rich intershell , and helium shell - flash products which would become visible at a now hydrogen - deficient surface .
the principles behind the ltp had been previously predicted by @xcite and demonstrated by @xcite .
in fact , two cases had been identified , which were later described more fully by @xcite . in a _
late _ thermal pulse , generally on the constant - luminosity part of the post - agb contraction curve , there is no mixing between h - rich and he - rich matter ( and hence no production of s - process elements ) at the time of the flash ; helium is dredged up as the star cools and surface convection penetrates into the helium - rich layers from above . in
a _ very late _ thermal pulse , generally when the star is on or almost on the white dwarf cooling track , the helium - shell flash is more violent and flash - driven convection mixes helium- and carbon - rich material directly to the surface .
both models have some attraction for the origin of rcb stars but have more weaknesses than strengths @xcite .
they are important in discussing the origin of `` born - again '' stars like fgsge ( late thermal pulse , @xcite ) , v4334sgr ( very - late thermal pulse @xcite ) , and hydrogen - deficient pn central stars and pg1159 stars @xcite . at around the same time as the ltp model emerged , @xcite and @xcite realised that a binary system containing two white dwarfs with a period less than about two hours should lose orbital energy by the radiation of gravitational waves @xcite .
subsequently the white dwarfs would collide in a timescale less than the hubble time . in the case of small mass ratios
, the collision will likely be in the form of stable roche - lobe overflow , identified as a possible origin for the amcvn stars , which are low mass hydrogen - deficient white - dwarf binary systems @xcite . with a mass ratio closer to unity
, the rate of mass loss would be such that the rate of radius increase of the mass - losing white dwarf ( @xmath12 ) leads to run - away mass transfer , the destruction of the less massive white dwarf , and its immediate accretion onto the surface of its more massive companion .
the overall outcome depends on the composition of the white dwarfs and the total mass of the merged product .
for example , a helium white dwarf ( he - wd ) accreted onto a carbon - oxygen white dwarf ( co - wd ) with a total mass less than the chandasekhar mass results in the ignition of shell - helium burning at the helium / carbon boundary , and the expansion of the star to become a supergiant , with a hydrogen - poor surface , as in the rcb stars . a more massive merger could produce a type ia supernova .
this connection beween rcb stars and sn ia , already recognised by @xcite , makes the study of double white - dwarf mergers , and of rcb and ehe stars , of considerable interest @xcite .
since the predictions by @xcite and @xcite , there have been several strands to calculations of white - dwarf mergers .
briefly , these include : * hydrodynamic calculations .
early simulations demonstrated that the donor white dwarf would disrupt on a dynamical timescale and rapidly form a hot corona and a keplerian disk surrounding the accretor @xcite .
only recently have large grids of simulations been completed , and only recently have simulations included a detailed calculation of nucleosynthesis during the merger .
an important question is whether these mergers are `` hot '' or `` cold '' ; i.e. is the initial helium - shell ignition sufficiently hot and of sufficient duration to drive nucleosynthesis ( e.g. @xmath13 ) that would explain the more exotic aspects of ehe and rcb chemistry .
a significant result is that , because the donor has further to fall into the potential well , mergers with large mass ratios ( @xmath14 ) are hotter and therefore more likely to produce @xmath9n+@xmath7 products than mergers with smaller mass ratios . *
post - merger evolution .
@xcite showed that a 0.6 or 0.7@xmath15 co - wd accreting between 0.1 and 0.3@xmath15 of helium at half the eddington rate would ignite said helium and expand to produce a giant with the mass , luminosity , temperature and basic composition of an rcb star .
after shell helium exhaustion , these models contracted with properties corresponding to observations of ehe stars .
the detailed surface chemistries could be explained in an _ ad hoc _ fashion using a simple recipe based on the pre - merger composition of the various hydrogen - rich , helium - rich and carbon - rich layers of the progenitor white dwarfs . with a decade of additional abundance data for both ehe and rcb stars
, @xcite extended this simple recipe ; in particular they suggested that the fluorine , neon and phosphorus overabundances could originate in the agb intershell of the co - wd , with @xmath16 coming from a pocket close to the carbon - helium boundary in the co - wd ( in their models ) .
this , together with near - normal abundances of magnesium , points to progenitor main - sequence stars with initial masses in the range 1.93@xmath15 .
@xcite , on the other hand , argue that most of the excess exotics can be produced during a `` hot '' merger . * population statistics .
the expectation of producing sufficient dwd mergers to populate the galaxy with rcbs and ehes is obtained by comparing the expected numbers of double white dwarfs , with appropriate masses and periods , with the gravitational - radiation orbital - decay times .
the galactic merger rate can then be multiplied by the predicted lifetimes of post - merger stars as rcbs and ehes .
the first qualitative estimates were made by @xcite , and confirmed by @xcite to give a co+he merger rate of @xmath17 @xmath18 . with a contraction ( thermal ) timescale of @xmath19 yr , there should be @xmath20 ehes in the galaxy , consistent with the observed number .
however , @xcite showed that binary - star population synthesis ( bsps ) contains large error margins , so the apparent agreement should be taken with a pinch of salt .
a significant problem still to be resolved is that the bsps calculations show most co+he mergers at the current epoch should come from relatively young systems ( @xmath21 gyr ) , which is inconsistent with the galactic bulge distribution of ehe and rcb stars . *
dynamical stability .
mass ratio plays a key role in whether a double white dwarf binary merges or transfers matter slowly when one component fills its roche lobe @xcite ; the primary factor is the rate of change of semi - major axis compared with the expansion rate of the mass - losing star , which gives a critical ratio @xmath22 @xcite . the question is closely connected with the efficiency of mass transfer and factors such as synchronicity and nuclear reactivity .
it is important , as it strongly affects the population statistics , but appears far from resolved .
* angular momentum .
the coalescence of two stars in a short - period orbit raises the questions of ( a ) whether the accretor can continue to accrete as it spins up and ( b ) whether the total angular momentum of the product would cause it to break up .
assuming angular momentum to be conserved within the system , there is an expectation that a contracting rcb star would spin up to beyond breakup @xcite .
however if the surface velocity is close to the kepler velocity , angular momentum is transported efficiently from the star to the disk so that accretion continues as long as matter around the star exists @xcite , and viscous disks are known to be very efficient transporters of angular momentum outward and of mass inward @xcite . * progenitors .
the question of whether low - mass short - period double white - dwarf binaries capable of forming ehe stars are produced by binary - star evolution has been emphatically answered by the discovery of significant numbers of such systems @xcite .
curiously , there is a connection between the dwd merger model and the single - star evolution models of @xcite and others .
@xcite extended the work of @xcite on double he - wd mergers to investigate the formation of helium - rich sdo and sdb stars , which are essentially low - mass helium main - sequence stars .
@xcite found that , in a few cases , hesdo stars formed from the most massive double he - wd mergers would become carbon - rich helium giants during a he - shell burning phase after core - helium depletion .
however , the he+he channel provides less than 1% of the rcb and ehe stars provided by the co+he channel .
diagram for ehe variables , adapted from @xcite , including the position of the eddington limit ( assuming thomson scattering : dashed ) and the loci of stars with given luminosity - to - mass ratios ( solar units : dotted ) .
stars above the boundaries shown for metallicities @xmath23 ( dot - dash ) are predicted to be unstable to pulsations @xcite .
ellipses ( coloured in electronic version ) identify three groups of pulsating helium stars . in the electronic version ,
pvteli variables are shown in purple , pvtelii variables in blue , and v652her variables in green .
non - variables are black .
, width=453 ]
@xcite made the first detection of variability in an ehe , the hot star hd160641 = v2076 oph which showed a brightening by 0.1 mag over seven hours .
the discovery of a 0.108d pulsation period in v652her @xcite prompted searches for short - period variability in other ehes .
@xcite reported irregular small - amplitude variations of weeks to months in hd168476 = pvtel , thus codifying the pvtel variables as a class .
@xcite reported a short - period variation in bd@xmath24 = dnleo which could not be verified @xcite . from 1983 1988 , a st andrews / saao campaign made discoveries of variability in six ehe stars , namely fqaqr , noser , v2205oph , v2244oph and v1920cyg @xcite . as with dnleo , variability was not confirmed in popper s star hd124448 = v821cen @xcite . in the following decade
, variability was discovered in v4732sgr , v5541sgr and v354nor @xcite , and most recently in the enigmatic mvsgr @xcite . following a prediction by @xcite
, @xcite discovered 0.1d pulsations in bxcir .
the distribution of variable and non - variable ehes is shown in fig .
[ f : puls ] .
@xcite gives a complete list of properties including original catalogue numbers , gcvs variable star designations , approximate periods , temperatures , and gravities .
the first discovery of the st andrews / saao campaign concerned fqaqr , which showed a 21.2d sinusoidal oscillation with an amplitude of 0.4mag in @xmath25 and 0.05mag in @xmath26 .
@xcite associated this with pulsation .
together with periods for other ehes , a pulsation - period effective - temperature relation of the form @xmath27 was apparent . for stars of the same luminosity
, this corresponds to the period mean - density relation for classical radial pulsators . on the basis of this inference , @xcite calculated a series of linear non - adiabatic pulsation models for low - mass high - luminosity stars in the temperature range 7000 to 30000k .
they showed that , for sufficiently high values of @xmath28 , opacity - driven strange - mode radial pulsations would be excited , consistent with the periods observed in the cooler pvtel variables ( fig . [
f : puls ] ) .
@xcite calls these `` type i '' pvtel variables . at @xmath29k ,
both v2205oph and v2076oph were apparently multiperiodic on timescales longer than that consistent with a fundamental radial pulsation , and had been inferred to be g - mode non - radial pulsators @xcite .
together with v5541sgr , @xcite calls these `` type ii '' pvtel variables .
the short - period variability in the lower luminosity star v652her was not explained until the introduction of opal opacities @xcite , when @xcite showed that , in the absence of hydrogen , the @xmath1-bump opacity mechanism could easily drive radial pulsations with normal ( solar - like ) metallicities .
@xcite puts bxcir and v652her in the same class , but they would be better labelled v652her variables ( fig .
[ f : puls ] ) . what promised to be a class of simply periodic variables with a strict period - temperature relation and the possibility of measuring direct radii using baade s method turned out to be a chimra .
a second season of observations of fqaqr did not reveal a unique period , with 21.5d and 23.0d being possible @xcite .
@xcite monitored both fqaqr and noser for five consecutive seasons and , from the fourier power spectrum , reported the `` apparent presence of several periods but , if real , none seems to persist for more than one season . '' data from superwasp ( unpublished ) , and a wavelet analysis of the kilkenny et al .
data are similarly ambiguous .
the only recurring signal occurs at a period of around 20d , but is not coherent over a long period of time .
the absence of a regular period is not fatal for extracting stellar radii , so long as the oscillation is radial and angular radii and radial velocities can be measured simultaneously .
radial - velocity amplitudes of a few kms@xmath30 were measured in pvtel , fqaqr and v2244oph at the same time as angular radius measurements were obtained with _ iue _ @xcite .
relative phases were consistent with a radial pulsation . in two cases ,
the resulting radius and mass measurements were consistent with theoretical expectation , but the relative mass errors did not provide a strong constraint for establishing an evolutionary origin . the aperiodicity of fqaqr is reflected in observations of other ehe stars ( cf . * ? ? ? * ) .
@xcite considerably extended the photometry and spectroscopy of v2076oph . instead of recovering the periods of 0.7 and 1.1d reported by @xcite , they reported that `` conventional fourier analysis fails to reveal coherent frequencies '' and suggest that the light curve `` could be a result of random variations '' . on the other hand ,
the high - resolution spectrum of v2076oph shows prominent line - profile variability ( lpv ) on a timescale of a few hours which is symptomatic of a non - radial pulsation . of note
is that , although lpvs are visible in all hei lines , they are only seen in the heii4686 line , and not in other heii lines ( fig . 2 :
such intriguing results are almost impossible to interpret in isolation from photometric monitoring and over such a short time interval . in
linear non - adiabatic pulsation analyses of hydrogen - deficient stellar envelopes .
the ordinate is @xmath31 in solar units , and the abscissa is @xmath32 .
the plot is virtually invariant to the mass , at least in the range @xmath33@xmath15 .
the solid squares correspond to ehe stars .
white means no unstable modes except for @xmath34 where model envelopes are difficult to compute .
surface gravity contours for @xmath35 are represented as broken lines .
, width=453 ] @xcite extended their investigations of pulsation stability as a function of hydrogen abundance ( mass fraction : @xmath36 ) to a larger range of effective temperature ( @xmath37 ) , luminosity - to - mass ratio ( @xmath28 ) and @xmath36 than before @xcite .
models for @xmath38 were computed .
figure [ f : puls ] shows the instability domain for @xmath39 and solar metallicity . at large @xmath36 , the classical cepheid instability strip is clearly seen . at high @xmath28 strange modes
are excited .
as @xmath36 is reduced , the @xmath1-bump instability finger starts to develop , and a strip of intermediate - order radial modes develops blueward of the classical instability strip .
the strange modes are less sensitive to @xmath36 , and at low @xmath36 correspond to the observed locations of the variable ehe stars ( fig .
[ f : ehe_puls ] ) . at low metallicity
, the @xmath1-bump finger is much diminished , and the lower edge of the strange - mode domain moves to higher @xmath28 ratio ( fig .
[ f : puls ] ) , so the absence of variability in metal - poor ehes is anticipated @xcite . with no hydrogen ,
the morphology of the classical strip again changes and other pockets of instability appear , in particular one at @xmath40k and @xmath41@xmath42/@xmath15 .
it remains to be seen whether any real hydrogen - deficient stars can be identified with such models .
strange - mode oscillations in stars with high @xmath28 ratios have been identified for some time @xcite .
they appear theoretically in non - adiabatic pulsation analyses and have no corresponding modes in the adiabatic approximation .
in particular , they appear to be associated with stellar envelopes where high opacities in the ionisation zones also lead to a density inversion effectively creating a radiation - pressure dominated cavity in the stellar interior .
they are discussed at greater length by , _
inter alia _ ,
@xcite and @xcite .
evidence that non - radial oscillations may be present in the hottest ehes led @xcite to find a large number of unstable opacity - driven g - modes in models for v2076oph
. it will be interesting to learn whether extreme non - adiabacity and strange - mode characteristics or frequent switching between closely - spaced modes is responsible for the absence of long - lived coherent periods in these stars .
bd@xmath43v652her is an extreme helium star and a b - type giant discovered by @xcite . with a nitrogen - rich , carbon - poor surface , it pulsates with a period of 0.108d which is decreasing at a substantial rate ( cf .
* ; * ? ? ?
this period decrease implies a radius contraction that , together with a radius @xmath44@xmath45 and mass @xmath46@xmath15 @xcite , in turn implies that v652her is evolving to become a hot subdwarf within @xmath47yr @xcite .
single - star models can not explain the origin of v652her , which has an almost completely cno - processed atmosphere @xcite .
on the other hand , models for the post - merger evolution of two helium white dwarfs very successfully match nearly all of the observational properties @xcite , as well as demonstrating that such mergers will become helium - rich hot subdwarfs @xcite .
@xcite demonstrated that @xmath1-bump opacity instability drives fundamental - mode radial pulsations in v652her .
the radial - velocity curve shows that , for nine tenths of the cycle , the surface layers are nearly ballistic @xcite .
close to minimum radius , the surface acceleration is so large that the atmosphere may be shocked @xcite .
since hydrogen - deficient atmospheres are in general more transparent than hydrogen - rich ones , v652her provides a unique opportunity to study the dynamical behaviour of a pulsating atmosphere at greater depths ( or densities ) than is usually the case .
suitable observations could test non - linear hydrodynamic models for pulsation @xcite .
@xcite present a more extended account .
high - speed spectroscopy of v652her was obtained with the subaru high dispersion spectrograph .
2 nights of observations cover over six pulsation cycles with a temporal resolution of 174 seconds .
these observations aim to identify structure in the line cores around minimum radius , to resolve the passage of the wave through the photosphere using ions with different ionisation potentials , to determine whether this generates a shock front , or whether the photosphere adjusts subsonically , and to establish how close the `` free - fall '' phase is to a true ballistic trajectory .
the interpretation of the observations requires coupling a hydrodynamic model of the pulsation to a radiative transfer code , so that the emergent spectrum can be computed realistically . a summary of progress
is given by @xcite .
extreme helium stars form a group of some 15 low - mass supergiants of spectral types a and b. their helium - dominated atmospheres are extremely hydrogen poor ( @xmath48 ) and carbon - rich ( @xmath49 ) .
evidence from kinematics , surface composition and distribution in effective temperature and surface gravity points to a strong link with the cooler rcrb stars , and to their origin in the merger of a helium white dwarf with a carbon - oxygen white dwarf .
most ehe stars are photometric variables with amplitudes around one tenth of a magnitude .
most are also radial - velocity variables with amplitudes of a few kms@xmath30 .
the timescales of these variations as a function of effective temperature are consistent with the stellar dynamical timescales and hence with being due to pulsations .
theoretical models show most ehes to be unstable to opacity - driven radial pulsations . however , in all but two cases ( v652her and bxcir ) , the observed variations are _ not strictly _ periodic .
additional work is necessary to better characterise the periods and amplitudes and to obtain radii using the baade - wesselink method .
v652her and bxcir are less luminous than the majority of ehes .
v652her can be explained by the merger of two helium white dwarfs evolving to become a helium - rich subdwarf .
the carbon - rich bxcir is harder to explain , but could be similar .
both stars pulsate radially with periods of 0.1d .
the radial - velocity curve of v652her is nearly ballistic , with a very steep acceleration phase .
new observations and models provide a unique dataset and toolkit for exploring the physics of these pulsations , c.s . , hill , p.w . , & morrison , k. 1986 , in : k. hunger , d. schnberner , n. kameswara rao ( eds . ) , _ hydrogen - deficient stars and related objects _ , proc
. iau colloqium no . 87 , astrophysics and space sci .
library , vol . 128 ( dordrecht : d. reidel publishing co. )
p. 95 , c.s . ,
shibahashi , h. , kurtz , d. , elkin , v. , montas - rodrguez , p. , & saio , h. 2013a , in : h. shibahashi , a.e .
lynas - gray ( eds . ) , _ progress in physics of the sun and stars : a new era in helio- and asteroseismology _ , _ asp - cs _ , in press , c.s . , shibahashi , h. , kurtz , d. , elkin , v. , montas- rodrguez , p. , & saio , h. 2013b , in : v. van grootel , e. green , g. fontaine , s. charpinet ( eds . ) , _ hot subdwarf stars and related objects _ , _ asp - cs _ , in press , d. 1986 , in : k. hunger , d. schnberner , n. kameswara rao ( eds . ) , _ hydrogen - deficient stars and related objects _ ,
iau colloqium no .
87 , astrophysics and space sci .
library , vol .
128 ( dordrecht : d. reidel publishing co. ) p. 471 | stars consume hydrogen in their interiors but , generally speaking , their surfaces continue to contain some 70% hydrogen ( by mass ) throughout their lives . nevertheless , many types of star can be found with hydrogen - deficient surfaces , in some cases with as little as one hydrogen atom in 10000 . amongst these , the luminous b- and a - type extreme helium stars are genuinely rare ; only @xmath0 are known within a very substantial volume of the galaxy .
evidence from surface composition suggests a connection to the cooler rcrb variables and some of the hotter helium - rich subdwarf o stars .
arguments currently favour an origin in the merger of two white dwarfs ; thus there are also connections with amcvn variables and type ia supernovae .
pulsations in many extreme helium stars provide an opportune window into their interiors .
these pulsations have unusual properties , some being `` strange '' modes , and others being driven by @xmath1-bump opacities .
they have the potential to deliver distance - independent masses and to provide a unique view of pulsation physics .
we review the evolutionary origin and pulsations of these stars , and introduce recent progress and continuing challenges . |
a number of studies have noted that there are more spheroid stars in the virgo constellation than were expected from smooth spatial models of the spheroid star density .
the first indication of spheroid substructure in this region of the sky came from the @xcite discovery of five excess rr lyrae stars from quasar equatorial survey team ( quest ) survey data .
although a small number , it was a significant overdensity in a small area .
the quest collaboration later identified an excess of 23 rr lyrae variables with this clump .
these rr lyrae stars have apparent magnitudes @xmath3 , and are within the galactic coordinates @xmath4 , @xmath5 .
the collaboration later dubbed this feature the
@xmath6 clump " @xcite .
subsequent spectra of the rr lyraes and bhb stars in this clump @xcite revealed a moving group at ( ra , dec ) @xmath7 with a line - of - sight , galactic standard of rest velocity of @xmath8 and a distance of 19 kpc from the sun .
they suggest that this moving group be called the virgo stellar stream ( vss ) .
the existence of an overdensity of stars in the stellar spheroid near @xmath9 was independently confirmed by @xcite using measurements of blue turnoff stars from the sloan digital sky survey ( sdss ; york et al .
2000 ) , and identified as s297 + 63 - 20.5 . turnoff . in the present paper
we correct the name to be s297 + 63 - 20.5 ; both names will refer to the same overdensity of stars . ]
the estimated distance of this structure from the sun , calculated from @xmath10 turnoff stars , is @xmath11 kpc , but the color - magnitude diagram of this region does not have a tight main sequence , indicating that the structure is likely to be dispersed in distance .
this overdensity appears to match the position of the @xmath6 clump " , but is somewhat offset from the vss , which is at @xmath12 . because the mapping between the physical structure on the sky and its nomenclature in the literature for the apparent excess in the number of spheroid stars in the virgo constellation is unclear , we will refrain for now from identifying the s297 + 63 - 20.5 stars with the vss , though they are at about the same distance and position in the sky .
we will return to the question of which named spheroid structures are related to each other in the discussion section of this paper .
@xcite also identify a structure in virgo using the stellar catalogs from the 2mass survey , and connect it to m giants tens of kiloparsecs above the galactic plane that are in the heart of the descending , foreshortened northern loop , " which is their terminology for the leading tidal tail of the sgr dwarf galaxy .
any overdensity in this part of the sky is close to the orbital plane of the sgr dwarf galaxy , and might be related to the sgr dwarf tidal debris .
later papers , for example @xcite , show both the leading and trailing tidal tails of the sgr dwarf looping around through a similar position on the sky , near the identified overdensities in virgo , before passing through the galactic plane very close to the sun . since the sagittarius tidal tails are quite prominent in the spheroid , and the observed overdensity or overdensities in the virgo region are near the orbital plane of the sagittarius dwarf , a connection between these structures is worthy of exploration . @xcite identify a large overdensity ( the virgo overdensity , vod ) in the virgo constellation , which may or may not be related to the previously named substructures . using photometric parallaxes for a large number of sdss stars , they show that this large overdensity is @xmath13 kpc above the plane and covers @xmath14 sq . deg . of sky .
they detect no downturn in the star counts toward lower galactic latitudes , indicating that the structure could continue further into the galactic plane than the sdss observations probe . at that time
the sdss had not yet probed galactic latitudes lower than about @xmath15 near @xmath16 . @xcite
interpret the overdensity as a significant spheroid substructure that might be an invading dwarf galaxy .
the estimated distance to the vod is smaller than the distance to the vss or s297 + 63 - 20.5 .
it is unclear whether the nearer distance is real , or if distance uncertainties are large enough that the vod , vss , and s297 + 63 - 20.5 could describe the same physical structure . in
2 and 3 of this paper we will discuss s297 + 63 - 20.5 only ; the relationship of this overdensity to the vod and vss will be deferred to the discussion .
more recently , @xcite suggest that the invading dwarf galaxy might be the sgr dwarf galaxy leading tidal tail , and show that the @xcite model passes through the observed location of the vod , and that the stellar density of the vod is similar to the model predictions .
they suggest that if the sgr dwarf tidal stream passes through the solar neighborhood , it will be detected as an excess of stars with negative radial velocities in the northern galactic hemisphere .
this paper treats the vod as a separate entity from the vss , but located in a similar position in space .
the vod is described by these authors as likely to be associated with the sgr leading arm , and therefore infalling , and the vss is a group of stars nearly coincident in sky position , but with a different vertical and kinematic structure . because the sgr dwarf galaxy tidal tails can help constrain the shape of gravitational potential of the milky way and its dark matter halo , they are the topic of considerable study , independent from their possible connection with the observed overdensity in virgo .
one can in principle , by building ever more sophisticated equipment , know the position and velocity of every star in the milky way .
stars in tidal streams are the only stars which one can even in principle know about their positions and velocities in the past .
stars in tidal streams all resided , at one time in the past , in the progenitor satellite dwarf galaxy or star cluster .
it is this property of tidal streams that allows us to use them to constrain the milky way potential . because the sgr dwarf tidal tails are the most prominent tidal debris in the spheroid , and because the sgr dwarf orbital plane is approximately perpendicular to the galactic plane
, observations of the sgr tidal debris potentially provide a strong indication of the shape of the gravitational potential surrounding the milky way .
interestingly , a comparison of disruption models with stellar positions and velocities in the sgr tidal tails has produced a contradictory result .
the leading and trailing sgr tidal tails do not lie on the same plane @xcite ; they are tilted in opposite directions with respect to the orbital plane of the sgr dwarf galaxy .
their tilt suggests that the galactic potential is oblate @xcite .
however , the radial velocities of what appear to be some leading tidal tail stars can only be fit to disruption models with prolate galactic potentials @xcite .
the leading tidal tail stars that are not compatible with oblate galactic potentials are near the various identified substructures in virgo .
one must consider the possibility that the milky way potential is more complex than the models , or that some sgr leading tidal tail stars might be misidentified .
@xcite seems to suggest the latter .
figure 3 of their paper shows radial velocities for leading tidal tail debris that has @xmath17 measurements like those of an oblate dark matter halo , and stars with @xmath18 km s@xmath19 that they interpret as arising from an ancient episode of tidal stripping .
@xcite , however , show the leading tidal stream from sgr transitioning seamlessly into a clump of stars with @xmath20 km s@xmath19 .
it is this latter clump of sgr leading tidal tail stars that can only be fit with a prolate halo model .
the data in both papers come from 2mass selected m giants , and the velocity dispersions of the stars in the vicinity of the virgo constellation , with @xmath20 km s@xmath19 , are much larger than other places along the tidal debris stream .
another complication that potentially blurs our understanding of the nature and extent of the overdensity ( or overdensities ) in virgo is the suggestion that the smooth population of stars in the spheroid might be triaxial @xcite .
a triaxial spheroid can produce a stellar density that is not symmetric around the center of the galaxy , as viewed from our vantage point near the sun .
this idea was proposed because several authors see more spheroid stars in quadrant iv ( @xmath21 ) than we see in quadrant i ( @xmath22 ) , and because the lowest density of spheroid stars appears to be at a galactic longitude slightly less than @xmath23 @xcite .
a triaxial spheroid might explain how an overdensity could cover a large region of the sky and not immediately disperse ; the overdensity in virgo might be partially composed of an increased density of spheroid stars from the smooth component .
figure 3 of @xcite shows that no model has been found that completely accounts for the observed overdensity of stars near s297 + 63 - 20.5 , so although a triaxial spheroid might fit the tails of an overdensity in virgo , we would still expect to find a more spatially concentrated component of tidal debris with coherent velocities . a key distinction between explaining an asymmetry with a ` local ' stream vs. a ` global ' structure such as a triaxial spheroid is the difference in observed velocity dispersion of member stars .
stars in a stream must have a relatively small velocity dispersion ( @xmath24 ) , given the origin of the stream and given the fact that stream remains coherent as it moves through the galaxy s potential .
on the other hand , a triaxial spheroid structure in the halo may have a very large velocity dispersion ( @xmath25 ) , but evidence for its member stars will be spread around the whole of the halo . in this paper
we attempt to disentangle the components of the spheroid substructure in virgo using sdss photometry of blue horizontal branch ( bhb ) and f turnoff stars to build density maps of the tidal streams and the stellar spheroid itself .
we show that the sgr dwarf leading tidal tail , which is in the north galactic cap , arcs over ( farther from the galactic plane ) the s297 + 63 - 20.5 overdensity , and over the position of the sun in the galactic plane .
if one extrapolates the path of the sgr leading tidal tail below @xmath26 , it passes through the galactic plane 15 kpc or more outside the solar circle .
this calls into question both the identification of s297 + 63 - 20.5 stars with the sgr leading tidal tail and the identity of the m giant stars identified with the leading tidal tail of sagittarius that have velocities in conflict with oblate models of the galactic potential .
we show that there is a clear peak in the density of @xmath27 f turnoff stars near @xmath28 that is at least @xmath29 across , but the total extent on the sky and distance range can not be constrained by photometry alone .
the densities of f turnoff stars over a four magnitude range ( @xmath30 ) are not consistent with an axisymmetric spheroid .
if the absolute magnitude range of the f turnoff stars is 1.5 magnitudes , the spread in distance is a factor of 3 , or in this case 10 - 32 kpc . in every direction that we probe at these implied distances ,
there are more stars in quadrants iii & iv than there are in the symmetric position ( with the same galactic latitude but with longitude @xmath31 ) in quadrants ii & i. if the spheroid is axisymmetric , then the number counts should differ only where there is a significant tidal substructure .
either there is an enormous substructure centered in virgo but covering a large fraction of the galaxy or the smooth portion of the spheroid is not axisymmetric .
we then measure the velocities of the f turnoff stars associated with s297 + 63 - 20.5 and find that their line - of - sight , galactic center of rest velocity is @xmath32 . throughout this paper ,
the symbol @xmath17 will refer to the line - of - sight component of the velocity , as determined from radial velocity , transformed to the galactic standard of rest , using a solar motion of @xmath33 km s@xmath19 @xcite .
the measured @xmath17 is not consistent with the leading tidal tail of sgr , which has negative @xmath17 , in support of our claim that s297 + 63 - 20.5 is not spatially coincident with the sgr stream .
this velocity is significantly higher than the measured @xmath17 for the vss , but not so different that the two structures could not be related in some way . in the direction @xmath34
, we see evidence for peaks with @xmath35 and @xmath36 .
the first of the two peaks appears to have a bluer turnoff than the latter peak , and could potentially be related to the 2mass m giant stars that have previously been assigned to the sgr leading tidal tail .
@xcite , @xcite , and @xcite show that both the leading and trailing tails of the sgr dwarf spheroidal galaxy can be traced in sdss color - selected bhb and f turnoff stars . since this work ,
much more sdss data have become available , covering nearly the entire north galactic cap .
we will use the bhb and f turnoff stars in these data to trace the sagittarius dwarf galaxy tidal tails , and explore their connection with the s297 + 63 - 20.5 overdensity in virgo .
the data presented in this paper come from the sdss dr5 @xcite plus about 200 square degrees of imaging data from sdss ii that fills in the gap in dr5 photometry near the galactic pole ( to be released as part of sdss - ii dr6 ) .
we will be analyzing @xmath37 photometry @xcite of 8500 sq .
degrees in the north galactic cap .
details of the sdss survey geometry may be found in @xcite and @xcite .
other sdss technical information can be found in : @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
blue horizontal branch ( bhb ) stars are important for tracing spheroid structure because they are intrinsically bright and because they are all approximately the same absolute magnitude so their distances can be estimated from apparent magnitudes to create a three dimensional picture of the spheroid .
bhb stars are typically found in older stellar populations such as are found in the spheroid , but are not found in all old populations . even in the spheroid , there are a significant number of high surface gravity stars with colors that suggest a spectral type of a. these stars are either blue stragglers or young stars that have been stripped from a star - forming region such as a dwarf galaxy , and are approximately two magnitudes fainter than bhb stars . from sdss
photometry we can select stars with colors consistent with spectral type a , and then we further divide this sample into the stars whose photometry is more consistent with high surface gravity and the stars whose photometry is more consistent with a bhb star .
the photometric separation of a stars by surface gravity is not perfect , but is good enough to reveal important spheroid substructures .
photometrically selected a - colored stars were chosen from the sdss database in a similar manner to @xcite .
the stars were selected by the criteria : primary , " which eliminates duplicates , classified as star " because the point - spread function was consistent with a point source , and having colors @xmath38 , and @xmath39 .
the subscript ` 0 ' indicates that the magnitudes were corrected by the @xcite reddening as implemented for sdss data release 5 ( dr5 ; adelman et al .
this correction is appropriate given the large distance of the bhb stars .
we further separate higher surface gravity blue straggler ( bs ) stars from the lower surface gravity blue horizontal branch stars using the color selection criteria of figure 10 in @xcite .
[ sagxy ] shows the spatial positions of the bhb stars in the plane of the sgr dwarf orbit , under the assumption that all of the stars have absolute magnitude @xmath40 @xcite .
we selected only bhb stars whose projected position was within 15 kpc of the sgr dwarf orbital plane , as defined in @xcite , and rejected all stars within @xmath41 of the globular clusters m53 , ngc 5053 , ngc 4147 , and ngc 5466 .
the @xmath42 and @xmath43 coordinates are in the sgr orbital plane , and the galactic plane is at @xmath44 .
one can faintly see the newly discovered bootes dwarf galaxy ( belokurov et al .
2006b ) near @xmath45 kpc .
this figure can be compared with figure 11 of @xcite , which is reproduced in fig .
[ f11 ] of this paper , and shows the positions of 2mass selected m giant stars in the sgr tidal tails . in fig .
[ sagxy ] , one can clearly see the leading tidal tail of the sgr dwarf tidal stream , 30 - 40 kpc above the galactic plane .
a piece of the trailing tidal tail is in the upper left corner .
the center of the leading tidal tail appears to arc over the solar position , and to descend outside the solar circle if one extrapolates the trajectory down toward the plane .
the stars in sgr stream directly over the sun in fig .
[ sagxy ] appear to be in a very broad region of the sky , which is not consistent with results of models of sgr tidal debris , for example those of @xcite .
since we have not surveyed the entire volume within 15 kpc of the sgr dwarf orbital plane , there are non - trivial selection effects near the edges of the data , which in particular make it difficult to study the structure within 20 kpc of the sun in this figure .
the leading tidal tail appears unexpectedly broad in this projection as it arcs over the sun , and the extrapolation of the leading tidal tail toward the galactic plane is somewhat ambiguous . since viewing the data in the observed ( magnitude ) rather the derived ( distance ) space often yields higher contrast on any structures present in the halo , we show in figure [ glambda ] the @xmath46 magnitude vs. @xmath47 angular position along the sgr tidal stream , separately for bhb and bs stars within 15 kpc of the sgr dwarf orbital plane .
the left panel contains the same stars as fig .
[ sagxy ] , but in apparent magnitude versus angle along the sgr orbit .
the angle @xmath47 is measured from the center of the sgr dwarf galaxy , as viewed from the sun , back along the direction of the trailing tidal tail , as defined by @xcite .
this is the longitude part of a @xmath48 system in which the equator is rotated to line up with the sgr dwarf orbital plane .
a comparison of the left and right panels shows that the bhb / bs separation is quite good but not perfect .
very few of the bhb stars in the leading tidal tail of sgr are present in the right panel showing high surface gravity ( blue straggler ) stars .
we have not analyzed the concentration of stars brighter than @xmath49 .
these stars are saturated in the sdss images , and more care is required to determine how well color separation works with the interpolated magnitudes of these stars .
this figure can be compared with the 2mass m giant stars in the sgr tidal tails of figure 8 in @xcite and figure 2 of @xcite .
the @xmath47 axis is flipped from previous papers , but matches the orientation of the stream in figure [ sagxy ] .
there is a density peak in the left panel of figure [ glambda ] , centered at @xmath50 , which we initially thought was related to at least one of the previously discovered overdensities in virgo .
however , figure [ glambda2 ] shows that this overdensity is on the wrong side of the galaxy for it to be in virgo .
figure [ glambda2 ] splits the data from the left panel of figure [ glambda ] into two pieces : the left panel of figure [ glambda2 ] shows stars with @xmath51 and to the right shows that with @xmath52 . since @xmath53 is in approximately the same direction as the solar motion ( galactic @xmath54 ) , the left panel shows primarily stars in quadrants iii and iv , and the right panel shows primarily stars in quadrants i and ii .
the overdensity at @xmath50 is in the right panel in fig .
[ glambda2 ] , which contains primarily stars with @xmath55 .
the origin of this overdensity is still uncertain ( but see belokurov et al .
it could be a single debris structure or a superposition of separate streams .
it is notable that several known globular clusters conspire to be in approximately the same position in this diagram ; possibly there is debris associated with these globular clusters that is as yet unidentified .
note from figures [ glambda ] and [ glambda2 ] that the trailing tidal tail shifts from positive @xmath56 to negative @xmath56 with increasing @xmath47 .
in the left panel of figure [ glambda2 ] , one notes a low contrast faint extension of this tidal tail down to @xmath57 . in the right panel of figure [ glambda2 ]
one sees the leading tidal tail at positive @xmath56 .
this also decreases substantially in density as one proceeds along it to the left , and appears to be close to @xmath58 so it is split between the left and right panels in figure [ glambda2 ] . in both
the leading and trailing tidal tails , we see the pile - up of debris near apogalacticon . in this projection , since the sagittarius dwarf galaxy is @xmath59 below the galactic plane , material falling down on us from close to the north galactic pole in this orbital plane would fall toward us at @xmath60 .
since the leading tidal tail arcs past this angle , it likely passes through the galactic plane outside the solar circle .
since there are many more f turnoff stars in any stellar population than there are bhb stars , these f stars are in principle better tracers of substructure .
the disadvantages of using f turnoff stars is that they are 3 or 4 magnitudes fainter than the horizontal branch , and have a broader spread in absolute magnitude ( so they are less accurate distance indicators ) .
the sagittarius dwarf galaxy and tidal stream have a particularly blue turnoff which makes them easy to separate from other galactic populations . in this section we present data which follows the leading tidal tail from the sagittarius dwarf galaxy and shows that it misses the position of the s297 + 63 - 20.5 overdensity and also passes over the position of the sun before falling down on the galactic plane outside the solar circle .
we then present polar plots of f star density in the north galactic cap in the magnitude ranges @xmath61 , @xmath27 , and @xmath62 .
the polar plots show that the sagittarius leading tidal tail is fainter and offset from the s297 + 63 - 20.5 overdensity , and also show the extent of the s297 + 63 - 20.5 overdensity . fig .
[ f11 ] shows our detections of sagittarius dwarf tidal debris from f turnoff stars in sdss stripes 13 and 15 - 23 , along with the previous detections of sgr debris from bhb stars ( figure 3 of newberg et al .
2003 ) and 2mass m giants @xcite .
the new sgr debris positions were determined by selecting by eye the center of the highest density peak in a histogram of @xmath46 versus angle along the sdss stripe for blue f stars ( @xmath63 ) .
these positions were then converted to @xmath64 , assuming @xmath65 .
this absolute magnitude was determined in @xcite by comparison with the sgr bhb stars , so the sdss bhb star distances should be on the same scale as the sdss f turnoff star distances , though there could be an overall scale error .
note that the sagittarius m giants have systematically smaller heliocentric distances than either the a stars or the f stars @xcite .
the debris was identified as sgr tidal debris because it was contiguous with the overdensity in the adjacent stripe .
there was some freedom in choosing the stream positions , as the profile of the overdensity was often asymmetric or bifurcated ( belokurov et al .
the larger open squares in figure [ f11 ] show the positions of the higher density tidal debris .
if there was a second piece of tidal debris near the main sgr leading tidal tail in any stripe , it is shown in figure [ f11 ] as a smaller open square .
the positions of the sgr f turnoff star detections are presented in table 1 .
[ f11 ] clearly demonstrates that the leading tidal tail of the sagittarius dwarf will pierce the galactic plane outside the solar circle .
this is in contrast to the 2mass m star observations shown on the same figure .
it is known that the stellar population in the tidal tails is more metal poor and older than the stars in the core of the sgr dwarf galaxy @xcite , so it is reasonable to expect that the stellar populations would differ as a function of distance along the stream .
however , one does nt expect that stellar population differences would produce this much of a distance difference between estimates from f turnoff stars and m giants .
another possibility that should be considered is that the m giants near the end of the leading tidal tail are not all related to sgr .
[ f11 ] also shows that although s297 + 63 - 20.5 , indicated with a large cross , is close to the plane of the sagittarius dwarf orbit , s297 + 63 - 20.5 is much closer to the sun than the sgr leading tidal debris .
any debris that is in the sgr dwarf orbit is plausibly connected to this dwarf galaxy , but this significant overdensity in virgo is not an extension of the sgr leading tidal tail , and there is no obvious connection with sgr . the distance discrepancy between s297 + 63 - 20.5 and the sgr leading tidal tail is demonstrated in another way by looking at polar plots of f star densities in the north galactic cap at different apparent magnitudes .
we selected the stellar data from sdss dr6 with @xmath66 and @xmath67 .
the @xmath68 range was selected to be centered on the turnoff color of s297 + 63 - 20.5 , which is measured as 0.26 in @xcite . by excluding stars stars redder than @xmath69 ,
one avoids thick disk turnoff stars , even if they exist at these faint magnitudes .
[ polarplot ] ( lower panel ) shows f stars with @xmath27 ; this magnitude range brackets the estimated apparent magnitude of the turnoff of s297 + 63 - 20.5 .
the figure includes photometric data from sdss , as well as some extra photometric data obtained as part of the sloan extension for galactic understanding and evolution ( segue ) , which consists of @xmath70-wide scans at constant galactic longitude which appear as radial extensions to our sky coverage of the north galactic cap in this figure . the stars in fig .
[ polarplot ] trace structure in the spheroid of the milky way at distances @xmath71 kpc from the sun . a larger density of f stars is represented by a darker shading in the equal - area polar plot , centered on the north galactic cap .
the sagittarius leading arm is clearly visible running from @xmath72 to @xmath73 .
more prominent is the s297 + 63 - 20.5 overdensity , centered near @xmath16 , and hugging the lower edge of the data .
note that although there could be some overlap in the spatial positions of some of the stars , the denser portions of the sgr stream are separate from s297 + 63 - 20.5 . while most of the stars in the s297 + 63 - 20.5 are contained within an area of the sky 15@xmath74 in diameter
, the tails of the distribution could cover half of the north galactic cap . in order to show conclusively that the polar plot cuts through s297 + 63 - 20.5 , we perform the mathematical equivalent of folding the data from the top half of the lower panel in fig . [ polarplot ] down over the line described by @xmath75 , @xmath76 , and subtracting it from the data on the bottom half .
in other words , data between @xmath77 and @xmath78 is subtracted from the symmetrical point with the identical @xmath79 at @xmath80 and the residual is binned and plotted .
the result , displayed in the top panel in fig .
[ polarplot ] , shows a localized structure near @xmath81 , which is at the same position as s297 + 63 - 20.5 , within the errors of each measurement .
this folding of the data should leave zero residual ( except where there are streams or other halo substructure ) if the spheroid is axisymmetric . if the spheroid is asymmetric , then more stars should have been subtracted over a wide area of the upper panel , but it is difficult to make a smooth spheroid density function that would include a strong peak like the one we see at @xmath16 .
to clarify the relationship between s297 + 63 - 20.5 and sgr , we show in fig .
[ polarplot2 ] the f star polar plots for stars selected with @xmath61 and @xmath62 .
the upper panel shows the brighter stars .
there is still a clear excess of stars at @xmath82 compared to the position in the sky that is symmetric with respect to the galactic center .
for example , there are many more stars at @xmath83 than there are at @xmath84 .
however , there is not a clear peak in the star counts . the asymmetry in star counts
could either be an asymmetry in a smooth distribution of stars or a dispersed stream over a large area of sky .
the lower panel is dominated by a more distant piece of the sagittarius dwarf leading tidal tail , crossed at @xmath85 with the much fainter orphan stream @xcite .
there are excess stars near @xmath28 , but they are confused with and overwhelmed by stars from the sgr tidal tails at this magnitude
. a connection between the sgr tidal tail and s297 + 63 - 20.5 can not be completely ruled out , but the bulk of the turnoff stars in the sgr stream are at least a magnitude fainter ( @xmath86 ) than the bulk of the turnoff of s297 + 63 - 20.5 ( @xmath87 ) .
models of sgr debris such as that of @xcite show some debris stars near the position of s297 + 63 - 20.5 , but not a strong peak of debris .
either the models do not adequately match the stellar debris , or s297 + 63 - 20.5 was not stripped from sgr .
the polar plots show in a compelling way that the sgr leading tidal tail intersects neither the sun nor the s297 + 63 - 20.5 overdensity in virgo .
the turnoff stars in the sgr dwarf tidal stream are fainter than @xmath88 near @xmath9 - they show up only in the faintest polar plot on the lower panel of fig .
[ polarplot2 ] .
the sgr leading tidal tail is also separated from s297 + 63 - 20.5 by about ten degrees in the sky ( with significant overlap ) .
as the leading tidal tail descends closer to the sun we see the stream moving toward the anticenter ( fig .
[ polarplot ] ) . in other words
, it passes over the sun in the north galactic cap .
close to the anticenter , the sgr stream is fainter than @xmath89 ( compare fig .
[ polarplot ] with fig .
[ polarplot2 ] near the anticenter ) . near @xmath90
the sgr stream is visible in all of the f turnoff star polar plots , but the star density is the greatest at this position for @xmath91 . since the the str stream turnoff at @xmath92 is at @xmath93 , its implied distance is 18 kpc from the sun .
the sgr stream is therefore at least 15 kpc from the solar position and well beyond the solar circle as it descends toward the galactic plane .
we have already remarked that s297 + 63 - 20.5 seems to extend over a large range of magnitudes and a significant fraction of the sky .
we now study whether that large extent could be partially explained by an asymmetric ( triaxial ) smooth distribution of spheroid stars .
[ fcountsplates ] shows quantitative comparisons between the numbers of f turnoff stars as a function of magnitude near s297 + 63 - 20.5 and those at a symmetric position in the milky way , with the same galactic latitude but on the opposite side of the galactic center .
if the spheroid is axisymmetric , then one could find the number of stars in s297 + 63 - 20.5 from the difference between the star counts at @xmath9 and the star counts at @xmath94 .
in particular , fig .
[ fcountsplates ] shows a comparison of star counts at two places in virgo : @xmath12 ( upper panel ) and @xmath34 ( lower panel ) , with the corresponding position on the sky on the other side of the galactic center .
these two positions were selected because they match the positions of segue spectroscopic fields in which we will analyze radial velocity kinematics later in this paper .
each sky area probed has a radius of @xmath95 , which matches the footprint of a segue spectroscopic plate .
in addition , f stars are selected with a color range : @xmath96 to match the color range with the highest concentration of f stars in s297 + 63 - 20.5 .
this redder color range will include a significant number of thick disk stars at magnitudes brighter than @xmath97 , but at fainter magnitudes there will be little contamination .
since the thick disk ( unlike the halo ) , is axisymmetric , there should be no effect on the difference in bright star counts due to thick disk stars between the virgo selected regions and the symmetric reference regions . comparing the star counts at @xmath12 , which is centered on the central knot of the vss , with the star counts at @xmath98 in the upper panel of fig .
[ fcountsplates ] , one sees that the star counts start to diverge at about @xmath99 magnitude in @xmath46 .
by @xmath93 , there are not quite a factor of two more stars in the vss field than in the comparison field . in the magnitude range @xmath100 , there are 560 stars in the @xmath12 field and only 357 stars in the corresponding reference field at @xmath98 . of the stars in the vss field ,
36@xmath1016% are part of s297 + 63 - 20.5 . in the magnitude range @xmath102 , there are 369 stars in the vss field and 206 stars in the reference field . of the fainter stars in the vss field ,
44@xmath1017% are part of s297 + 63 - 20.5 .
this means that if one selected a random turnoff star in the vss field at @xmath93 , there is nearly a 50% chance of it being a star from s297 + 63 - 20.5 .
the @xmath34 field in virgo diverges from the reference field at @xmath103 at the even brighter magnitude of @xmath104 ( lower panel of fig .
[ fcountsplates ] ) . about a quarter of the stars with @xmath105
are part of the s297 + 63 - 20.5 excess in virgo . in the magnitude range @xmath100 , there are 699 stars in the virgo field , and only 440 in the reference field . of the stars in the virgo field ,
37@xmath1015% are part of s297 + 63 - 20.5 . in the magnitude range @xmath102 , there are 453 stars in the virgo field and 238 stars in the reference field . of the fainter stars in the virgo field ,
47@xmath1016% are part of s297 + 63 - 20.5 . at a magnitude of @xmath93
, more than half of the stars in the @xmath106 field should be part of the excess .
if the excess population has a unique kinematic signature , as is expected for a stream or satellite , it should be quite prominent in any sample of faint f turnoff stars .
we will return to these fractions when we examine the kinematic distribution of selected f turnoff star spectra in these fields .
the wide range of magnitudes and sky coverage makes the overdensity in virgo difficult to rationalize with our expectations for tidal debris , even before we begin looking at the spectroscopy results .
we consider a non - axisymmetric spheroid model as a possible alternative explanation for the excess star counts in the virgo region . to compare the observed star counts with those predicted by the best currently available asymmetric model of the spheroid , the galactocentric triaxial hernquist model of @xcite
is chosen . in this model ,
the center of the spheroid distribution is fixed at 8.5 kpc from the sun with an assumed absolute magnitude of blue f turnoff stars of @xmath107 . if the absolute magnitude is incorrect , then the overall scale of all lengths and distances in the model will be off by the same factor .
see @xcite and @xcite for parameter definitions and values for the triaxial hernquist model .
the model was generated to fit sdss f turnoff star counts with color @xmath108 , which is a bit bluer than the data at @xmath109 . to account for this
, the normalization of the model is adjusted upward by a factor of 1.5 so that it matches the star counts as well as possible .
we can not correct for the fact that the model does not include the thick disk stars at brighter magnitudes , so we will have to ignore the difference between the star counts and the model brighter than about @xmath97 .
these triaxial spheroid model curves are plotted in fig .
[ fcountsplates ] for comparison with the f turnoff star counts .
the model is a reasonable fit to the data for @xmath110 . at brighter magnitudes the thick disk dominates , and at fainter magnitudes
there are fewer stars than the model predicts .
the discrepancy at faint magnitudes is probably due to a deficiency in the models , and possibly the lower completeness of the fainter stellar data .
note that if one compares the star counts at @xmath12 and @xmath34 with the model , the star counts begin to diverge somewhere between @xmath111 and @xmath112 .
there is a much smaller excess in the number of stars at magnitudes brighter than @xmath112 . even using the current best fit triaxial spheroid model , however
, one can not account for all of the excess stars , particularly those with magnitudes @xmath27 . in the next section the two possibilities for explaining the stellar count excess
are explored using additional information obtained from a sample of spectra from the same population of stars .
if the spheroid is axisymmetric , and the s297 + 63 - 20.5 structure is tidal debris , then 37@xmath1014% of the stars with @xmath100 should have coherent velocities ( i.e. show a small @xmath113 dispersion ) , and 46@xmath1015% of the stars with @xmath102 should have coherent velocities . if the spheriod is asymmetric , then the fraction of stars with coherent velocities is expected to be lower . with the best fit model presented , one estimates that at @xmath12 , 64 of 560 , or 11% of stars with @xmath100 should have coherent velocities , and 159 of 369 , or 43% of stars with @xmath114 should have coherent velocities . in the direction @xmath34 , 76 of 699 , or 11% of the stars with @xmath100 should be coherent , and 127 of 453 , or 28% of stars with @xmath114 should be coherent .
the model numbers are an estimate , and will change somewhat as new non - axisymmetric spheroid models are developed .
since s297 + 63 - 20.5 was discovered in f turnoff stars , and is prominent against the background of smooth spheroid stars in that population , one would like to measure the velocities of these same types of stars if possible . since the overdensity is most noticeable in quite faint f turnoff stars ( with @xmath115 ) , this is not an easy task .
essentially zero f turnoff stars in this magnitude range were observed in the primary sdss galaxy survey .
we are currently in the middle of sdss ii , which includes three survey projects , including the sloan extension for galactic understanding and exploration ( segue ) .
the segue portion of sdss ii is designed to reveal velocity substructure in the milky way thick disk and spheroid through analysis of radial velocities and stellar properties from spectroscopy of @xmath116 galactic stars .
segue spectra are identical to sdss spectra , in that they are observed with the same instrument and processed with the same ( evolving ) software ; the spectra cover 3700 - 9300 with a resolution of @xmath117 2000 .
the difference is that sdss targeted a very few galactic stars , most of which are bhbs or bright f subdwarfs that were used as standard stars , while segue targets galactic stars exclusively .
while the sdss obtained galaxy spectra in all areas of the sky with imaging data , segue samples the sky with a set of pencil beam spectral surveys in 200 target fields , spaced roughly 10 to 20 degrees apart over 3/4 of the sky . in the segue survey ,
two 640 fiber plates of spectra @xcite are obtained in each target field .
the brighter targets are put on one plate which is exposed about 45 minutes .
stars fainter than about @xmath118 are assigned to the second plate , which is exposed for 1.5 to 2 hours .
the faint plates reach about @xmath10 , which is just at the bright end of the s297 + 63 - 20.5 turnoff population .
targets for segue spectroscopic fibers are chosen by color from the sdss imaging data to represent a variety of stellar types at a variety of distances sampling the thick disk and spheroid of the milky way .
we have already obtained two pairs of segue spectroscopic plates in directions that probe the s297 + 63 - 20.5 overdensity .
these plate pairs are numbered 2689/2707 ( bright / faint ) at @xmath34 and 2558/2568 at @xmath12 .
the second plate pair is centered on the vss rr lyrae star overdensity . during dr6 spectro test processing of globular cluster calibration stars with
known cataloged velocities , a universal offset of @xmath119 was noted in the wavelength solutions ; this zero point correction was added to all segue radial velocities used in this paper .
most of the faint stars of interest , with @xmath120 , are selected as f subdwarf , " blue horizontal branch , " or f / g " spectroscopic target types .
the bhb category goes as red as @xmath121 and as faint as @xmath122 .
these stars are randomly sampled to favor lower surface gravity stars using a photometric measure of surface gravity for a - colored stars called the v - index @xcite to estimate surface gravity .
the f / g category randomly samples stars with @xmath123 and @xmath124 .
the faint subdwarf category favors f stars with low metalicity by favoring stars that are at the very bluest tip of the stellar locus in @xmath125 and @xmath68 , using a measure of position along the stellar locus in color space called @xmath126 @xcite .
the criteria for selecting each segue category are complex , and include techniques for varying the likelihood of selecting a star for spectroscopy as a function of apparent magnitude , so that we select fewer of the more numerous , fainter stars .
the broader goal is to sample all of the components of the milky way , so less common components are favored in the selection process .
[ seguets ] shows the color magnitude diagram of the bluer sdss stars within @xmath95 of the center of the segue plate pair 2689/2707 , along with the colors and magnitudes of the stars with @xmath127 and @xmath128 for which we have obtained segue spectra with the two segue plate pairs 2689/2707 and 2558/2568 .
this is the entire region from which the spectra in this paper are selected .
the symbols show that the spectra are spread in color and magnitude , and do not follow the color distribution of the spheroid stars .
we must select stars in narrow color and magnitude ranges that correspond to a component we intend to sample ; we can not rely on the fact that one component has many more stars to guarantee that it will dominate the sample . the stars that we expect to best sample the turnoff of s297 + 63 - 20.5 are those with @xmath96 and @xmath27 .
the segue spectra only go as faint as @xmath129 , and most of the spectra have @xmath130 , so we sample only the brightest end of the magnitude range .
the right panels of fig .
[ fstarrv ] show @xmath17 for stars from segue faint plates 2707 and 2568 ( available publicly with the release of sdss dr7 ) that have the colors and magnitudes of s297 + 63 - 20.5 f turnoff stars , and radial velocity errors , as measured by cross - correlation with elodie standard templates @xcite , of less than 20 km s@xmath19 . in both plates
, there is a clear excess of faint stars with positive radial velocities . in the combined panel ( bottom
right of fig .
[ fstarrv ] ) , there is a five sigma peak in the bin @xmath131 .
there are 30 stars with positive @xmath17 and only 13 stars with negative @xmath17 .
if all of the excess stars are associated with s297 + 63 - 20.5 , then 40@xmath10111% of the fainter f turnoff stars could be associated with a single tidal debris structure in virgo .
this fraction is consistent with estimates from photometry , whether or not the spheroid is symmetric . in the symmetric spheroid case , we expected @xmath132% of the excess stars with @xmath102 to be in the excess population .
if the triaxial spheroid model is correct , the percentage in the excess population is about 35% for both plates combined .
our measured value of @xmath133% is consistent with either of these options .
therefore , from the faint ( @xmath134 ) spectroscopic and photometric data combined we conclude : the majority of the stars in the s297 + 63 - 20.5 overdensity are consistent with the presence of single substructure with a coherent velocity of @xmath135 and relatively small dispersion , as we would expect for tidal debris . in the left panels of fig .
[ fstarrv ] , we show @xmath17 for the stars with @xmath100 . in these plots , there is no excess in the @xmath136 bin .
one plate has more positive velocities , and the other has more negative velocities . in the combined panel
( lower left of fig .
[ fstarrv ] ) , the only significant peak has @xmath137 . thus the brighter f stars , in the left panels of fig
. [ fstarrv ] , have distinctly different kinematics than the fainter stars in the right panels of the same figure . out of 72 brighter stars , a few at best are part of the moving group we identified in fainter stars .
the symmetric spheroid assumption indicated that we should have found 37% , which is 27 stars , in a peak near @xmath138 . the most we can imagine assigning to a peak are about seven stars from the plate at @xmath12 that are around @xmath139 .
the moving group is ten percent or less of the stars with @xmath100 .
this differs at the @xmath140 level with our expectation of 27 stars present in a peak if the halo is axisymmetric .
this measurement of brighter ( @xmath141 ) star kinematics and photometry thus favors a triaxial spheroid . in the symmetric spheroid model ( plus individual star stream peaks ) , we expected @xmath142% of the stars to be coherent in velocity , and we only found 11% at the most .
the triaxial spheroid model predicted that 11% of the stars would be coherent in velocity , which is in good agreement with the observed fraction .
there is one significant peak in the velocities of the brighter f turnoff stars , but it has a very negative @xmath17 .
we identify as a peak " any bin with more than a @xmath143 excess in the expected number of stars , where @xmath144 , @xmath145 is the expected number of stars , and @xmath146 is the total number of stars in the 100 square degree region .
the limit for a bin to be identified as a peak is shown by the dotted curve in fig.[fstarrv ] . in plate 2707 , there are five stars with @xmath147 where we would have expected only one , and three stars in plate 2568 where we would have expected only two .
this amounts to an excess of about five stars in the summed histogram , lower left of fig .
[ fstarrv ] .
the mean velocity of this peak is @xmath148 . since @xcite measure a turnoff color for s297 + 63 - 20.5 of @xmath149
, one would not expect many stars in this structure to have measured colors of @xmath150 in the sdss .
we show measurements of @xmath17 for these bluer stars in fig .
[ fstarrv2 ] . in the fainter ( @xmath151 )
set of stars , there are maybe two excess stars in the bin one to the right of the moving group associated with the center of s297 + 63 - 20.5 , but no significant velocity peaks are observed . in the left panels of fig .
[ fstarrv2 ] , we see two marginally significant peaks : one at nearly the same velocity as the @xmath152 peak in fig .
[ fstarrv2 ] , and one at in the bin with @xmath153 .
we looked at the individual stars in positive velocity peak , and most of them are at the very faint and very red edges of the selection criteria .
possibly , they represent blue stragglers that are at the same distance but have brighter absolute magnitudes than the f turnoff stars in the moving group identified in the fainter panels of fig
. [ fstarrv ] .
the peak with negative radial velocity is at @xmath154 , but note that there are more stars on the negative velocity side of this peak than on the positive side . if most of the brighter stars in this color range are associated either with the positive or the negative @xmath17 peak , then the gaussian distribution from the expected spheroid would be much lower and the center of the peak would move one bin to the left .
the center of the peak is nearly between two bins .
this peak is more pronounced in plate 2707 than in plate 2568 .
once this peak is identified , one can see that it also present at some level in the fainter panels on the right of fig .
[ fstarrv2 ] . since the stars in this peak are at about the same distance and line - of - sight velocity as stars which other authors have attributed to the sgr dwarf leading tidal tail , they will be analyzed at greater length in the discussion section of this paper .
there was a time when any significant clump of tidal debris in the spheroid was suspected of belonging to the one known example of present - day merging in our galaxy @xmath155 the sagittarius dwarf tidal stream .
we now know there are many tidal debris streams , though the only currently known tidal stream as large as that of the sagittarius dwarf is the monoceros stream in the galactic plane @xcite .
we identify stars in the tidal tails of the sagittarius dwarf spheroidal galaxy because either their locations and radial velocities are contiguous to those of other known pieces of sagittarius debris or because their locations and velocities match those of models . in this paper
, we have presented a strong argument that s297 + 63 - 20.5 is not part of the sgr leading tidal tail because the turnoff stars in s297 + 63 - 20.5 are much brighter than those in the leading tidal tail debris in the same direction in the sky , and the stars in s297 + 63 - 20.5 have positive @xmath17 while sgr leading tidal debris would have negative @xmath17 .
moreover , @xcite showed that the turnoff of s297 + 63 - 20.5 is redder than that of the sgr tidal stream anywhere it has been detected in the sky .
although @xcite were able to generate a model that shows the leading tail of the sagittarius stream going through the position of the s297 + 63 - 20.5 , they did so at the cost of not fitting some of the more distant spatial detections of that same tidal tail . in this paper , ( see 2 ) , we locate additional pieces of what appears to be the leading tidal tail of sagittarius in sdss f turnoff stars , and show that it does not intersect the position of s297 + 63 - 20.5 .
since the sgr tidal debris may be bifurcated , lumpy , wrap multiple times around the milky way , and have different stellar populations in different places , it is difficult to prove that any spheroid substructure , especially one that happens to be in the sgr orbital plane , is unrelated to this tidal disruption event . since the s297 + 63 - 20.5 debris is outgoing in velocity , it is possibly related to the sgr trailing tidal tail .
however , no existing model shows a lump of material near s297 + 63 - 20.5 .
more extensive modeling and mapping of the sgr tidal debris is required to determine definitively whether they are related to each other .
what is much murkier is the relationship between s297 + 63 - 20.5 and the other substructures that have been discovered in the virgo region , and have been identified as overdensities in virgo .
these include the vss ( aka the [email protected] clump " ) , and the vod .
the vss is in the same general area of the sky as s297 + 63 - 20.5 , and at a similar calculated distance .
they both have positive @xmath17 , though the vss has a measured @xmath17 of @xmath157 compared to our measurement of @xmath158 for s297 + 63 - 20.5 .
it is possible that membership or measurement errors of a few stars in either the s297 + 63 - 20.5 sample or the vss sample could explain the difference in measured velocity .
the central knot of the vss is also a bit offset ( @xmath159 ) from the center of s297 + 63 - 20.5 .
one wonders whether the clump of rr lyrae stars that defines the vss could be a substructure within s297 + 63 - 20.5 , such as a disrupted gc within a disrupted dwarf galaxy . or
, we may find as the number of known stars in the vss grows , it might resemble the s297 + 63 - 20.5 structure more closely .
given the prevalence of velocity substructures in even the small amount of data analyzed here , it can not be completely ruled out that the vss and s297 + 63 - 20.5 might be chance superpositions . however , that possibility seems fairly unlikely since the velocities are quite similar .
we expect that as more data is analyzed , the s297 + 62 - 20.5 overdensity will become known as the virgo stellar stream ( vss ) .
it is somewhat difficult to assess the relationship between s297 + 63 - 20.5 and the vod . in the @xcite paper ,
their first step is to determine the distance to each star in the sdss photometric database .
all subsequent analyses are carried out with these spatial positions .
the vod has a distance of @xmath160 kpc , which is a little closer than our estimate of 18 kpc for s297 + 63 - 20.5 , which is at a similar estimated distance as the vss .
however , our distance errors to s297 + 63 - 20.5
are large , and there is no guarantee it is the same as the vss . in particular , there is no guarantee that our absolute magnitude estimate of @xmath107 , which was derived for sgr tidal debris with similar colors , is appropriate for s297 + 63 - 20.5 .
the factor of 2 density excess for the vod is similar to the density excess that we see in comparisons of the number of f turnoff stars near @xmath28 and @xmath161 . also supporting the identification of s297 + 63 - 20.5 with
the vod is the position in the sky and the large sky area covered .
if there were two separate overdensities , one at 10 kpc and one at 18 kpc , we would expect to see a separate vod structure in the top panel of fig .
[ polarplot2 ] , with an apparent magnitude around @xmath162 .
all we see in that figure is a general asymmetry in the star counts - no detected peak density .
this suggests that either our technique of estimating distance gives different results from @xcite or that the turnoff of the vod is redder than @xmath69 .
probably , they are the same overdensity measured with the same star catalog , but it is difficult to know for sure .
what is needed is a representation of the vod stars that allows us to identify which stars are members of the structure , such as a color magnitude diagram and a map of angular position on the sky .
then we would be able to check whether we are looking at the same stars in the same catalog or not .
the controversy regarding the oblateness of the milky way gravitational potential hinges on the radial velocity data for m giants in the part of the leading tidal tail that is closest to the sun , @xmath163 .
figure 12 from @xcite shows the model distances and velocities of sagittarius debris for prolate , spherical , and oblate dark matter halos , along with data for 2mass m stars .
there are a large number of stars near @xmath164 that have velocities of @xmath165 km s@xmath19 and distances from the sun of about 15 kpc .
it is precisely these stars that do not fit the published models , unless one assumes a very prolate dark matter halo @xcite .
the stars with @xmath166 also have a larger velocity dispersion than sgr debris stars at other places in the stream .
the last column in table 1 gives the position of the sgr leading tidal tail in terms of @xmath47 ; at @xmath164 the distance to the sgr stream is 36 kpc .
m giant distances in @xcite are systematically @xmath167 less than ours , as determined from a comparison of m giant stars and a / f stream position of the leading tidal tail stars near apogalacticon @xcite . applying this scale difference
, the m giant stars at @xmath164 should be at 31 kpc on the m giant distance scale .
there are very few stars in the @xcite sample that are that far away .
[ f11 ] shows where the discrepancy arises . if one looks directly up from the sun in fig .
[ f11 ] , following the direction of @xmath168 , the sgr f turnoff stars are estimated to be more than 30 kpc above the galactic plane . in contrast , the m giant stars are scattered from 10 kpc to 35 kpc in the same direction .
the m giant stars are at about the same distance , and in about the same direction , as the s297 + 63 - 20.5 overdensity . in fact
, we detect a radial velocity peak for f turnoff stars near s297 + 63 - 20.5 that has a velocity of @xmath169 , which is very similar to the m giant velocities .
there is an excess of stars in the bin with @xmath170 in every histogram in fig .
[ fstarrv2 ] .
there are 15 of these bluer f turnoff stars ( @xmath150 ) with @xmath171 where we expected to see less than seven .
the excess in this bin is more than 3 sigma , and there are additional extra blue f turnoff stars with more negative @xmath17 .
there is also velocity peak in rr lyrae stars near the vss , at about the same distance and direction .
figure 2 of @xcite shows spectroscopy of 28 bhb and rr lyrae stars from the quest survey . in the @xcite paper , a very tight group of stars at 18.5 - 20 kpc from the sun and within @xmath172 of each other in the sky is identified .
the mean velocity in the galactic rest frame of these stars is 99.8 km s@xmath19 .
additional stars as close as 16 kpc and as distant as 24 kpc have similar enough radial velocities that they might be considered to be part of the same kinematic structure .
the remaining stars are not consistent with a gaussian distribution , and we note that there is a very significant peak in the lower panel of their figure 2 that is very near @xmath173 . if one considers all of the radial velocities obtained by @xcite , there are 12 stars in a peak at nearly 100 km s@xmath19 , nine stars in a peak at @xmath174 km s@xmath19 , and 7 stars with very large negative or positive radial velocities that are consistent with neither peak , nor with any existing model of spheroid kinematics .
there is insufficient information in the @xcite paper to adequately analyze the nine stars with velocities toward the sun , but they likely cover an area @xmath29 on the sky or larger , and at least the distance range @xmath175 kpc .
these stars are very likely associated with our f turnoff star peak at @xmath169 and with the stars that @xcite assigns to the sgr leading tidal tail .
figure 3 of @xcite also show stars in the same part of the sky , @xmath176 , and with similar velocities , @xmath177 km s@xmath19 .
interestingly , they identify these stars as part of an ancient tidal debris stream that was stripped from sgr @xmath178 gyr ago . although this presents a possible explanation , it does not explain why the dispersion of the stars appears so large in this portion of the sky , and is at odds with the general assumption that 2mass giants favor younger , more recently stripped portions of the debris .
figure 12 of @xcite shows a subset of stars in this region that have a very narrow velocity dispersion and very nearly match the expected position of the ancient tidal debris if one assumes an oblate dark matter halo . one wonders whether a simpler explanation for the debris with @xmath20 km s@xmath19 might be the presence of a previously unidentified spheroid component .
the fact that the stars in this infalling structure have a blue turnoff supports , but does not prove , the claim that they are part of the sgr dwarf tidal stream .
however , they appear to be closer than the main part of the tidal tail . further investigation is required to determine whether they are part of the sgr dwarf tidal debris , and whether they represent the population of stars they are being compared to in the tidal disruption models .
if we knew the spatial position of every spheroid star , or at least a representative sample of spheroid stars , we could use this information to construct a density model for this milky way component . but
given that there are very large known overdensities in the spheroid , what would the parameters in this model be ?
@xcite show that all smooth models are poor fits to the distribution of stars in the spheroid .
we have chosen to model the spheroid as a set of significant overdensities that are coherent in position and velocity , plus a roughly smooth spheroid distribution .
overdensities include globular clusters , dwarf galaxies , and large chunks of associated debris from these objects .
thus , we are now in the process of identifying large , coherent debris structures so that we can fit the remaining stars to a smooth spheroid model ; and fitting smooth spheroid models so we can identify overdensities that might be large , coherent debris structures .
there is no controversy over the asymmetry of the star counts in the stellar spheroid stars in the north galactic cap @xcite , though there is a question of whether there is an asymmetry in the south @xcite .
an important question is whether the asymmetry is due to a single large debris structure added to an otherwise axisymmetric distribution of stars .
the data in this paper suggest that is not the case .
the s297 + 63 - 20.5 is a very significant , coherent overdensity that contains about half the spheroid stars at @xmath93 and @xmath9 .
however , we have shown that it is localized in magnitude and velocity , and does not explain the asymmetry at brighter magnitudes .
it is possible that there is more than one structure , or maybe a set of structures , that conspire to make the north galactic cap appear asymmetric .
it seems at least as likely that the spheroid itself is not symmetric about an axis through the center of the milky way and perpendicular to the galactic plane .
we have shown that our existing triaxial spheroid model is not perfect , but goes some way toward predicting the number of spheroid stars in the observations .
we hope the model will improve as more large , coherent spheroid substructures are identified and as our fitting procedures evolve .
the sgr leading tidal tail is traced through the north galactic cap over the center of the milky way , over s297 + 63 - 20.5 , over the sun , and at @xmath26 it is heading toward the galactic plane at a galactic longitude of @xmath92 .
we expect that it pierces the galactic plane well outside the solar circle .
we question whether some of the m giant stars that have been compared to sgr dwarf disruption models are at the right distance to be members of the sgr dwarf leading tidal tail .
the main part of the sgr dwarf leading tidal tail is not spatially coincident with the s297 + 63 - 20.5 overdensity in virgo .
the stars in s297 + 63 - 20.5 have line - of - sight , galactic standard of rest velocities @xmath179 .
the velocity dispersion is difficult to estimate because we do not know for sure which stars are part of the structure , but velocity dispersions of 10 @xmath180 to 30 @xmath181 are not unreasonable .
since the sgr dwarf tidal tail would have negative velocities in this part of the sky , we determine that s297 + 63 - 20.5 is not associated with the leading tidal tail . because the color of the turnoff , the density structure , and the @xmath17 velocities do not match any known or expected sgr debris , we suggest that it is instead part of a distinct merger event .
the virgo stellar stream ( vss , duffau et al . 2006 ) is at the same distance from the sun as s297 + 63 - 20.5 and has a similar but not identical position in the sky and in @xmath17 .
it could be a the same as s297 + 63 - 20.5 or a structure within s297 + 63 - 20.5 , but the relationship is not certain .
the virgo overdensity ( vod , juri et al . 2006 ) and s297 + 63 - 20.5 are in the same position in the sky and have a similar number of stars in excess of the smooth spheroid , but they are at a different distances from the sun .
probably , there is a scale difference between our distance measurements and the vod is the same as s297 + 63 - 20.5 .
we show that the number densities of f turnoff stars are not symmetric about the galactic center at @xmath182 , and that this discrepancy is not due to s297 + 63 - 20.5 . either the spheroid is asymmetric about the galactic center , or there are additional substructures that conspire to be on the same side of the galaxy as s297 + 63 - 20.5 .
finally , we note that in every figure showing the spatial and velocity distribution in the spheroid , there are hints of extra unidentified substructure . in particular , we note an unexplained overdensity of bhb stars at @xmath183 , and two additional moving groups near s297 + 63 - 20.5 with velocities @xmath184 and @xmath35 .
hjn acknowledges funding from the national science foundation ( ast-0307571 , ast-0607618 , ast-0612213 ) , the nasa ny space grant , and john huberty .
tcb acknowledges partial funding for this work from national science foundation grants ast 04 - 06784 and phy 02 - 16783 : physics frontiers center / joint institute for nuclear astrophysics ( jina ) .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , cambridge university , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
adelman - mccarthy , j. , et al .
2007 , apjs , in press .
abazajian et al .
2003 , aj , 126,2081 bell , e. f. et al .
2007 , apj , submitted bellazzini , m. , newberg , h. j. , correnti , m. , ferrar , f. r. , & monaco , l. 2006 , , 457 , l21 belokurov , v. et al .
2006a , apj 642 , l137 .
belokurov , v. et al .
2006b , apj 647 , l111 .
belokurov , v. et al .
2007a , apj 658,337 belokurov , v. et al . 2007b , apj letters , in press , astro - ph/0701790 chou , m .- y . ,
majewski , s. r. , et al .
2006 , , submitted , astro - ph/0605101 dehnen , w. & binney , j. j. 1998 , , 298 , 387 duffau , s. , zinn , r. , vivas , a. k. , carraro , g. , mendez , r. a. , winnick , r. , & gallart , c. 2006 , , 636 , l97 fellhauer , m. , et al .
2006 , , 651 , 167 fukugita , m. , ichikawa , t . , gunn , j. e. , doi , m. , shimasaku , k. , schneider , d. p. 1996
, , 111 , 1758 gunn , j. e. et al .
1998 , , 116 , 3040 gunn , j. e. et al .
2006 , , 131 , 2332 helmi , a. et al . 2005 , , 586 , 195 helmi , a. 2004 , , 610 , l97 hogg , d. w. , finkbeiner , d. p. , schlegel , d. j. , & gunn , j. e. 2001 , , 122 , 2129 ibata , r. a. , irwin , m. j. , lewis , g. f. , ferguson , a. m. n. , & tanvir , n 2003 , mnras 340 , l21 ivezi , z. , et al .
2004 , astronomische nachrichten , 325 , 583 johnston , k. v. , law , d. r. , & majewski , s. r. 2005 , , 619 , 800 juric , m. , et al .
2006 , astro - ph/0510520 law , d. r. , johnston , k. v. & majewski , s. r. 2005 , , 619 , 807 lenz , d. d. , newberg , h. j. , rosner , r. , richards , g. t. , & stoughton , c. 1998 , , 119 , 121 majewski , s. r. , skrutskie , m. f. , weinberg , m. d. , and ostheimer , j. c. 2003 , , 599 , 1082 martinez - delgado , d. , penarrubia , j. , juric , m. , alfaro , e. j. , ivezic , z. 2007 , ap .
j. suppl . , 660 , 1264 martinez - delgado , d. , gomez - flechoso , m. a. , aparicio , a. , carrera , r. 2004 , , 601,242 moultaka , j. , ilovaisky , s. a. , prugniel , p. , & soubiran , c. 2004 , pasp , 116 , 693 monaco , l. , bellazzini , m. , bonifacio , p. , buzzoni , a. , ferraro , f. r. , marconi , g. , sbordone , l. , & zaggia , s. 2006 , a&a , 464 , 201 newberg , h. , yanny , b. , et al .
2002 , , 569 , 245 newberg , h. , yanny , b. , et al .
2003 , , 596 , l191 newberg , h. , & yanny , b. 2005 , in asp conf .
ser . 338 : astrometry in the age of the next generation of large telescopes , ed .
p. k. seidelmann & a. k. b. monet , 210 , astro - ph/0502386 newberg , h. , & yanny , b. 2006 , in jpc conf .
ser . : physics at the end of the galactic cosmic ray spectrum , ed .
g. thomson & p. sokolsky , astro - ph/0507671 odenkirchen , m. et al .
2006 , , 126 , 2385 parker , j. e. , humphreys , r. m. , & larsen , j. a. 2003 , , 126 , 1346 pier , j. r. , munn , j. a. , hindsley , r. b. , hennessy , g. s. , kent , s. m. , lupton , r. h. , and ivezi , z. 2003 , , 125 , 1559 savage , c. , newberg , h. j. , freese , k. , & gondolo , p. 2006
, journal of cosmology and astroparticle physics , 7 , 3 schlegel , d.j . ,
finkbeiner , d.p . ,
& davis , m. 1998 , apj , 500 , 525 smith , j. a. et al .
2002 , , 123 , 2121 stoughton , c. , et al .
2001 , , 123 , 485 tucker , d. , et al .
2006 , astronomische nachrichten , 325 , 583 vivas , a. k. et al .
2001 , , 554 , l33 xu , y. , deng , l. c. & hu , j. y. 2006 , , 368 , 1811 xu , y. , deng , l. c. & hu , j. y. 2007 , , in press yanny , b. , newberg , h. j. , et al .
2000 , , 540 , 825 yanny , b. , newberg , h. j. , et al .
2003 , , 588 , 841 york , d.g .
2000 , , 120 , 1579 zinn , r. , vivas , a. k. , gallart , c. & winnick , r. 2004 , asp conf .
series , 327 , 92 rrrrrrrr 13 & 7.5 & 308.7 & 70.2 & 22.00 & 36 & 195 & 253 + 15 & 12.5 & 277.1 & 73.9 & 21.85 & 34 & 186 & 242 + 16 & 15.0 & 245.4 & 69.0 & 21.60 & 30 & 175 & 231 + 17 & 17.5 & 233.2 & 65.9 & 21.52 & 29 & 170 & 225 + 18 & 20.0 & 221.7 & 57.2 & 21.35 & 27 & 160 & 216 + 19 & 22.5 & 215.9 & 51.0 & 21.12 & 24 & 153 & 209 + 20 & 25.0 & 212.1 & 46.7 & 21.12 & 24 & 149 & 205 + 21 & 27.5 & 209.2 & 41.0 & 21.02 & 23 & 143 & 200 + 22 & 30.0 & 207.2 & 33.7 & 20.68 & 20 & 136 & 193 + 19 & 22.5 & 229.1 & 75.6 & 21.75 & 32 & 179 & 232 + 20 & 25.0 & 217.2 & 72.9 & 21.45 & 28 & 175 & 226 + 21 & 27.5 & 209.0 & 67.5 & 21.25 & 26 & 170 & 220 + 22 & 30.0 & 204.0 & 62.4 & 21.65 & 31 & 165 & 218 + 23 & 32.5 & 201.8 & 53.2 & 21.40 & 28 & 156 & 209 + | we investigate the relationship between several previously identified galactic halo stellar structures in the direction of virgo using imaging and spectroscopic observations of f turnoff stars and blue horizontal branch stars from the sloan digital sky survey ( sdss ) and the sloan extension for galactic understanding and exploration ( segue ) .
we show that the sagittarius dwarf leading tidal tail does not pass through the solar neighborhood ; it misses the sun by more than 15 kpc , passing through the galactic plane outside the solar circle .
it also is not spatially coincident with the large stellar overdensity s297 + 63 - 20.5 in the virgo constellation .
s297 + 63 - 20.5 has a distinct turnoff color and kinematics .
faint ( @xmath0 ) turnoff stars in s297 + 63 - 20.5 have line - of - sight , galactic standard of rest velocities @xmath1 , opposite in sign to infalling sgr tail stars . the path of the sgr leading tidal tail is also inconsistent with the positions of some of the nearer stars with which it has been associated , and whose velocities have favored models with prolate milky way potentials .
we additionally show that the number densities of brighter ( @xmath2 ) f turnoff stars are not symmetric about the galactic center , and that this discrepancy is not primarily due to the s297 + 63 - 20.5 moving group . either the spheroid is asymmetric about the galactic center , or there are additional substructures that conspire to be on the same side of the galaxy as s297 + 63 - 20.5 .
the s297 + 63 - 20.5 overdensity in virgo is likely associated with two other previously identified virgo substructures : the virgo stellar stream ( vss ) and the virgo overdensity ( vod ) .
however , the velocity difference between the vss and s297 + 63 - 20.5 and the difference in distance estimates between the vod and s297 + 63 - 20.5 must be reconciled . |
blue horizontal branch ( bhb ) stars have long been important objects of study as tracers of the galactic potential @xcite .
their constant absolute magnitude allows accurate space positions to be determined .
historically , many researchers have used samples of bhb stars , or other standard - candle stars such as rr lyraes , to study the local halo , i.e. the vicinity of the solar neighborhood where typical distances are of order the solar galactocentric radius @xmath10 .
these samples can be used to constrain the kinematics of the local halo @xcite . the distant halo ( i.e. typical distances @xmath11 ) , however , is less well constrained ( but see @xcite ) .
isolating a sample of distant bhb stars would enable study of the structure and kinematics of the galactic halo on larger scales ( * ? ? ?
* hereafter paper ii ) , and estimates of the mass of the galaxy .
the sloan digital sky survey ( sdss ) , currently halfway through a photometric and spectroscopic survey of a quarter of the sky , offers an excellent opportunity to isolate and study bhbs .
the sdss spectroscopic survey targets objects between about @xmath12 = 15.5 to 19 ( the bright limit is set to avoid contamination in adjacent spectroscopic fibers at the slit end of the fiber bundle ) , although a significant number of stars outside this range is also observed
. thus bhb stars at distances from @xmath13 to @xmath14 from the sun appear in the spectroscopic survey .
this paper presents @xmath15 bhb stars selected from sdss that probe the outer halo .
the plan of the paper is as follows .
pertinent characteristics of the sloan digital sky survey are summarized in [ s : sdss ] .
[ s : color_cut ] describes the preliminary color cut .
[ s : balmer_cut ] discusses spectroscopic bhb classification criteria . for the faint stars , the spectroscopic classification breaks down
so we resort to a more stringent color cut described in
[ s : fainter_stars ] .
radial velocity errors are evaluated in
[ s : duplicates ] . [ s : absmag ] discusses the calculation of bolometric corrections for accurate distances , and evaluates magnitude errors .
some sanity checks are presented in
[ s : sanity_checks ] , including an analysis of the limited proper motion data available , and a map of the galactic halo which shows the conspicuous sagittarius tidal stream ( figure [ f : map_3d ] ) .
the results are summarized in [ s : summary ] .
the sample is available in table [ t : data ] in the electronic version of this paper .
the sloan digital sky survey ( sdss : * ? ? ? * ) is a project to image about 1/4 of the sky at high galactic latitude to a depth of about 21 to 22 magnitudes in five broad bands covering the entire optical range and to obtain spectra of about a million galaxies and 100,000 quasars selected from the imaging data .
while galaxies and quasars are the primary targets of the spectroscopic survey , the sdss also obtains a vast amount of stellar data .
stellar spectra are acquired by several routes : observations of calibration stars , observations of various classes of star to be used as `` filler '' observations when a given region of sky has too low a density of primary targets , and via objects targeted as candidate galaxies and quasars that turn out to be stars .
the sdss uses a dedicated 2.5 meter telescope and a large format ccd camera @xcite at the apache point observatory in new mexico to obtain images almost simultaneously in five broad bands ( @xmath16 ) centered at 3551 , 4686 , 6166 , 7480 , and 8932 respectively @xcite .
the imaging data are automatically processed through a series of software pipelines which find and measure objects and provide photometric and astrometric calibrations to produce a catalogue of objects with calibrated magnitudes , positions and structure information .
the photometric pipeline photo ( * ? ? ? * and in preparation ) detects the objects , matches the data from the five filters , and measures instrumental fluxes , positions and shape parameters . the last allow the classification of objects as `` point source '' ( compatible with the point spread function [ psf ] ) , or `` extended . ''
the instrumental fluxes are calibrated via a network of primary and secondary stellar flux standards to @xmath17 magnitudes @xcite .
we use the `` psf magnitude , '' which represents the magnitude of the best - fit psf to the object , as determined by photo version 5.4.25 ( 2003 june ) .
the psf magnitudes are currently accurate to about 2% in @xmath18 , and @xmath19 and 3 - 5% in @xmath20 and @xmath21 for bright ( @xmath22 20 mag ) point sources .
the sdss is 50% complete for point sources to @xmath23 , and the full - width at half - maximum ( fwhm ) of the psf is about @xmath24 @xcite .
the data are saturated at about 14 mag in @xmath12 , @xmath25 and @xmath19 and about 12 mag in @xmath20 and @xmath21 .
astrometric calibration is carried out using a second set of less sensitive ccds in the camera , which allows the transfer of astrometric catalogue positions to the fainter objects detected by the main camera .
absolute positions are accurate to better than @xmath26 in each coordinate @xcite .
the sdss spectroscopic survey is carried out by targeting objects selected from the imaging catalogues according to pre - determined and carefully tested criteria @xcite .
the targeted stars , quasars and galaxies have limiting magnitudes of around @xmath27 mag @xcite .
bhb stars appear in the sdss spectroscopic sample via two main routes . at redshifts around 2.7
, the quasar color locus crosses the bhb color locus @xcite and many bhb stars are targeted as candidate quasars .
bhb stars are also photometrically selected to serve as filler observations in regions of the sky with a low density of primary targets .
the sky locations of the target objects are mapped to a series of plug plates which feed a pair of double spectrographs ( a. uomoto , s. smee et al .
unpublished ) via optical fibers .
the spectrographs jointly observe 640 objects ( including sky positions and standard stars ) simultaneously , and provide continuous wavelength coverage from about 3800 to 9500 at a resolution of about @xmath28 .
the nominal total exposure time for each spectroscopic plate is 45 minutes ( 15 minutes @xmath29 3 exposures ) , plus observations of flat field and arc standard lamps for every plate .
the entrance aperture of the fibers is @xmath30 .
the spectra are matched to the photometric objects using a video mapping system for the plug plates ( d. schlegel and d. finkbeiner , unpublished ) .
the spectra are recorded on ccd detectors .
the data are optimally extracted , the sky spectrum is subtracted , and wavelength , flat - field and spectrophotometric calibrations are applied automatically by the spectro-2d software pipeline , written by d. schlegel and s. burles ( unpublished ) .
each spectroscopic plate is a circle of diameter @xmath31 on the sky .
the present work is based on the data of @xmath32 unique spectroscopic plates taken to date , through 2003 aug 6 .
in addition , @xmath33 plates were observed redundantly , but in defining the bhb sample we exclude redundant observations by adopting the plate with the highest s / n .
however , we use the information from these redundant observations in [ s : duplicates ] below to evaluate velocity errors .
the radial velocity is found by fitting the spectrum to template spectra , using an automated fitting program ( * spectro_brightstars * ) written by d. schlegel ( unpublished ) .
the basic procedure is as follows .
@xcite have published a catalogue of about a thousand spectra of stars with a wide range of properties ( spectral types , metallicities etc . ) observed at high spectral resolution with the elodie spectrograph .
one version of the elodie spectra is binned to r @xmath34 20,000 and spectrophotometrically calibrated to about 5% .
* brightstars * was used to select template spectra from the sdss data which match the elodie spectra .
these template spectra are typically composites from many sdss spectra , and are set up to provide relatively coarse sampling of the spectral type grid .
the template spectra are aligned to zero heliocentric velocity . each sdss spectrum
is then compared with the template spectra in wavelength space , and the template and wavelength shift that give the best match is automatically measured .
the stars in the present sample were almost all best - fit with spectral types between b6 and a0 .
the radial velocities are corrected to the heliocentric standard of rest , and throughout this paper are denoted @xmath35 .
bhb stars have very characteristic colors in the high - galactic latitude stellar sample observed by the sdss imaging survey , and photometric selection of candidate objects is straightforward .
bhb stars are bluer in @xmath36 than most halo stars because the halo main sequence turnoff lies at spectral type f / g .
the large balmer jump in bhb low - gravity stars gives them redder @xmath37 colors than other blue objects such as low redshift quasars and white dwarfs , and indeed the sdss @xmath20 and @xmath12 filters are specifically designed to provide optimal photometric separation of low - redshift quasars @xcite .
the sdss @xmath36 vs. @xmath37 color - color diagram in figure [ f : ugr_g18 ] shows all objects in the spectroscopic sample of @xmath32 plates used herein that were subsequently identified as stars by * brightstars * , which is accurate enough that quasars and galaxies are virtually nonexistent in the sample .
( throughout this paper all magnitudes are corrected for galactic extinction using the maps of @xcite . )
the region occupied by bhb stars is indicated with a `` color cut box , '' with boundaries @xmath38 .
the boundaries of this selection box are similar to those used by @xcite , and figure [ f : ugr_bhb18 ] ( see [ s : fainter_stars ] ) shows that our final sample of bright bhb stars is well contained within the boundaries .
stars that lie in this region of color - color space pass what we will refer to as the `` color cut . '' although bhb stars are effectively selected through the color cut described above , there remains a significant degree of contamination from two other types of star .
the locus of cooler f stars is actually quite outside of the color cut box , to the red in @xmath36 @xcite , but there are so many f stars that they leak into the bhb candidate sample , via intrinsic variations and photometric errors , with numbers comparable to the number of bhb stars . to identify these cooler stars
merely requires a measure of temperature independent of @xmath39 or @xmath40 .
somewhat more challenging to identify are the a - type stars with higher surface gravity @xcite .
these are probably field blue stragglers , which are some 2 magnitudes less luminous @xcite .
while the balmer jump as measured by @xmath37 can separate out many of these stars , some remain in the sample .
they can be removed using their broader balmer lines as described in the next section .
the upper panel of figure [ f : bhb_spectrum ] presents a typical bhb spectrum , showing the deep balmer lines of an a - type star . in this section
we describe two methods for analyzing these lines : the widely used @xmath41 method @xcite and the scalewidth - shape method @xcite .
we focus on the h@xmath42 and h@xmath43 lines because the the h@xmath44 and h@xmath45 lines have lower continuum levels , and the lines above h@xmath43 are much more closely spaced , making the determination of the continuum difficult @xcite . to analyze the lines , the continuum must first be fitted and divided out . for the @xmath41 method ,
following @xcite , we take the continuum to range from 4000 to @xmath46 with @xmath47 masks covering the positions of the h@xmath42 and h@xmath43 lines .
we then divide the spectrum by the best - fit sixth - order legendre polynomial to this extracted continuum .
finally , the spectrum is boxcar - smoothed with a boxcar length of 5 pixels .
an example normalized and extracted sed is shown in the lower panel of figure [ f : bhb_spectrum ] .
similarly , for the scalewidth - shape method , we take the continuum extraction rule of @xcite ( which ranges from 3863 to @xmath48 ) and fit to a fourth - order legendre polynomial .
after the continuum is divided out with this prescription , the sed looks very similar to the example sed in the lower panel of figure [ f : bhb_spectrum ] .
the following terminology is used in this section and throughout .
a `` method '' refers to a general algorithm , which may or may not be desirable for isolated use , which employs one or more `` criteria '' for discriminating bhb stars .
the @xmath41 method and the scalewidth - shape method are presented in the following .
a `` cut '' is a method or combination of methods that we _ actually _ use to eliminate contaminants from the sample .
the combination cut discussed below incorporates the @xmath41 and scalewidth - shape methods .
since bhb stars have low surface gravity , their balmer lines are narrower than the smaller main sequence blue stragglers .
the @xmath41 method discriminates bhb stars from bs stars by determining the value of @xmath41 , the width in angstroms of the balmer line at 80% of the continuum @xcite .
figure [ f : d02_demo ] demonstrates how this parameter can discriminate a bhb from a bs star .
although simple , this method has been found to be fairly robust and can provide a better measurement of the linewidth than other more complicated procedures ( e.g. * ? ? ?
the @xmath41 method nominally only applies to objects with a - type spectra .
however , as mentioned in @xmath49[s : color_cut ] , a large contaminant of our sample is f stars .
since f - type stars have weaker balmer lines than a - type stars , one way of measuring temperature is by measuring the depth of the h@xmath42 or h@xmath43 line .
here we define @xmath50 as the flux relative to the continuum at the minimum of the line .
thus , higher values of @xmath50 for the balmer lines signify cooler stars .
a plot of @xmath41 vs. @xmath50 for stars passing the color cut is shown in figure [ f : d02_v_fm ] . in this figure , only stars with @xmath12 magnitudes @xmath51 are shown , so that the trends in the plot are not obscured by the noise in the data of fainter stars ( see figure [ f : g18_breakdown ] , to be discussed in
[ s : criteria ] ) .
the concentration of stars near @xmath52 represents the bhb stars , and those stars with larger values of @xmath41 are bs stars .
figure [ f : d02_v_fm ] also suggests that the locii of bs and bhb stars propagate to larger @xmath50 and smaller @xmath41 until the two locii converge at about @xmath53 . in other words , for a given @xmath54 , the distribution of stars in @xmath41 is bimodal , but the stars are most cleanly separated for the smaller values of @xmath50 and not separated at all when @xmath55 .
even though figure [ f : d02_v_fm ] does nt show the trend clearly , we return to this point in [ s : criteria ] and show that the stars with smaller values of @xmath41 but with @xmath56 are the other horizontal branch stars ; rr lyraes , for example , fall into this region of the horizontal branch . because horizontal branch stars are evidently more difficult to discriminate from main sequence stars at lower temperatures , we adopt the criteria @xmath57 for likely bhb stars for the h@xmath43 @xmath41 method .
@xcite have proposed the scalewidth - shape method for discriminating bhb stars based on the srsic profile for balmer lines @xcite : @xmath58}.\ ] ] here , @xmath59 is the normalized flux and @xmath60 is the wavelength in @xmath61 .
we fit the normalized extracted spectrum ( where the bounds of extraction are given by @xcite ) to the srsic profile with three free parameters : @xmath62 and @xmath63 .
the parameter @xmath64 is assumed to be the nominal location of the balmer line corrected for redshift , which has already been determined .
after fitting both balmer lines of every star to the three - parameter ( @xmath65 ) srsic profile , there is a reasonably small dispersion in the determination of @xmath66 : for the @xmath67 line the average is @xmath68 and the root - mean - square ( rms ) is @xmath69 ; for @xmath70 the average is @xmath71 and the rms is @xmath72 .
we fix @xmath66 at its average value among stars passing the bhb criteria of the h@xmath43 @xmath41 method , which is not objectionable because the @xmath41 method is independent of the scalewidth - shape method .
the balmer lines are then refitted to the two - parameter ( @xmath73 ) srsic profile .
this refitting is important because errors in @xmath66 are correlated with the errors in @xmath74 and @xmath63 @xcite . for this work
we adopt @xmath75 for the h@xmath42 line and @xmath76 for the h@xmath43 line .
the optimum value of @xmath66 most likely depends on the resolution of spectroscopy and similar factors .
fitting to the srsic profile thus provides the two parameters @xmath74 and @xmath63 for both the h@xmath42 and h@xmath43 lines , for every star .
@xcite show that bhb and bs stars separate rather cleanly in a plot of @xmath74 against @xmath63 .
furthermore , they show that @xmath63 is a measure of temperature : cooler stars have smaller values of @xmath63 .
we concentrate on the set of brighter stars with @xmath1 because the noise in the spectra of fainter stars throws large errors in the parameter determinations for the @xmath41 method and the scalewidth - shape method .
this point is illustrated in figure [ f : g18_breakdown ] .
we will return to the selection of bhbs from the fainter star sample in
[ s : fainter_stars ] .
figure [ f : compare1 ] shows the parameters of the sample of bhb candidates with @xmath1 as determined by the two linewidth analysis methods described in the preceding sections .
each star is represented by the same color in all four panels . in this figure
the h@xmath43 @xmath41 method is represented in the lower left panel , and stars which pass the bhb criteria for this method are colored blue , stars which are likely blue stragglers are green , and all others ( cooler stars ) are red .
it is readily apparent that the four _ independent _ parameter spaces map bhb stars , bs stars , and cool stars to different regions .
the exact bhb criteria for each method were determined through a combination of adjusting the bounds of the criteria by eye and comparing the results of one method to the others .
for instance , the scalewidth - shape method separates bhb and bs stars cleanly enough that there is a natural gap between the two distributions in a plot of @xmath74 vs. @xmath63 , as shown in figure [ f : compare1 ] .
a parabola is qualitatively drawn to divide the two distributions .
similarly , with regard to the @xmath41 method , the distribution in @xmath41 of the stars with low values of @xmath50 naturally separates into the two populations of bhb and bs stars .
the issue of where to draw the temperature criterion ( @xmath50 for the @xmath41 method and @xmath63 for the scalewidth - shape method ) is more sensitive because the `` natural '' gap is harder to see in the @xmath41 method and seems to be nonexistent in the scalewidth - shape method .
in fact , the upper right panel of figure [ f : compare1 ] hints that the division between stars with low surface gravity and high surface gravity continues even for much cooler stars than bhb stars .
the cooler population ( @xmath77 ) with lower surface gravity ( lower @xmath78 ) is thus the horizontal branch stars which are not blue , e.g. rr lyraes .
this @xmath74 vs. @xmath63 plot corroborates the evidence in
[ s : d02_method ] that non - blue horizontal branch stars form a locus in the @xmath79 plane which extends out to cooler temperatures corresponding to @xmath55 .
figure [ f : compare_rhb ] shows that the stars of the population that has smaller values of @xmath78 also have smaller values of @xmath41 . with this information
, we deduce that the bhb stars are located in the clump in the h@xmath43 @xmath80 plane centered on @xmath81 and we set the temperature criterion at @xmath82 . a subsequent inspection of figure [ f : compare1 ]
reveals that most bhb stars have @xmath83 , so we adopt this as the temperature criterion for the h@xmath42 scalewidth - shape method .
table [ t : criteria ] lists the adopted bhb selection criteria for each method .
we define the combination cut as those stars which pass the criteria for ( the h@xmath42 @xmath41 and h@xmath43 scalewidth - shape methods ) or ( the h@xmath43 @xmath41 and h@xmath42 scalewidth - shape methods ) .
we assume that the h@xmath42 @xmath41 and h@xmath43 scalewidth - shape methods are completely independent of each other , as are the h@xmath43 @xmath41 and h@xmath42 scalewidth - shape methods .
this assumption is reasonable because no information is shared in the analyses of either line , and the @xmath41 and scalewidth - shape methods measure fundamentally different properties of the spectral energy distribution . under this assumption ,
the contamination fraction for the combination cut would be @xmath84 , where @xmath85 is the contamination fraction of a single method . in appendix [ s : contamination ] we show that the contamination fraction of any one of the four methods is probably @xmath86 , so @xmath87 under these assumptions .
however , to be more conservative , the fact that @xmath88 for all four individual methods suggests that @xmath89 is probably @xmath4 or less ; in paper ii we use @xmath90 to study errors introduced by contamination into the kinematic analysis .
as shown in figure [ f : g18_breakdown ] , the linewidth analysis methods can not be expected to perform accurately for fainter stars than @xmath91 ; in fact the number of @xmath3 stars that are identified differently in any two methods ( @xmath92 ) is of the same order as the number of @xmath3 stars identified as bhb stars in one method ( @xmath93 ) . because these fainter stars have noisier spectra , not only are the algorithmic bhb criteria unreliable , but eye - inspection would be difficult at best
. however , an inspection of the results of the combination cut applied to the bright ( @xmath1 ) stars , as shown in figure [ f : ugr_bhb18 ] , reveals that bhb stars , bs stars , and cooler stars occupy different regions of color - color space , although there is significant overlap . in this
figure bhb stars are defined as those which pass the combination cut ; ( green ) bs stars are defined as those which do not pass the combination cut but lie in the bs region for the h@xmath43 @xmath41 method ; and red stars are everything else . the thick curved line with periodic circles will be explained in [ s : absmag ] .
the bounds of this particular plot are the same as the color cut discussed in [ s : color_cut ] , so it is evident that the bhb stars are more likely to be found in the region enclosed by the `` piano '' shape indicated than outside of it . since sdss photometry can be expected to be accurate to a fainter magnitude ( @xmath94 , * ? ? ?
* ) than can the combination method ( see figure [ f : g18_breakdown ] ) , we select bhb stars with @xmath3 solely on the basis of color . the new color cut is indicated by the piano shape in figure [ f : ugr_bhb18 ] ( and is given by the region @xmath95 excluding the elliptical subregion @xmath96 ) and will be referred to here as the `` stringent '' color cut .
what are the contamination and incompleteness for the stringent color cut with no spectroscopic analysis ? assuming that the combination cut is 100% accurate for @xmath1 , there are @xmath0 bhb stars , @xmath97 of which do not pass the stringent color cut .
there are @xmath98 stars that pass the stringent color cut that are not classified as bhb stars by the combination method .
most of these ( @xmath99 ) are classified as blue stragglers by the h@xmath43 @xmath41 method .
thus the contamination for the stringent color cut alone is @xmath100 .
since the spectroscopic plates overlap on the sky , sometimes objects are targeted for spectroscopy on two different plates . of the @xmath101 stars passing the color cut ( from _ unique _ plates ) , @xmath102 are duplicate observations of the same star on different plates .
we remove these @xmath102 redundant observations from our bhb sample .
in addition , from the data of the @xmath33 non - unique plates ( see [ s : sdss ] ) , there are @xmath103 stars passing the color cut which duplicate observations of stars from the unique plates .
each spectroscopic observation of a star provides an independent measurement of radial velocity @xmath35 and , for the bright ( @xmath1 ) stars , an independent classification as a bhb star or non - bhb star .
figure [ f : duplicates ] shows the @xmath104 stars which have multiple spectra and therefore independent determinations of @xmath35 .
the ordinate axis shows the difference in the radial velocity determinations @xmath105 between each duplicate star and its corresponding star in our purified sample of unique stars .
stars which are classified as bhbs for both observations are colored blue , stars which are classified as non - bhbs for both observations are colored red , and stars which have conflicting classifications between the two observations are colored green .
brighter than @xmath106 , there are @xmath107 stars which are doubly classified as bhbs , @xmath108 stars which are doubly classified as non - bhbs , and @xmath109 stars which have conflicting classifications .
this is consistent with the @xmath4 contamination determined in appendix [ s : contamination ] , assuming that there is about an equal number of contaminants in the bhb sample as the number of bhb stars which are not classified as such ( i.e. , assuming @xmath110 ) .
figure [ f : duplicates ] also shows the dependence of velocity errors on magnitude ; as expected , fainter stars tend to have greater velocity errors , presumably due to noiser spectra .
the rank correlation coefficient of @xmath111 is @xmath112 .
the velocity differences determined this way turn out to be consistent with the errors derived from the spectral fits for only about half of the stars , independent of the magnitude or whether or not a star is a bhb or bs star .
this comparison implies that the errors have been underestimated and this is currently under investigation .
the rms of @xmath105 in figure [ f : duplicates ] is @xmath113 .
clearly the errors in radial velocity measurements are not gaussian , but a meaningful approximate result may still be obtained if they are assumed to be so .
since the distribution of @xmath105 is based on the errors of two measurements , the error of one measurement is approximately @xmath114 .
in order to refine our photometric distance estimates , we have computed theoretical colors and bolometric corrections for bhbs .
we integrate the latest kurucz model atmospheres ( downloaded from www.kurucz.harvard.edu ) against the sdss filter response functions at a standard airmass of 1.2 atmospheres , in the manner described by @xcite . unfortunately
, the latter authors report only colors , so it was necessary to repeat their calculations in order to obtain the bolometric corrections . apart from wavelength ,
the kurucz spectra are functions of three parameters : effective temperature @xmath115 , surface gravity @xmath116 , and metallicity relative to solar @xmath117 $ ] .
to derive absolute magnitudes and bolometric corrections , it is necessary to assume a relationship between the luminosity @xmath118 and @xmath115 .
the models of @xcite for the theoretical zero - age horizontal branch ( zahb ) at @xmath117\le-1.5 $ ] can be summarized roughly by @xmath119 over the range @xmath120 . from a detailed study of the horizontal branch of m5 , however , @xcite find that luminosity varies more weakly with @xmath115 ; in fact their results appear to be consistent with constant luminosity and mass , @xmath121 with a scatter of no more than @xmath122 and @xmath123 , respectively , over the same temperature range .
we adopt the mean values in equation ( [ e : mlrels ] ) , so that @xmath124 . to give an idea of the theoretical uncertainties , we note that if in fact the relation ( [ e : zahb ] ) holds for our hb stars , then by adopting ( [ e : mlrels ] ) we have overestimated their distance moduli by an average of @xmath125 over the indicated range of @xmath115 , and by neglecting the variation of @xmath118 with @xmath115 ,
we have increased the scatter in distance modulus by @xmath126 , which should be taken in quadrature with the photometric measurement error .
the results of this exercise are shown in table [ t : thb ] .
evidently , for the colors of interest to us , @xmath127 , and the influence of metallicity is slight if @xmath117\le -1 $ ] .
the sequence of theoretical horizontal branch stars , parametrized by temperature , is projected onto the @xmath128 plane in figure [ f : ugr_bhb18 ] .
the five black circles on the thick line denote , starting at the top and progressing down and to the left , stars with absolute magnitudes @xmath129 . for every observed bhb star in our sample ( colored blue in the figure )
we define the most probable absolute magnitude associated with its @xmath128 colors by simply finding the absolute magnitude of the point on the theoretical track that is closest to the observed star in this color - color space .
the stringent color cut ( piano shape in figure [ f : ugr_bhb18 ] , see [ s : fainter_stars ] ) was defined
before the theoretical track was calculated , so the fact that the theoretical track satisfies the stringent color cut is a meaningful consistency check . as a further consistency check
, we have isolated several galactic globular clusters observed by the sdss photometric survey to date and plotted their color - magnitude diagrams ( cmds ) .
pal 3 , ngc 5904 , and ngc 6205 do not show a discernable horizontal branch , so they are not considered here .
the @xmath12 vs. @xmath36 cmds of pal 5 @xcite and ngc 2419 are presented in figure [ f : gc_cmd ] , where the distance modulus , derived from cluster distances of @xmath130 and @xmath131 for ngc 2419 and pal 5 , respectively @xcite , has been subtracted from the dereddened apparent magnitude @xmath12 .
stars within @xmath132 of the center of ngc 2419 and within @xmath133 of the center of pal 5 were selected .
the sdss automated photometry avoids analyzing very crowded fields , so the stars presented here , though within the stated radii , tend to lie in the outskirts of the clusters
. vertical lines at @xmath134 indicate our color cut . the theoretical horizontal branch sequence is also plotted in both panels .
the horizontal branch stars of ngc 2419 appear to be systematically fainter than the theoretical prediction , while those of pal 5 seem to be systematically brighter ; however , these systematic differences could easily be due to inaccurate distances @xmath135 . by selecting from these two globular clusters the stars passing the color cut ( @xmath38 ) and lying in the region of the cmd for the horizontal branch ( @xmath136 )
, we compare the directly observed absolute magnitude @xmath137 to the indirectly obtained theoretical absolute magnitude based on the @xmath128 colors . for ngc 2419 , the rms of the differences of
the two among the 94 stars is .134 magnitudes , and the root - sum - square ( rss ) is .167 magnitudes , @xmath138 , @xmath139 . for pal 5 ,
the rms is .126 magnitudes and the rss is .202 magnitudes .
the difference between the rms and rss is due to the nonzero mean , or systematic error as discussed above .
therefore , the magnitude error of a star , with absolute magnitude determined by comparison to the theoretical horizontal branch sequence described above , is of order .2 magnitudes
. the error may be as small as @xmath140 magnitudes if the only cause of the systematic difference between directly observed absolute magnitudes and our indirectly obtained absolute magnitudes is incorrect distances for pal 5 and ngc 2419 .
as figure [ f : ugr_bhb18 ] shows , bhb stars determined by the combination cut are in general redder in @xmath39 and bluer in @xmath40 than other objects such as blue stragglers .
this result is consistent with the work of @xcite , so it is reassuring . in this section we present several more sanity checks which verify that our sample of bhb stars is probably accurate .
bhb stars are halo stars , and as such they should have lower metallicity than the disk star contaminants .
the most promising spectral feature to use for a metallicity determination is the caii k line at @xmath141 ( the h line , at @xmath142 , is not used because it is too close to h@xmath143 ) .
unfortunately , it is impractical to use the caii k line as a _ reliable _ indicator of metallicity in individual sdss spectra because much higher spectral resolution would be necessary .
however , the following exercise is executed anyway as a sanity check . by employing a continuum - subtraction algorithm very similar to
that described in [ s : d02_method ] for the h@xmath42 and h@xmath43 balmer lines , we find the value of @xmath50 for the caii k line ; smaller values indicate larger metallicity . figure [ f : caii ] presents the distributions of this parameter for our sample of bhb stars and non - bhb stars passing the color cut .
the figure shows different distributions for bright ( @xmath1 ) and faint ( @xmath3 ) stars because of the difference in s / n .
reassuringly , our bhb stars tend to have lower metallicities .
the proper motions of objects in sdss are found automatically by matching the stellar positions to those in the astrometric catalogue usno - a v2.0 @xcite derived from the digitized palomar observatory sky survey ( poss ) plates , for objects bright enough to have a poss counterpart .
the normalized distribution of the proper motion of spectroscopically confirmed qsos is shown in figure [ f : pm_dist ] as a thick line .
these alleged proper motions of qsos are obviously due to measurement errors , both from poss and sdss ( primarily from the former ) . if both measurements follow bivariate gaussian errors with radially symmetric variances @xmath144 and @xmath145 , then the distribution of the separation of the measurements ( i.e. , proper motion ) also follows a bivariate gaussian with variance @xmath146 .
thus , the probability density distribution of observing a star with true proper motion @xmath147 at an observed proper motion @xmath148 is @xmath149 where @xmath150 is the modified bessel function of the first kind . since @xmath151 for the qso distribution , the second moment of the distribution of @xmath148 should equal @xmath152 for gaussian statistics .
this determination yields @xmath153 , and the resulting theoretical error distribution function from equation [ e : proper_motion ] is shown in figure [ f : pm_dist ] as a thin line .
we have estimated the statistical intrinsic proper motions of our sample using the derived photometric distances .
the transverse velocity of each star was drawn 1000 times at random from an isotropic velocity distribution with dispersion @xmath154 in one dimension and corrected for an assumed solar velocity @xmath155 @xcite .
these choices were based on the kinematic results presented in paper ii .
the distribution of intrinsic proper motions so derived , @xmath156 , is shown in figure [ f : pm_dist ] as the dotted line .
the theoretical proper motion distribution for bhb stars with measurement errors should follow the form @xmath157 and is drawn in figure [ f : pm_dist ] as the dashed line .
the distribution of _ measured _ proper motions of our bhb sample is also shown .
this distribution is clearly not very much broader than that of the qsos ; the second moment yields @xmath158 , which is actually too small to account for the combination of measurement error ( @xmath153 ) and intrinsic proper motion ( @xmath159 ) ! in contrast , the proper motion distribution of a sample of faint ( here @xmath160 ) f stars from sdss is also shown , with a much broader distribution .
this finding demonstrates that contamination in the bhb sample from f stars must be very small .
we estimate that the f star contamination is less than @xmath161 among the 837 stars in our sample that were matched to usno - a . without better understanding of systematic errors
, we can not claim to have detected proper motion of the bhb sample .
however , the usno - b catalog , which has appreciably better data quality than usno - a , will soon be provided as the first epoch for proper motions , so we defer further analysis to later work .
figure [ f : map_3d ] shows two views of a three - dimensional representation of the locations of the bhb stars .
each star is colored according to its line - of - sight radial velocity with the solar velocity of @xmath162 @xcite added .
these two views compose a stereograph which can be viewed by focusing the eyes on a point midway to the paper .
the three - dimensional positions of the stars are derived from their angular positions and photometric distances ( [ s : absmag ] ) , assuming @xmath163 .
one could search for substructure in the galactic halo by a careful examination of figure [ f : map_3d ] , and indeed a clump appears in the upper left part of the plot ( at @xmath164 ) .
this clump belongs to the leading arm of the sagittarius dwarf tidal stream @xcite .
it is interesting to investigate this structure in velocity phase space , as shown in figure [ f : sgr_d_v_v ] .
the sagittarius stream approximately follows the great circle on the celestial sphere with pole @xmath165 or @xmath166 , as determined from the distribution of sdss candidate rr lyrae stars .
practically the same position is implied by the distribution of 2mass m giants @xcite .
this great circle intersects the footprint of the sdss in three places .
these three intersections with bhb stars ( two in the north and one in the south , with a strip width of 10 degrees ) are represented in the three panels in figure [ f : sgr_d_v_v ] .
one can see a clump in the phase space of the northern subsample representing the sagittarius stream ( left plot ) ; indeed it is the same clump in the upper left of figure [ f : map_3d ] . the coherent structure in the middle panel ( at @xmath167 )
is also associated with the sgr dwarf tidal stream ( trailing arm ) .
this is the same clump that was discovered by @xcite using a - colored stars , and is also detected in the velocity - distance distribution of sdss candidate rr lyrae stars @xcite .
the fact that we see such structure in the bhb sample strongly indicates that our bhb absolute magnitudes and selection criteria are reliable .
using broadband colors and spectroscopic surface - gravity indicators , we have selected @xmath15 blue horizontal branch stars from the sdss .
for the @xmath0 bright stars ( @xmath5 ) the contamination is probably @xmath168 , and for the @xmath2 faint bhb stars ( @xmath7 ) the contamination is probably @xmath169 .
the absolute magnitude of bhb stars was determined from globular cluster color - magnitude diagrams , and an absolute - magnitude - color relation established .
this was used to estimate photometric distances , with errors @xmath170 mag .
the stars also have measured radial velocities with errors @xmath8 .
proper motions determined by matching against the digitally scanned palomar sky survey ( usno - a , * ? ? ?
* ) also indicate that contamination from the disk is small .
we emphasize , however , that no kinematic information was used to select the sample .
the data can be downloaded as table [ t : data ] in the electronic version of this paper .
we thank a. gould , b. paczyski , r. lupton , and j. hennawi for helpful discussions , and m. strauss for a close reading of a draft of this paper .
grk is grateful for generous research support from princeton university and from nasa via grants nag-6734 and nag5 - 8083 .
funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , the university of pittsburgh , princeton university , the united states naval observatory , and the university of washington .
there is no independently confirmed subsample of bhb stars , but we can estimate the accuracy of our spectroscopic selection methods by intercomparing their results . after applying our color cut ,
we are left with @xmath171 candidates bright enough to be subjected to our spectroscopic selection criteria . of these
, @xmath172 are true bhb stars , and the other @xmath173 are potential contaminants .
let @xmath174 be the probability that method @xmath19 fails to identify a true bhb as such , and @xmath175 be the probability that it mistakenly accepts an individual member of the non - bhb group .
then the expected number of stars accepted by method @xmath19 is @xmath176 similarly , the numbers of stars selected jointly by methods @xmath19 & @xmath177 , or by @xmath19 , @xmath177 , & @xmath178 , are @xmath179 respectively , assuming that these methods are truly independent .
it is convenient to define @xmath180 so that the number of contaminants among the @xmath181 stars accepted by method @xmath19 is @xmath182 . then @xmath183n_in_j \\ \label{e:3rdorder } n_{i\cap
j\cap k}&= & \left[\frac{(1-\kappa_i)(1-\kappa_j)(1-\kappa_k)}{n^2 } ~+\frac{\kappa_i\kappa_j\kappa_k}{c^2}\right]n_in_jn_k\end{aligned}\ ] ] as described in [ s : balmer_cut ] , we have four methods .
therefore we have six double intersections ( [ e:2ndorder ] ) and four triple intersections ( [ e:3rdorder ] ) with which to solve for the six unknown parameters @xmath184 that they contain .
so the problem is overdetermined , allowing us perhaps to test the assumption of independence .
perfect consistency is not to be expected since even for fixed values of the above parameters , the numbers @xmath185 are subject to statistical fluctuations .
if the contamination fractions @xmath186 are reasonably small it is sufficient that @xmath187then the terms involving @xmath173 in eqs .
( [ e:2ndorder])-([e:3rdorder ] ) are unimportant , so that @xmath188 this approach yields the results shown in table [ t : accuracy ] , and also an estimate for the true number of bhbs tested ( _ i.e _ , not including those too faint for spectroscopic selection ) : @xmath189 .
note that there are three possible choices of @xmath190 with which to estimate @xmath186 from eq .
( [ e : nratios ] ) , all of which are shown in the table to give an idea of their consistency .
it can be seen that the contamination and incompleteness fractions for each are indeed small , @xmath191 , if these methods are indeed independent as we have assumed . in support of the latter , the differences among the estimates in columns 2 - 4 are @xmath192
as would be expected by chance since , for example , binomial statistics predict that the number of false negatives for method @xmath19 should fluctuate by @xmath193 , or about 7 out of 772 if @xmath194 .
( for clarity ) which were subsequently confirmed as stars .
the main sequence runs from the center of this diagram towards the upper right , and the large balmer jump of a - colored stars places them in the offshoot , where our `` color cut '' selection box is drawn .
note that the relative densities of stars in this plot are affected by selection bias ; for example , more stars appear here which lie in the more `` interesting '' regions of color - color space for qsos . ] and @xmath41 as determined by the h@xmath43 @xmath41 method for stars that pass the color cut with @xmath1 .
the trail of stars with @xmath195 are too cool to be bhb stars , and the concentration of stars with @xmath196 is due to blue stragglers with higher surface gravity . the @xmath197 stars that lie well above the main locus for @xmath195 are placed there by poor parameter determinations due to missing spectroscopic data at the location of the h@xmath43 line . ]
@xmath41 method ( left panels ) and of the h@xmath42 scalewidth - shape method ( right panels ) for five @xmath12 magnitude bins , labelled .
errors evidently become too large to safely discriminate bhb stars from other types of stars at about @xmath198 . ]
method ( left panels ) and the scalewidth - shape method ( right panels ) for the h@xmath42 line ( top panels ) and the h@xmath43 line ( bottom panels ) .
each star is represented by the same color in all four plots , so the compatibility among all four methods is seen . here
stars are colored according to their classification by the @xmath41 method for the h@xmath43 line : bhb stars are blue , bs stars are green , and the cooler stars are red .
only stars with @xmath1 are plotted here . ]
@xmath41 method for stars with @xmath1 .
stars are colored according to their location on the @xmath199 diagram for the h@xmath42 scalewidth - shape method : stars for which @xmath200 are colored blue ; all others are colored red .
( the temperature criterion is not used . )
thus stars which have low values of @xmath74 from the scalewidth - shape method also have low values of @xmath41 , and are likely horizontal branch stars . ] but zoomed into the color cut box and with stars colored according to their determined classifications by the combination cut . only stars with @xmath1
are shown .
bhb stars are blue , bs stars are green , and all others ( f stars , cooler horizontal branch stars ) are red .
note that the three samples do comprise different regions of color space , suggesting that purely photometric separation criteria are possible , but that the overlap between the regions is quite significant .
the sequence of theoretical horizontal branch models discussed in [ s : absmag ] is plotted as the thick curved line .
the five black circles on the thick line denote , starting at the top and progressing down and to the left , stars with absolute magnitudes @xmath129 . ] from duplicate observations of the same stars vs. @xmath12 magnitude .
the velocity error increases for faint stars , as expected .
stars are colored blue if both observations indicate a bhb star ( using the combination cut for @xmath5 and the stringent color cut for @xmath7 ) , red if both indicate a non - bhb star , and green if the two observations are in conflict .
the reason why there are some conflicting classifications for @xmath7 is because the two spectroscopic observations sometimes actually reference different photometric observations when available . ] parameter for the caii k line for bhb stars and non - bhb stars .
although bhb stars clearly have larger values of @xmath50 and thus lower metallicity implying that they are halo objects , they are not so cleanly separated from the non - bhb stars . this is probably because sdss spectroscopy does not have high enough resolution and s / n to probe the caii k line as accurately as it can the balmer h@xmath42 and h@xmath43 lines . ] .
our sample of @xmath15 bhb stars is plotted as jagged ( not boxy ) for clarity .
the dotted line is the expected distribution of true proper motions for the bhb sample .
the dashed curve is the expected _ measured _ distribution of bhbs given the theory of measurement errors described in the text .
the close match between the qsos and bhb stars qualitatively indicates that contamination from foreground objects is low . ] .
the left plot , from the north galactic cap , clearly shows a clump in phase space at @xmath203 which represents the sagittarius stream . the middle plot , from the south , can also be seen to show similar structure .
the right plot is from the north but the number of stars is too small to make any conclusions about the detection of the sagittarius stream in this region of space . ]
lll @xmath41 method ( h@xmath42 ) & @xmath204 & @xmath205 + @xmath41 method ( h@xmath43 ) & @xmath206 & @xmath205 + scalewidth - shape method ( h@xmath42 ) & @xmath207 & @xmath208 + scalewidth - shape method ( h@xmath43 ) & @xmath209 & @xmath210 + cccccc 7000 & 2.900 & 0.7115 & 1.0335 & 0.0945 & 0.0736 + & & 0.6862 & 1.0836 & 0.0999 & 0.0724 + [ 1.0ex ] 8000 & 3.132 & 0.5712 & 1.1792 & -0.1313 & -0.2700 + & & 0.5494 & 1.1944 & -0.1291 & -0.2724 + [ 1.0ex ] 9000 & 3.336 & 0.6311 & 1.0695 & -0.2300 & -0.4301 + & & 0.6121 & 1.0732 & -0.2317 & -0.4351 + [ 1.0ex ] 10000 & 3.520 & 0.8000 & 0.8911 & -0.2742 & -0.5023 + & & 0.7753 & 0.8853 & -0.2780 & -0.5093 + [ 1.0ex ] 11000 & 3.685 & 0.9954 & 0.7381 & -0.2999 & -0.5449 + & & 0.9672 & 0.7239 & -0.3034 & -0.5512 + [ 1.0ex ] 12000 & 3.836 & 1.1884 & 0.6139 & -0.3203 & -0.5785 + & & 1.1582 & 0.5950 & -0.3232 & -0.5836 + [ 1.0ex ] 13000 & 3.975 & 1.3735 & 0.5103 & -0.3382 & -0.6075 + & & 1.3425 & 0.4881 & -0.3404 & -0.6115 + llllclllllclllllclllcllllclcllll 145.82320 & -1.17255 & 237.13901 & 36.66249 & & 17.89 & 16.68 & 16.85 & 16.95 & 17.04 & & 17.65 & 16.50 & 16.73 & 16.86 & 16.97 & & 0.0005883 & 24.71 & 15.34 & & 27.3646 & 0.3316 & 25.6610 & 0.2890 & & 0.9099 & & 11.5214 & 1.1931 & 11.2208 & 1.1947 + 160.36295 & -1.05517 & 249.66811 & 47.91172 & & 18.08 & 16.83 & 16.91 & 16.94 & 17.06 & & 17.80 & 16.62 & 16.76 & 16.83 & 16.98 & & 0.0005751 & 33.98 & 16.36 & & 25.9615 & 0.3286 & 28.0140 & 0.2953 & & 0.8329 & & 11.0010 & 1.1813 & 11.9872 & 1.0941 + 161.66508 & 0.06255 & 249.84621 & 49.62371 & & 18.55 & 17.20 & 17.36 & 17.47 & 17.50 & & 18.34 & 17.05 & 17.25 & 17.39 & 17.44 & & 0.0005553 & 32.86 & 19.98 & & 24.1510 & 0.3700 & 27.3331 & 0.3313 & & 0.8483 & & 10.4411 & 1.0110 & 10.8343 & 0.9828 + 161.08659 & 1.00213 & 248.17850 & 49.86035 & & 18.64 & 17.37 & 17.40 & 17.51 & 17.55 & & 18.41 & 17.20 & 17.28 & 17.42 & 17.48 & & 0.0008044 & 109.62 & 21.27 & & 23.5976 & 0.3580 & 27.3562 & 0.3028 & & 0.8724 & & 9.3291 & 1.0594 & 11.6956 & 0.9662 + 161.20377 & 0.68167 & 248.66331 & 49.72445 & & 17.88 & 16.60 & 16.72 & 16.81 & 16.89 & & 17.65 & 16.43 & 16.60 & 16.72 & 16.82 & & 0.0006674 & 67.72 & 14.98 & & 26.9836 & 0.3368 & 26.8800 & 0.3101 & & 0.7589 & & 10.5769 & 1.0842 & 11.7013 & 1.0022 + ccccc 0 & 0 & 0 & 0 & 2338 + 0 & 0 & 0 & 1 & 741 + 0 & 0 & 1 & 0 & 746 + 0 & 0 & 1 & 1 & 683 + 0 & 1 & 0 & 0 & 779 + 0 & 1 & 0 & 1 & 661 + 0 & 1 & 1 & 0 & 658 + 0 & 1 & 1 & 1 & 620 + 1 & 0 & 0 & 0 & 777 + 1 & 0 & 0 & 1 & 670 + 1 & 0 & 1 & 0 & 692 + 1 & 0 & 1 & 1 & 642 + 1 & 1 & 0 & 0 & 673 + 1 & 1 & 0 & 1 & 620 + 1 & 1 & 1 & 0 & 630 + 1 & 1 & 1 & 1 & 595 + lllllll @xmath211 & 0.0706 & 0.0640 & 0.0486 & 0.0611 & 0.0552 + @xmath212 & 0.0995 & 0.0789 & 0.0977 & 0.0920 & 0.0840 + @xmath213 & 0.0283 & 0.0131 & 0.0312 & 0.0242 & 0.0573 + @xmath214 & 0.0173 & 0.0381 & 0.0360 & 0.0305 & 0.0696 + | we isolate samples of @xmath0 bright ( @xmath1 ) and @xmath2 faint ( @xmath3 ) high - galactic latitude blue horizontal branch stars with photometry and spectroscopy in the sloan digital sky survey ( sdss ) .
comparison of independent photometric and spectroscopic selection criteria indicates that contamination from f and blue - straggler stars is less than @xmath4 for bright stars ( @xmath5 ) and about @xmath6 for faint stars ( @xmath7 ) , and this is qualitatively confirmed by proper motions based on the usno - a catalog as first epoch .
analysis of repeated observations shows that the errors in radial velocity are @xmath8 .
a relation between absolute magnitude and color is established using the horizontal branches of halo globular clusters observed by sdss .
bolometric corrections and colors are synthesized in the sdss filters from model spectra .
the redder stars agree well in absolute magitude with accepted values for rr lyrae stars .
the resulting photometric distances are accurate to about @xmath9 magnitudes , with a median of about 25 kpc .
modest clumps in phase space exist and are consistent with the previously reported tidal stream of the sagittarius dwarf galaxy . the sample is tabulated in electronic form in the online version of this article , or by request to the authors . |
type ia supernovae ( sne ia ) are a powerful tool to investigate the expansion history of the universe , because their peak luminosity is as bright as a galaxy and is known as an excellent standard candle . since the discovery of the acceleration of the universe based on the observations of sne ia , higher than ever accuracy of their peak luminosity is needed to investigate various problems in cosmology @xcite .
we started a project to improve the accuracy of the calibration of the peak luminosity of sne ia by measuring accurate distances to nearby resolved galaxies that host sne ia .
we derive accurate distances to the sn ia host galaxies using the method to measure the luminosity of the tip of the red giant branch ( trgb ) @xcite .
we presented the result of the first target , m101 , a well - known spiral galaxy hosting sn 2011fe that is the nearest sn ia since 1972 ( lee & jang 2012 ( paper i ) ) .
this paper is the second of the series , presenting the results for m66 and m96 in the leo i group .
m66 ( ngc 3627 , sab(s)b ) and m96 ( ngc 3368 , sab(rs)ab ) are nearby bright spiral galaxies hosting sne ia : sn 1989b in m66 @xcite and sn 1998bu in m96 @xcite .
m66 has been host to other three sne as well : sn ii 1973r @xcite , sn imposter sn 1997bs @xcite , and sn ii - l 2009hd @xcite .
they are considered to be the members of the compact leo i group that includes three subgroups : the leo triplet ( m66 , m65 , and ngc 3628 ) , the m96 group ( including m96 ( ngc 3368 ) , m95 ( ngc 3351 ) , and m105 ( ngc 3379 ) ) , and the ngc 3607 group @xcite . the leo i group has played an important role as a stepping stone for calibration of the secondary distance indicators , because it includes both early and late type galaxies at the distance closer than the virgo cluster and because it hosts sne ia . in particular m66 and m96 have been used as important calibrators for the absolute magnitudes of sne ia and the tully - fisher relation @xcite .
@xcite derived a value for the distance to the leo i group , @xmath15 ( @xmath16 mpc ) , from the mean of the known distances to five brightest galaxies in the group ( m66 , m95 , m96 , m105 , ngc 3351 and ngc 3377 ) .
often the member galaxy candidates without known distances are assumed to be at the same distance , but it is still important to derive a precise distance to each member galaxy candidate for investigating various aspects of the member galaxies .
unfortunately recent estimates of the distances to m66 and m96 based on resolved stars show a large range @xcite .
@xcite found 68 cepheids in m66 from @xmath1 and @xmath2 images obtained with the _ hubble space telescope ( hst)_/wide field planetary camera 2 ( wfpc2 ) and derived a distance modulus of @xmath17 from the photometry of 25 good cepheids .
later cepheid estimates range from @xmath18 @xcite to @xmath19 @xcite , showing as much as 0.8 mag differences . on the other hand
, @xcite presented a distance modulus @xmath20 using the trgb method from @xmath1 and @xmath2 images obtained with the @xmath21/advanced camera for surveys ( acs ) .
furthermore @xcite presented an even smaller trgb distance estimate , @xmath22 .
thus there is a significant difference between the cepheid distances and trgb distances as well as among the estimates of each method . in the case of m96
, @xcite found 7 cepheids from @xmath21/wfpc2 @xmath1 and @xmath2 images and derived a distance modulus of @xmath23 .
later cepheid estimates showed a significant spread , ranging from @xmath24 @xcite to @xmath25 @xcite .
surprisingly @xcite presented a much smaller trgb distance estimate @xmath26 derived from the @xmath21 images .
thus the difference between the cepheid distances and trgb distance is as much as 0.3 to 0.7 mag and the range of the cepheid distances is about 0.4 . in this study
we use the well - known trgb method to estimate the distances to m66 and m96 from the images available in the @xmath21 archive .
the trgb method is an efficient and precise primary distance indicator for resolved galaxies so that it is an excellent tool for calibration of more powerful distance indicators such as sn ia and tully - fisher relations @xcite .
section 2 describes how we derive photometry of the point sources in the images and 3 presents color - magnitude diagrams of the resolved stars in each galaxy , and derive distances to each galaxy using the trgb method .
we discuss implications of our results in 4 , and summarizes primary results in the final section .
[ cols="^,^,^,^,^,^,^ " , ]
we present @xmath0 photometry of the resolved stars in two spiral galaxies m66 and m96 that host sne ia in the leo i group , derived from @xmath21/acs @xmath1 and @xmath2 images
. then we estimate the distances to these two galaxies applying the trgb method to this photometry .
we summarize main results in the following .
* most of the resolved stars in the selected regions of m66 and m96 are red giants , allowing us to determine the distances to these galaxies . * the @xmath3-band magnitudes of the trgb
are found to be @xmath4 for m66 and @xmath5 for m96 .
these trgb magnitudes yield distance modulus @xmath27 for m66 and @xmath28 for m96 .
this result shows that m66 and m96 are the members of the same group . *
the absolute maximum magnitudes of the sne ia are derived from the previous photometry and the distance measurement in this study , as listed in tables 5 and 6 .
similarly we derive nir magnitudes for sn 1998bu : @xmath29 , @xmath30 , and @xmath31 . * combining the results for sn 1989b and sn 1998bu with those for sn 2011fe in m101 based on the same method given in @xcite
, we obtain an estimate of the hubble constant , @xmath32 km s@xmath11 mpc@xmath11 .
this paper is based on image data obtained from the multimission archive at the space telescope science institute ( mast ) .
the authors would like to thank won - kee park for technical support in image processing .
this work was supported by the national research foundation of korea ( nrf ) grant funded by the korea government ( mest ) ( no .
2012r1a4a1028713 ) .
barone - nugent , r. l. , lidman , c. , wyithe , j. s. b. , et al .
2012 , , 425 , 1007 9 bellazzini , m. , ferraro , f. r. , sollima , a. , pancino , e. , & origlia , l. 2004 , , 424 , 199 bennett , c. l. , larson , d. , weiland , j. l. , et al . 2012 , arxiv:1212.5225 burns , c. r. , stritzinger , m. , phillips , m. m. , et al .
2011 , , 141 , 19 ciardullo , r. , feldmeier , j. j. , jacoby , g. h. , et al .
2002 , , 577 , 31 ciatti , f. , & rosino , l. 1977 , , 56 , 59 de vaucouleurs , g. 1975 , in stars and stellar systems 9 : galaxies and the universe , ed .
a. sandage , m. sandage , & j. kristian ( chicago : univ .
chicago press ) , 557 freedman , w. l. , & madore , b. f. 2010 , , 48 , 673 freedman , w. l. , madore , b. f. , scowcroft , v. , et al .
2012 , , 758 , 24 gibson , b. k. , stetson , p. b. , freedman , w. l. , et al .
2000 , , 529 , 723 gibson , b. k. , & stetson , p. b. 2001 , , 547 , l103 | m66 and m96 in the leo i group are nearby spiral galaxies hosting type ia supernovae ( sne ia ) .
we estimate the distances to these galaxies from the luminosity of the tip of the red giant branch ( trgb ) .
we obtain @xmath0 photometry of resolved stars in these galaxies from @xmath1 and @xmath2 images in the _ hubble space telescope _ archive . from the luminosity function of these red giants we find the trgb @xmath3-band magnitude to be @xmath4 for m66 and @xmath5 for m96 .
these values yield distance modulus @xmath6 for m66 and @xmath7 for m96 .
these results show that they are indeed the members of the same group . with these results
we derive absolute maximum magnitudes of two sne ( sn 1989b in m66 and sn 1998bu in m96 ) .
@xmath8-band magnitudes of these sne ia are @xmath90.2 mag fainter than sn 2011fe in m101 , the nearest recent sn ia .
we also derive near - infrared magnitudes for sn 1998bu .
optical magnitudes of three sne ia ( sn 1989b , sn 1998bu , and sn 2011fe ) based on trgb analysis yield a hubble constant , @xmath10 km s@xmath11 mpc@xmath11 .
this value is similar to the values derived from recent wmap9 results , @xmath12 km s@xmath11 mpc@xmath11 . and
from planck results , @xmath13 km s@xmath11 mpc@xmath11 , but smaller than other recent determinations based on cepheid calibration for sne ia luminosity , @xmath14 km s@xmath11 mpc@xmath11 . |
in the last decade , embedded stellar disks or disky distorsions have been found in many elliptical and s0 galaxies ( scorza & bender 1995 hereafter sb95 ; seifert & scorza 1996 ) .
they are characterized by a smaller scalelength and higher central surface brightness with respect to the large kpc - scale disks typical of lenticular and spiral galaxies .
the existence of embedded disks gives further observational support to the idea that the disky ellipticals are the continuation of the sequence ranging from i m s through spirals to s0 s ( kormendy & bender 1996 and references therein ) .
this suggests also a continuity in the formation history , whereby one or several parameters of the protogalaxy vary smoothly ( e.g. van den bosch 1998 ) .
subarcsecond resolution provided by state - of - the - art ground and space - based telescopes have revealed even smaller stellar disks in the nuclear region of nearby galaxies . to date
the smallest disks we know have scalelengths of few tens of pc and they have been identified in the nucleus of a handful of early - type disk galaxies : the s0 s ngc 3115 ( sb95 ; lauer et al .
1995 ) , ngc 4342 ( van den bosch , jaffe , & van der marel 1998 ; scorza & van den bosch 1998 ) , and s0 ngc 4570 ( van den bosch et al .
1998 ; scorza & van den bosch 1998 ; van den bosch & emsellem 1998 ) , and the sa ngc 4594 ( burkhead 1986 , 1991 ; kormendy 1988 ; emsellem et al . 1994 , 1996 ) .
in addition , nuclear stellar disks of gas , dust and stars have been found in the e3 galaxy ngc 5845 ( kormendy et al . 1994 ) and in the dwarf e2 galaxy ngc 4486a ( kormendy et al .
although the photometric parameters are not known , these stellar disks appear to be similar to those of ngc 3115 and ngc 4594 . this phenomenon may be more common if some of the e and s0 galaxies with disky isophotes observed by ravindranath et al .
( 2001 ) and rest et al .
( 2001 ) turn out to host a nuclear stellar disk .
the presence of nuclear disks raises the question about the epoch ( i.e. coeval or not with that of the host ) and mechanism ( i.e. external or internal origin ) of their formation .
indeed , it is possible that they formed with a different mechanism than the one that formed normal s0 disks since their size is one order of magnitude smaller .
the blue colour of the nuclear disks of ngc 4342 , ngc 4570 and ngc 4486a suggests they are made of younger stars with respect to the bulk of their host galaxy ( van den bosch et al .
1998 ; kormendy et al .
2002 ) . in the framework of galaxy formation via hierarchical merging acquired gas may end up forming some of these nuclear stellar disks . on the other hand
, they could be built up from gas transported towards the galaxy center during the secular evolution of a bar .
this seems to be the case of ngc 4570 ( van den bosch & emsellem 1998 ) , where the observed features in photometry and kinematics correspond with the position of the main resonances of a small bar .
furthermore , any model for the formation of nuclear disks has to account for supermassive black holes , which are supposed to reside in every galaxy and whose gravitational influence may extend to pc scale of the disks .
galaxies hosting nuclear disks are not exceptions in this context as ngc 3115 , ngc 4342 , ngc 4594 , and ngc 5845 are known to harbour a central supermassive black hole ( kormendy et al .
1996a , b ; cretton & van den bosch 1999 , kormendy & gebhardt 2002 ) . to date studies on nuclear stellar disks considered mainly elliptical or s0 galaxies but there is no a priori reason why they should not be present also in spiral galaxies .
would this be true , such nuclear disks would be the result of star formation by gas either driven toward the center by a bar instability or acquired from the outside . for these reasons
we undertook a search for nuclear stellar disks in spiral galaxies based on hst / wfpc2 archive images . in this paper
we present as a result of our investigation three new cases of spiral galaxies hosting a nuclear disk embedded in their main outer disk and bulge .
we selected rc3 ( de vaucouleurs et al .
1991 ) galaxies classified as spiral ( @xmath0 ) , and with @xmath1 .
we restricted our sample to galaxies within @xmath2 mpc ( for @xmath3 ) to be able to detect 10-pc scalelength nuclear disks on wfpc2/pc frames out to about @xmath4@xmath5 scalelengths .
we did not exclude low - inclined galaxies from the sample since nuclear disks might have a different inclination with respect to their hosts .
we searched the hst science archive looking for all the available images .
we realized that wfpc2/f606w was the only camera / filter combination producing an homogeneous sample large enough for our purposes ( 159 galaxies ) .
this resulted in a sample of 112 objects after we rejected those frames where galaxy nucleus was out or too close to the chip edge as well as those with too short exposure times .
it has to be noted that all but 3 of the selected galaxies belong to 4 snap observing programs : ( 1 ) 57 objects were taken from the sample of disk galaxies ranging from e / s0 to spirals observed in program i d 5446 ( pi : g. illingworth ) ; ( 2 ) 33 are unbarred sa / sb spirals from program i d 6359 ( pi : m. stiavelli ) ; ( 3 ) 14 spirals were taken from the nearby seyfert galaxies observed in program i d 5479 ( pi : m. malkan ) ; ( 4 ) 5 are seyfert spirals from program i d 8597 ( pi : m. regan ) . in our final sample the hubble types ranging from sa to scd are well represented as well as the number of barred spirals roughly equals that of unbarred ones , although images were collected from hst programs based on different selection criteria .
the on - the - fly calibrated wfpc2/f606w images of the 112 sample galaxies were retrieved from the hst archive .
the different images of the same target were aligned and combined .
cosmic ray events were removed with iraf tasks crrej and imedit .
the conversion to the johnson system has been calculated using synphot in stsdas .
since this correction depends on the spectral energy distribution of the object , it has been calculated using the kinney et al .
( 1996 ) spectral templates . as a first step to identify candidates hosting a nuclear disk we construct unsharp - masked images for all the sample galaxies , dividing each wfpc2/f606w frame by itself after being convolved with a circular gaussian of @xmath6 2,4 , and 6 pixels , respectively ( fig .
1 ) . the advantage of this procedure is to quickly enhance any surface - brightness fluctuation and non - circular structure extending over a spatial region comparable with the @xmath7 of the smoothing gaussian . in the present context
this enables first to set apart 74 galaxies whose nucleus is strongly affected by dusts , preventing any further analysis .
this is the case of ngc 1637 shown , as an example , in fig .
it should be noted that , in principle , one can check for the presence of dust lanes using images with different filters which are not available for the whole galaxy sample .
however adopting different values for @xmath7 in the unsharp - mask allows to clearly reveal dust lanes , when present .
successively , on the unsharp - masked images of the 38 remaining galaxies , we looked for the highly flattened nuclear structures which are possibly inclined nuclear disks .
such flattened structures are not artifacts of the unsharp - masking procedure since they are always associated to a central increase of ellipticity as measured by performing the isophotal analysis using the iraf task ellipse .
this subset of 38 galaxies ranges from sa to sm and from @xmath8 to @xmath9 as shown in fig . 2 and represents our qualified sample for which we are confident we are able to detect an inclined nuclear disk , if present .
we found three objects , namely ngc 1425 , ngc 3898 and ngc 4698 , which show this flattened nuclear structure ( fig .
1 ) . a positive fourth cosine fourier coefficient ( describing the disky deviation of the isophotes from pure ellipses )
confirms the presence of a nuclear stellar disk in inner regions of these early - type spirals ( fig .
3 ) . in fig .
1 we show the case of ngc 4539 as an example of those galaxies where we did not identify any nuclear flattened structure . once the existence of the nuclear disks is established , we derived their photometrical properties by using the method described by sb95 . when adopting this technique to study the innermost regions of galaxies it is essential to restore the images from the effects of the hst point spread function ( psf ) in order to properly derive the nuclear disks parameters , as shown by scorza & van den bosch ( 1998 ) .
such deconvolution was performed through the richardson - lucy method by means of the iraf task lucy .
although susceptible to noise amplification , this algorithm has been proved by van den bosch et al .
( 1998 ) to lead to a restored surface - brightness distribution comparable to the one obtained by means of a multi - gaussian representation ( monnet , bacon , & emsellem 1992 ) .
we decided to deconvolve the images with a number of iterations between 3 and 6 .
a larger number of iterations does not affect the result of the decomposition but amplify the noise .
we believe that the results obtained by van den bosch et al .
( 1998 ) are directly applicable to our case : we are dealing with images obtained with similar or longer integration times , galaxies with less steep surface - brightness profiles and nuclear disks with equal or larger scalelengths . for each given image and nucleus position on the pc ( ngc 3898 , ngc 4698 ) or
wf2 ccd ( ngc 1425 ) we adopted a model psf calculated using the tiny - tim package ( krist & hook 1999 ) .
no correction for telescope jitter was necessary .
the sb95 method consists in the iterative subtraction from the galaxy image of a thin disk model .
the parameters of such a disk are varied until the departures from perfect ellipses are smallest ( i.e. @xmath10 and @xmath11 are nearly zero ) . for the disk component we assumed an exponential surface brightness profile , with central surface brightness @xmath12 , radial scalelength @xmath13 , and an inclination given by @xmath14 .
we verified that the parameters of nuclear disks resulting from the photometric decomposition are not affected by small changes in psf ( e.g. its generation in different chip positions ) .
the results of the photometric decomposition of the surface - brightness distribution of ngc 1425 , ngc 3898 and ngc 4698 are shown in fig . 3 and fig . 4
3 we plot the ellipticity , position angle , @xmath15 , and @xmath16 fourier coefficient radial profiles before and after the nuclear disk subtraction . in fig .
4 the same is true for the galaxy isophotes .
the photometric parameters derived for the nuclear disks are given in tab .
we discuss the individual objects .
* ngc 1425 . * to obtain a residual bulge with nearly elliptical isophotes we needed to subtract two exponential thin disks of different scalelengths , central surface brightnesses and inclinations .
if this difference in inclination is real we are facing two structures in an unstable configuration . as discussed by scorza & van den bosch ( 1998 ) for the the nuclear and main disks of ngc 4342 , it is more likely we are looking at two thick disks of different thickness ( e.g. , the exponential spheroid disks of van den bosch & de zeeuw 1996 ) rather than two infinitesimally thin disks of different inclination .
such a different thickness may be interpreted as due to the presence of two distinct exponential disks .
alternatively we may think we are facing a non - exponential disk with tickness varying with radius .
considering the two disk hypotesis , we note that the photometric properties of the larger - scale disk are close to those observed for disky ellipticals ( fig . 5 ) . *
ngc 3898 . *
the galaxy hosts a small ( @xmath17 pc ) nuclear disk that is noticeable in particular from the fourier @xmath15 coefficient positive values , observed within @xmath18 . in terms of physical size
this is the smallest disk so far detected .
it possibly corresponds to the somewhat steeper rise of the stellar rotation curve measured in the inner @xmath19 and the two relative maxima in the stellar velocity dispersion at @xmath20 by vega beltrn et al .
( 2001 ) but not considered in the dynamical modelling of pignatelli et al .
( 2001 ) . *
* although sandage & bedke ( 1994 ) include this galaxy in the carnegie atlas of galaxies as an example of the early - to - intermediate sa type , ngc 4698 has photometrical and kinematical properties which are uncommon among spiral galaxies .
it shows a clear geometrical decoupling between bulge and disk and hosts a kinematically isolated core ( bertola et al .
the nuclear stellar disk corresponds to this isolated core , which is rotating perpendicularly with respect to the galaxy main disk ( bertola et al .
we have provided evidence for the presence of a nuclear disk in 3 early - type spirals , namely ngc 1425 , ngc 3898 and ngc 4698 over a qualified sample of 38 objects .
the photometric properties of these 20-pc scale exponential disks are consistent with those of the 4 nuclear disks so far detected in disk galaxies of even earlier morphological type ( fig .
we have not found nuclear disks neither in sbc
sm nor in barred galaxies although these classes represents the majority of our sample . to further address the demography of nuclear disks we need to apply the relatively easy approach adopted here ( based on unsharp - masking ) on a larger sample of galaxies , imaged at high spatial resolution in near - infrared pass - bands to deal with central dusts which prevented us to exclude the presence of nuclear disks in a large fraction ( roughly two thirds ) of our wfpc2/f606w sample galaxies . without drawing general conclusions from such a limited number of galaxies ,
nevertheless we have the indication that the presence of nuclear disks is restricted to s0 s and bulge - dominated unbarred spiral galaxies . in the framework of massive bulge formation through a process of hierarchical clustering merging
, the nuclear disks may be the final result of dissipational and star - formation processes subsequent to a second acquisition event .
to date however , there were only photometric and kinematical evidences that nuclear disks can also be formed via secular evolution of a bar ( e.g. ngc 4570 ) . with ngc 4698 we showed for the first time that second events indeed represent a viable mechanism to build a nuclear disk in the center of disk galaxies .
indeed the nuclear disk of ngc 4698 is geometrically ( this paper ) and kinematically ( bertola et al . 1999 ) decoupled in an orthogonal way with respect to the host galaxy .
this phenomenon can be hardly explained without invoking the acquisition of external material from the galaxy outskirts ( see bertola & corsini 2000 ) .
cllccccccccc 1425 & sb(s ) & sb(r ) & 13.1 & @xmath21 & 63 & 129 & & 16.90 & 26 & 70 & 137 + & & & & & & & & 18.53 & 162 & 66 & 135 + 3898 & sab(s ) & sa & 16.4 & @xmath22 & 54 & 107 & & 15.36 & 18 & 73 & 102 + 4698 & sab(s ) & sa & 12.6 & @xmath23 & 52 & 170 & & 17.27 & 32 & 74 & 71 + | we report evidence for nuclear stellar disks in 3 early - type spirals , namely ngc 1425 , ngc 3898 and ngc 4698 , revealed by wfpc2/f606w images out of a sample of 38 spiral galaxies , selected from the _ hubble space telescope _
data archive . adopting the photometric method introduced by scorza & bender we derived their central surface brightness and scalelength by assuming them to be infinitesimally thin exponential disks .
no nuclear disk was found in barred galaxies or galaxies of hubble type later than sb .
the external origin of the disk in ngc 4698 is strongly suggested by its orthogonal geometrical decoupling with respect to the host galaxy . |
it is likely that gas observed in the central regions of elliptical galaxies is dominated by magnetic stresses .
the chaotic , non - equilibrium appearance of dusty clouds of cool gas apparent in hst observations in the cores of many elliptical galaxies ( e.g. van dokkum & franx 1995 ) is often ascribed to the effects of a recent merger .
however , it may be more likely that the disorganized appearance of these clouds reflects the well known inability of magnetic fields to reach quiescent configurations in differentially rotating , gravitationally bound fluids .
in addition it is plausible that strong magnetic forces in the gas that cools near the centers of elliptical galaxies may influence the initial mass function or other aspects of star formation associated with galactic cooling flows .
the interstellar medium in ellipticals and its magnetic field can be regarded as a result of cumulative stellar mass loss from an evolving population of old stars .
gas ejected from orbiting stars thermalizes a significant fraction of its kinetic energy and settles into a quasi - hydrostatic equilibrium with temperatures near the virial temperature of the galactic potential , @xmath3 k. the hot gas then loses energy by emitting the thermal x - rays that are observed and slowly sinks in the galactic potential where it is compressed and maintained approximately isothermal . although these subsonic inflows are commonly referred to as galactic `` cooling flows , '' the cooling occurs only in the central regions where radiative losses finally overwhelm compressional heating .
the hot interstellar medium in elliptical galaxies operates like a magnetic distillery , first amplifying fields introduced in stellar ejecta then concentrating this magnetic energy in high density gas near the center of the galactic cooling flow where intense radiative cooling occurs .
small fields contained in winds and planetary nebulae expelled from red giants are thought to be amplified further by turbulent motions in the interstellar medium ( ism ) . even in the presence of
turbulent amplification magnetic stresses at large galactic radii are expected to be small compared to the pressure in the hot interstellar gas .
however , the interstellar gas compresses as it slowly flows toward the galactic center and the field strength must also grow by flux conservation @xmath4 .
the ratio of magnetic to ( isothermal ) gas pressure also increases with gas density , @xmath5 . finally , as gas cools near the central parts of the cooling flow , the thermal support of the gas drops precipitously and magnetic stresses are distilled out to support the remaining ionized gas against gravity and external gas pressure .
soker & sarazin ( 1990 ) discussed the spherical , steady state evolution of completely disordered , frozen - in fields in cluster - scale cooling flows .
they chose the field strength to be 1 @xmath6 g at a radius of 50 kpc . within @xmath7 kpc in their models the magnetic pressure grew by compression in the cooling flow to be comparable to the gas pressure , but further amplification was not thought to be possible because of loss of field by reconnection . since the cooling flow accelerates as the galactic center
is approached , soker & sarazin noted that an initially random field would develop a large scale radial component .
more recently christodoulou & sarazin ( 1996 ) computed the time - dependent evolution of large scale toroidal and axial fields in cluster cooling flows adapted to cylindrical geometry .
although @xmath8 may be small relative to the local gas pressure , the origin of @xmath9 g fields assumed by soker & sarazin and christodoulou & sarazin at the start of their calculations must be understood . primordial field production prior to galaxy formation results in much smaller fields ( @xmath10 g ) ( rees 1987 ; lesch & chiba 1995 ) .
but the high metallicity in gas in clusters of galaxies implies that much of this gas was enriched and expelled by type ii supernovae out of galaxies in which interstellar field growth mechanisms were already available .
thus cluster - scale magnetic fields probably have their origin in fields generated in the ism of cluster galaxies .
our interest here is to explore the growth of interstellar magnetic fields in individual elliptical galaxies in a self - consistent fashion beginning with small seed fields contained in stellar ejecta .
we assume that stellar dynamos are efficient in producing fields in stellar envelopes .
stellar mass loss eventually transports these fields into the ism where the field strength is greatly diminished by adiabatic expansion .
however , recent theoretical work ( lesch & bender 1990 ; moss & shukurov 1996 ) has suggested that some the kinetic and thermal energy that accompanies stellar mass loss processes in ellipticals can also drive turbulent motions in the ism which in turn amplify the magnetic field to @xmath11 g . in this paper
we shall assume with these authors that relatively small interstellar magnetic fields can grow by this turbulent dynamo .
we describe how large scale fields may develop from random fields , but we suppose for simplicity that most of the field remains random on small spatial scales ( @xmath12 kpc say ) .
the growth of magnetic forces in the ism may ultimately be limited by field reconnection which we allow for in an approximate way .
unfortunately , the physical processes that control the turbulent growth of galactic magnetic fields are poorly understood and controversial . in our galaxy
the mean field @xmath13 dynamo has been used to describe the amplification of small seed fields to equipartition with interstellar turbulence ( ruzmaikin , shukurov , & sokoloff 1988 ) .
equipartition of magnetic and turbulent energy densities is in fact observed .
fields in spiral disks can grow both by turbulence and differential rotation .
enhancement of very small seed fields ( @xmath14 g ) by the turbulent dynamo process is attractive because amplification by differential rotation in the galactic disk is limited by the relatively few rotations ( 50 - 100 times ) experienced by the galaxy during its lifetime .
however , several authors ( vainshtein & cattaneo 1992 ; kulsrud & anderson 1992 ) have argued that the linear turbulent dynamo can not amplify very small seed fields to equipartition values .
kulsrud & anderson assert , for example , that the field strength can grow exponentially with time to equipartition only on small spatial scales where it is dissipated into heat while the mean field at intermediate and large spatial scales grows to saturation far below equipartition values .
the ( damping ) reaction of the field on the background hydrodynamic turbulence is also considered .
these authors claim that the failure of dynamo theory to account for fields of equipartition strength implies that galactic fields must be primordial . but
fields expected in proto - galactic gas are likely to be much lower than those currently observed ( lesch & chiba 1995 ; beck et al . 1996 ) . field ( 1995 )
has reviewed this apparent difficulty with the mean field dynamo theory and suggests that a more appropriate application of the kulsrud - anderson formalism would be to study the _ steady - state _ distribution of spectral field components in response to a constant input of turbulent energy on large scales .
but an extension of the theory to include non - linear terms may also be necessary . in their recent comprehensive review of galactic fields
beck et al .
( 1996 ) note the apparent inadequacy of the linear theory in addressing turbulent growth ; a non - linear turbulent dynamo theory would include inverse cascade of small scale field energy to larger scales as well as non - linear aspects of small scale dissipation . finally , beck et al .
( 1996 ) note that fields observed in the galactic ism are not concentrated to small spatial scales as the linear turbulent dynamo theory would predict . in view of these difficulties and possible shortcomings of mean field dynamo theory ,
we adopt here a simplified model that allows turbulent growth of mean interstellar fields in elliptical galaxies . we are encouraged to take this deliberately optimistic approach by analogy with equipartition fields observed in the ism of our own galaxy and by the arguments of moss & shukurov ( 1996 ) and others that turbulent dynamos are likely to be effective in elliptical galaxies . by assuming the validity of turbulent dynamos
, we can estimate the ( maximum ) field strengths expected in the ism of elliptical galaxies from internal processes alone ; no _ ad hoc _ initial galactic field is required .
we have been motivated to study the evolution of magnetic fields in ellipticals by our recent work on the nature of cooling flows in slowly rotating ellipticals ( brighenti & mathews 1996 ) .
we showed there that a small rotation of the stellar component , similar to those observed in most large ellipticals ( @xmath15 km s@xmath16 ) , causes the interstellar gas to cool onto a very large disk comparable in size to the half - light radius .
most of the new gas reaching the disk arrives at its outer edge so that the cooling disk slowly grows with the age of the galaxy .
as hot gas approaches the relatively high density regions near the outer edge of the cooling disk , its thermal x - ray image resembles a bright disk or torus in the equatorial plane . near the equatorial plane the gas spins up to the circular velocity which is much larger ( @xmath17 km s@xmath16 ) than that of the stellar system .
we estimated that a small , dynamically insignificant magnetic field would amplify by a factor of @xmath18 in the differentially rotating flow as hot gas approaches the edge of the cooling disk .
we speculated that the corresponding increase in magnetic forces , by @xmath19 , might be sufficient to limit the maximum compression and x - ray visibility of the hot torus just beyond the edge of the cooling disk .
if interstellar magnetic fields are sufficiently large , the x - ray images of rotating ellipticals may appear to be more circular than those predicted by brighenti & mathews ( 1966 ) .
this type of magnetic influence on the x - ray image may be necessary to account for the apparent circularity of _ einstein _ and rosat images of several bright ellipticals in the virgo cluster .
alternatively , the less flattened x - ray appearance of these galaxies could simply result from a fortuitous orientation of galactic rotation axes along the line of sight . in any case , the growth and evolution of magnetic fields in spherical , non - rotating galaxies which we discuss here is a necessary first step toward a more complete study of field amplification and x - ray morphology in slowly rotating ellipticals .
our approach to field amplification is similar to that of moss and shukurov ( 1996 ) who describe the growth of magnetic fields in the turbulent interstellar environment energized by supernova explosions and the orbital energy of stellar ejecta .
we extend the discussion of moss and shukurov to consider how the field is further increased ( i ) by compression as the cooling flow approaches the galactic core and ( ii ) by the greater interstellar turbulence expected in past times when the stellar mass loss and supernova rates were both larger .
our calculation is necessarily approximate since many theoretical aspects of the turbulent dynamo process and field reconnection are based on dimensional arguments .
our models are exploratory , not definitive .
we begin with a brief review of magnetic fields observed in elliptical galaxies then proceed to the description of our time - dependent models and the resulting magnetic field distributions .
the common signposts for the presence of magnetic fields in spiral galaxies polarization of aligned dust grains , polarized emission from relativistic electrons and zeeman splitting are not available in ellipticals due in part to the relative absence of dust ( tsai & mathews 1996a ; b ) , the infrequency of supernova explosions and the absence of distributed cold hi gas . by far
the best means of detecting fields in ellipticals is through the differential faraday depolarization of radio sources by fields within the hot , x - ray emitting cooling flow gas which lies along the line of sight . in
most fr ii - type double radio sources the jet that powers the radio lobes is visible at radio frequencies on only one side
. laing ( 1988 ) first noted that the radio polarization was systematically higher on the jet - side which is also the near side , assuming that radiation from the jet is made more visible by relativistic doppler beaming . if near and far lobes are otherwise identical , the depolarization of the more distant lobe must arise in the galactic ism , not within the lobes or in a shell of gas adjacent to the radio lobe .
the body of evidence supporting this correlation has grown considerably , both for lobes in large ellipticals and in small clusters of galaxies ( garrington et al .
1988 ; garrington , conway , & leahy 1991 ; garrington & conway 1991 ) . the depolarization asymmetry is particularly strong for projected lobe separations of galactic dimension , @xmath20 kpc . by comparing the relative faraday depolarization of two lobes it is possible to determine the mean product of the electron density and net line of sight field strength to the center of the galaxy .
a typical value is @xmath21 @xmath22 @xmath6 g , implying @xmath23 @xmath6 g for plasma densities characteristic of the hot interstellar gas .
these values are based on models in which the magnetic field is uniform in many small cells which are randomly oriented along the line of sight through the hot ism ( burn 1966 ) .
randomly directed fields are also implied by the patterns of differential radiofrequency faraday rotation observed at high spatial resolution across the face of extended radio lobes or regions ( e.g. strom & jgers 1988 ; owen , eilek , & keel , 1990 ; clarke , burns , & norman 1992 ) .
magnetic fields even larger than those indicated above may be present if they are tangled on spatial scales much less than the radio resolution limit . finally , greenfield , roberts , & burke ( 1985 ) observed differential radio polarization in two gravitationally - lensed images of a distant quasar which they attribute to magnetic fields in the hot ism of the intervening cd galaxy lens .
we conclude that the hot gas in elliptical galaxies that participates in galactic scale cooling flows contains magnetic fields that are comparable to those in spirals and that a significant component of this field is spatially disordered .
the equations that describe the evolution of hydromagnetic cooling flows in elliptical galaxies are the usual conservation equations modified with appropriate source terms : @xmath24 @xmath25 = -{\bf \nabla}p + { 1 \over 4 \pi } ( { \bf \nabla \times b}){\bf \times b } - \rho { \bf \nabla } \phi - \alpha \rho _ * ( { \bf v } - { \bf v}_*),\eqno(3.2)\ ] ] and @xmath26 = { p \over \rho } \left [ { \partial \rho \over \partial t } + ( { \bf v } \cdot { \bf \nabla } ) \rho \right ] - { \rho^2 f l(t ) \over m_p^2}\ ] ] @xmath27 the galactic potential @xmath28 is appropriate for a stellar density distribution @xmath29 and a dark matter halo . except for the magnetic force term in the equation of motion ( 3.2 ) and heating due to reconnection of magnetic fields , the last term in equation ( 3.3 ) ,
these equations are identical to those used by brighenti and mathews ( 1996 ) . in the absence of plasma resistivity and field sources ,
the magnetic field evolves according to the usual frozen - in condition : @xmath30 however , for applications to the interstellar medium of elliptical galaxies , it is necessary to include source terms due to fields ejected by evolving stars , turbulent growth , and the loss of field due to reconnection .
these modifications will be developed in the following sections .
although the non - magnetic source terms in equations ( 3.1 ) - ( 3.3 ) have been described in detail by brighenti and mathews ( 1996 ) , for completeness we shall briefly review them again here .
new gas is introduced into the interstellar medium by mass ejection from normally evolving stars formed in a single burst at time @xmath31 .
the specific rate of mass loss is proportional to the local stellar density @xmath32 ( gm @xmath22 s@xmath16 ) where @xmath33 ( mathews 1989 ) . here
@xmath34 gyr represents the present time and @xmath35 s@xmath16 , valid for a wide variety of initial mass functions ( mathews 1989 ) . on average
this new gas appears in the galaxy at zero velocity relative to the net local stellar velocity @xmath36 which is non - zero only for rotating galaxies .
the negative source term in equation ( 3.2 ) represents the momentum drag on the local gas flow @xmath37 as new gas created with velocity @xmath36 is accelerated to the local flow velocity @xmath37 .
equation ( 3.3 ) for the conservation of specific thermal energy @xmath38 ( @xmath39 is the proton mass ) contains a negative term representing optically thin radiative losses by thermal emission as computed by raymond , cox & smith ( 1976 ) : @xmath40 ergs s@xmath16 gm@xmath16 where @xmath41 is related to the mean molecular weight by @xmath42 ; we shall assume @xmath43 .
additional source terms in equation ( 3.3 ) describe the heating of the gas by the thermal and kinetic energy of stellar ejecta , including supernova explosions .
after new gas enters the flow with mean velocity @xmath36 it is accelerated to the local gas velocity , dissipating energy at a rate @xmath44 where @xmath45 and @xmath46 is the corresponding specific thermal energy .
the term @xmath47 in the source term of equation ( 3.3 ) represents the work done by the stellar ejecta on the local ambient gas as it expands from an initially dense cloud toward pressure equilibrium with the interstellar gas .
the solar or subsolar iron abundance observed in the interstellar medium of ellipticals indicates that the ( type ia ) supernova rate is low and that it has been low in recent gyrs ( loewenstein and mathews 1991 ; loewenstein et al .
1994 ) . observations of the current supernova rate in elliptical galaxies also suggest rather low values , @xmath48 snu ( van den bergh , mcclure , & evans 1987 ; cappelaro et al .
1993 ) where 1 snu corresponds to one supernova per @xmath49 @xmath50 every 100 years . the past rate of supernova explosions is even more uncertain , but it is generally assumed that the rate was higher when the galaxy was younger .
we shall assume a power law time dependence of the form @xmath51 where @xmath50 is in units of the solar value @xmath52 ergs s@xmath16 and the reduced hubble constant is assumed to be @xmath53 .
the snu parameter defined by equation ( 3.7 ) can not be compared directly with observed values since other uncertain parameters ( such as @xmath54 ) also appear in this same equation .
recent calculations of spherical cooling flows ( mathews 1997 ) have shown that global thermal instabilities or `` galactic drips '' can occur in isolated ellipticals if the supernova rate parameter snu @xmath55 .
therefore we shall assume snu = 0.066 and @xmath56 here although the iron production may still be somewhat larger than observed ( mathews 1997 ) .
the time variation for @xmath57 is similar to that adopted by ciotti _
( 1991 ) but less steep .
we choose a king model for the stellar density @xmath58^{-3/2}\ ] ] where @xmath59 is the stellar core radius .
most of the gravitational potential arises from a dark halo having an approximately isothermal distribution : @xmath60^{-1}.\ ] ] parameters describing the spherical galaxy model listed in table 1 have been chosen so that the galaxy lies on the fundamental plane ( see tsai & mathews 1996a and brighenti & mathews 1996 for details ) .
the stellar temperature @xmath61 can be found by solving the equation of stellar hydrodynamics in spherical symmetry ( binney & tremaine 1987 ; mathews 1988 ) .
for this purpose we assume a velocity ellipsoid of the form @xmath62 where @xmath63 and @xmath64 are the transverse and radial stellar velocity dispersions respectively .
we choose @xmath65 , but the evolution of the ism is not sensitive to the value of @xmath66 over a wide range @xmath67 .
in spherical symmetry , the magnetic field can have radial and tangential components , @xmath68 and equation ( 3.4 ) becomes @xmath69 and @xmath70 where @xmath71 is the lagrangian or comoving derivative and @xmath72 is the radial velocity of the cooling flow .
the ( source - free ) equation of continuity has a similar form : @xmath73 dividing equations ( 4.1 ) and ( 4.2 ) by ( 4.3 ) , we find @xmath74 and @xmath75 where @xmath76 therefore each component of the field evolves with the local gas density as a power law having an exponent that depends on the structure the local velocity field as characterized by the exponent @xmath77
. in general @xmath77 can be a function of galactic radius .
for example in a homologous , hubble - type flow , @xmath78 and @xmath79 is spatially uniform . in this case
@xmath80 , @xmath81 so the total magnetic field @xmath82 evolves according to @xmath83 in typical galactic cooling flows the gas velocity slowly accelerates toward the galactic center , @xmath84 where @xmath85 . for this range of velocity fields
the radial component @xmath86 grows faster than @xmath87 toward the ( high density ) center of the cooling flow ( e.g. @xmath88 , @xmath89 for @xmath90 ; @xmath91 , @xmath92 for @xmath93 ) . as a result of this unequal growth , an initial field random on small scales in non - rotating ellipticals
will develop radial fields having large scale coherence as the flow approaches the galactic core .
differential shear in slowly rotating ellipticals is expected to convert disordered fields into coherently toroidal configurations ( brighenti & mathews 1996 ) . in either case , however , reconnection may be expected to occur in regions of random field ; when such disordered fields are stretched by the global flow , reconnection could still be important since field lines in adjacent flux tubes will tend to be antiparallel .
the fraction of the field contained in random and coherent ( but antiparallel ) components in galactic cooling flows can be estimated by comparing the bulk flow velocity @xmath94 and the mean turbulent velocity @xmath95 ; both velocities are subsonic .
whenever @xmath96 is less than the equipartion field @xmath97 and @xmath98 we expect that the field geometry will be continuously randomized by local turbulence ; should @xmath99 the field will be stretched with the fluid and develop large scale coherence .
the ratio of characteristic timescales for turbulence and bulk flow @xmath100 can be used to define a parameter @xmath101 that increases with the development of large scale fields . here
@xmath102 or @xmath103 in strong or weak turbulence respectively .
in general we may imagine that the individual field components have a ( source - free ) evolution described by @xmath104 b_r^2 { 1 \over \rho}{d \rho \over dt}\ ] ] and @xmath105 b_t^2 { 1 \over \rho}{d \rho \over dt}.\ ] ] when continuous randomization obtains , @xmath106 implies @xmath107 and equation ( 4.8 ) holds ; in the following discussion we shall assume this limit , i.e. the field is randomized on timescales short compared to the ( quite long ) flow time in the interstellar medium .
the validity of this assumption can be estimated by evaluating @xmath108 from the computed cooling flow solutions .
while the flux - freezing condition ( 4.8 ) describes the evolution of a completely disordered field , additional source and sink terms must be included to account for the origin , turbulent growth and loss of the field : @xmath109 the second term on the right represents the growth of the field due to turbulent amplification ; the unit step function [ @xmath110 @xmath111 ; @xmath112 @xmath113 ensures turbulent growth only when @xmath114 .
this term deals in a schematic fashion with a subject of considerable complexity .
normally the turbulent growth of the field would be represented using the ( linear ) mean field dynamo equation ( see beck et al .
1996 and references therein ) , however in view of the difficulties described earlier in generating equipartition with the mean field equation , we prefer the simplified approach in equation ( 4.9 ) . the third term on the right accounts for the loss of field by reconnection .
the last two terms on the right represent original seed fields from stellar mass loss and galactic events at time @xmath115 which create non - stellar turbulence in the interstellar gas .
these terms are discussed in more detail in the following . aside from the possibility of primordial fields which are very small in any case ( beck et al .
1996 ) the most credible source of magnetic field in the interstellar medium are the fields created by dynamos in stellar interiors which are expelled into the ism when red giant stars lose their envelopes .
the relevant magnetic fields are not those observed in stellar photospheres or chromospheres , but the mean fields @xmath116 that are likely to exist throughout ( convective ) red giant envelopes before ejection from the star .
guided by the magnetic flux observed in white dwarfs , moss and shukurov ( 1996 ) adopted @xmath117 in cgs units as a typical value , corresponding to @xmath118 g and @xmath119 gm @xmath22 ; we adopt this same value here .
after the stellar envelope has expanded to the local density of interstellar gas @xmath120 conserving magnetic flux , the seed field will be very small : @xmath121 g. an additional source of seed fields may arise from strong turbulent mixing following a significant merger of another galaxy with the elliptical at time @xmath115 or after some other global environmental disturbance .
following such an event , we assume that the entire ism becomes strongly turbulent , amplifying any seed fields until the magnetic energy density is some fraction @xmath122 of the thermal energy density in the gas .
such a hypothetical turbulent event is represented schematically in equation ( 4.9 ) by a magnetic energy density proportional to the local gas pressure , @xmath123 imposed suddenly at time @xmath115 . the growth of magnetic flux by plasma turbulence in the interstellar medium of elliptical galaxies has been discussed in detail by moss and shukurov ( 1996 ) ; we shall adopt many of their suggestions here .
the two principal internal sources of interstellar turbulence are the mixing of gas ejected from orbiting stars and the creation and buoyant evolution of ( type ia ) supernova remnants ( mathews 1990 ) .
we assume that the random vortical motion associated with these sources amplifies the interstellar magnetic field .
the energy density in the field is expected to grow until it becomes comparable with that in the turbulence and then saturates at that value .
the equipartition field is defined by @xmath124 the turbulent energy density is written as a fraction @xmath125 of the local thermal energy density ( @xmath126 is the isothermal sound speed and @xmath95 is an rms speed characterizing local turbulent motion ) .
normally , we expect @xmath127 and therefore @xmath128 .
ideally , a separate dynamical equation should be introduced to describe the creation , transfer and evolution of turbulent energy density .
however , we avoid these complications here by assuming that the turbulent energy density is generally small ( @xmath129 ) and that the principal catalytic role of turbulent energy is to convert sources of interstellar energy into magnetic energy density .
in addition we shall assume for simplicity that the turbulence generated by stellar ejecta and supernovae is homogeneous and isotropic ; this ( somewhat optimistic ) assumption is discussed by moss and shukurov ( 1996 ) .
the turbulent field is produced by a superposition of kinetic energy created by stellar ejecta and supernova remnants .
the stellar contribution to the local turbulent energy density is given by the product of the rate of energy production by mass ejection from evolving ( red giant ) stars @xmath130 ( ergs @xmath22 s@xmath16 ) , the fraction @xmath131 of this energy that goes into turbulence , and the characteristic eddy turnover time @xmath132 : @xmath133 the characteristic rms turbulent velocity is @xmath134 although the full range of turbulent velocities must extend up to the local stellar velocities .. the typical eddy size @xmath135 should be comparable to the dimension of a stellar envelope ( or planetary nebula ) after it has expanded to pressure equilibrium ( at @xmath136k ) with the local interstellar gas .
for this purpose ( following moss & shukurov 1996 ) we use the estimate of this length scale determined by mathews ( 1990 ) , @xmath137 cm , where @xmath79 is the local interstellar density .
we assume that the turbulent mixing time is the eddy turnover time of the largest eddies , @xmath138 ; this is conservative since the turnover rate is faster for smaller eddies .
equation ( 4.11 ) can be solved at each galactic radius for the stellar equipartition parameter @xmath139 where @xmath140 .
for the elliptical described in table 1 @xmath141 k ( for @xmath142 ) and @xmath143 gm @xmath22 .
field - free cooling flows in this galaxy result in gas density and temperature distributions given by @xmath144 gm @xmath22 and @xmath145 k at time @xmath146 gyr .
these relations result in a stellar equipartition parameter @xmath147 valid for @xmath142 .
the corresponding turbulent velocity induced by stellar ejecta , @xmath148 km s@xmath16 , is indeed very subsonic ( @xmath149 km s@xmath16 ) .
the turnover time for the largest eddies is @xmath150 years . in a similar manner type ia supernovae feed energy into the interstellar medium at a rate @xmath151 ( ergs @xmath22 cm@xmath16 ) and a fraction @xmath131 of this is assumed to go into turbulent motions .
the eddy mixing time is @xmath153 where @xmath154 . for supernova - driven turbulence
the ( largest ) eddy dimensions are comparable to the size of the hot bubble formed in the interstellar medium following each supernova event , @xmath155 cm , based on the blast wave calculations of mathews ( 1990 ) .
combining these expressions , the equipartition parameter for supernova induced turbulence is @xmath156 for the bright elliptical shown in table 1 we find @xmath157 ( valid for @xmath142 ) , which is somewhat larger than the stellar contribution to turbulence for a given @xmath131 .
the corresponding rms turbulent velocity and turbulent mixing time are @xmath158 km s@xmath16 and @xmath159 years .
the combined turbulent mixing time required in equation ( 4.9 ) is found from @xmath160 and the total equipartition field is given by equation ( 4.10 ) with @xmath161 .
obviously the introduction of turbulent energy by supernovae occurs at the location of these events , not in a spatially smooth manner as represented by the turbulent source term in equation ( 4.9 ) .
nevertheless we shall assume that the turbulence is driven in a smooth , homogeneous manner .
the applicability of this assumption is discussed by moss & shukurov ( 1996 ) .
in addition to using stellar processes to produce vortical turbulence , moss and shukurov ( 1996 ) also invoke `` acoustic turbulence , '' a general field of compressional ( sound ) waves or noise associated with stellar mass loss and ( particularly ) supernova explosions . although the coupling of this acoustical energy density to the usual vortical turbulence is non - linear and inefficient , moss and shukurov claim that acoustical wave energy contributes substantially to the turbulent growth of the magnetic field and that these waves ultimately steepen into shocks .
however , in the absence of reasonably strong magnetic fields , a condition that may obtain at early times , viscosity and thermal conductivity in the hot interstellar medium are expected to damp acoustic waves within a wavelength of their source .
compression waves should damp before they steepen into shocks .
for example , damping due to plasma viscosity @xmath6 exponentially reduces the flux of plane parallel sound waves of frequency @xmath162 propagating in the x - direction by @xmath163 , where @xmath164 ( e.g. zeldovich and raizer 1966 ) .
damping by thermal conductivity @xmath165 is comparable .
evidently acoustic waves generated by either stellar mass loss or supernova events are strongly damped and thermalized within a single wavelength of their point of origin in the interstellar medium . both @xmath6 and @xmath165
may be reduced by the presence of a magnetic field , but a reduction factor of @xmath166 would be required for acoustic damping to be negligible , requiring the presence of significant magnetic fields at early times . in view of this complication
we do not consider turbulent field generation by the acoustical wave field here . the expected dominance of random magnetic fields in elliptical galaxies implies that the occurrence of counter - directed nearby fields should be common .
in such regions thin boundary layers are created where the finite resistivity converts field energy to other forms of energy : thermal , bulk kinetic energy , accelerated particles , etc .
( soker & sarazin 1990 ; lesch & bender 1990 ; jafelice & friaca 1996 ) .
the physics of field reconnection in the hot ism of ellipticals galaxies is discussed in some detail by jafelice and friaca where many earlier references are cited . from dimensional and physical arguments ,
the flow into the reconnection regions is expected to occur at some fraction of the local alfven velocity @xmath167 .
we shall follow the simple representation used by these authors for the field reconnection time scale in equation ( 4.9 ) : @xmath168 here @xmath169 is a single parameter that describes the reconnection efficiency .
in addition we shall assume with these authors that all of the field energy lost by reconnection is entirely converted to thermal energy ; so a term @xmath170 must appear in the thermal energy equation ( 3.3 ) .
the principal parameters related to the turbulent dynamo are @xmath131 and @xmath171 both of which are very uncertain .
summarizing the previous results , the equations describing the time - dependent evolution of spherical cooling flows containing a self - generated , disordered magnetic field are : @xmath172 @xmath173 @xmath174 + { b^2 \over 8 \pi \tau_{rec } } .
\eqno(4.18)\ ] ] the source of thermal energy for the gas is determined by @xmath175\ ] ] which allows a fraction @xmath176 of the energy from stellar processes to be thermalized when @xmath177 . when @xmath1 all the stellar energy is assumed to heat the gas either by direct heating or by eventual dissipation of the turbulent component .
the magnetic field evolves according to equation ( 3.9 ) which we write as @xmath178 we consider two simple representations ( @xmath179 ) for the turbulent growth of the magnetic field : @xmath180 and @xmath181 for case i the magnetic field can increase only on the time scale @xmath182 that characterizes turbulent growth .
the step function @xmath183 guarantees that the field can not grow beyond equipartition . in case
i some part of the power density @xmath184 is allocated to increase the field strength and the rest is assumed to maintain the local turbulent energy density .
alternatively , the case ii option allows stellar and supernova energy to convert directly to either magnetic or thermal energy density , depending on the value of @xmath183 . the turbulent energy density , which still can be regarded as an intermediate stage , is not explicitly considered .
the field grows at a rate @xmath185 until it reaches local equipartition then growth suddenly saturates .
when equations ( 4.18 ) and ( 4.19 ) are added in the case ii approximation , the total energy is explicitly conserved either as thermal or magnetic energy . in the following we consider solutions to both sets of equations , cases i and ii .
we have computed evolving magnetic cooling flows in the galaxy described in table 1 using a lagrangian code .
the primary parameters that define each flow are snu , @xmath131 and @xmath169 .
our cooling flow solutions are not sensitive to the value of the seed field parameter @xmath186 since for all cases considered the exponential growth of field in the turbulent dynamo is rapid compared to the local gas flow time ( @xmath187 ) in the ism .
the parameters @xmath122 and @xmath115 ( see eqn . 4.19 ) can be used to simulate the global turbulence expected following a galactic merger or some other short - lived source of turbulence . each calculation
is begun at time @xmath188 gyr when it is assumed that the galactic winds ( driven by type ii supernovae ) have reversed to become cooling flows ( david _ et al .
_ 1990 ; 1991 ) .
interstellar turbulence is turned on at a later time ( @xmath189 gyr ) to avoid numerical interference that may occur if the gas and the magnetic field are initiated at the same time .
the calculations are stopped at 15 gyr which we regard as the present time .
our elliptical galaxy is assumed to be perfectly isolated apart from the environmental influences represented by the parameters @xmath122 and @xmath115 . in figure 1
we compare several cooling flow models at time @xmath146 gyr .
the non - magnetic flow shown in figure 1a - c resembles spherical cooling flows that have been discussed in detail elsewhere ( e.g. brighenti & mathews 1996 ; mathews 1997 ) .
the parameters used for this non - magnetic cooling flow are : @xmath190 and snu @xmath191 .
some numerical irregularity is visible in figures 1a and 1b for @xmath192 because of the smaller number of computational zones in the galactic core .
most of this numerical noise arises when central zones are removed as they cool below @xmath193 k. we recognize that the x - ray surface brightness @xmath194 corresponding to our model in figure 1a - c decreases somewhat more steeply with galactic radius than that typically observed ( trinchieri , fabbiano , & canizares 1986 ) .
the traditional correction for this discrepancy has been to remove gas from the flow , assuming that stars or some other invisible objects are formed from the mass that `` drops out '' ( e.g. stewart et al .
1987 ; white & sarazin 1987 ; thomas et al .
our models would adjust in a similar manner if mass were removed .
we have not included drop out in our calculations for several reasons : ( i ) we wish to illustrate the possible dynamic influence of magnetic forces without additional uncertain complications , ( ii ) in those ( perhaps unlikely ) situations when the magnetic field is strong , its influence on the solution ( and @xmath194 ) is very similar to mass drop - out , and ( iii ) thermal instabilities , often invoked to justify mass drop out , are generally arrested at a very early stage in the presence of modest magnetic fields .
the corresponding case i magnetic cooling flow evolution with parameters @xmath195 snu @xmath191 , @xmath196 cm , and @xmath197 is illustrated by the solid lines in figure 1d - f .
our choice of a large @xmath169 effectively shuts down the reconnection process , maximizing the computed field strength .
even though half of the energy from stellar sources is available to drive the turbulent dynamo , the gas density , temperature and pressure are all essentially unchanged from the field - free cooling flow ( figure 1a - c ) .
the magnetic pressure @xmath8 ( long dashed line in fig .
1f ) is never more than a few percent of the gas pressure although it is somewhat larger at small galactic radii .
but the field strength indicated in figure 1f ( long - dashed curve ) is not small , reaching 1 @xmath6 g at @xmath198 kpc and 10 @xmath6 g at 1 kpc .
of course these fields are upper limits because we assume that the dynamo is efficient and our choice of large @xmath169 effectively deactivates field reconnection .
we find that the results in figure 1f are essentially unchanged as long as @xmath199 cm . the magnetic field throughout the galaxy slowly decreases with @xmath169 ;
in figure 1f we show that the field is reduced only by a factor of @xmath200 ( short - dashed line ) when @xmath201 cm .
unfortunately the physical significance of this is unlear . the reconnection process ,
if it is important , is expected to occur throughout the flow in very thin boundary layers separating regions of adjacent antiparallel fields .
such a physically complex process can not be adequately modeled with the single parameter @xmath169 .
also illustrated in figure 1f ( dotted line ) is the magnetic pressure at @xmath146 gyr that results when @xmath131 is increased to @xmath202 with the other parameters unchanged ( snu @xmath191 , @xmath196 cm , and @xmath197 ) .
the field is slightly larger in this case because of the somewhat greater turbulent energy that existed early in the cooling flow evolution .
we have also computed cooling flow solutions with the same parameters but using the case ii equations .
we find that the results are essentially identical to those of the case i solutions .
as explained below , the insensitivity to the specific representation ( equation 4.20a or b ) for the turbulent growth of the field arises because for either case the field grows to equipartition early in the calculation in a time short compared to the local flow time ( @xmath187 ) . of particular interest for understanding the amplification of disordered fields in galactic cooling flows
are the radial dependences of the thermal , magnetic and turbulent energy densities .
using the approximate power law variations discussed in 4.3.2 , the thermal energy density should vary as @xmath203 .
for turbulence driven by either stars or type ia supernovae , the turbulent energy density varies as @xmath204 and the equipartition field must vary in the same manner , @xmath205 .
however , the field compression that accompanies the cooling flow causes a slightly steeper variation , @xmath206 .
two important conclusions follow from these spatial dependencies : ( 1 ) the increase in the magnetic field strength toward the galactic center can be dominated by compression in the cooling flow , not local turbulence , and , if so , ( 2 ) the magnitude of the interstellar field depends entirely on the most distant radius in the galactic ism where the turbulent dynamo process amplified stellar seed fields at some previous time . in figure 2a
we plot the spatial variation of the thermal , magnetic and turbulent energy densities for the standard magnetic cooling flow solution ( shown in figure 1d - f with solid lines ) at times @xmath207 and @xmath208 gyr ; these contours are essentially identical for case i and case ii equations .
the slopes of the various energy densities ( in the linear parts of the log - log contours ) are in excellent agreement with the approximate variations estimated above . at either 12 or 15
gyrs @xmath209 is almost 10 times larger than @xmath210 throughout most of the galactic volume .
as expected from the previous estimate , the local turbulent energy density plays no explicit role in field amplification since @xmath211 .
this excess of the field above the local equipartition value occurs because the field at @xmath146 gyr is a relic of a historical field which developed when the turbulence was larger . at an early time in the evolution of this cooling flow equipartition ( @xmath212 )
was reached in time @xmath213 throughout the galaxy , but as the calculation proceeded the turbulent energy decreased ( @xmath214 as expected from the scaling relations in 4.3.2 when @xmath215 ) .
the super - equipartition field is maintained and amplified by compression in the cooling flow .
since @xmath216 increases with radius , the current galactic field is most sensitive to past turbulent conditions in the outermost parts of the galactic flow . in this regard
we note the small decrease in @xmath209 in figure 2a that occurs near the outer boundary @xmath217 of the galaxy between 12 and 15 gyrs . in this part of the flow the gas density drops slightly during this time interval due to
the ( losing ) competition between the decreasing rate that gas is introduced by evolving stars and the depletion of gas by global inflow toward the galactic core .
as @xmath79 decreases , the magnetic field must also decrease ( @xmath107 ) .
some additional loss of local field near @xmath217 may occur because the local field is being diluted by new gas that only contains the very small seed field yet turbulent amplification is not possible until @xmath96 drops below the equipartition value .
figure 2b shows how the frozen - in field has been advected inward toward the galactic center , creating and maintaining the expected slope @xmath218 throughout the interstellar medium .
this also illustrates how the field throughout the x - ray bright central parts of the ism depends critically on the historical turbulent growth in very distant regions . in view of the relationship between the global field and past turbulence near the outer edge of the galaxy
, it might be thought that the growth of the field would be sensitive to the outer boundary conditions in the flow , but this is not the case .
for example we repeated the standard calculation shown in figure 1d - f ( solid lines ) with an open boundary condition at @xmath219 which allows gas to flow out beyond the galaxy into the surrounding vacuum .
after @xmath146 gyrs the flow variables and the field strength distribution within the galaxy were essentially identical to those shown in figure 1d - f .
evidently the gas density and pressure near @xmath220 is maintained by local stellar mass loss , not by the small amount of gas that flows out of the galaxy when an open boundary is allowed .
finally , we have performed some calculations with non - zero @xmath122 to simulate an event , such as a galactic merger , that introduces global turbulence throughout the ism at some time @xmath115 in the past . in figure 1g - i
we illustrate the magnetic cooling flow at time @xmath146 gyr following a disturbance in the the interstellar gas that occurred at @xmath221 gyr when the turbulent energy density was raised to one fourth of the thermal energy density ( @xmath222 ) throughout the galaxy . as a result of this event , the magnetic energy density at 15 gyr ( fig .
1i ) is @xmath223 times larger than it would have been otherwise ( fig .
this illustrates that present day magnetic fields in elliptical galaxies can be significantly increased by past events that stirred up the outer parts of the interstellar gas and which expended energies that are small compared to the total binding energy of the interstellar gas . for the cooling flow in figure 1g
- i the magnetic energy density @xmath224 is imposed impulsively at time @xmath115 and subsequent growth of @xmath8 is due to compression in the global flow . by time @xmath146 gyr figure 1i shows that magnetic pressure dominates the flow dynamics within about 5 kpc of the galactic center while the gas density remains monotonic , implying that the magnetic core does not become buoyant .
however , the magnetic stresses are sufficient to slow down the local flow in @xmath225 kpc allowing an accumulation of gas near this radius where some radiative cooling and mass deposition occurs .
this is reminisent of the `` galactic drip '' phenomenon in which gas in non - magnetic cooling flows can cool far from the galactic center ( mathews 1997 ) , but the increase in gas density and thermal instability shown in figure 1i has a purely magnetic origin .
when @xmath122 is lowered to 0.15 , @xmath226 occurs at a smaller radius , @xmath227 kpc , and radiative cooling beyond the galactic core no longer occurs .
some caution must be exercised in interpreting the non - stellar turbulent growth imposed at @xmath115 in terms of a galactic merger . for galactic mergers
in which the galactic mass is increased by a relatively small factor , @xmath228 percent , mathews and brighenti ( 1997 ) have shown that enough energy can be released by type ii supernovae from newly created stars to remove most or all of the preexisting galactic interstellar gas .
such an energy release obviously depends on the assumed initial mass function for the newly formed stars .
we do not consider these complications here .
in this study of the evolution of interstellar magnetic fields in elliptical galaxies we have shown that the observed fields can be understood as a natural result of internal evolutionary processes : stellar seed fields are first amplified by a turbulent dynamo then by compression in the galactic cooling flow .
we have assumed that stellar dynamos occur in virtually all stars and that these stars supply tiny seed fields to the ism as they undergo mass loss during normal red giant evolution .
our optimism concerning the efficacy of the turbulent dynamo mechanism and the relative inefficacy of field reconnection have been essential in obtaining central field strengths comparable to those observed , @xmath229 @xmath6 g .
our models extend a modified version of the turbulent dynamo theory proposed by moss and shukurov ( 1996 ) to include the important effects of compressional field amplification in the global galactic cooling flow and the evolutionary relationship between currently observed fields and those created by galactic turbulence many gyrs ago . as the field advects from the outer regions of the galaxy to the galactic core
the field compresses by @xmath230 and the magnetic energy density increases by @xmath231 . in our models
the magnetic energy density exceeds the local turbulent energy density ( @xmath211 ) throughout most of the interstellar gas so local turbulence plays little or no role in field amplification .
when @xmath211 the field may quench some of the turbulent activity although we have not allowed for this feedback in our models ; it is also unclear whether the random character of the field ( on scales of @xmath135 or @xmath232 ) can be fully maintained in this limit . for most of the magnetic cooling
flows described here we have assumed that half of the stellar energy goes ( at least initially ) into turbulence ( @xmath195 ) . although the ratio @xmath131 of turbulent energy to immediately thermalized energy resulting from stellar mass loss processes is uncertain , our results suggest that the resulting field strength does not depend strongly on @xmath131 .
how sensitive are our results , particularly the condition @xmath211 , to the specific galaxy model we have adopted , a simple king stellar distribution with an isothermal dark halo ? to address this question it is useful to express the radial variation of the magnetic and turbulent energy densities in terms of the gas and stellar densities . for either stellar or supernova turbulence
we find @xmath233 and @xmath234
. therefore the ratio of these energy densities @xmath235 .
clearly the radial variation of the gas density depends on both the stellar and dark matter density profiles , so the interpretation of this last relation is complex . to make further progress ,
suppose that we adopt the point of view that the x - ray and optical surface brightness profiles are the same , i.e. @xmath236 .
canizares , fabbiano , & trinchieri ( 1987 ) found this to be the case for three bright ellipticals in virgo , but more recent rosat observations show some deviations from exact proportionality .
however , if @xmath236 then @xmath237 and @xmath238 so that @xmath209 _ increases _ very slowly relative to @xmath210 with increasing galactic radius , opposite to the variation shown in figure 2a .
we may expect , therefore , that the ratio @xmath239 could either slowly increase or decrease with galactic radius , depending on the details of the galactic model .
however , regardless of these details it is clear that the interstellar magnetic field can increase toward the galactic center _ no slower _ than @xmath107 as shown in figure 2b .
finally we note that in our models the @xmath239 increases with the cooling inflow most strongly in the outermost parts of the galaxy where the power law scaling breaks down and where the observational uncertainty is greatest .
if interstellar fields in elliptical galaxies have their origins in the seed fields of stellar ejecta , the physical scale @xmath240 of regions of faraday coherence may be smaller than previously supposed ( burn 1966 ; garrington & conway 1991 ) .
regions in the hot interstellar gas having masses comparable to those of stellar envelopes , @xmath241 @xmath242 , have diameters of only @xmath243 pc .
however , faraday - coherent regions of parallel field lines will be even smaller because of the violent instabilities that accompany stellar mass loss .
since the observed faraday depolarization parameter is proportional to @xmath244 , the field implied by any observation will be larger as @xmath240 is reduced .
this difficulty is not alleviated by large scale fields that can develop in ( non - hubble ) spherical or differentially rotating non - spherical cooling flows , since such an `` aligned '' field will still be counter - directed on small mass scales .
the scale of field coherence @xmath240 can increase if reconnection is efficient , but this would also decrease the overall magnetic energy density . clearly ,
more attention must be given to the likelihood of field reconnection in cooling flow environments .
\(1 ) self - generated interstellar magnetic fields in elliptical galaxies can be comparable to those suggested in faraday depolarization studies provided ( i ) the turbulent dynamo process can amplify the mean field to equipartition with the turbulent energy density and ( ii ) field reconnection is not dominant throughout most of the ism .
\(2 ) using our galactic model , a significant additional amplification of the overall field occurs by compression in the galactic cooling flow ; the magnetic energy density is increased by factors of @xmath231 .
\(4 ) the exact value of the field in ellipticals is sensitive to interstellar turbulence at large distances from the galactic center and at distant times in the past . for this reason we may expect some variability in field strengths among elliptical galaxies , depending on the environmental history of the outer galaxy .
\(5 ) in our spherical cooling flow models the magnetic energy density or pressure in self - generated cooling flows ( @xmath197 ) never dominates the galactic gas pressure ( nor becomes buoyant ) so we do not expect the x - ray images to be strongly influenced by the field . in more realistic rotating cooling flows ,
however , the magnetic pressure resulting from differential rotation may decrease the x - ray brightness near the cooling flow disk ( brighenti & mathews 1996 ) .
+ balbus , s. a. 1991 , apj , 372 , 25 + beck , r. , brandenburg , a , moss , d. , shukurov , a. , & sokoloff , d. 1996 , _ ann .
_ , 34 , 155 + binney , j. & tremaine , s. 1987 , galactic dynamics ( princeton : princeton univ . press ) , 195 + brighenti , f. & mathews , w. g. 1996 , apj , 470 , 747 + burn , b. j. 1966 , mnras , 133 , 67 + canizares , c. , fabbiano , g. , & trinchieri , g. 1987 , apj , 312 , 503 + cappelaro , e. , turatto , m. , benetti , s. , tsvetkov d.yu . , bartunov , o. s. , & makarova i. n. , 1993 , a&a , 273 , 383 + christodoulou , d. m. & sarazin , c. l. 1996 , apj , 463 , 80 + clarke , d. a. , burns , j , o. , & norman , m. l. 1992 , apj , 395 , 444 + david l. p. , forman , w. , & jones , c. 1990 , apj , 359 , 29 + david l. p. , forman , w. , & jones , c. 1990 , apj , 369 , 121 + donnelly , r. h. , faber , s. m. , & oconnell , r. m. 1990 , apj , 354 , 52 + field , g. b. 1995 , in _ the physics of the interstellar medium _ , asp conf . series vol
80 , 1 + garrington , s. t. , conway , r. g. & leahy , j. p. , 1991 ,
mnras , 250 , 171 + garrington , s. t. & conway , r. g. , 1991 , mnras , 250 , 198 + garrington , s. t. , leahy , j. p. , conway , r. g. & laing , r. a. , 1988 , nature , 331 , 147 + greenfield , p. e. , roberts , d. h. , & burke , 1984 , apj , 293 , 370 + jafelice , l. c. & friaca , a. c. s. 1996 , mnras , 280 , 438 + kim , dong - woo , & fabbiano , g. 1995 , apj , 441 , 182 + kulsrud , r. m. & anderson , s. w. 1992 , apj , 396 , 606 + laing , r. a. , 1988 , nature , 331 , 149 + lesch , h. & bender , r. 1990 , a&a 233 , 417 + lesch , h. & chiba , m. 1995 , a&a , 297 , 305 + mathews , w. g. 1988 , aj , 95 , 1047 + mathews , w. g. 1990 , apj , 354 , 468 + mathews , w. g. 1997 , aj , ( in press ) + mathews , w. g. & brighenti , f. 1997 , in _ the nature of elliptical galaxies , proc . of the second stromlo symposium _ , eds . m. arnaboldi , g. s. da costa & p saha , ( san francisco : asp ) , a.s.p .
( in press ) + moss , d. & shukurov , a. 1996 , mnras , 279 , 229 + owen , f.n . ,
eilek , j. a. & keel , w. c. , 1990 , apj , 362 , 449 + rees , m. 1987 , qjras , 28 , 197 + ruzmaikin , a. a. , shukurov , a. m. & sokoloff , d. d. 1988 , _ magnetic fields of galaxies _ , ( dordrecht : kluwer ) + soker , n. & sarazin , c. l. 1990 , apj , 348 , 73 + stewart , g. c. , canizares , c. r. , fabian , a. c. , & nulsen , p. e. j. 1987 , apj , 278 , 536 + strom , r. g. & jgers , 1988 , a&a , 194 , 79 + thomas , p. a. , fabian , a. c. , & nulsen , p. e. j. 1987 , mnras , 228 , 973 + trinchieri , g. , fabbiano , g , & canizares , c. r. 1986 , apj , 310 , 637 + tsai , j. c. & mathews , w. g. 1996a , apj , 448 , 84 + tsai , j. c. & mathews , w. g. 1996b , apj , 468 , 571 + vainshtein , s. i. & cattaneo , f. 1992 , apj , 393 , 165 + van den bergh , s. , mcclure , r. d. , & evans , r. 1987 , apj , 323 , 44 + van dokkum , p. g. & franx , m. 1995 , aj , 110 , 2027 + white , r. e. iii , & sarazin , c. l. 1987 , apj , 318 , 612 + zeldovich , ya . b. , & raizer , yu .
p. 1966 , physics of shock waves and high - temperature hydrodynamic phenomena , ( new york : academic press ) , 74 @xmath59 & 311.59 pc@xmath245 & 5.088 kpc@xmath246 & 6.22
kpc @xmath217 & 113.1 kpc@xmath247 & @xmath248 gm @xmath22@xmath249 & @xmath250 gm @xmath22@xmath251 & @xmath252 @xmath242@xmath253 & 9 @xmath251@xmath50 & @xmath254@xmath255 & 9.14@xmath256 & 351 km s@xmath16@xmath257 & 40.0 + * figure 1 : * three vertical arrays of figures show cooling flow conditions in the interstellar gas after evolving to @xmath146 gyr .
solid lines in each plot refer to the gas density , temperature or pressure corresponding to these parameters : ( a - c ) , nonmagnetic cooling flow with @xmath195 and snu @xmath191 ; ( d - f ) , magnetic cooling flow with @xmath195 , snu @xmath191 , @xmath196 and @xmath197 ; ( g - i ) , magnetic cooling flow with @xmath195 , snu @xmath191 , @xmath196 , @xmath222 and @xmath221 gyr . long - dashed lines in the density plots show the variation of the ( unnormalized ) stellar density .
long - dashed lines in the temperature plots show the variation of stellar temperature .
long - dashed lines in the pressure plots show the magnetic pressure .
plot ( f ) also shows the magnetic pressure with @xmath258 ( dotted line ) and with some field reconnection @xmath201 ( short - dashed line ) .
see text for further details .
* figure 2 : * ( a ) variation of energy densities with galactic radius at @xmath207 gyr ( _ thick lines _ ) and @xmath146 gyr ( _ thin lines _ ) : @xmath259 ( _ solid lines _ ) , @xmath209 ( _ dotted lines _ ) , and @xmath210 ( _ dash - dotted lines _ ) .
( b ) variation of magnetic energy density with gas density at @xmath207 gyr ( _ dashed line _ ) and @xmath146 gyr ( _ solid line _ ) .
the flow parameters are : @xmath195 , snu @xmath191 , @xmath196 and @xmath197 . | interstellar magnetic fields in elliptical galaxies have their origin in stellar fields that accompany normal mass loss from an evolving population of old stars .
it is likely that these seed fields are amplified by interstellar turbulence which in turn is driven by stellar mass loss and supernova events . since the local turbulent velocity is likely to exceed the global velocity of the interstellar cooling flow , magnetic fields
are expected to be disordered and this is indicated by numerous observations .
further amplification occurs as these tangled fields are compressed in the inward moving cooling flow of the interstellar gas . near the centers of galactic cooling flows , where the gas radiates away most of its thermal and gravitational energy , magnetic stresses
are expected to dominate .
we study here the time - dependent growth of interstellar magnetic fields in elliptical galaxies and demonstrate that fields similar in strength to those observed can be generated solely from natural galactic processes . although interstellar turbulent velocities and therefore the equipartition field @xmath0 are larger near the galactic center , the interstellar field throughout elliptical galaxies can be determined by the outermost turbulent regions in the interstellar gas .
this occurs whenever the increase in the field strength due to compression in the galactic cooling flow exceeds the rate of field amplification by local turbulence , i.e. when @xmath1 .
the magnitude of the interstellar field depends on the most distant radius in the galactic ism at which the turbulent dynamo process can successfully have amplified stellar seed fields . because of the long hydrodynamic flow times in galactic cooling flows , currently observed magnetic fields may result from periods of turbulent field amplification that occurred in the outer galaxy in the distant past . in our most optimistic turbulent dynamo models in which field reconnection
is ignored , tangled interstellar magnetic fields of @xmath2 g are typical ; these are consistent with observed fields .
however , the sensitivity of the galactic magnetic field to poorly known historical turbulent conditions in the outer galaxy complicates estimates of field strengths that can be produced internally .
internal fields in ellipticals may also result from ancient galactic mergers or from shear turbulence introduced at the boundary between the interstellar gas and ambient cluster gas .
# 1#2#3#4 # 1#2 |
the study of the properties of liquids confined at the nanometer scale is a topic of high interest for its technological , experimental and theoretical implications .
furthermore , confinement plays an important role in hydrated biological systems and organic solvents @xcite .
structural , thermodynamical and dynamical properties of a liquid can change near an interface ( solid , liquid , etc . )
@xcite . when the surface - to - volume ratio is large , at least along one direction as for the slit pore geometry , the effect of the confining surfaces has to be taken into account .
experiments and simulations on nanoconfined fluids show that molecules arrange in layers parallel to the surface .
the effect becomes stronger for decreasing temperature or increasing density , until the fluid eventually solidifies .
the nature of the solid , as an amorphous or a crystal , can depend on the interparticle potential and the confinement conditions @xcite . in different cases
the role of the interfaces can results in complex behaviors , e.g. a persisting fluid mono - layer around a spherical impurity while the rest of the system is in a polycrystal or glassy state @xcite , or persisting amorphous water mono - layer near an hidrophilic disordered surface @xcite . in a recent experiment @xcite , kaya and
coworkers found that a thin - film water on a @xmath0 surface remains in a high density liquid form for temperatures ranging from ambient ( 300k ) to supercooled ( 259k ) .
the result is unexpected because , based on thermodynamics arguments @xcite , kaya and coworker would expect that the templating effect of @xmath0 on the structure of the water film should promote the tetrahedral structure of the low - density liquid .
moreover , the presence of an interface can promote the heterogeneous nucleation of the crystal .
however , recent experiments and simulations based on crystallographic analysis of liquids on crystals @xcite , or on colloidal self - assembly @xcite , show that the traditional theories need to be revised in these cases .
confined fluids , under suitable conditions of density and temperature , can spontaneously develop patterns , e.g. stripes , and different mesophases or more complex structures @xcite .
the confinement affects also the dynamics of the liquid .
it has been found that near an interface there is a reduction of the local diffusivity of the liquid @xcite .
computer simulations can help in interpreting the experimental results for nanoconfined fluids that are difficult to understand @xcite .
different numerical approaches based on first principles simulations can give detailed informations , but are limited by their high computational cost .
classical molecular dynamics of empirical fluid models employ parameters tested for the bulk case that not necessarily hold in confinement .
the difficulty to adopt these bulk fluid models to the case of confinement leads to a variety of simulation results concerning the aggregation state of the fluid near the surfaces that in principle are model - dependent .
it is , therefore , useful to develop coarse - grained models that allow for analytic calculations @xcite and more efficient simulations @xcite , like isotropic pairwise core - softened potentials , and that could allow us to better understand common features of fluids under confinement and the basic mechanisms of complex phenomena emerging in these systems , like pattern formation , e.g. stripes and different mesophases .
this has been confirmed in a recent work @xcite in which the authors claim that the origin of quasi - crystals could be understood in the context of a coarse - grained model by the competing effect of the hard and soft core radius of interacting particles .
here we focus on the study of nanoconfined anomalous fluids , relevant for biological and technological applications @xcite , by means of molecular dynamics simulations of a system of identical particles interacting through the continuous shouldered well potential ( csw ) , an isotropic pairwise core - softened potential with a repulsive shoulder and an attractive well @xcite . the csw fluid
is confined in a slit pore obtained by a solvophilic structured wall , and a solvophobic wall with no structure , as described in sec.[sec : simulations ] .
the csw model is suitable for studying globular proteins in solution @xcite , specific colloids @xcite , and liquid metals @xcite , and displays water - like anomalies @xcite .
in particular the csw reproduces density , diffusion and structure anomalies following the water hierarchy @xcite and displays a liquid - gas and a liquid - liquid ( ll ) phase transition , both ending in critical points @xcite .
here we focus on structural and dynamical properties of the csw model in confinement and show the role that the characteristic length scales of the inter - particle potential have in the structuring of the fluid .
we consider a system of @xmath1 identical particles interacting by means of the csw potential confined between two parallel walls .
the csw potential is defined as @xcite @xmath2+\left(\dfrac{a}{r}\right)^{24}\ ] ] where @xmath3 is the diameter of the particles , @xmath4 and @xmath5 are the distance of the attractive minimum and the repulsive radius , respectively , @xmath6 and @xmath7 are the energies of the attractive well and the repulsive shoulder , respectively , @xmath8 is the variance of the gaussian centered in @xmath4 and @xmath9 is the parameter which controls the slope between the shoulder and the well at @xmath5 .
the parameters employed are the same as in refs.@xcite : @xmath10 , @xmath11 , @xmath12 , @xmath13 . in order to reduce the computational cost ,
we impose a cutoff for the potential at a distance @xmath14 . in the present simulations we use @xmath15 that allows to better evidence the anomalies in density , diffusion and structure @xcite . in our simulations
we consider a @xmath16 ensemble system composed of @xmath17 particles at fixed temperature @xmath18 and volume @xmath19 .
the temperature of the thermal bath @xmath18 is kept constant by rescaling the velocity of the particles at each time step by a factor @xmath20 , where @xmath21 is the instantaneous kinetic temperature ( allen thermostat ) @xcite .
pressure , temperature , density and diffusion constant are all expressed in internal units : @xmath22 , @xmath23 , @xmath24 , and @xmath25 , respectively .
the equation of motion are integrated by means of the velocity verlet method @xcite , using the time - step @xmath26 defined in units of @xmath27 ( that corresponds to @xmath28 for water - like molecules and to @xmath29 for argon - like atoms @xcite ) .
we performed the same check as in ref .
@xcite in order to verify that the value used for @xmath30 is small enough to satisfy the energy conservation of the system .
the confining parallel walls are placed along the @xmath31 axis at a separation distance @xmath32 .
the solvophilic wall is composed of a triangular lattice of csw particles quenched with the position of the centers placed at @xmath33 . the lattice constant is @xmath34 .
the solvophobic wall has no structure and is obtained by imposing the repulsive potential @xmath35 where @xmath36 , and @xmath37 @xcite . in the following we consider the parameter @xmath38 .
we adopt periodic boundary conditions in the @xmath39 and @xmath40 directions . in order to compute the effective density @xmath41
, we need to compute the effective volume @xmath42 accessible to particles @xmath43 , where @xmath44 is the section of the simulation box and @xmath45 is the effective distance between the plates . by considering the quenched particles forming the philic wall and the repulsive strength of the potential of the phobic wall , we obtain @xmath46 @xcite .
the effective density is @xmath47 . to explore different densities for the confined system in the @xmath16 ensemble
, we change @xmath48 and @xmath49 .
we keep @xmath32 and @xmath1 constant to exclude finite - size effects when we compare results for different densities . for each density , we equilibrate the system by annealing from @xmath50 .
for each temperature , the system is equilibrated during @xmath51 time steps .
we observe equilibrium after @xmath52 time steps for the range of @xmath53 and @xmath18 considered here .
all our results are averaged over @xmath54 independent samples . in order to check the stability of the system
, we verify that the energy and the pressure are equilibrated . to compute the pressure
we follow the approach of refs.@xcite . due to the inhomogeneous nature of the system ,
the pressure is a tensor @xmath55 where @xmath56 is the kinetic contribution as in an ideal gas , @xmath57 is the unit tensor , @xmath58 is the potential contribution due to the interparticle interaction , and @xmath59 is the contribution of the walls . at equilibrium
the system is mechanically stable if @xmath60 . considering the planar symmetry of the system
, the normal ( or orthogonal ) and tangential ( or parallel ) component of the pressure can be written respectively as : @xmath61 and @xmath62 , while mixed components are zero , i.e. , no shear forces are present @xcite .
to verify the stability of the system , we computed the normal component of the pressure as a function of @xmath31 @xmath63 the kinetic part is @xmath64 .
for the potential part we used the todd , evans and daivis formulation of the pressure @xcite @xmath65\right\rangle\ ] ] where @xmath66 is the z - component of the interaction force between particles @xmath67 and @xmath68 .
the products of heaviside step functions , @xmath69 , select couple of particles that lie in different semispace respect to the plane parallel to the walls with coordinate @xmath31 .
the walls contribution to the pressure can be computed from eq .
[ equ : pressure_pot ] , with @xmath70 for @xmath71 , as @xcite @xmath72\right\rangle\ ] ] where @xmath73 and @xmath74 are the interaction forces of the particle @xmath67 with the philic wall and the phobic wall , respectively .
normal pressure profile @xmath75 , for densities @xmath76 and temperatures @xmath77 .
the effective densities are : @xmath78 , @xmath79 , @xmath80 , @xmath81 .
each component of the normal pressure is displayed .
in particular , @xmath82 is the contribution of the philic wall , @xmath83 is the contribution of the phobic wall , @xmath84 is the kinetic contribution as in an ideal gas , @xmath85 is the potential contribution due to the interparticle interaction.,width=604 ] the different components compensate each other in order to keep the normal pressure constant ( fig.[fig : pressures ] ) . in particular
, the philic wall contributes with a positive term @xmath86 from @xmath87 to @xmath88 and with a negative term from @xmath88 to @xmath89 , due to the interaction force between the particles of the fluid and those of the philic wall .
the phobic wall contributes with a positive term @xmath90 from @xmath91 to @xmath92 due to the repulsive potential .
the kinetic pressure @xmath93 is a signature of the density profile , while the potential pressure @xmath94 can contribute with a positive or negative term depending which part of the interparticle interaction , repulsive or attractive , respectively , dominate . after having verified that the system is stable ,
i.e. , @xmath95 , we compute the normal pressure as @xmath96 where @xmath97 and @xmath98 are z - coordinates sufficiently near the philic wall and the phobic wall , respectively , for which the kinetic @xmath99 and the potential @xmath100 part of the pressure are zero .
as pointed out in the introduction ( sec.[sec : introduction ] ) , the confinement can modify the structure of a fluid resulting in an inhomogeneous density profile . in a slit pore geometry , near the confining walls , particles form layers parallel to the walls as the temperature
is decreased or the density is increased , as shown by our calculations for the density profile @xmath101 ( fig . [
fig : density ] ) . to establish the aggregation state for each layer ,
we compute , layer by layer , the lateral radial distribution function @xmath102 ; the 2d voronoi tessellation of each layer ; the mean square displacement ( msd ) and the survival probability ( sp ) of molecules in each layer .
all these quantities together , as we discusse in the following , allow us to identify the presence and coesistence of the solid , heterogeneous fluid and homogeneous fluid .
our analysis shows that the system organizes forming layers parallel to the solvophilic wall at any temperature @xmath18 and average density @xmath53 . at high @xmath18 and low @xmath53
we find only one well defined layer and no layering near the solvophobic wall , while the whole system is in the fluid state . by decreasing @xmath18 and increasing @xmath53 , the number of layers increases up to eleven .
furthermore , at higher value of @xmath53 , the layers appear also near the solvophobic wall .
however , for high enough @xmath18 and @xmath53 we observe that away from the walls the system is in the fluid state .
nonetheless , the walls affect the density profile , changing system density and aggregation state , over an extension that in our case can be up to @xmath103 layers that depends on temperature and average density of the system .
the in - layer radial distribution function @xmath104 for the @xmath105-th layer ( with @xmath106 ) is computed as @xmath107\ ] ] where @xmath108 and @xmath109 are the density of particles and the z - coordinate of the layer @xmath105 , respectively , @xmath110 is the transverse distance between two particles in the same layer ( and in internal units is @xmath111 ) . the heaviside step functions , @xmath69 , select couple of particles that lie in the layer @xmath105 of width @xmath112 .
the @xmath104 is proportional to the probability of finding a molecule in the layer @xmath105 at a distance @xmath113 from a randomly chosen molecule of the same layer @xmath105 .
the definition of the layer @xmath105 in which lies a particle @xmath67 is once for all established according to the value of the z - coordinate of the particle @xmath67 as : for the first layer @xmath114 , with @xmath115 , and @xmath116 for the others .
this is a natural choice because particles at low temperatures or high densities tend to stratify in layers whose interdistance is approximately equal to @xmath117 . at low density ( @xmath118 , fig.[fig : gr_rho011 ] )
, we observe that the system is in a fluid state for high temperature ( @xmath119 ) in any layer .
the @xmath104 of the layer @xmath120 shows a first peak around the shoulder radius ( @xmath121 ) and a second peak around the attractive well radius ( @xmath122 ) , while for the other layers only the second peak is present .
we interpret this difference between the first and the other layers as the consequence of a `` templating '' effect of the solvophilic wall that at high @xmath123 is observed only on the first layer . at @xmath124
the system shows the same behavior observed for higher temperatures , except that the layer near the solvophilic wall develops patterns .
this behavior is reminiscent of what has been observed in monolayers with an interparticle potential composed by a hard core and a soft repulsive shoulder @xcite . at @xmath125
the layer near the solvophilic wall is still showing patterns , while the layer @xmath126 is forming crystal patches .
this is evident from the analysis of the @xmath127 that goes to zero for @xmath128 and @xmath129 at @xmath125 , consistent with an incipient triangular crystal with lattice step given by the interaction potential attractive distance @xmath130 .
the other layers are in a fluid state . at @xmath131
the first layer is forming a hexagonal crystal .
although the hexagonal crystal in @xmath120 does not overlap exactly with the triangular wall structure , the comparison of the @xmath132 and @xmath133 of the wall shows a strong correlation between the two structures , suggesting a `` templating '' effect .
this effect due to the attraction to the solvophilic wall is so strong at @xmath131 that is forcing particles to be at their repulsive distance .
the high - energy cost of the resulting honeycomb lattice forming in the layer @xmath120 is compensated by the large number of attractive interactions between the particles at @xmath120 and those of the wall ( @xmath134 ) from one hand , and between the particles themeselves at @xmath120 from the other hand .
this free energy minimization process is analyzed in sec.[sec : structural ] in the discreet potential approximation to understand the stripe phase formation .
the triangular structure that was incipient for @xmath126 at high @xmath123 , for @xmath131 is well defined for @xmath126 and @xmath135 , with defects in the layer @xmath135 .
this triangular structure is the dual lattice of the @xmath120 hexagonal layer and its formation is the consequence of a `` molding '' effect of the layer @xmath120 onto the layer @xmath126 .
note that while the wall ( @xmath134 ) layer has a templating effect on the @xmath120 layer , the @xmath120 layer has a molding effect on the @xmath126 layer .
the difference between the two cases is due to the smaller density of the @xmath120 layer with respect to that of the wall .
the smaller density does not allow to compensate the high energy cost of the propagation of the hexagonal crystal to the layer @xmath126 .
on the other hand , the triangular crystal of the @xmath126 layer is energetically favorable , because the particles are all at the attractive distance , and at this temperature can propagate to the @xmath135 layer again with a `` templating '' effect . the layer @xmath136 is made of a few triangular crystallites immersed in the fluid , while the other layers are in a fluid state . at @xmath137
both the templating and the molding effect are stronger .
in particular the template of the @xmath126 layer propagates over all the six layers that are formed at this density and temperature . at intermediate density ( @xmath138 , fig.[fig : gr_rho022 ] ) ,
for @xmath119 and @xmath124 we observe the same qualitative behavior as for the low density case .
for @xmath125 the layer @xmath120 has less tendency to form patterns respect to the low density case , and the layer @xmath126 to order in a crystal structure .
therefore , the confined system is more fluid at this density than at lower density .
we understand this result as a consequence of the larger hydration at higer density . at @xmath131 the first layer
has partially crystallized in the hexagonal and partially in the triangular structure following the template of the wall .
therefore , the templating effect is now stronger then the corresponding case at lower density .
the hexagonal crystal shows now a preferred direction of symmetry .
this direction propagates to the layer @xmath126 , where we observe stripes along the preferred direction .
the stripes propagate up to @xmath136 layer , while the other layers are in a fluid state .
the peak of @xmath104 at @xmath139 ( that corresponds to the average second nearest neighbor distance ) is a signature of the stripe phase formation . at @xmath137 the preferred direction in the deformation of the hexagonal crystal for @xmath120
is more evident and we observe a clear stripe phase for the layers from @xmath126 to @xmath136 , with a peak of @xmath104 at @xmath139 more pronounced than the case at @xmath131 . the other layers form a triangular crystal at the attractive distance . at high density ( @xmath140 , fig.[fig : gr_rho030 ] ) , for @xmath119 we observe the same qualitative behavior as for the lower density cases . at @xmath124
we found that the only difference with the lower density case is that the layer @xmath120 is forming crystallites following the template of the wall . at @xmath125
the layer @xmath120 has a different and incipient crystal structure ( kagome lattice ) with defects that is better defined at lower @xmath18 .
this is evident from the analysis of the @xmath132 that goes to zero for @xmath121 and @xmath141 at @xmath125 .
the layer @xmath126 shows patterns very close to the stripe configuration .
the corresponding @xmath127 goes to zero for @xmath142 and shows a peak for @xmath139 .
these characteristics of the @xmath143 are typical of a stripe phase .
the layer @xmath135 and @xmath136 still show patterns close to the stripe phase , but in a less pronounced way . from the layer @xmath144 to the @xmath145 the pattern is vanishing .
the layer @xmath146 is showing an incipient triangular crystal with lattice step given by the interaction potential attractive distance @xmath130 .
this is evident from the analysis of the @xmath147 that approaches zero for @xmath128 and @xmath129 at @xmath125 . at @xmath131 the layer @xmath120 is forming a kagome crystal with defects .
the layers from @xmath126 to @xmath145 show a stripe phase and the layer @xmath146 is forming a triangular crystal with defects . at @xmath137 the kagome crystal of layer
@xmath120 has no defects .
the layers from @xmath126 to @xmath145 are in a stripe phase and the layer @xmath146 is forming a well defined triangular crystal .
as discussed above , the layer @xmath120 close to the solvophilic wall ( @xmath134 ) is subjects to the templating effect for all densities at low temperatures . in fig.[fig : layers_gr_snap ] we compare the @xmath148 and the snapshots of the first layer for @xmath137 at several densities . in order to compare layers , that correspond to different densities , between them
, we considered a portion of each layer of the same size ( @xmath48x@xmath49 of the system at @xmath140 ) . for densities
@xmath149 the first layer is forming a distorted hexagonal lattice characterized by a @xmath148 with a first peak at @xmath150 and vanishing for @xmath121 and @xmath141 . by increasing @xmath53
we observe a progressive shift to higher values of @xmath113 of all the peaks of @xmath148 , but the first that , instead , is becoming more pronounced as a consequence of a better local order . for densities between @xmath151 and @xmath152 the first layer shows a polycrystal phase with coexistence of triangular and square lattices .
this corresponds to an intermediate stage toward the well defined kagome lattice that is formed for @xmath153 , as showed by the splitting of the second peak of @xmath148 into two close peaks at @xmath154 and @xmath155 . in order to characterize space - dependent diffusion properties of our system , we compute the mean square displacement ( msd ) associated to each layer of the slit .
we observe that , except for low temperatures , a particle can visit different layers in which the aggregation state can change from homogeneous to heterogeneous liquid and vice versa .
for this reason we calculate the msd only for those particles that remain in a layer over the entire time interval under consideration and we average over all possible time interval . therefore , the msd associated to each layer @xmath105 is defined as @xmath156 where @xmath157 is the time spent in the layer @xmath105 by a particle that entered in the layer at time @xmath158 . in according to the standard definition of the msd , @xmath159 where @xmath160 is the lateral , or parallel , diffusion coefficient and @xmath161 the diffusion exponent .
the value @xmath162 means that the system is arrested , as in a solid state where particles can only vibrate around theirs equilibrium positions ; for @xmath163 the system is subdiffusive corresponding in general to particles diffusing in complex structures ( with non trivial microscopic disorder ) ; @xmath164 is the standard diffusive behavior as in a normal fluid state . for @xmath165
the system is superdiffusive . on the other hand , for early times free diffusion
we expect the ballistic regime with @xmath166 .
our analysis ( fig.[fig : diff_layers ] ) shows that the in - layer msd has always a ballistic regime for @xmath167 .
the corresponding mean displacement is approximately half particle diameter at high @xmath18 and low @xmath53 and weakly decreases for increasing @xmath53 and decreasing @xmath18 corresponding to the expected decrease of the mean free path of the particles . for @xmath168
all the layers reach the diffusive ( @xmath164 ) behavior for long times . by decreasing the temperature
the behavior of the layers becomes more heterogeneous .
in particular , we observe that the layer @xmath120 near to the solvophilic wall slows down in a sensible way with respect to the layers at @xmath169 and becomes arrested for @xmath170 . at these temperatures the other layers , including the one near the solvophobic wall , are diffusive at low densities .
however , at @xmath140 and @xmath131 all the layers develop the plateau in the msd typical of glassy dynamics .
this behavior is reminiscent of the caging effect in glasses where the plateau in the msd is followed by a diffusive regime . here ,
instead , at @xmath131 and @xmath140 we observe that for all the layers but the one near the solvophobic wall ( @xmath146 ) , after the plateau , the dynamics enters in a superdiffusive regime with @xmath171 .
this effect is related to the presence of defects and of a nonuniform stress field , as discussed in sec.[sec : structural ] .
for @xmath137 we observe that all the layers are arrested at @xmath118 . at this low density
the slit is only partially filled ( fig.[fig : diff_layers ] ) . at @xmath137 and @xmath172 and @xmath138
also the layer @xmath146 near the solvophobic wall is present and it is characterized by a larger msd with respect to the other layers and by a diffusive regime at long times .
the other layers have an arrested dynamics . at @xmath137 and @xmath140 all the layers from @xmath120 to @xmath145 have a superdiffusive regime at long times , while the layer @xmath146 reaches the diffusive log - time regime .
however , its msd is smaller than that of the other layer for very long times ( @xmath173 ) . as the time proceeds , the average in eq .
[ equ : msd ] for the msd is performed on a decreasing number of particles because some of them can leave the layer . _ a priori _ this reduction of the statistics is not homogeneous
, that means that in general there can be a correlation between particles that leave the layer and theirs properties , as theirs velocity components .
therefore , for @xmath174 , low enough @xmath53 , and for the most diffusive layers there is a time , @xmath175 , after which the in - layer msd ( fig.[fig : diff_layers ] ) is not well defined . to estimate @xmath175 as function of @xmath18 and @xmath53 for different layers we analyse the population relaxation of particles in each layer . in particular , we compute the survival probability ( sp ) function , @xmath176 , which is the probability that a given particle stay in the layer @xmath67 for a time interval @xmath177 .
the sp can be calculated as @xmath178 where @xmath179 is the number of particle in the layer @xmath67 at time @xmath180 and @xmath181 is the number of particle that do not leave the layer @xmath67 during the time interval @xmath182 $ ] .
the sp give an indication of the time interval @xmath175 over which the msd is well defined .
we observe that @xmath183 has an exponential decay in our simulations ( fig.[fig : stau ] ) .
we , therefore , define @xmath175 as the characteristic decay time @xmath184 .
this choice is consistent with the observation that the msd in fig.[fig : diff_layers ] is well defined when @xmath185 .
we observe that for @xmath186 and all the densities and for @xmath131 and @xmath187 , the sp decay is slower for the layer @xmath120 near the solvophilic wall and becomes faster for the layers away from the two walls .
when the layer @xmath146 near the solvophobic wall is present , we observe that it has a decay in sp slower than those layers that are farther away from the wall . at @xmath137 for all densities , and at @xmath131 for @xmath140 , there is no decay in sp , consistent with the crystallization of the layers .
the non monotonic behavior of @xmath175 is reported in fig.[fig : tau_layers ] as a function of layers for different densities and temperatures . by comparing fig.[fig : tau_layers ] and fig.[fig : density ] we observe that @xmath175 increases when the layers are more structured in the @xmath31 direction . hence , both walls facilitate the stratification of the fluid , although the structureless phobic wall does it in a less strong way with respect to the structured solvophilic wall .
characteristic time decay @xmath175 as a function of fluid layer for density @xmath188 and temperature @xmath77 . for clarity ,
the points for @xmath124 are shifted up by 100 units.,width=377 ] in sec.[sec : msd ] , analysing the msd layer by layer , we have seen ( fig.[fig : diff_layers ] ) that for high densities and low temperatures , after a plateau , the dynamics can enter in a superdiffusive regime with @xmath171 . in this section
we show how this behavior is due to the formation of liquid `` veins '' in such layers .
the formation of liquid veins is of particular interest in ice during the freezing of water .
recently , experiments and simulations showed the presence of liquid water between nanometer - sized ice crystal @xcite . in polycrystalline systems ,
the liquid is found along intergranular junctions , as grain boundaries ( see @xcite and references therein for the case of water ) .
residual stress in these polycrystal structures can be localized along integranular junctions , and can results in an effective force that acts on fluid particles present in these junctions .
the origin of the residual stress in our system is due to the fact that when the fluid solidifies as the temperature is decreased , the minimization process of the free energy take place locally , instead of globally . in glass forming liquids ,
this effect is caused by a fast cooling , while in our system it is due to the layering of the fluid caused by the confinement . in fig.[fig : snap_z1 - 2 - 3_rand1 - 2 - 3 ] we show the spatial configuration of the first three layers of the system close to the philic wall , for three different runs ( i.e. for three different realization of initial conditions ) at @xmath137 and @xmath140 .
we observe that particles in the first layer ( @xmath120 ) are characterized by the same msd , while particles in other layers ( @xmath189 ) can have different msd .
in particular in some configurations , we observe veins with mobility higher than the rest of the system ( fig.[fig : config_t_z2-z3 ] ) .
we analyzed the trajectories of the particles of these specific realizations of the system ( fig.[fig : config_t_z2-z3 ] ) .
for these cases we find that these particles with a msd higher than the majority belong to the same stripe and diffuse along the stripe itself .
we observe that the majority of particles in the layer @xmath126 ( and in a less evident way for the layer @xmath135 ( fig.[fig : config_t_z2-z3]b ) , remain spatially localized during the entire simulation , except those belonging to two stripes moving in the same direction as along stripe veins ( fig.[fig : config_t_z2-z3]a ) .
we observe a similar situation for the layer @xmath135 , but here all the particles are more mobile and the particles in the veins move in opposite directions .
further analysis , that goes beyond the goals of the present work , is necessary to understand the effect of the vicinity of the solvophilic wall and if the veins are related to point - like defects as seems to be suggested by fig.[fig : config_t_z2-z3 ] .
spatial configuration of the first three layers of the system close to the solvophilic wall , for three different runs ( i.e. for three different realization of the initial conditions ) at @xmath137 and @xmath140 .
the size of circles , representing particle positions , are chosen to be equal to the particles hard core diameter @xmath3.,width=415 ] single - particle msd for layers and runs that are in fig.[fig : snap_z1 - 2 - 3_rand1 - 2 - 3 ] . for run # 1 ,
red and green colors are used for those particles belonging to veins performing a dynamics different from the rest of the particles in the layer.,width=226 ] in - layer trajectories for particles in the second ( a ) and third ( b ) layer of run # 1 in fig.[fig : snap_z1 - 2 - 3_rand1 - 2 - 3 ] . in red and green
we show the trajectories of representative particles belonging to stripes veins .
the msd of these particles is represented with the same color code in fig.[fig : diff_z1 - 2 - 3_rand1 - 2 - 3 ] .
we apply periodic boundary conditions for @xmath190 where @xmath191 ( dashed lines ) and for @xmath192 where @xmath193.,title="fig:",width=94 ] in - layer trajectories for particles in the second ( a ) and third ( b ) layer of run # 1 in fig.[fig : snap_z1 - 2 - 3_rand1 - 2 - 3 ] . in red and green
we show the trajectories of representative particles belonging to stripes veins .
the msd of these particles is represented with the same color code in fig.[fig : diff_z1 - 2 - 3_rand1 - 2 - 3 ] .
we apply periodic boundary conditions for @xmath190 where @xmath191 ( dashed lines ) and for @xmath192 where @xmath193.,title="fig:",width=94 ] our msd and sp analysis show that at low @xmath18 and high @xmath53 there are layers taht behave as a solid
. however , by looking only at the msd and sp is not possible to establish if a solid layer is in an amorphous , crystal or polycrystal state @xcite . in order to better understand the structure of solid layers , we computed the standard 2d voronoi tessellation ( useful to identify defects present in the crystal structures , as vacancies , frenkel - like , dislocations and grain boundaries ) , and a modified version of it ( suitable to identify distorted crystal structures ) . with this analysis
we can also disentagle the role that the three relevant length scales ( the diameter of the particles @xmath3 , the repulsive radius @xmath5 , and the attractive minimum @xmath4 ) , giving rise to two competing length scale @xmath194 and @xmath195 , play in the determination of layer s structure . indeed , the interdistance between two adjacent layers is @xmath196 , while when stripes form in a specific layer for intermediate densities , just to consider a specific case , particles within a stripe are compressed at a distance @xmath197 , while the distance between stripes depends on @xmath196 and @xmath198 , as discussed in the last part of this section . in the standard voronoi tessellation
we construct polygons centered around particles forming a lattice whose edges are crossed in their middle point by the edges of the voronoi cells .
this procedure garantees that each voronoi cell represent the proper volume of each particle . to better visualize the result
, we represent voronoi cells having a different number of edges with different colors . to reduce the noise in our analysis
we adopt also a modified version of the voronoi tessellation in which we associate a color to a polygon in according to the number of edges of the polygon that have a length @xmath199 larger than that of the average edge lenght calculated over the specific polygon itself .
this procedure allows us to better visualize polycrystal structures despite the presence of small lattice deformations .
we compute the voronoi tessellation for low density ( @xmath118 ) and high density ( @xmath140 ) at low temperature ( @xmath137 ) and very low temperature ( @xmath200 ) , for three different realizations of initial conditions ( figs.[fig : voronoi_rho0.11]a , b , [ fig : voronoi_rho0.30]a , b , and figs.[fig : voronoi_rho0.11_mod2]a , b , [ fig : voronoi_rho0.30_mod2]a , b in supplementary material ) .
the configurations at temperature @xmath200 are obtained by annealing configurations equilibrated at @xmath137 with an annealing rate of @xmath201 . at low density ( @xmath118 ) , the first layer at @xmath137 ( fig.[fig : voronoi_rho0.11_mod2 ] ) is in a frustrated solid state that by annealing toward @xmath200 ( fig.[fig : voronoi_rho0.30_mod2 ] ) becomes a frustrated polycrystal .
the very low-@xmath18 polycrystal has point and line defects as grain boundaries dividing a deformed honeycomb lattice ( the deformed green triangles ) from a stripe phase ( the stretched hexagonal cyan polygons ) . for both
considered @xmath123 the other layers are organized in a triangular lattice ( where each particle is surrounded by a hexagonal cyan polygon ) .
we only observe defects , such as dislocations ( run # 2 in fig.[fig : voronoi_rho0.11_mod2]a and fig.[fig : voronoi_rho0.11]a ) that are not eliminated by annealing ( run # 2 in fig.[fig : voronoi_rho0.11_mod2]b ) . at high density ( @xmath140 ) , the first layer at @xmath137 ( fig.[fig : voronoi_rho0.30_mod2]a ) is in a polycrystal state with defects . at @xmath200 ( fig.[fig : voronoi_rho0.30_mod2]b )
we observe two principal crystal grains : a triangular lattice ( cyan polygons ) and a kagome lattice with defects ( blue rhombouses ) . at @xmath137 ,
the layers @xmath189 present a zigzagging stripe structure with orientation and angles that can change from run to run . at @xmath200 the stripe structure of these layers becomes more regular .
these observations emphasize that the increase of density induces an increase of disorder in the solid layers , propagating from the layer @xmath120 to the other layers and up to the layer @xmath146 .
the formation of crystal defects during the annealing , and the fact that they are different for different initial conditions , indicate that the system can not reach easely the global minimum of the free energy landscape , corresponding to the crystal configuration , but is trapped in local minima due to the slowing down of the dynamics and the templating effect of the solvophilic wall .
in particular , the mismatch of the wall structure with the bulk crystal structure induces a frustrating effect that is more evident near the wall ( in layers @xmath120 and @xmath126 ) for increasing density . to understand the formation of stripes , we follow the same approach as in refs @xcite to show that
, if the principal contribution to the minimization of the free energy comes from the energetic therm , in the discreet potential approximation , under suitable conditions of density and temperature , particles organize in straight or zigzagging stripes . in the discreet potential approximation
the energetic terms can be reduced to the soft core ( @xmath7 ) and the attractive well ( @xmath6 ) . as a consequence
the energetic contribution coming from the interaction with particles of the adjacent layers is approximately constant when the layer density is fixed .
therefore , the energetic cost of stripes formation is determined only by the contribution of the in - layer particle interactions .
in particular , if a layer has a triangular structure , as the stable configuration of layers @xmath2022 , ... , 11 at low density ( fig.[fig : voronoi_rho0.11_mod2]a , b ) , then for sufficiently high density the layer will prefer to form stripes .
2d schematic representation of particles composed by an hard core ( in dark blue ) of size @xmath3 , a soft corona ( in cyan ) of size @xmath5 and an attractive external corona ( in green ) that extends uo to the potential cutoff @xmath203 .
the equilateral and isosceles triangles represent the unitary cell of the triangular and straight - stripes lattice , respectively .
an example of ( maximal ) zigzagging - stripes of the same density of straight - stripes is also shown.,width=264 ] consider our fluid made of particles with a hard core surrounded by a soft corona and an external attractive corona ( fig.[fig : stripes ] ) .
if @xmath204 is the lattice constant of the triangular structure at the soft - corona distance , then the density of the layer is @xmath205 , where @xmath206 is the number of particles present in this layer . if we allow the triangular lattice to deform in order to minimize the energy of the layer , the new unit cell will be composed by the isosceles triangle in which one side is equal to @xmath39 and the other two are equals to @xmath40 with @xmath207
the fact that the density does nt change implies that @xmath208 .
the energy per particle of the layer is @xmath209u_x+2[(\sqrt{n_l}-1)/\sqrt{n_l}]u_y$ ] .
for @xmath210 it becomes @xmath211 , where @xmath212 and @xmath213 are the energy associated to the interaction between the particle along the @xmath39 and @xmath40 side of the triangle respectively . in the discreet potential approximation it is @xmath214 for @xmath215 , @xmath216 for @xmath217 and @xmath218 for @xmath219 ( note that we obtain the same result if instead of @xmath203 we consider any value between @xmath5 and @xmath203 .
indeed , the present approach has been applyed to show the stability of stripes cnofiguration for a pure repulsive potential model @xcite ) .
the same holds for @xmath213 substituting @xmath39 with @xmath40 . for sufficiently high density ,
i.e. for @xmath220 or @xmath221 , the energies per particle associated to a layer formed by equilateral or isosceles triangles are @xmath222 or @xmath223 , respectively . therefore ,
under these conditions , the layer will prefer to form stripes .
furthermore , from geometric consideration it is possible to conclude that the zigzagging - stripe lattice can be obtained as a deformation of the straight - stripe lattice without changing the density and keeping the energies per particle @xmath224 @xcite . in general , many different zigzagging stripe lattices are possible all with comparable energy per particle ( fig.[fig : stripes ] ) . considering the stripes that form in the layers @xmath225 for system density @xmath140 and temperature @xmath137 , the resulting average in - layer density is @xmath226 , and @xmath227 .
hence , from the previous equation for @xmath228 , we find @xmath229 . in view of all the approximation made
, we consider this value consistent with @xmath230 of the distance between the closest particles belonging to two adjacent stripes in the same layer .
by considering many layers of a confined anomalous fluid @xcite we show that the effect of the structured solvophilic wall can extend up to the entire slit pore . in particular , we study structural and dynamical properties of a monocomponent anomalous liquid under confinement .
the fluid has two characteristic distances and can be considered as a coarse - grained model for globular proteins @xcite , colloidal systems @xcite or , to some extent , liquid metals with water - like anomalies @xcite .
we perform molecular dynamics simulations of the fluid in a slit pore with a solvophilic wall and a solvophobic wall .
the solvophilic wall has structure while the solvophobic one has no structure .
we observe that the molecules organize in an inhomogeneous way , forming layers that are parallel to the surfaces , with higher density near the solvophilic surface with respect to the center of the slit pore . for sufficiently high densities ,
for which the fluid occupy entirely the pore , we observe an increase of density also close to the solvophobic surface , but in a less prominent way .
these results are consistent with experimental and theoretical works for nanoconfined fluids . at low temperature
we observe coexistence between the homogeneous liquid , heterogeneous liquid and solid phase of the fluid .
the influence of the structured solvophilic surface on the solid layers can extend as far as the sixth hydration layer at low @xmath18 and high @xmath53 .
in particular , we find a strong correlation between the structure of the solvophilic surface and that of the first layer suggesting a `` templating '' effect . indeed
, the large density of solvophilic surface particles allows the formation of a first layer at high density .
the high energy cost of this first layer is compensated by the large number of attractive interactions between the particles of the first layer and those of the surface .
further energy gain comes from an extra energy term due to the first in - layer particle interactions . moving further from the wall
, we find that the first layer has a `` molding '' effect on the second layer .
this is because the density of the first layer is smaller than that of the surface and is not high - enough to propagate its template .
nevertheless , the low - density second layer is in condition to template the third layer replicating its structure and inducing a long - range effect that can , eventually , involve the whole system . from the calculation of the mean square displacement and the voronoi tessellation
we conclude that at low temperature the first layer close to the solvophilic surface is a polycrystal with two competing phases that generate low - energy states with high degeneracy and very slow dynamics . at low densities
the two competing phases are stripes and honeycomb lattice , while at high densities are triangular and kagome lattice .
we understand this result as a consequence of the high density ( triangular ) structure of the solvophilic wall with a lattice step that corresponds to the hard repulsive distance of the solvent , and the strong wall - solvent attractive interaction .
these properties of the wall generate in the first solvent layer local regions with density and energy that are higher than the average of the layer . as a consequence ,
other regions within the layer have density and energy below the average , giving rise to a competing crystal structure .
our results remind us of some recent experiments and simulations for a thin - film of water on @xmath0 surface for which the authors found a very high density first interfacial layer for all temperatures , while they would expect , from thermodynamic arguments , a lower density liquid at supercooled conditions @xcite .
in other recent experiments and simulations @xcite , the authors pointed out that it is necessary to revise the theory of heterogeneous nucleation when the crystalization induces a non - zero entropy at zero temperature and the system initially is far from equilibrium .
apart from the layers close to the two surfaces , we observe that the structure of each layer mainly depends on its density . in the case of the stripe phase , using simple geometrical and energetic considerations , we find that straight and zigzagging stripes are the stable configurations for intermediate densities .
furtheremore , analysing the mean square displacement layer by layer , we observe layers at high densities and low temperatures with a caging - like behavior characterized by a ballistic dynamics followed by an arrested state ( plateau ) and a superdiffusive regime with a diffusion exponent @xmath171 .
our analysis shows that this behavior is due to the formation of liquid veins within the stripe phase .
in particular we observe that each vein can behave differently from the others diffusing in one of the two possible directions along the stripes and having a different diffusion exponent @xmath171 .
we rationalize the different possible values of @xmath161 as a consequence of the presence of residual stress that could introduce an effective force acting on the fluid . under suitable conditions ,
e.g. a constant effective force along the stripe , the particles in the vein could perform a biased one - dimensional random walk characterized by an exponent of the msd that approaches the ballistic value ( @xmath166 ) .
the behavior of these veins can be analyzed in a more quantitative way by computing , for example , the temporal autocorrelation function , the intermediate scattering function @xcite , the relative displacement of nearest neighbors or the particles displacements following the lindemann criterion @xcite .
we will present this analysis in future works .
our results show that the dynamical slowing down of the anomalous solvent near the solvophilic wall does not imply by necessity the complete freezing of the first hydration layers , because at low @xmath18 and high @xmath53 we observe largely heterogeneous dynamics in three layers with the formation of liquid veins within a frozen matrix of solvent . therefore , under these considerations the partial freezing of the first hydration layer does not correspond necessarily to an effective reduction of the channel section in terms of transport properties , at variance with the conclusions of ref.@xcite .
we thank y. shokef for useful discussion on stripes formation .
12 m .- c .
bellissent - funel , s .
h. chen , and j .-
zanotti , physical review e * 51 * , 4558 ( 1995 ) .
s. martin and d. dennis j. , j. chem .
phys . * 109 * , 5596 ( 1998 ) .
m. r. jay , j. sony , n. r. aluru , and e. jakobsson , nano letters * 3 * , 589 ( 2003 ) .
j. mittal , t. m. truskett , j. r. errington , and g. hummer , phys .
lett . * 100 * , 145901 ( 2008 ) .
g. cicero , j. c. grossman , e. schwegler , f. gygi , and g. galli , journal of the american chemical society * 130 * , 1871 ( 2008 ) . a. de virgiliis , r .
l. c. vink , j. horbach , and k. binder , phys .
e * 78 * , 041604 ( 2008 ) .
f. mallamace , c. corsaro , m. broccio , c. branca , n. gonzlez - segredo , j. spooren , s .- h .
chen , and h. e. stanley , proc .
usa * 105 * , 12725 ( 2008 ) .
n. giovambattista , p. j. rossky , and p. g. debenedetti , j. phys .
b * 113 * , 13723 ( 2009 ) .
r. mancinelli , s. imberti , a. k. soper , k. h. liu , c. y. mou , f. bruni , and m. a. ricci , j. phys .
b * 113 * , 16169 ( 2009 ) .
p. gallo , m. rovere , and s .- h .
chen , journal of physics : condensed matter * 22 * , 284102 ( 2010 ) .
w. rzysko , a. patrykiejew , s. sokolowski , and o. pizio , j. chem . phys . * 132 * , 164702 ( 2010 ) . s. han , m. y. choi , p. kumar , and h. e. stanley , nat . phys . * 6 * , 685 ( 2010 ) .
t. m. truskett , p. g. debenedetti , and s. torquato , j. chem . phys .
* 114 * , 2401 ( 2001 ) .
f. de los santos and g. franzese , j. phys .
b * 115 * , 14311 ( 2011 ) .
s. k. schnell , t . j. h. vlugt , j .-
simon , d. bedeaux , and s. kjelstrup , chem .
* 504 * , 199 ( 2011 ) .
s. k. schnell , t. j. h. vlugt , j .- m .
simon , d. bedeaux , and s. kjelstrup , molecular physics * 110 * , 1069 ( 2011 ) .
r. r. nair , h. a. wu , p. n. jayaram , i. v. grigorieva , and a. k. geim , science * 335 * , 442 ( 2012 ) .
d. r. paul , science * 335 * , 413 ( 2012 ) . a. l. ferguson , n. giovambattista , p. j. rossky , a. z. panagiotopoulos , and p. g. debenedetti , j. chem .
phys . * 137 * , 144501 ( 2012 ) .
m. c. stewart , and r. evans , phys .
e * 86 * , 031601 ( 2012 ) .
l. b. krott , and m. c. barbosa , j. chem . phys . *
138 * , 084505 ( 2013 ) .
g. schir , a. cupane , e. vitrano , and f. bruni , j. phys .
b * 113 * , 9606 ( 2009 ) .
s. karan , s. samitsu , x. peng , k. kurashima , and i. ichinose , science * 335 * , 444 ( 2012 ) .
f. biedermann , m. vendruscolo , o. a. scherman , a. de simone , and w. m. nau , j. am .
soc . * 135 ( 39 ) * , 14879 ( 2013 ) .
g. franzese , and v. bianco , food biophysics * 8 * , 153 ( 2013 ) . n. giovambattista , and p. j. rossky , and p. g. debenedetti , annual review of physical chemistry * 63 * , 179 ( 2012 ) .
k. b. jinesh , and j. w. m frenken , phys .
* 101 * , 036101 ( 2008 ) .
k. koga , x. c. zeng , and h. tanaka , phys .
* 79 * , 5262 ( 1997 ) .
r. zangi , and a. e. mark , phys .
lett . * 91 * , 025502 ( 2003 ) .
j. slovk , k. koga , h. tanaka , and x. c. zeng , phys .
e * 60 * , 5833 ( 1999 ) . v. w. a. de villeneuve , r. p. a. dullens , d. g. a. l. aarts , e. groeneveld , j. h. scherff , w. k. kegel , and h. n. w. lekkerkerker , science * 309 * , 1231 ( 2005 ) .
d. t. limmer , and d. chandler , j. chem . phys . *
137 * , 044509 ( 2012 ) .
s. kaya , d. schlesinger , s. yamamoto , j. t. newberg , h. bluhm , h. ogasawara , t. kendelewicz , g. e. brown , l. g. m. pettersson , and a. nilsson , sci . rep . * 3 * , 1074 ( 2013 ) .
v. sadtchenko , p. conrad , and g. e. ewing , j. chem . phys . * 116 * , 4293 ( 2002 ) . m. c. foster , and g. e. ewing , j. chem . phys . * 112 * , 6817 ( 2000 ) .
s. j. cox , s. m. kathmann , j. a. purton , m. j. gillan , and a. michaelides , phys .
* 14 * , 7944 ( 2012 ) .
q. chen , s. c. bae , and s. granick , nature * 469 * , 381 ( 2011 ) .
g. malescio , and g. pellicane , nature mater .
* 2 * , 97 ( 2003 ) .
m. a. glaser , g. m. grason , r. d. kamien , a. komrlj , c. d. santangelo , and p. ziherl , eur .
. lett . * 78 * , 46004 ( 2007 ) .
y. von hansen , s. gekle , and r. r. netz , phys .
lett . * 111 * , 118103 ( 2013 ) .
y. zhang , a. faraone , w. a. kamitakahara , k .- h .
liu , c .- y .
mou , j. b. leo , b. juscelino s. chang , and s .- h .
chen , proc .
usa * 108 * , 12206 ( 2011 ) .
a. k. soper , proc .
usa * 108 * , e1192 ( 2011 ) .
y. zhang , a. faraone , w. a. kamitakahara , k .- h .
liu , c .- y .
mou , j. b. leo , b. juscelino s. chang , and s .- h .
chen , proc .
usa * 108 * , e1193 ( 2011 ) .
j. bai , and x. c. zeng , proc .
usa * 109 * , 21240 ( 2012 ) .
e. a. jagla , phys .
e * 58 * , 1478 ( 1998 ) .
g. franzese , g. malescio , a. skibinsky , s. v. buldyrev , and h. e. stanley , nature * 409 * , 692 ( 2001 ) .
g. franzese , g. malescio , a. skibinsky , s. v. buldyrev , and h. e. stanley , phys .
e * 66 * , 051206 ( 2002 ) .
a. skibinsky , s. v. buldyrev , g. franzese , g. malescio , and h. e. stanley , phys .
e * 69 * , 061206 ( 2004 ) .
y. zhenyu , v. s. buldyrev , n. giovambattista and h. e. stanley , phys .
lett . * 95 * , 130604 ( 2005 ) .
l. xu , p. kumar , s. v. buldyrev , s .- h .
chen , p. h. poole , f. sciortino , and h. e. stanley , proc .
usa * 102 * , 16558 ( 2005 ) .
a. b. de oliveira , p. a. netz , t. colla , and m. c. barbosa , j. chem
* 124 * , 084505 ( 2006 ) .
a. b. de oliveira , a. n. paulo , t. colla , and m. c. barbosa , j. chem
. phys . * 125 * , 124503 ( 2006 ) .
g. franzese , and h. e. stanley , j. phys . : condens .
matter * 19 * , 205126 ( 2007 ) .
b. jr m. ney , e. salcedo , and m. c. barbosa , j chem phys * 131 * , 094504 ( 2009 ) .
b. guillot , j. mol . liq . * 101 * , 219 ( 2002 ) .
t. dotera , t. oshiro , and p. ziherl , nature * 506 * , 208 ( 2014 ) .
z. ronen , journal of physics : condensed matter * 16 * , s5371 ( 2004 ) .
d. g. lev , k. e. gubbins , r. radhakrishnan , and m. sliwinska - bartkowiak , reports on progress in physics * 62 * , 1573 ( 1999 ) .
g. franzese , journal of molecular liquids * 136 * , 267 ( 2007 ) .
a. b. de oliveira , g. franzese , p. a. netz , and m. c. barbosa , j. chem .
phys . * 128 * , 064901 ( 2008 ) .
p. vilaseca , k. a. dawson , and g. franzese , soft matter * 9 * , 6978 ( 2013 ) .
h. lwen , physica a : statistical mechanics and its applications * 235 * , 129 ( 1997 ) .
a. h. marcus , and s. a. rice , phys .
e * 55 * , 637 ( 1997 ) .
p. bladon , and d. frenkel , phys .
* 74 * , 2519 ( 1995 ) .
t. bryk , s. de panfilis , f. a. gorelli , e. gregoryanz , m. krisch , g. ruocco , m. santoro , t. scopigno , and a. p. seitsonen , phys .
lett . * 111 * , 077801 ( 2013 ) .
g. malescio , g. franzese , a. skibinsky , s. v. buldyrev , and h. e. stanley , phys .
e * 71 * , 061504 ( 2005 ) .
p. vilaseca and g. franzese , j. chem . phys . *
133 * , 084507 ( 2010 ) .
m. p. allen , and d. j. tildesley , _
computer simulation of liquids _ ( oxford university press , 1989 ) .
f. varnik , j. baschnagel , and k. binder , j. chem . phys . * 113 * , 4444 ( 2000 ) .
j. s. rowlinson , and b. widom , _ molecular theory of capillarity _
( clarendon oxford , 1982 ) .
b. d. todd , d. j. evans , and p. j. daivis , phys .
e * 52 * , 1627 ( 1995 ) .
d. banerjee , s. v. bhat , and d. leporini , liquid polymorphism * 152 * , 1 ( 2013 ) .
s. j. gerbode , u. agarwal , d. c. ong , c. m. liddell , f. escobedo , and i. cohen , phys .
lett . * 105 * , 078301 ( 2010 ) .
y. han , y. shokef , a. m. alsayed , p. yunker , t. c. lubensky , and a. g. yodh , nature * 456 * , 7224 ( 2008 ) .
y. shokef , a. souslov , and t. c. lubensky , proc .
usa * 108 * , 11804 ( 2011 ) .
s. t. cui , p. t. cummings , and h. d. cochran , j. chem
. phys . * 114 * , 7189 ( 2001 ) .
a. jabbarzadeh , p. harrowell , and r. i. tanner , j. chem . phys . * 125 * , 034703 ( 2006 )
. j. r. bordin , j. s. andrade jr . ,
a. diehl , and m. c. barbosa , j. chem . phys . * 140 * , 194504 ( 2014 ) .
a. a. milischuk , v. krewald , and b. m. ladanyi , j. chem . phys . * 136 * , 224704 ( 2012 ) . t. s. ingebrigtsen , and j. r. errington , t. m. truskett , and j. c. dyre , phys .
lett . * 111 * , 235901 ( 2013 ) .
k. zahn , r. lenke , and g. maret , phys .
. lett . * 82 * , 2721 ( 1999 ) .
e. g. solveyra , e. llave , v. molinero , g. j. a. a. soler - illia , and d. scherlis , j. phys .
chem c * 117 * , 3330 ( 2013 ) . | confinement can modify the dynamics , the thermodynamics and the structural properties of liquid water , the prototypical anomalous liquid . by considering a general anomalous liquid , suitable for globular proteins , colloids or liquid metals , we study by molecular dynamics simulations the effect of a solvophilic structured and a solvophobic unstructured wall on the phases , the crystal nucleation and the dynamics of the fluid .
we find that at low temperatures the large density of the solvophilic wall induces a high - density , high - energy structure in the first layer ( `` templating '' effect ) . in turn
, the first layer induces a `` molding '' effect on the second layer determining a structure with reduced energy and density , closer to the average density of the system .
this low - density , low - energy structure propagates further through the layers by templating effect and can involve all the existing layers at the lowest temperatures investigated .
therefore , although the high - density , high - energy structure does not self - reproduce further than the first layer , the structured wall can have a long - range effect thanks to a sequence of templating , molding and templating effects through the layers .
we find dynamical slowing down of the solvent near the solvophilic wall but with largely heterogeneous dynamics near the wall due to superdiffusive liquid veins within a frozen matrix of solvent .
hence , the partial freezing of the first hydration layer does not correspond necessarily to an effective reduction of the channel section in terms of transport properties . |
although machine learning approaches have achieved success in many areas of natural language processing , researchers have only recently begun to investigate applying machine learning methods to discourse - level problems ( litman 1994 , andernach 1996 , reithinger & klesen 1997 , wiebe et al .
1997 , dieugenio , moore , & paolucci 1997 ) .
an important task in discourse understanding is to interpret an utterance s * dialogue act * , which is a concise abstraction of the speaker s intention ; figure [ ex - das ] presents a hypothetical dialogue that has been labeled with dialogue acts . recognizing dialogue acts is critical for discourse - level understanding and can also be useful for other applications , such as resolving ambiguity in speech recognition .
however , computing dialogue acts is a challenging task , because often a dialogue act can not be directly inferred from a literal interpretation of an utterance . [ cols="^,^,<,^",options="header " , ] the effect of this modified filter varies dramatically , removing 23% ( 3224 ) to 72% ( 10,237 ) of the 14,231 phrases , as shown in figure [ filter - results ] .
however , figure [ modified - filter ] shows that , as expected , using the filter does not cause the accuracy to decrease .
in addition , it allows the system to maintain a high accuracy with fewer phrases .
in particular , dcp s accuracy is significantly higher than all s accuracy when using only 5% ( 712 ) of the phrases in all .
this suggests that the filter is effectively removing redundant phrases , to produce a more parsimonious set of phrases .
this paper presented an investigation of various methods for selecting useful phrases .
we argued that the traditional method of selecting phrases , in which a human researcher analyzes discourse and chooses general cue phrases by intuition , could miss useful phrases .
to address this problem , we introduced _
automatic _ methods that use a tagged training corpus to select phrases , and our experimental results demonstrated that these methods can outperform the manual approach .
another advantage of automatic methods is that they can be easily transferred to another tagged corpus .
our experiments also showed that the effectiveness of different methods on the dialogue act tagging task varied significantly , when using relatively small sets of phrases .
the method that used our new metric , dcp , produced significantly higher accuracy scores than any of the baselines or traditional metrics that we analyzed .
in addition , we hypothesized that repetitive phrases should be eliminated in order to produce a more concise set of phrases .
our experimental results showed that our modified lexical filter can eliminate many redundant phrases without compromising accuracy , enabling the system to label dialogue acts effectively using only 5% of the phrases .
there are a number of research areas that we would like to investigate in the future , including the following : we intend to experiment with different weightings of unsoundness and incompleteness in the dcp metric ; we believe that the simple lexical filter presented in this paper can be enhanced to improve it ; we would like to study the merits of enforcing frequency thresholds for methods that have a frequency bias ; for the semantic - clustering technique , we selected the clusters of words by hand , but it would be interesting to see how a taxonomy , such as wordnet , could be used to automate this process ; since all of the experiments in this paper were run on a single corpus , in order to show that these results may generalize to other tasks and domains , it would be necessary to run the experiments on different corpora .
the members of the verbmobil research group at dfki in germany , including norbert reithinger , jan alexandersson , and elisabeth maier , generously granted us access to the verbmobil corpora .
this work was partially supported by the nsf grant # ger-9354869 . , b. , j. moore , and m. paolucci .
1997 . learning features that predict cue usage . in _ proceedings of the 35th annual meeting of the association for computational linguistics and the 8th conference of the european chapter of the association for computational linguistics_. 8087 .
madrid , spain . , p. , d. byron , and j. allen .
1998 . identifying discourse markers in spoken dialog . in _ applying machine learning to discourse processing :
papers from the 1998 american association for artificial intelligence spring symposium_. 4451 .
stanford , california . ,
d. 1994 . classifying cue phrases in text and speech using machine learning . in _ proceedings of the twelfth national conference of the american association for artificial intelligence_. 806813 .
seattle , washington . , l. and m. marcus .
1994 . exploring the statistical derivation of transformation rule sequences for
part - of - speech tagging . in _ proceedings of the 32nd annual meeting of the association for computational linguistics_. 8695 .
las cruces , new mexico . balancing act workshop . , k. , s. carberry , and k. vijay - shanker .
1998a . computing dialogue acts from features with transformation - based learning . in _ applying machine learning to discourse processing : papers from the 1998 american association for artificial intelligence spring symposium_. 9097 .
stanford , california . , k. , s. carberry , and k. vijay - shanker .
dialogue act tagging with transformation - based learning . in _ proceedings of the 17th international conference on computational linguistics and the 36th annual meeting of the association for computational linguistics_. 11501156 .
montral , qubec , canada . , j. , t. ohara , k. mckeever , and t. hrstroem - sandgren .
an empirical approach to temporal reference resolution . in _ proceedings of the second conference on empirical methods in natural language processing_. 174186 .
providence , rhode island . ,
i. and j. pearl .
comprehension - driven generation of meta - technical utterances in math tutoring . in _ proceedings of the sixth national conference of the american association for artificial intelligence_. philadelphia , pennsylvania . | we present an empirical investigation of various ways to _ automatically _ identify phrases in a tagged corpus that are useful for dialogue act tagging .
we found that a new method ( which measures a phrase s deviation from an optimally - predictive phrase ) , enhanced with a lexical filtering mechanism , produces significantly better cues than manually - selected cue phrases , the exhaustive set of phrases in a training corpus , and phrases chosen by traditional metrics , like mutual information and information gain . |
the navier stokes equations have wide applications in many scientific and engineering fields , such as ocean currents , weather forecast , and air flow around a wing .
regardless of the wide range of their applications , whether the navier
stokes equations admit a unique solution is not known yet in three - dimensional domains for general smooth initial data .
existence of global weak solutions has been proved by leray and hopf a long time ago @xcite , but it turns out to be very difficult to prove the smoothness of the weak solutions .
many recent efforts have been devoted to global well - posedness for small initial data @xcite and blowup examples of some navier stokes type equations @xcite . driven by the various applications ,
many numerical methods have been proposed for solving the navier
stokes equations , such as the finite element methods @xcite , finite difference methods @xcite , spectral methods @xcite , the lagrange galerkin method @xcite , and the projection method for time discretization @xcite .
on one hand , conventionally , convergence of the numerical solutions were all proved by assuming that the navier
stokes equations have a unique smooth solution . without assuming well - posedness of the navier
stokes equations , convergence of the numerical solutions can not be proved a priori . on the other hand , numerical evidence of existence of blowup solutions for the navier
stokes equations is inadequate . in most practical situations , the numerical solution looks smooth and does not change much if the mesh is refined .
a natural question is , given a bounded numerical solution , what can we say about the smoothness of the true solution without assuming well - posedness of the navier
stokes equations ?
we answer this question in this paper : for any given smooth initial data , if the numerical solution is bounded in a certain norm for a small mesh size , then the navier stokes equations have a unique smooth solution and , simultaneously , the numerical solution successfully approximates the solution of the navier
stokes equations .
it is remarkable that we only need one numerical solution to draw the conclusion , instead of a sequence of numerical solutions .
to illustrate our idea , we consider the navier - stokes equations @xmath0 in a convex polyhedron @xmath1 with the dirichlet boundary condition @xmath2 on @xmath3 and the initial condition @xmath4 ( where @xmath5 is the viscosity constant ) , and we impose the condition @xmath6 for uniqueness . for simplicity ,
we focus on a simple linearized finite element method for the discretization of the navier stokes equations .
we hope that our methodology can be extended to other numerical methods as well as other nonlinear time - evolution equations for which global existence , uniqueness and regularity are unknown .
let the domain @xmath7 be partitioned into quasi - uniform tetrahedra @xmath8 , @xmath9 , and denote by @xmath10 the spatial mesh size .
for any positive integer @xmath11 , let @xmath12 be the space of globally continuous piecewise polynomials of degree @xmath11 subject to the partition of the domain , and let @xmath13 be the subspace of @xmath12 consisting of functions which are zero on the boundary @xmath3 . for any nonnegative integer @xmath14 , we denote by @xmath15 the conventional sobolev space of functions defined on @xmath7 , and denote by @xmath16 the subspace of @xmath17 consisting of functions whose traces on the boundary are zero .
let @xmath18 and @xmath19 so that @xmath20 and @xmath21 are finite element subspaces of @xmath22 and @xmath23 , respectively , with the following approximation property : @xmath24 & \inf_{q_h\in v_h}\|q - q_h\|_{l^2 } \leq c\|q\|_{h^{l+1}}h^{l+1 } , \quad & & \forall~q\in h^l , \quad\ , 0\leq l\leq 1,\end{aligned}\ ] ] for some positive constant @xmath25 which is independent of the mesh size @xmath26 .
it is well known that the following `` inf - sup condition '' @xcite is satisfied by the finite element spaces @xmath20 and @xmath21 : @xmath27 this condition guarantees the existence , uniqueness and stability of the finite element solution , but it will not be used explicitly in this paper .
we remark that other conforming finite element spaces satisfying the three conditions above can also be used in the subsequent analysis .
let the time interval @xmath28 $ ] be partitioned uniformly into @xmath29 , and denote @xmath30 .
for any sequence of functions @xmath31 , @xmath32 , ... , @xmath33 , we define @xmath34 for any given @xmath35 , we look for @xmath36 and @xmath37 as the solution of the following linearized finite element equations @xmath38 & \big(\nabla\cdot \u^{n+1}_h , q_h\big)=0 , \quad \forall\ , q_h\in v_h , \label{femeq2 } \end{aligned}\ ] ] where @xmath39 is simply the lagrange interpolation of @xmath40 .
note that for any given initial data @xmath39 and mesh sizes @xmath41 , @xmath42 , the linearized equations - have a unique finite element solution @xmath43 , @xmath44 , which satisfies the discrete energy estimate : @xmath45 this finite element scheme has been studied in @xcite , where error estimates of the finite element solution were presented by assuming that the navier
stokes equations have a smooth solution . here
we have a different purpose in using the numerical scheme - .
we define the numerical solution @xmath46 \,\,\,\mbox{and}\,\,\ , x\in\omega , \end{aligned}\ ] ] and present our main result in the following theorem .
[ mainthm ] _ for any @xmath47 there exist positive constants @xmath48 and @xmath49 such that if a numerical solution @xmath50 defined by satisfies @xmath51 and @xmath52 then the navier
stokes equations - possess a unique solution satisfying @xmath53 the positive constants @xmath48 and @xmath49 are decreasing functions of @xmath54 and independent of @xmath55 . _
theorem [ mainthm ] states that , when solving the navier
stokes equations , we do not need to assume existence , uniqueness or regularity of the solution .
instead , if we have a initial data @xmath40 and a numerical solution @xmath50 , one can pick up @xmath54 satisfying and refine the mesh according to .
if the conditions - are satisfied by one numerical solution , then one can say that , for the given initial data the navier stokes equations admit a unique solution satisfying .
conversely , if the navier
stokes equations have a smooth solution satisfying , then there exists a positive constant @xmath54 such that implies . from the proof of the theorem in the next section
, one can see that the numerical solution successfully approximates the exact solution in the sense that @xmath56 clearly , the order of this a posteriori error estimate can be improved .
it is not our purpose to present optimal - order error estimates here . in this paper , we only prove existence of such constants as @xmath48 and @xmath49 .
the @xmath57 norm used in may be replaced by some other norm .
it is interesting to find useful expressions for @xmath48 and @xmath49 in terms of @xmath54 via more delicate analysis .
then theorem [ mainthm ] can be used in practical computations of the navier
stokes equations , and can be viewed as an a posteriori error estimate .
it is possible to extend theorem [ mainthm ] to other nonlinear time - evolution equations for which global existence , uniqueness and regularity of the solution are unknown but local existence , uniqueness and regularity are known for smooth initial data . for such equations ,
our method proves global uniqueness and regularity of the solution as well as convergence of the numerical solutions in an a posteriori way .
this can be viewed as an improvement of the traditional approach on error estimates of numerical solutions ( which is based on global well - posedness assumptions that are not proved yet ) .
it is well known that a solution with the regularity is unique .
it suffices to prove the regularity of the solution . in the rest part of this paper ,
we denote by @xmath58 a generic positive constant which may depend on the parameters @xmath59 and @xmath60 , but is independent of @xmath61 , @xmath14 , @xmath62 , @xmath26 and @xmath55 .
[ prop1 ] there exists a decreasing function @xmath63 and an increasing function @xmath64 such that if @xmath65 and the solution of - satisfies @xmath66 , then @xmath67 where the functions @xmath68 and @xmath69 do not depend on @xmath55 .
lemma [ lem0 ] concerns a sobolev embedding inequality and a well - known sobolev interpolation inequality , which can be found in @xcite .
based on the proof of lemma [ lem1 ] and lemma [ lem2 ] below , one can choose @xmath77^{9}]^2}$ ] and @xmath78 , where @xmath79 and @xmath80 are some positive constants .
these expressions for @xmath81 and @xmath76 are not sharp and may be improved .
_ proof of lemma [ lem2]._@xmath86 it is well known that if @xmath87 and @xmath66 , then the navier
stokes equations have a strong solution .
for the convenience of the readers , we present a priori estimates here , where all the positive constants are independent of @xmath55 .
differentiating with respect to @xmath94 , we obtain @xmath95 integrating the equation above against @xmath88 , we obtain @xmath96 and from we further derive that @xmath97 the last two inequalities imply @xmath98 which further reduces to @xmath99^{10 } { \nonumber}\\ & \leq c+c(\|\u\|_{l^\infty(0,t;l^4)}+\|\u_0\|_{h^2})^{18 } \end{aligned}\ ] ] where we have used in estimating @xmath100 . substituting the estimate above into , we further derive @xmath101 _ proof of lemma [ lem1]._@xmath86 it is well known that the navier
stokes equations admit local @xmath109 solutions for a given @xmath109 initial data . if we let @xmath110 then from we know that @xmath111;h^1)$ ] and
so @xmath112 .
from we see that @xmath113t_*^{1/2}\\ & \leq \|\u_0\|_{h^1 } + [ c_1^*+c_1^*(c_0\|\u\|_{l^\infty(0,t_*;h^1 ) } + \|\u_0\|_{h^2})^{9}]t_*^{1/2 } , \end{aligned}\ ] ] which implies @xmath114 ^ 2 } \geq \frac{1}{[c_1^*+c_1^*[c_0+(c_0 + 1)\|\u_0\|_{h^2}]^{9}]^2 } .\end{aligned}\ ] ] therefore , lemma [ lem1 ] holds with @xmath77^{9}]^2}$ ] .
we introduce the stokes ritz projection operator @xmath115 by @xmath116 and impose the condition @xmath117 for uniqueness .
this linear projection operator has the approximation property @xmath118 for any divergence - free vector field @xmath119 and scalar field @xmath120 ; see @xcite for the proof of the case @xmath121 ; the case @xmath122 can be proved by using the inverse inequality and the bramble
hilbert lemma .
this approximation property , together with the inverse inequality @xmath123 will be used in the analysis , and we also need the discrete gronwall s inequality @xcite : [ gronwall ] _ let @xmath62 , @xmath124 and @xmath125 , @xmath126 , @xmath127 , @xmath128 , for integers @xmath129 , be nonnegative numbers such that @xmath130 and suppose that @xmath131 for all @xmath132 .
then @xmath133 _ to prove theorem [ mainthm ] , we use mathematical induction on @xmath134 since @xmath135 is the lagrange interpolation of @xmath40 , we have @xmath136 for some positive constant @xmath137 .
thus our assumption holds for @xmath138 when @xmath139 .
we shall prove that if holds for some nonnegative integer @xmath14 , then it also holds when @xmath140 is replaced by @xmath141 . to simplify the notations ,
we define @xmath142 . from the induction assumption
we see that @xmath143 .
therefore , when @xmath144 lemma [ prop1 ] implies @xmath145 under this regularity , the solution @xmath146 satisfies the variational equations @xmath147 for @xmath148 , where @xmath149 , \\ & { \bf f}^{n+1}=\frac{1}{2}\u^{n+1}\otimes(\u^{n+1}-\u^{n } ) , \end{aligned}\ ] ] are the truncation errors due to the time discretization , satisfying @xmath150 let @xmath151 and @xmath152 .
then the difference between - and - gives @xmath153 & \big(\nabla\cdot \e^{n+1}_h , q_h\big)=0 , \label{erreq2 } \end{aligned}\ ] ] substituting @xmath154 into the equation and using the identities @xmath155 we obtain @xmath156 note that @xmath157 & \sum_{n=0}^{k}\tau |{\bf e}^{n+1}|_{h^{-1}}^2 + \sum_{n=0}^{k}\tau|{\bf f}^{n+1}|_{l^2}^2 \leq c\phi(m)^4\tau^2 , \\[5pt ] & \sum_{n=0}^{k}\tau ( \| \u_h^{n+1}\|_{l^3}^2+\|\nabla \u^{n+1}\|_{l^2}^2h+ \| \u^{n+1}\|_{l^4}^2h^{1/2}+\| \u_h^{n}\|_{l^4}^2h^{1/2 } ) \sum_{m = n}^{n+1}(\|\u^{m}\|_{h^2}^2+\|p^{m}\|_{h^1}^2 ) \\
& \leq \bigg[\tau\| \u_h^{0}\|_{l^4}^2h^{1/2}+c\sum_{n=0}^{k}\tau ( \| \nabla\u_h^{n+1}\|_{l^2}^2 + \|\nabla \u^{n+1}\|_{l^2}^2 ) \bigg ] \big(\|\u \|_{l^\infty(0,t_{k+1};h^2)}^2+\|p\|_{l^\infty(0,t_{k+1};h^1)}^2\big ) \\ & \leq \bigg[c\| \u_{0}\|_{h^2}^2+c\sum_{n=0}^{k}\tau \|\nabla \u^{n+1}\|_{l^2}^2 \bigg ] \phi(m)^2 \\ & \leq \bigg[c\|
\u_{0}\|_{h^2}^2 + c\|\nabla\u \|_{l^2(0,t_{k+1};l^2)}^2 + c\tau^2 \|\partial_t\nabla\u\|_{l^2(0,t_{k+1};l^2)}^2 \bigg]\phi(m)^2 \\ & \leq c\phi(m)^4 , \end{aligned}\ ] ] where we have used in estimating @xmath158 and used the expression @xmath159 in estimating @xmath160 . by using the last three inequalities and summing up for @xmath161 ( with @xmath162 ) , we obtain @xmath163 by using the expression @xmath164
, we derive that @xmath165 we see that when @xmath166 we have @xmath167 , satisfying the condition of lemma [ gronwall ] , and so @xmath168 substituting the estimate above into , we obtain @xmath169 from the last inequality we derive that , for any @xmath170 , @xmath171 and for any @xmath172 $ ] we have @xmath173 and @xmath174 which implies that @xmath175 when @xmath176 we have @xmath177 and this completes the mathematical induction on .
overall , by mathematical induction , when the mesh conditions , and are satisfied , we have @xmath178 and , as a consequence of lemma [ lem2 ] , the solution @xmath146 possesses the regularity .
theorem [ mainthm ] is proved with @xmath179 and @xmath180 r. bermejo , p.g .
del sastre , and l. saavedra , _ a second order in time modified lagrange
galerkin finite element method for the incompressible navier stokes equations _ , siam j. numer .
, 50 ( 2012 ) , pp .
30843109 .
j. donea , s. giuliani , h. laval , and l. quartapelle , _ finite element solution of the unsteady navier stokes equations by a fractional step method _ , comput
( 1982 ) , pp .
5373 .
heywood and r. rannacher , _ finite element approximation of the nonstationary navier stokes problem .
i. regularity of solutions and second - order spatial discretization _
, siam j. numer .anal . , 19 ( 1982 ) , pp .
275311 .
heywood and r. rannacher , _ finite - element approximation of the nonstationary navier stokes problem part iv : error analysis for second - order time discretization _ , siam j. numer .anal . , 27 ( 1990 ) ,
353384 .
r. rannacher , _ finite element methods for the incompressible navier stokes equations _ , in `` fundamental directions in mathematical fluid mechanics '' , pp . 191293 , edited by g.p .
galdi , j. heywood and r. ranancher .
birkhuser , basel , boston , berlin , 2000 . | we prove that if the initial data is smooth enough and a finite element solution of the three - dimensional navier
stokes equations is bounded in a certain norm with a small mesh size , then the navier
stokes equations have a unique smooth solution . |
despite the fact that dsrs suffer from the lack of a consistent and well - established mathematical structure similar to what exists for sr , @xmath0-minkowski spacetime and @xmath0-poincar are likely the richest frameworks which can be assigned to dsrs , at least so far .
therefore , in the present paper , we just adopt the dsr interpretation proposed in magueijo - smolin model @xcite as one of the conceivable phenomenological frameworks of @xmath0-minkowski spacetime . as an interesting and promising feature of the mentioned framework
, it can be used to explain astrophysical data received from grbs ( note that gamma - ray bursts ( gbrs ) along with supernovae , neutron stars and black holes are four main pillars of relativistic astrophysics ) @xcite . with these preliminaries and in the context of some prevalent phenomenological issues through combination of the modified mass shell condition ( [ e1-a ] ) and @xmath0-deformed lorentz transformations ( [ e1-b ] ) and ( [ e1-c ] ) , we examine and derive some phenomenological consequences of a @xmath0-minkowski relativistic model .
specifically , we treat the elastic scattering problem with the planck energy cutoff by focusing on some relativistic kinematical properties of this scattering problem .
we are looking for the mentioned objectives in these steps : in section 2 we study the effects of the planck energy upper bound , @xmath0 , on the elastic scattering process where no new particles are created during the process .
section 3 is devoted to the investigation of the _ rapidity _ as one of the most important kinematical variables involving in the heavy ion collisions in the presence of the @xmath0 deformation of the minkowski spacetime .
the paper follows with a conclusion along with a brief remark in section 4 .
in this section , we investigate the elastic scattering processes where no new particles are created during the process in a @xmath0-deformed minkowski space - time . to this end
, we firstly present an explicit relation for @xmath0-deformed energy and momentum .
suppose a clock that is fixed at the position @xmath39 of a rest frame , @xmath40 .
if the clock emits signals in a regular time interval @xmath41 , then according to the modified lorentz transformations ( [ e1-b ] ) , an observer located in the moving system @xmath42 , measures this time interval as follows @xmath43\;.\ ] ] since the clock is fixed at the system @xmath40 , then @xmath44 , which gives @xmath45 this equation can be rewritten in terms of the proper time @xmath46 as @xmath47 in this framework we have @xmath48 where @xmath49 is the classical velocity and @xmath17 marks the rest mass .
now , inserting the @xmath0-deformed relativistic momentum ( [ e2 - 3 ] ) into the modified dispersion relation ( [ e1-a ] ) , we derive the following equation @xmath50=0\;.\ ] ] solving this equation we find @xmath51\right]^{1/2}}{2 \left(1-\frac{m^{2}c^{4}}{\kappa^{2}}\right)}\;.\ ] ] the positive sign in this solution is acceptable since for the case with @xmath52 and @xmath53 , naturally the @xmath0-deformed energy ( [ e2 - 5 ] ) should reduce to the relativistic result @xmath54 . by applying the approximation @xmath55 , eq .
( [ e2 - 5 ] ) can be rewritten as @xmath56 where @xmath57}\;,\ ] ] is a dimensionless running constant .
note that @xmath58 . in order to study the classical collision process
, the center of mass ( cm ) coordinate system is an appropriate tools to derive many of kinematic relations , while in the relativistic framework it is meaningless to speak about the cm system . in sr theory ,
mass and energy are related so that it is common in sr kinematics that one uses a center of momentum coordinate system instead of the cm . of course , in the center of momentum coordinate as cm , the total linear momentum of the system is zero . in the context of the relativistic collision ,
the laboratory coordinate system is related to the inertial system @xmath40 and the center of momentum system @xmath42 ( the moving system ) via a lorentz transformation . in dsr framework
we follow the standard procedure with the difference that laboratory system @xmath40 and the moving system @xmath42 are related now by the @xmath0-deformed lorentz transformations .
so , if a particle of rest mass @xmath59 which moves in one dimension collides elastically with a particle of the rest mass @xmath60 , then in the center of momentum system , we have denote the magnitude of the 3-momentum of particles @xmath61 in the cm system so that @xmath62 . an observer located in the cm
frame sees these two particles are moving towards each other so that depending on the direction , one has @xmath63 or @xmath64 . ]
@xmath65 at the first glance this equality seems to be misleading .
it is obvious that from the deformed lorentz transformations one naturally gets the deformed dispersion relation ( [ e1-a ] ) ( since this transformation keeps ( [ e1-a ] ) invariant ) .
therefore , one has also the deformed conservation of energy - momentum so that this relation then governs on the nature of how particles behave in collisions .
more precisely , one expects that in the standard momentum conservation law ( [ e2 - 8 ] ) , @xmath0-deformation must be regarded . while this is sensible in essence , as we show equation ( [ e2 - 8 ] ) is still valid up to the first order correction . + by using the relation @xmath66 , the space components of the momentum 4-vector in ( [ e2 - 8 ] ) can be written as @xmath67 since @xmath68 , @xmath69 and also @xmath70 . according to the deformed lorentz transformation ( [ e1-c ] ) and using ( [ e2 - 6 ] ) ,
the transformation of the momentum @xmath71 ( from @xmath40 to @xmath42 ) is written as @xmath72 finally by replacing this relation into the left hand side of relation ( [ e2 - 8 * ] ) , we arrive at the following relation @xmath73 where this relation can be solved for @xmath74 and @xmath75 in terms of @xmath76 to find @xmath77 and @xmath78 in the absence of the @xmath0 deformation of the minkowski space - time background i.e. for @xmath29 , then @xmath79 and @xmath80 .
so , the above equations reduce to the flowing special relativistic counterparts @xmath81 and @xmath82 now we write the transformation equations for momentum components between the moving system @xmath42 and the laboratory system @xmath40 after the scattering .
it is obvious that after scattering we have both @xmath83 and @xmath84 components of the momentum .
the @xmath83 and @xmath84 components of the momentum after scattering in the laboratory system @xmath40 read as follows @xmath85 and @xmath86 by introducing the angle of scattering @xmath87 in the laboratory frame and by dividing ( [ e2 - 15 ] ) with ( [ e2 - 14 ] ) , we get @xmath88 where by inserting @xmath89 , it takes the following form @xmath90 by the same procedure , for the recoiled particle we find respectively @xmath91 and @xmath92 introducing an angle of scattering @xmath93 in the laboratory frame for the recoiled particle and then by dividing ( [ e2 - 19 ] ) with ( [ e2 - 18 ] ) , we find @xmath94 for the special case with @xmath95 , i.e. with @xmath96 we find @xmath97 therefore , for the case @xmath95 , the angles @xmath87 and @xmath93 take the following forms respectively @xmath98 and @xmath99 so , the angle between the directions of the scattered and recoiled particles is given by @xmath100 . through the relation @xmath101
, one finds that for @xmath102 the angle @xmath103 has the maximum @xmath104 as expected , for the case @xmath38 , eq .
( [ e2 - 24 ] ) recovers its relativistic counterpart as @xmath105 therefore , one finds that in the @xmath0-deformed non - commutative geometry , the maximum amount of the included angle @xmath103 is dependent on the two dimensionless parameters @xmath76 and @xmath106 . in order to have a qualitative understanding of eq .
( [ e2 - 24 ] ) , we plot variation of @xmath107 in terms of @xmath76 in figure 1 .
if we set the @xmath106 to be a running constant , this figure gives a qualitative description of the situation .
for any given value of @xmath76 , by increasing the value of @xmath106 the maximum included scattering angle @xmath107 increases .
so , we can say that correction of @xmath106 due to the @xmath0-deformed minkowski spacetime results in a shift of the maximum included scattering angle @xmath107 . in what follows , to see another prediction of the setup , we assume that the recoil angle @xmath93 is negligible
. then we track the effect of @xmath0 modification on the kinetic energy @xmath108 of the scattered particle 1 by the target particle 2 .
using the inverse @xmath0-modified lorentz transformation ( [ e1-c ] ) , we have @xmath109 where @xmath110 and @xmath111 represent the total energy of particle 1 after collision in the laboratory and center of momentum frames , respectively . by substituting @xmath112 and @xmath113 into eq .
( [ e2 - 26 ] ) , we get @xmath114 by introducing @xmath115 and @xmath116 as the kinetic energies of the particle 1 before and after scattering respectively , then one can write @xmath117 using equation ( [ e2 - 20 m ] ) , this equation can be rewritten as follows @xmath118 now we introduce a scattering angle @xmath119 in the center of momentum frame in terms of the @xmath87 in the laboratory frame . for this purpose , by squaring eq .
( [ e2 - 22 ] ) , we find @xmath120 where by definition @xmath121 the solution of equation ( [ e2 - 30 ] ) , after some rearrangement , reads as follows @xmath122}\times \nonumber\\ & & \left\{\big[4\alpha^{6}f^{2}(\gamma_{1})-8\alpha^{8}f^{2}(\gamma_{1})+4\alpha^{4}f^{4}(\gamma_{1 } ) + 4\alpha^{8}f^{4}(\gamma_{1})\big]\right .
\nonumber\\ & & \left .
\cos^{2}(2\psi)+ 8\alpha^{8}f^{4}(\gamma_{1})\cos(2\psi)-4\alpha^{6}f^{2}(\gamma_{1 } ) \right .
\nonumber\\ & & \left .
+ 4\alpha^{8}f^{2}(\gamma_{1})-4\alpha^{4 } f^{4}(\gamma_{1})+4\alpha^{6}f^{4}(\gamma_{1})+\right .
\nonumber\\ & & \left
. 4\alpha^{8}f^{4}(\gamma_{1})\right\}^{1/2}\end{aligned}\ ] ] through a straightforward calculation , one can show that equation ( [ e2 - 32 ] ) in the limit of @xmath53 recovers its special relativistic counterpart as follows @xmath123 finally , by inserting this relation along with ( [ e2 - 21 ] ) into equation ( [ e2 - 29 ] ) , we are able to show the qualitative behavior of @xmath124 in terms of the laboratory scattering angle @xmath87 , as shown in figure 2 . as we see in this figure , contrary to our expectation based on the sr theory ( @xmath79 ) and even classical mechanics , the ratio @xmath124 in terms of @xmath87 results in the simple relation @xmath125 which the laboratory scattering angle @xmath87 is limited to the interval @xmath126 . for more details
one can see classical mechanics " text book such as ref .
if @xmath127 then @xmath128 , namely , no collision takes place since the speed remains constant before and after the collision and no momentum is transferred to particle-2 .
however , if @xmath129 then @xmath130 ( @xmath131 ) and particle-2 is forwardly scattered with the same kinetic energy before owned by particle-1 . ]
( that is , @xmath132 , in eqs . ( [ e2 - 29 ] ) and ( [ e2 - 33 ] ) ) , for the case @xmath127 we have @xmath133 . here , we are confronted with an unusual status in which as soon as we turn on @xmath0-spacetime deformation , the nature of two - body collision alters from elastic to inelastic one .
it is important to stress that the overall behavior of the curves in figure 2 are independent of the values of the dimensionless relativistic parameter @xmath134 .
+ at this point a question arises whether it is possible to remove these unusual phenomenological effects via @xmath0 deformation of the momentum conservation law ( [ e2 - 8 ] ) .
to check this situation , inspired by @xcite we modify eq .
( [ e2 - 8 ] ) as follows @xmath135 let s focus carefully on this modified momentum conservation law in the cm system .
since in the adopted dsr proposal ( the ms proposal ) one has the spacetime translational invariance , the energy and momentum are conserved .
however , dsr proposals are non - linear in essence .
this means that 4-momentum of a system of two particles is not just a linear summation of two particles momentums . as a consequence , in passing from sr to dsr , i.e. from linear to non - linear relativity
, the additivity of energy and momentum is replaced with non - additivity character . for more insight on this issue
we refer the reader to @xcite . now
the modified relation ( [ e2 - 34 ] ) can be rewritten as follows @xmath136 where @xmath137 . up to the first order of the modification ( i.e. @xmath138 ) ,
the second and third terms on the right hand side of this relation can be neglected .
this means that , up to the first order correction , momentum conservation law ( [ e2 - 8 ] ) is still valid and consequently the obtained unusual outcomes are direct consequences arising from extended phenomenological framework at hand .
+ as the final remark in this section , we note that the apparent energy non - conservation appeared above has been conjectured in nc manifolds and could be due to a discrete ( underlying ) time evolution @xcite .
so , noether s theorem should be correspondingly reformulated for the case of nc spacetime translations .
in this section we firstly present an important kinematical variable , the rapidity , that relates the momentum of the particle to the dynamics of a heavy - ion reaction .
there is the possibility of creation of several particles after collision .
the momentum of each particle can be decomposed into a longitudinal component @xmath139 and a transverse component @xmath140 with reference to the collision axis .
the longitudinal momentum of a particle , due to its dependence on the velocity of the cm frame with respect to the laboratory frame , is not so conventional variable in literature . besides , in order to analyze some experimental outputs , it is essential to be able to view them from the cm frame s perspective .
as has been mentioned previously , we use the word `` cm frame '' for center of momentum frame in this context . by introducing a kinematical variable as the rapidity @xmath84 , one is able simply to choose or change the reference frame .
this is arising from the fact that unlike the velocity , the variable @xmath84 is defined in special relativity to be additive under a sequence of the lorentz transformations along the same direction . in
what follows , we answer the question whether the variable @xmath84 still remains additive under successive @xmath0-deformed lorentz transformations ( [ e1-c ] ) . in three dimensional space , modified on shell condition ( [ e1-a ] ) , can be rewritten as follows @xmath141 so that @xmath142 is known as transverse mass . by writing eq .
( [ e3 - 1 ] ) as @xmath143 , then the variable @xmath84 can be defined in terms of the energy and momentum as follows @xmath144 therefore , for a @xmath0-deformed special relativity scenario with dispersion relation as eq .
( [ e1-a ] ) , one finds that the rapidity / energy - momentum relation reads as eq .
( [ e3 - 3 ] ) which is similar to its relativistic counterpart .
so , the relation between velocity and rapidity now is obtained from eq .
( [ e3 - 3 ] ) as @xmath145 this relation gives the rapidity @xmath84 as follows @xmath146 since @xmath147 to investigate the additivity of rapidity under a sequence of the @xmath0-deformed lorentz transformations , we consider the transformation of the momentum vector under a change of the reference frame along the collision axis .
it is obvious that under such a transformations , @xmath140 and @xmath148 are unchanged .
therefore , under a @xmath0-deformed lorentz transformation , the energy and longitudinal component of the momentum transform as @xmath149 note that here a prime marks the quantities that are measured by an observer in the laboratory system which moves with the velocity @xmath150 with respect to the cm frame of reference in which the energy @xmath15 and momentum @xmath139 are measured .
putting eq .
( [ e3 - 1 ] ) along with @xmath151 and @xmath152 derived from eq .
( [ e3 - 4 ] ) into eq .
( [ e3 - 7 ] ) , we get @xmath153 therefore , the rapidity @xmath154 as seen in the laboratory frame can be read as follows @xmath155 while this relation reflects the fact that rapidity is not an additive quantity under a sequence of the @xmath0-deformed lorentz transformations , for the case @xmath79 ( that is , under standard lorentz transformations ) this relation restores the additivity property , i.e , @xmath156 . therefore , in passing from the standard sr to @xmath0-deformed sr , one loses the additivity nature of rapidity .
this is much similar to losing additivity rule of speeds in passing from galilean relativity to special relativity .
note that relation ( [ e3 - 9 ] ) is a direct result of the nonlinear representation of lorentz transformations .
technically , as is shown in ref .
@xcite , the magueijo - smolin proposal of dsr can be realized via deformed translation invariance without need to deformation of the ls ( this is also the case for all other proposals of dsr , see @xcite ) . the @xmath0-minkowski spacetime modified translation generator in the ms model of dsr
is given by @xmath157 . in this respect ,
the origin of deviations from sr observed in this paper comes back to the deformed translation invariance since this really controls how momentum of particles behaves in collisions .
noncommutative geometry ( spacetime ) is undoubtedly one of the richest framework among other alternatives to pursue physics at the planck scale . in this paper , we were concerned on a particular form of noncommutative spacetimes which highly takes care of lorentz symmetry and known as @xmath0-minkowski spacetime .
the noncommutative parameter @xmath0 can be conceived as quantum gravity scale where high energy physics is trying to unravel its mysteries .
the mentioned noncommutative parameter is observer independent which leads to a natural connection between @xmath0-minkowski spacetime and doubly special relativity theories ( dsrs ) .
therefore , noncommutative geometry parameter @xmath0 is responsible for cutoff energy scale featured in dsr theories @xcite as the seconded invariant . in the new framework ,
we are faced with dsr formalism as a nonlinear @xmath0 extension of the lorentz transformations and dispersion relations of special relativity . in this work , through focusing on the magueijo - smolin version of dsr theories @xcite , we have investigated how such energy upper bound will affect some relativistic kinematical parameters in a typical collision of particles . as a first step , by applying @xmath0-deformed lorentz transformations ( [ e1-b ] ) , ( [ e1-c ] ) along with modified mass shell condition ( [ e1-a ] ) , we have investigated the effect of the planck energy cutoff on the elastic scattering processes for the case that no new particles are produced in the process .
we have shown ( as figure 1 indicates ) that the existence of @xmath106-term caused by the @xmath0-minkowski spacetime , leads increment of maximum included scattering angle @xmath107 relative to the standard case . while the classical and special relativistic kinematics indicate that the maximum included scattering angle is bounded as @xmath158 , the @xmath0-deformed nonlinear extension of special relativity violates this bound since for angles larger than @xmath159 , there is probability of detection of scattered particles . in the same way , by assuming that the recoil angle @xmath93 is negligible , we have treated the effect of the planck energy upper bound ( @xmath0 modification ) on the kinetic energy @xmath108 of the scattered particle .
surprisingly , in contrast to classical mechanics and even special theory of relativity , we have observed that for the case @xmath127 and @xmath160 , @xmath133 as figure 2 shows .
considering the fact that for the case @xmath79 one recovers the standard result @xmath161 , in the context of the @xmath0 deformed extension of the special relativity one tempts to think that spacetime seems to have some sort of dissipative effects at quantum gravity scale .
the apparent energy non - conservation appeared in our context could be due to a discrete ( underlying ) time evolution as shown in ref .
it is also expected that the noether s theorem must be correspondingly reformulated in the present case of nc spacetime translations . as an important result
, we have shown that unlike the special relativity , rapidity in @xmath0-minkowski spacetime is no longer an additive quantity under a sequence of @xmath0-deformed lorentz transformations .
this is reminiscent of the fact that one loses the additive nature of speeds in passing from galilean relativity to special relativity .
we note that the observed deviations from sr in this paper originate from the deformed translation invariance , not necessarily to the deformed ls .
in fact , this deformed translation invariance is that controls how momentum of particles behaves in collisions .
the outputs of our study support the idea that the main feature of dsr theories is the deformed translation symmetry , not the deformed ls as has been discussed also in @xcite . finally , about the algebraic structure of @xmath0-minkowski spacetime if we look at the full algebra generated by boosts , rotations and translations , then the algebra is really the lorentz algebra .
depending on the choice of realization , even the translation part can be un - deformed and therefore having the full poincar algebra .
but this is only on the algebra level , that is what governs the one particle representations .
however the coalgebraic sector is deformed , leading to nontrivial multi - particle states , which are the ones appearing in collisions .
the @xmath0-minkowski algebra is compatible with deformed poincare algebra , usually @xmath0-poncar algebra ( but one can also go beyond this in general to deformations of igl - hopf algebra ) for which we have that the lorentz sector is intact , but depending on realization the translation part is deformed . only in the classical basis ( or natural realization ) the translation part is also intact . anyway , for any realization ( natural , bicrossproduct etc ) the coalgebraic sector is deformed , meaning that the coproduct of boosts , rotation and translation is not primitive , and this gives some new interesting phenomenon for the multiparticle states , and therefore collisions @xcite . + * acknowledgement * + we are indebted to two anonymous referees for detailed and very insightful comments .
99 s. m. carroll , g. b. field and r. jackiw , phys . rev .
d 41 , 1231 ( 1990 ) s. coleman and s. l. glashow , phys .
b * 405 * , 249 ( 1997 ) d. colladay and v. a. kostelecky , phys .
d * 55 * , 6760 ( 1997 ) r. m. mansouri , r. u. sexl , gen .
* 8 * , 497 ( 1977 ) ; ibid .
* 8 * , 515 ( 1977 ) j. a. lipa , j. a. nissen , s. wang , d. a. stricker , d. avaloff , phys .
* 90 * , 060403 ( 2003 ) .
g. amelino - camelia , j. ellis , n. e. mavromatos and d. v. nanopoulos , int . j. mod .
a * 12 * , 607 ( 1997 ) g. amelino - camelia , j. ellis , n. e. mavromatos , d. v. nanopoulos and s. sarkar , nature * 393 * 763 ( 1998 ) s. d. biller et al , phys
. rev . lett .
* 83 * , 2108 ( 1999 ) t. kifune , astrophys . j. lett . *
518 * , l21 ( 1999 ) r. aloisio , p. blasi , p. l. ghia and a. f. grillo , phys .
d * 62 * , 053010 ( 2000 ) r. j. protheroe and h. meyer , phys
b * 493 * , 1 ( 2000 ) j. alfaro , h. a. morales - tecotl and l. f. urrutia , phys .
* 84 * , 2318 ( 2000 ) g. amelino - camelia and t. piran , phys .
d * 64 * , 036005 ( 2001 ) g. amelino - camelia , nature * 408 * , 661 ( 2000 ) . s. sarkar , mod .
a * 17 * , 1025 ( 2002 ) d. v. ahluwalia , mod .
a * 17 * , 1135 ( 2002 ) t. jacobson , s. liberati and d. mattingly , phys .
d * 66 * , 081302 ( 2002 ) s. hossenfelder , l. smolin , phys .
canada * 66 * , 99 ( 2010 ) s. doplicher , k. fredenhagen , j. e. roberts , phys .
b * 331 * , 39 ( 1994 ) s. doplicher , k. fredenhagen , j. e. roberts , commun .
* 172 * , 187 ( 1995 ) m. r. douglas , and n. a. nekrasov , rev .
mod . phys . * 73 * , 977 ( 2001 ) .
a. pacho , journal of physics : conference series * 442 * , 012039 ( 2013 ) i. m. gelfand , m. a. naimark , math .
sbornik * 12 * , 2 ( 1943 ) n. seiberg and e. witten , jhep * 9909 * , 032 ( 1999 ) j. lukierski , h. ruegg , phys .
b * 329 * , 189 ( 1994 ) j. lukierski , h. ruegg , a. nowicki , v. n. tolstoi phys . lett .
b * 264 * , 331 ( 1991 ) s. majid , h. ruegg , phys . lett .
b * 334 * , 348 ( 1994 ) j. lukierski , a. nowicki , h. ruegg , phys .
b * 293 * , 344 ( 1992 ) s. meljanac , z. skoda , m. stojic , arxiv : math.qa/1409.8188 ( 2014 ) s. meljanac , z. skoda , d. svrtan , sigma * 8 * , 013 ( 2012 ) s. kresic - juric , s. meljanac , m. stojic , eur .
j. c * 51 * , 229 ( 2007 ) p. vitale , j. christophe wallet , jhep * 04 * , 115 ( 2013 ) a. gere , t. juric , j. christophe wallet , jhep * 12 * , 045 ( 2015 ) .
a. pacho , arxiv : math - ph/1112.5366 ( 2011 ) t. juric , s. meljanac , a. samsarov , j. phys .
( 2015 ) 1 , 012005 paolo aschieri , arxiv : hep - th/0703013 ( 2007 ) p. aschieri , m. dimitrijevic , p. kulish , f. lizzi , j. wess , `` noncommutative spacetimes '' , lecture notes in physics , vol .
774 isbn : 978 - 3 - 540 - 89792 - 7 , springer , ( 2009 ) a. agostini , g. amelino - camelia and f. dandrea , int
a * 19 * , 5187 ( 2004 ) m. dimitrijevic , l. jonke , l. moeller , e. tsouchnika , j. wess and m. wohlgenannt , eur .
j. c * 31 * , 129 ( 2003 ) g. amelino - camelia and m. arzano , phys .
d * 65 * , 084044 ( 2002 ) p. kosinski , j. lukierski , p. maslanka , phys .
d * 62 * , 025004 ( 2000 ) j. lukierski , h. ruegg and w. j. zakrzewski annals .
phys . * 243 * , 90 ( 1995 ) s. meljanac , a. samsarov , j. trampetic , m. wohlgenannt , jhep * 1112 * , 010 ( 2011 ) s. meljanac , a. samsarov , int . j. mod
a * 26 * , 1439 ( 2011 ) h. grosse , m. wohlgenannt , nucl .
b * 748 * , 473 ( 2006 ) l. freidel , j. kowalski - glikman , s. nowak , phys .
b * 648 * , 70 ( 2007 ) j. kowalski - glikman , a. walkus , mod .
a * 24 * , 2243 ( 2009 ) g. amelino - camelia , phys .
b * 510 * , 255 ( 2001 ) + g. amelino - camelia , int .
d * 11 * , 35 ( 2002 ) j. magueijo and l. smolin , phys .
lett . * 88 * , 190403 ( 2002 ) ; j. magueijo and l. smolin , phys .
d * 67 * , 044017 ( 2003 ) .
j. kowalski - glikman and s. nowak , int .
d * 12 * 299 ( 2003 ) s. das , s. ghosh , d. roychowdhury , phys .
d * 80 * , 125036 ( 2009 ) + s. das , d. roychowdhury phys .
d * 81 * , 085039 ( 2010 ) g. amelino - camelia and t. pairan , phys .
d * 64 * , 036005 ( 2001 ) s. corley , phys .
d * 57 * , 6280 ( 1998 ) s. corley and t. jacobson , phys .
d * 59 * , 124011 ( 1999 ) a. blaut , j. kowalski - glikman , d. nowak - szczepaniak , phys .
b * 521 * , 364 ( 2001 ) .
g. salesi , phys .
d * 85 * , 063502 ( 2012 ) j. kowalski - glikman arxiv : hep - th/0209264 ( 2002 ) s. sarkar , mod .
a * 17 * , 1025 ( 2002 ) r. aloisio , p. blasi , a. galante , p. l. ghia , a. f. grillo , astropart
* 19 * , 127 ( 2003 ) a. borowiec , kumar s. gupta , s. meljanac , a. pachol , europhys . lett . * 92 * , 20006 ( 2010 ) g. amelino - camelia , nature * 418 * , 34 ( 2002 ) j. kowalski - glikman , phys . lett .
b * 547 * , 291 ( 2002 ) n. r. bruno , g. amelino - camelia , j. kowalski - glikman , phys .
b * 522 * , 133 , ( 2001 ) j. kowalski - glikman , phys .
a * 286 * , 391 ( 2001 ) g. amelino - camelia , d. benedetti , f. dandrea , a. procaccini . class .
. grav . * 20 * , 5353 ( 2003 ) j. kowalski - glikman , arxiv : hep - th/0209264 ( 2002 ) j. kowalski - glikman , lect .
notes phys .
* 669 * , 131 ( 2005 ) g. amelino - camelia , j. kowalski - glikman , g. mandanici , a. procaccini , int . j. mod
a * 20 * , 6007 ( 2005 ) a. abdo et al .
, science * 323 * , 1688 ( 2009 ) g. amelino - camelia and l. smolin , phys .
d * 80 * , 084017 ( 2009 ) p. kosinski , j. lukierski , p. maslanka , j. sobczyk , mod .
lett . a * 10 * , 2599 ( 1995 ) j. kowalski - glikman , s. nowak , phys . lett .
b * 539 * , 126 ( 2002 ) j. lukierski and a. nowicki , acta phys .
b * 33 * , 2537 ( 2002 ) s. mignemi , phys . lett .
b * 672 * , 186 ( 2009 ) s. judes , m. visser , phys .
d * 68 * , 045001 ( 2003 ) s. ghosh , p. pal , phys .
d * 75 * , 105021 ( 2007 ) s. meljanac , a. samsarov , m. stojic , k. s. gupta eur .
j. c * 53 * , 295 ( 2008 ) t. r. govindarajan , kumar s. gupta , e. harikumar , s. meljanac , d. meljanac phys .
d * 77 * ( 2008 ) 105010 t. juric , s. meljanac , r. strajn int . j. mod
a * 29 * , 1450022 ( 2014 ) s. meljanac , a. pachol , a. samsarov , kumar . s. gupta phys
d * 87 * , 125009 ( 2013 ) s. meljanac , m. stojic eur .
j. c * 47 * , 531 ( 2006 ) n. jafari , a. shariati , aip conf .
proc . * 841 * , 462 ( 2006 ) g. amelino - camelia , symmetry * 2 * , 230 ( 2010 ) s. das , s. pramanik , s. ghosh , sigma * 10 * , 104 ( 2014 ) g. amelino - camelia and s. majid , int
a * 15 * , 4301 ( 2000 ) keith r. symon , mechanics " , isbn : 978 - 0201073928 , 3rd ed . , addison - wesley , ( 1971 ) + s. t. thornton and j. b. marion , classical dynamics of particles and systems " , isbn : 964 - 01 - 0783 - 2 , 3rd ed . ,
saunders college publishing / harcourt brace , ( 1988 ) .
a. p. balachandran , a. g. martins and p. teotonio - sobrinho , jhep * 0705 * , 066 ( 2007 ) this part of the text is taken directly from the referees report with minor changes . | the doubly special relativity ( dsr ) theories are suggested in order to incorporate an observer - independent length scale in special theory of relativity .
the magueijo - smolin proposal of dsr is realizable through a particular form of the noncommutative ( nc ) spacetime ( known as @xmath0-minkowski spacetime ) in which the lorentz symmetry is preserved . in this framework
, the nc parameter @xmath0 provides the origin of natural cutoff energy scale .
using a nonlinear deformed relativistic dispersion relation along with the lorentz transformations , we investigate some phenomenological facets of two - body collision problem ( without creation of new particles ) in a @xmath0-minkowski spacetime . by treating an elastic scattering problem , we study effects of the planck scale energy cutoff on some relativistic kinematical properties of this scattering problem .
the results are challenging in the sense that as soon as one turns on the @xmath0-spacetime extension , the nature of the two - body collision alters from elastic to inelastic one .
it is shown also that a significant kinematical variable involving in heavy ion collisions , the rapidity , is not essentially an additive quantity under a sequence of the nonlinear representation of the lorentz transformations .
+ * keywords * : doubly special relativity ; noncommutative spacetime ; nonlinear lorentz transformations ; elastic scattering ; rapidity . |
one of the currently most debated and consequential issues in astrophysics is the formation of massive elliptical galaxies ( see * ? ? ? * for a recent review ) . from the theoretical point of view
, cdm simulations can accurately predict the redshift evolution of dark matter haloes . specifically , on large scales cdm simulations have enjoyed great success in accounting for the growth through cosmic times of the structures - clusters , filaments and voids - starting from the extremely smooth initial conditions inferred from the cosmic microwave background ( e.g. * ? ? ? * ) .
we still lack , however , an understanding of the astrophysical processes which produce massive galaxies with the properties that these show in the local universe : gas cooling , star formation , stellar and agn feedback , are all poorly -if at all- understood processes , which are currently parametrized with _ ad hoc _ recipes in galaxy formation models .
simulations have been less successful on these lower , galactic scales ( @xmath17 kpc ) and it is not clear yet whether this is due to the just mentioned complexities of baryon physics , or to other reasons . from the observational point of view , the emerging picture is still uncertain . in the local universe ,
studies of the stellar populations of massive ellipticals indicate a formation epoch for the bulk of their stars at redshifts @xmath18 ; however , the old stellar ages at @xmath9 can not break the degeneracy between the mass assembly of old smaller sub - units relative to in - situ star formation @xcite .
this degeneracy can in principle be removed by observations of the evolution of the number density of ellipticals as a function of redshift ; however , small statistics , cosmic variance and details in the different analysis of the high@xmath4 samples have generated a debate as to whether massive ellipticals are already fully assembled by @xmath19 a fact which , if true , would require considerable rethinking of the currently favored galaxy formation models ( see , e.g. , bell et al .
2004a , and daddi et al .
2005 for examples of different views , and renzini 2006 for an extensive review and further references ) .
some studies have suggested that the majority of the most massive ( @xmath20 ) early - type galaxies ( etgs ) is assembled over the last few gyr from the dissipationless ( `` dry '' ) mergers of less massive etgs @xcite .
estimates of the merger rate since @xmath19 have been attempted by several authors ( e.g. van dokkum 2005 ; bell et al 2006a ; lin et al .
2004 ; kampczyk et al .
@xcite used the statistics of relict tidal interactions in the local universe and conclude that @xmath21 of bulge - dominated galaxies should have experienced a merger with mass ratio @xmath22 1:4 since @xmath23 . based on 6 etg - etg close pairs in the gems field , @xcite estimate that each present - day etg with @xmath24 has undergone @xmath250.52 major dry mergers since @xmath26 .
several other studies however dispute such high frequencies of dry mergers since @xmath27 .
for example , @xcite , based on the 3d two - point correlation function of etgs , conclude that only 20% of all @xmath28m@xmath29 galaxies have experienced a major merger since @xmath30 .
this result roughly agrees with the analysis of lin and collaborators who , from the count of close pairs in the deep2 sample , conclude that only @xmath31 of present - day @xmath32 galaxies have undergone a major merger over the same period . supporting evidence against a large contribution of dry mergers to
the formation of massive elliptical galaxies comes also from estimates of the dry merging rate at @xmath33 from the sdss database @xcite , which show a @xmath34 probability per gyr for an etg to merge with another etg .
the masjedi et al .
estimate implies a dry - merging rate much lower than the rate at which etg - hosting dark matter halos merge with one another @xcite , possibly highlighting a potential problem with cdm simulations , or in the way these simulations are related to the observed galaxy populations . in any event , there are clearly major discrepancies among current attempts at measuring the dry merger rate , which in part are the result of the uncertainty affecting the time it will take to a given galaxy pair to merge , or to some tidal debris to disappear , and in part to small - number statistics . in this paper
we follow a different approach in order to set limits to the role played by dry mergers in establishing the population of massive etgs in the local universe . here
we map the evolution with redshift up to @xmath11 of the number density of massive etgs .
since the bulk of stars in these galaxies formed at much higher redshifts ( see the extensive literature quoted in renzini 2006 ) , an increase in the number density of massive etgs since @xmath16 should be ascribed to dry mergers , alternatively , a lack of evidence for such an increase would allow us to set limits on the dry merging rate .
we use the data from the cosmos program @xcite to study the evolution of the luminosity function ( lf ) of the etgs up to redshift @xmath35 .
we base the analysis on the sample of @xmath36 , @xmath37 cosmos galaxies with reliable photometric redshift in the range @xmath38 presented in ( * ? ? ?
* hereafter paper i ) to extract a combined sample of 3980 morphologically- and/or photometrically - selected progenitors of early - type systems to study , in the @xmath9 to @xmath35 redshift window , their lf evolution in the rest - frame @xmath39band .
specifically , we select two complementary samples of possible progenitors of massive etgs , and study their individual and combined redshift evolution . the first sample is selected morphologically , using the classification of zest ( zurich estimator of structural types , paper i ) .
the second sample is selected photometrically , using the red - sequence identified , at all redshifts , on the rest
frame @xmath40 diagram .
both samples are further constrained by requiring that , by passive - evolution fading , the high@xmath4 progenitors of etgs evolve into @xmath9 systems that lie on the @xmath10 `` kormendy relation '' @xcite .
our double ( i.e. , morphological and photometrical ) selection is motivated by the fact that , at @xmath9 , elliptical galaxies are characterized by well defined stellar population dynamical / structural properties .
specifically , @xmath9 ellipticals _ both _
@xmath41 have uniformly old stellar populations with a small scatter in the observed colors , implying a small scatter in the formation epoch of their stellar populations ; _ and _ @xmath42 are described by regular surface brightness distributions , well represented by almost devaucouleurs density profiles , indicating a large role of violent relaxation in their dynamical history .
depending on the relative timing and importance of in - situ star formation versus stellar mergers , @xmath19 progenitors of ellipticals may thus appear as already morphologically - relaxed systems or possibly show passively - evolving stellar populations with irregular morphologies due to recent mergers .
therefore , to identify _ all _ plausible high@xmath4 progenitors of local etgs we consider the sample created by the union of the morphologically and photometrically selected galaxies .
this sample allows us to be comprehensive in the `` counting '' of the number density of massive galaxies at earlier epochs .
still , selection biases remain in our analysis : the photometric - selected sample could include galaxies which are actually not progenitors of ellipticals and have red colors simply because of the effects dust extinction . on the other hand , blue progenitors with merger morphologies would not enter in either of the samples .
we discuss later in the paper the impact of these biases on our conclusions .
the paper is organized as follows .
section 2 describes the data and the basic measurements ; section 3 describes in detail the sample selection criteria ; section 4 quickly summarizes the structural properties of the photometrically - selected etgs , and the color properties of the morphologically - selected sample ; section 5 presents our main analysis , i.e. , the evolution with redshift of the rest - frame @xmath1-band lf of the different and also combined samples of high-@xmath43 etg progenitors .
we discuss our results in section 6 and highlight a few concluding remarks in section 7 .
appendix a presents the tests performed on the lfs to account for the photometric redshift uncertainties , and appendix b presents the sdss - based @xmath9 comparison sample . throughout the paper we assume @xmath44 , @xmath45 , and @xmath46
km s@xmath47 mpc@xmath47 .
all magnitudes are ab magnitudes @xcite , unless otherwise specified .
the analysis presented in this paper uses the cosmos hubble space telescope ( hst ) advanced camera for survey ( acs ) f814w images ( hereafter @xmath48 ) , and ancillary ground - base uv to near - infrared data available for the cosmos field .
the ground - based @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 data were acquired with the subaru telescope , the @xmath55-band data with the canada
hawaii telescope , and the infrared @xmath56 images with the cerra tololo international observatory and kitt peak national observatory telescopes .
the observations and the data processing are described in detail in @xcite , @xcite , for the ground - based data ; and in @xcite for the hst - acs data .
photometric catalogs were created separately for the ground - based and the hst data .
the acs based catalog was generated by @xcite using sextractor @xcite in a two steps strategy , in order to correctly detect and deblend objects with a wide range of magnitudes and sizes .
extensive simulations presented in @xcite show that the acs catalog is at least 90% complete for objects smaller than 14 and @xmath37 ( roughly corresponding to a surface brightness limit of @xmath57 mag arcsec@xmath58 ) .
the ground - based catalog was generated using sextractor in dual - image mode .
the detection was performed in the original best seeing psf @xmath59band image , while the photometry was measured in the psf - matched images .
this process ensured that the photometry of nearby galaxies was optimally deblended in the final catalog .
magnitudes were measured within apertures of @xmath60 diameter ; and @xmath61 magnitude limits are 26.4 , 27.3 , 27.0 , 26.6 , 26.8 , 26.2 , 25,2 , and 21.6 in the @xmath55 , @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 , and @xmath56 filters , respectively @xcite . in this work , we present the results for the central area of the cosmos field of @xmath250.74 degrees@xmath3 covered by the 260 acs pointings acquired during hst cycle 12 observations .
we limit our study to the @xmath36 galaxies with total magnitude ( as measured by the sextractor mag_auto ) brighter than @xmath62 and photometric redshifts ( see next section ) in the redshift range 0.2@xmath63 .
we adopt the maximum likelihood photometric redshift estimates obtained for the cosmos galaxies by @xcite with the zurich extragalactic bayesian redshift analyzer ( zebra ) .
our photo@xmath4 were obtained using as a basic set of galaxy templates the empirical spectra by @xcite and @xcite , that cover the spectral types from elliptical to the star - forming galaxies . by means of an iterative technique
zebra automatically corrects their original set of templates to best represent the galaxy spectral energy distributions ( seds ) in different redshift bins ( to empirically take into account , e.g. , dust absorption effects and other possible inadequacies inherent in the original set of templates ) .
the availability of a `` training set '' of spectroscopically
derived @xmath64cosmos redshifts @xcite for a small fraction of the whole photometric sample under investigation allows zebra to achieve an optimal correction of the galaxy templates and a precise calibration of the photometric redshifts , and thus accurate photo@xmath4 estimates .
the resulting photometric redshifts have an accuracy of @xmath65 over the whole redshift range considered in the current analysis , relative to the @xmath43cosmos spectroscopic redshifts of @xmath66 galaxies .
tests on the @xmath67 spectroscopic sample artificially fainted to reproduce a @xmath68 sample , indicate photometric redshift errors @xmath69 .
we checked that our main results remain unchanged when : @xmath41 changing the redshift binning adopted by zebra for the template optimization , the minimum errors in the photometry allowed by zebra , the corrections applied to the photometric catalog ( i.e. , the pliantness parameter @xmath70 and the regularization parameter @xmath71 , that appear in the @xmath72 minimization approach of the zebra code ; feldman et al .
2006 ) ; @xmath42 using only galaxies with @xmath73 in the zebra fits , with @xmath74 the value corresponding to the 95@xmath75 percentile in an ideal @xmath72-distribution of a varying number of degrees of freedom ; or all galaxies in the sample , independent of their @xmath72 in the zebra photometric fits . for a total of @xmath76 of the galaxies that were detected in the original acs based catalogue it was not possible to derive photometric redshift estimate , due to either their absence , or their blending with other galaxies in the ground - based catalogue .
these galaxies were excluded from our analysis . the usable cycle 12 @xmath37 acs - based catalogue contains 32540 galaxies with photometric redshift in the range [ 0.2,1.0 ] and @xmath73 . to extract the sample of _ morphologically - selected _ etgs at high redshifts , and also to study later on in the paper the morphologies of the photometrically selected sample of etgs
, we used the zest classification presented in paper i. zest quantitatively describes the galaxy structure using three variables ( @xmath77 ; @xmath78 ; @xmath79 ) , obtained by performing a principal component analysis ( pca ) in the five - dimensional parameter space of asymmetry ( @xmath80 ) , concentration ( @xmath81 ) , gini coefficient ( @xmath82 ) , the 2nd order moment of the brightest 20% galaxy pixels ( m@xmath83 , e.g. , * ? ? ?
* ; * ? ? ?
* ) , and the ellipticity of the light distribution ( @xmath84 ) .
these non parametric diagnostics provide complementary , but also redundant information on galaxy structure . with the pca we found that the first three @xmath85 variables explain more than 90% of the variance in the original dataset , and thus completely describe the galaxy structure .
the morphological classification is performed in the space with axes @xmath86 . to associate a ( dominant ) morphological class to different regions in the @xmath87space
, the latter was partitioned into a regular 3d - grid with unit steps in each of the coordinates , and all the galaxies in the cosmos sample within each of the unit @xmath85-cubes were visually inspected . each galaxy was then assigned the morphological class of the @xmath87cube corresponding to its position .
the zest classification associates to each @xmath87cube a type ( @xmath88 for early - type galaxies ; @xmath89 for disk galaxies ; and @xmath90 for irregular galaxies ) .
furthermore , each @xmath87cube classified as type 2 is assigned a `` bulgeness '' parameter according to the median value ( @xmath91 ) of the distribution of @xmath92 sersic indices of all galaxies brighter than @xmath93 in that cube .
the two - dimensional gim2d fits were performed by @xcite on the cosmos - bright sample , i.e. , @xmath66 ; we verified in paper i that the pca provides consistent results if separately applied to the faint ( i.e. , @xmath37 ) and the bright ( i.e. , @xmath66 ) samples .
the `` bulgeness '' parameter is related to the @xmath94 ratiodisk galaxies , the @xmath92 sersic index of the _ global _ galaxy profile is found to be monotonically related to the galaxy @xmath94 ratio ( e.g. , * ? ? ?
] ( see paper i for detail ) and ranges from 0 ( bulge dominated disk galaxies , type 2.0 ) to 3 ( pure disk galaxies , type 2.3 ) .
we use @xmath91 to assign the bulgeness parameter to each type 2-classified pc - cube , according to the following scheme : bulgeness @xmath95 and @xmath96 respectively for @xmath97 , @xmath98 , @xmath99 , and @xmath100 .
the distributions of sersic @xmath92 for all galaxies classified as type 2 , splitted according to their bulgeness parameter are shown in figure 9 of sargent et al .
( 2007 ) . in paper
i we discuss in detail the uncertainties and the systematic errors in the measured structural parameters as a function of signal - to - noise ratio ( s / n ) , and to what extent the cosmos - calibrated zest morphological classification grid is affected by the s / n of the individual galaxies ; our tests show that the zest morphological classification is robust down to @xmath101 .
furthermore , the zest classification is substantially more efficient in disentangling different galactic types than simpler classification schemes based on , e.g. , a threshold in @xmath92-sersic index , or two / three of the non - parametric diagnostics mentioned above . in order to compute the rest
frame absolute @xmath39band magnitude ( @xmath102 ) , we define a color @xmath103correction @xmath104 , where @xmath1 is the magnitude at @xmath105 , and @xmath106 indicates the observed @xmath107band magnitude for a galaxy at redshift @xmath43 .
the central wavelength of the acs filter ( f814w ) corresponds to the central wavelength of the @xmath39band at redshift @xmath108 , therefore @xmath109 .
it follows that the rest
frame absolute magnitude @xmath102 can be expressed as : @xmath110 where @xmath111 is the luminosity distance at redshift @xmath43 .
the color @xmath112 depends on the galaxy spectral energy distribution . for each galaxy
we therefore use its zebra best fit template and the corresponding photometric redshift to compute the @xmath113 color . the rest
frame @xmath114 colors are derived from the ground - based photometry , by interpolating the observed seds at the wavelengths of the @xmath115 and @xmath116 filters , redshifted at the galaxy photometric redshift .
the adopted effective wavelengths of the @xmath115 and @xmath116 filters are 3841 , and 5479 , respectively .
typical accuracy in the rest - frame @xmath114 color is of 0.05 mag for @xmath117 , and 0.1 for @xmath118 .
the morphologically - selected sample was constructed by including all galaxies classified by zest as type 1 , i.e. , as early type galaxies , and all bulge - dominated type 2.0 disk galaxies .
the latter are located in volumes of @xmath85space that are adjacent to the type 1 early - type galaxies , and are therefore characterized by very similar structural properties ( see paper i ) .
indeed , type 2.0 galaxies have bulge properties very similar to the type 1 systems , and differ from these only for a clearly detected ( non dominant ) disk component . down to the magnitude limit of our analysis ( @xmath101 ) , in the fraction of the cosmos field under investigation , and in the redshift range @xmath119
there are a total of 2352 and 1352 galaxies , with good photometric redshift that are respectively classified as type 1 and type 2.0 galaxies by zest .
both samples have similar sersic @xmath92 distribution , with 80% and 70% of type 1 early - type and type 2.0 bulge dominated galaxies , respectively , having @xmath120 .
we highlight in the sections below further similarities between type 1 and type 2.0 galaxies .
the inclusion of these ( heavily ) bulge - dominated disk galaxies in our morphologically - selected sample of etgs also facilitates the comparison with other studies , since these have generally tended to include such systems in the analysis ( see , e.g. , * ? ? ?
etgs in the local universe follow well defined scaling relations .
of particular importance is the fundamental plane @xcite and , if the velocity dispersion is not available , its photometric projection , i.e. , the so - called `` kormendy - relation '' @xcite . the kormendy relation is a correlation between the half
light radius of the galaxy and the average surface brightness within the half light radius . selecting etg progenitors at high redshifts on the basis of their elliptical - like morphologies is equivalent to requiring that these progenitors are already fully assembled and dynamically - relaxed galaxies at those earlier epochs : they become @xmath9 early - types by evolution of their stellar populations ( and possibly some modest amount of additional star formation )
. therefore , in constructing a morphologically - selected sample of high@xmath4 etgs , we can apply an additional constraint , namely the surface brightness must be such that with plausible surface brightness evolution it evolves in to the kormendy relation .
clearly , surface brightness evolution may be uncertain , especially for blue objects with composite stellar population , however we can take a conservative approach and estimate a minimum evolution that leads to the minimum number of excluded objects . in figure [
fig : kormendy ] we show the average surface brightness within the half
light radius in the rest - frame @xmath39band as a function of the half
light radius for the type 1 ( early - type galaxies , left panel ) and type 2.0 ( bulge - dominated disk galaxies , right panel ) galaxies of our morphologically - selected sample .
we use different colors to represent galaxies with different zebra best fit to their seds ; in particular , red points are galaxies fitted by zebra with an `` elliptical - galaxy '' template , and blue points represent later spectral types @xcite . dotted and dot - dashed lines show the @xmath101 magnitude limit at the central redshift of the bin , for galaxies with the sed of a star - forming and an elliptical galaxy , respectively .
typical error - bars ( i.e. , estimated at the median size and magnitude of the sample galaxies ) are shown on the upper right part of each panel .
the errors are calculated by also taking into account the effects of the photometric redshift .
the half - light radii used in figure [ fig : kormendy ] are measured using a growth curve analysis on the acs @xmath48 images . by comparison with the half - light radii measured with gim2d by sargent et al .
( 2007 ) we found that psf effects ; which are accounted for with gim2d , start to be important only for galaxies with measured @xmath12102 . at these sizes , @xmath122 is typically overestimated by @xmath123% compared with the gim2d measurement .
we find similar results by analyzing a set of @xmath124 simulated galaxies with sersic index @xmath125 and 4 .
we decided not to correct for this effect , since the correction depends on the exact shape of the galaxy surface - brightness profile , which is unknown .
we note , however , that our sample includes only @xmath126% of galaxies with measured @xmath12102 , so the impact of the psf correction on the final selected sample of kormendy - consistent galaxies is minimal . in the figure
, the solid black line represents the best linear fit derived for the coma cluster galaxies @xcite , passively evolved at the central redshift of the bin assuming @xmath127 [ i.e. , the evolution of a single stellar population model with formation redshift @xmath128 ; @xcite ] .
the jrgensen et al .
( 1995 ) data were converted from johnson vega to ab magnitudes using the relation @xmath129 .
the solid thin lines represent the best fit relation at the two extremes of the redshift bins .
the dashed lines indicate a distance from the @xmath9 best fit line of twice its @xmath130 dispersion .
figure [ fig : kormendy ] shows that up to redshift @xmath19 , for a given size , objects that have a zebra elliptical - like classification tend to have , on average , higher surface brightness relative to those with later spectral types .
furthermore , the figure shows that , up to @xmath131 , many galaxies lie significantly below the local kormendy relation . a fraction of order @xmath2580% of these `` low - density '' systems
have the zebra seds of late - type , star - forming galaxies . we note that at redshifts higher than @xmath132 selection effects are responsible for the absence of galaxies with @xmath133 mag arcsec@xmath58 and radii of few kiloparsecs . in order to quantify the criterion for the exclusion of implausible progenitors of etgs on the basis of the @xmath9 kormendy relation ,
we followed the procedure described in @xcite , namely : * we assigned a fading rate of @xmath134 to each galaxy , based on the zebra best fit to its sed .
* we considered two extreme star formation histories ( sfhs ) , both with an initial formation redshift @xmath128 .
the first sfh describes the passive evolution of a single burst stellar population and was assigned to all galaxies with a zebra `` elliptical galaxy '' best - fit .
the second sfh was assigned to galaxies best fitted by zebra with a star - forming sed .
this sfh is described by a constant star formation rate of 1m@xmath135yr up to the redshift at which the galaxy is observed ( if larger than @xmath136 ) , and by zero sfr from the galaxy redshift ( or @xmath136 ) to @xmath9 .
the lowest redshift possible for the truncation of the star formation , i.e. , @xmath136 , was chosen in order for the @xmath9 stellar population of the so - evolved galaxies to have colors consistent with the observed scatter in the color
magnitude relation of nearby etgs @xcite .
the sfhs , i.e. , the @xmath134 fading rates , for intermediate zebra photometric types were derived by interpolating between the two extreme cases . the evolution of the sfhs was computed assuming a constant metallicity equal to the solar value . * finally , we excluded from the sample of `` kormendy - compatible '' , morphologically - selected progenitors of etgs all objects that , when evolved to @xmath9 with the above evolutionary tracks , were found to lie away from the local kormendy relation kormendy - relation , as we discussed in the text , is derived using galaxies belonging to the coma cluster .
@xcite -see also discussion in @xcite- found that there are very small differences in the fundamental plane derived for galaxies in cluster or in the field , and these differences remain small up to redshift @xmath137 @xcite .
] at a distance more than @xmath138 times its observed scatter . using a threshold of @xmath139 times the observed scatter would increase the number of galaxies included in the final sample of kormendy - compatible etgs by @xmath140% , without changing our main conclusions on the evolution of the number density of bright etgs .
this selection might in principle be affected by the specific value of the adopted formation redshift ( @xmath128 ) and the approach of using the zebra best fit templates to determine the @xmath39band surface brightness evolution of each galaxy .
however , neither of these choices affects our final conclusions .
first , if some of the galaxies which have a zebra elliptical - galaxy fit were actually reddened star - forming objects , then the number of rejected objects would be minimized . assuming a slower @xmath39band luminosity evolution ( i.e. , a higher formation redshift of , e.g. , @xmath141 ) would change the expected brightening at redshift @xmath35 by @xmath250.2 magnitudes , resulting in a final sample of kormendy - consistent etgs only @xmath126% larger than the sample obtained using @xmath128 .
our choice of using @xmath128 is supported by recent results , based on the evolution of the fundamental - plane , showing that @xmath142 early - type galaxies have @xmath143 @xcite , with no significant differences for etgs in cluster and field environment .
althought @xmath144 measured by @xcite is a luminosity weighted mean star - formation epoch , the star - formation timescale for massive etgs , estimated from @xmath9 ellipticals , is shorter than 1 gyr .
furthermore , @xcite and @xcite show that the star formation in massive galaxies is largely completed by @xmath145 . although the evolution of the stellar - population for less massive etgs is less constrained , there is evidence that less massive etgs evolve faster than high mass ones , both in the field and in cluster @xcite . in the light of our results
a mass independent formation redshift is thus a conservative approach , since it gives the smallest number of rejected low luminosity etgs .
the kormendy - cut increases the fraction of galaxies in the morphologically - selected ( type @xmath146 ) sample which also have the zebra sed of an elliptical galaxy : this increases from 53@xmath147 to 63@xmath147 .
particularly affected are the lowest redshift bins up to @xmath132 , in which the fraction of early - type morphologies with an early - type sed increases from 53@xmath147 to 70@xmath147 .
we note that about 46% of the excluded zebra late - type galaxies at redshifts @xmath148 are faint systems with @xmath149 . in the highest redshift bin ,
the fraction of excluded objects with zebra late - type seds reduces to 20% , and only 1% of the zebra elliptical type objects do not pass the kormendy - relation test .
a similar fraction of type 1 and type 2.0 galaxies are excluded on the basis of the kormendy relation constraint : the fraction of type 2.0 galaxies in the final morphologically - selected sample changes from 37% to 35% before and after the kormendy - cut , i.e. , it remains basically constant .
the exclusion of the `` kormendy - rejected '' objects implies a cut of about 20% in the original sample of zest morphologically - selected early type galaxies : our _ sample of morphologically - selected _ progenitors of etgs contains 2730 galaxies ( 1798 type 1 and 932 type 2.0 systems ) .
progenitors of @xmath9 massive etgs may appear , at @xmath19 , as morphologically - irregular , merging systems with old and passively evolving stellar populations .
these galaxies can be selected on the basis of their sed properties . to define our cosmos sample of photometrically - selected progenitors of @xmath9 etgs
, we used the red - sequence identified on the rest - frame @xmath7-@xmath8 color - magnitude diagram , as derived in the two - step procedure that we describe below . in the left panel of figure [ fig : cm ]
we show the rest - frame @xmath40 color magnitude diagram for the cosmos @xmath150 galaxy sample , split in four redshift bins .
the gray levels in each panel represent the weighted galaxy volume density in bins of 0.15 of @xmath8 magnitude and 0.04 of @xmath114 color ; the density in each color magnitude bin is computed as @xmath151 .
the @xmath152 value is the maximum volume within which a galaxy @xmath153 , with a given observed @xmath48 magnitude and spectral type , is detectable in the cosmos survey ; the weights @xmath154 are the corrections required so as to account for objects excluded either because no redshift is available , or because the redshift estimate is less accurate than the required threshold ( see section [ sec : photoz ] , and also paper i for details on the derivation of the @xmath154 values ) .
the grey scale bar at the top of each panel shows the density in mpc@xmath155 corresponding to the grey intensity levels .
the presence of radial color gradients in galaxies can have an effect on the measured colors and on the slope of the red sequence , depending on the size of apertures used to compute colors and magnitudes ( see , e.g. * ? ? ?
although not easy to quantify , an estimate of this effect can be evaluated as follows .
the galaxy colors in the cosmos photometric catalogue are computed within @xmath60 diameter apertures , regardless of galaxy redshift @xcite .
the local magnitude - size relation implies a half light radius of @xmath156 kpc at @xmath157 and of @xmath158 kpc at @xmath159 ; therefore , when no evolution is taken into account , the cosmos @xmath60 aperture correspond to @xmath160 and @xmath161 half light radii for bright and faint galaxies , respectively .
assuming a de vaucouleur profile and using the average @xmath114 color gradient derived by scodeggio ( 2001 ; @xmath162 ) , the @xmath114 colors computed within 1.5 , 2.0 , 2.5 , and 5 @xmath122 differ from the color computed within 0.3 @xmath122 respectively by @xmath163 , @xmath164 , @xmath165 , @xmath166 magnitudes . given the assumptions on the shape of the galaxy surface - brightness profile , on the value for the local color - gradient , and on the size - luminosity relation , the estimates of the effect of the color gradient on the measured colors can only be considered as indicative , and we therefore decided not to correct the data , but rather to consistently derive the slope of the red sequence from the observed cosmos color - magnitude diagram in each redshift bin .
color for the cosmos galaxies .
we divide all galaxies in four redshift bins ( one redshift per column ) and three magnitude bins of 1.5 magnitudes for each redshift ( as indicated on the right side of the plot ) .
the number in the upper - left corner of each panel indicates the median error on the @xmath114 colors in each magnitude redshift bin ( section [ sec : slope ] ) .
the dotted vertical line shows the ( @xmath114)@xmath167 color that we use to broadly isolate the red sequence in each redshift - magnitude bin ; the color of the peak in each bin ( ( @xmath114)@xmath168 ) is indicated by the vertical dashed line .
the histograms are normalized to the total number of objects in each magnitude - redshift bin.,width=340 ] in figure [ fig : newhistocolor ] we show the normalized distribution of the rest - frame @xmath114 color of all galaxies split in both redshift and absolute magnitude . in particular , we consider four redshift bins , one per column , and three bins of 1.5 magnitudes centered at the values indicated in the figure ( top : @xmath169 ; middle : @xmath170 ; bottom : @xmath171 ) .
all panels in the figure show a degree of bimodality between red and blue galaxies ( see e.g. , * ? ? ? * for a discussion of color bimodality at high redshift ) , albeit with varying degrees of sharpness .
we are nonetheless able to identify in each of the panels the color of the `` red peak '' , which we use as the starting point for the computation of the slope of the red sequence in each of the individual redshift bins .
specifically , in each magnitude redshift bin , we define @xmath172 as the color which broadly isolates the _ initial red sequence _ ; @xmath173 is defined as the color of the minimum in the valley between the blue and red peak in each of the color distributions , and it is indicated as a dotted line in each panel of figure [ fig : newhistocolor ] . the @xmath174 color is then computed as the median of all colors redder than @xmath173 .
the @xmath174 colors is indicated with a black dashed vertical line in each panel of figure [ fig : newhistocolor ] .
the number in the upper - left corner of each panel indicates the median error on the @xmath174 colors in each magnitude redshift bin ; in some bins , these are negligible .
these median errors are conservatively obtained considering the maximum error derived by interpolating the errors in each of the two adjacent passbands that are used to derive the rest - frame @xmath7 color of the galaxy .
we tested that varying the adopted color thresholds ( i.e. , @xmath174 ) within @xmath175 does not affect the resulting sample of photometrically selected etgs . we then compute the slope of the _ initial red sequence _ in each redshift bin by fitting a linear relation to the @xmath176 points ( and their errors ) .
the four slopes , as a function of increasing redshift , are : @xmath177 , @xmath178 , @xmath179 , and @xmath180 . within the errors , these values are consistent with the average of @xmath181 . given the measured uncertainties , we therefore choose to keep fixed in our analysis the average slope of the red - sequence to the `` consistency value '' of @xmath182(@xmath183 ) for all redshifts . for each galaxy in each redshift bin ,
a `` slope - corrected '' color is then obtained by setting @xmath184 .
the zero point of the initial red sequence at the mean redshift of each bin is thus defined as the color of the red peak of the @xmath185 distributions .
these zero point colors @xmath186 are equal to @xmath187 , @xmath188 , @xmath189 , and @xmath190 at the redshift values corresponding to the center of our bins , i.e. , @xmath191 , and @xmath192 , respectively .
the error - bars are computed by estimating the shifts in the red peak of the @xmath185 distributions induced by a change in the red - sequence slope of @xmath193 .
the above - defined initial red - sequence curves are shown , for each redshift bin , as solid lines in the left panel figure [ fig : cm ] ; in each panel , the dashed line is @xmath194 away from the initial red sequence .
the initial red sequence derived above coarsely mixes galaxies within broad redshift bins . to refine the evolution with redshift of the zero point , we plot in figure [ figumb_z ] the values of @xmath186 derived above as a function of redshift .
specifically , figure [ figumb_z ] shows the rest frame @xmath114 color as a function of redshift for an absolute @xmath116 magnitude of @xmath195 , passively evolved by assuming a single burst stellar population formed at redshift @xmath128 and metallicity @xmath196 .
the color evolution of the model is shown in figure [ figumb_z ] as a dashed line .
color of the initial red sequence calculated , as a function of redshift , at a magnitude @xmath197 passively evolved using a single - burst stellar population with formation redshift @xmath128 and metallicity @xmath198 .
the color evolution of this model is shown as a dashed line .
the error bars are those listed for the @xmath186 in section [ sec : zp ] .
the @xmath9 point is derived applying the same procedure adopted for the cosmos galaxies to the comparison sdss sample .
the thick solid line shows the linear fit to the color evolution of the zero point of the red sequence , derived including the sdss comparison data point .
the thin solid lines are 1 @xmath199 away from the best fit.,width=340 ] the @xmath9 point shown in figure [ figumb_z ] is obtained by applying to the comparison sdss sample described in appendix [ app : sdss ] ( and discussed in detail in * ? ? ?
* ) , the same procedure that we have described above for the cosmos galaxies .
the observed redshift evolution of the @xmath114 color of the red - sequence at @xmath200 is consistent with the evolution predicted for the single stellar population formed at redshift @xmath128 , that also reproduces the @xmath9 @xmath114 color of the red - sequence of the sdss etg sample .
a linear fit in the @xmath201 redshift range ( i.e. , including the sdss point ) to the @xmath186-@xmath43 relationship at the evolved @xmath169 magnitude gives : @xmath202 this best fit is shown in figure [ figumb_z ] as a thick solid line ; thin solid lines are located 1 @xmath199 away from the best fit .
this result is in agreement with the redshift evolution of ( @xmath114)@xmath203 found by @xcite .
we use this best fit , together with the slopes measured as described in section [ sec : slope ] , to define the final red sequence at any redshift @xmath43 .
we point out that the red sequence that is identified by our two - step procedure is more robust towards contamination by interlopers than a red sequence that is obtained , e.g. , by keeping the slope fixed to a value derived from local samples of etgs ( e.g. , * ? ? ?
using a non - optimally chosen slope introduces undesirable effects since this basically determines the fraction of faint galaxies in the resulting red sequence selected sample .
estimates for the slope of the @xmath9 red - sequence in local galaxy clusters range from @xmath204 to @xmath205 , depending on the aperture adopted for the measurements @xcite . assuming a too steep slope for the data under study would lead to the inclusion of a large fraction of faint blue interlopers .
the final red sequence defines our _ photometrically - selected sample _ of high-@xmath43 progenitors of @xmath9 etgs .
this sample includes all galaxies with @xmath114 color redder than the value corresponding to `` minus @xmath206 '' from the final red sequence at the appropriate redshift , i.e. , @xmath207 .
the sample contains 3844 galaxies with @xmath37 and photometric redshift in the range @xmath208 .
red - sequence selected galaxies in the redshift range @xmath209 down to @xmath210 in the combo17 survey , in an area similar to the cosmos area presented here .
if we consider the same redshift range , and relax our constrain on the accuracy of the photometric redshift we find @xmath211% more photometrically - selected cosmos etgs than combo17 .
this is consistent with our being and @xmath101 selected sample and their being a @xmath210 selected sample .
we also stress that ours and bell et al.s samples are not directly comparable , since bell et al . fixed the red - sequence slope to the value of @xmath165 found in local galaxy clusters , while we consistently compute the red - sequence slope , from the observed color magnitude diagram . ] the right panel of figure [ fig : cm ] highlights in color the galaxies which belong to this photometrically - selected sample , and specifically in red the galaxies that have a zebra elliptical - galaxy fit , and in blue those with zebra late - type spectral fits .
below ( section [ sec : kps ] ) we argue that it is sensible to apply the kormendy - test described above also to the so - derived photometrically - selected sample ; we will thus define as _ photometrically - selected sample _ those galaxies in the sample constructed so far that also pass the `` kormendy - test '' .
in figure [ fig : colors ] we show , separately for the considered four redshift bins , the @xmath7-@xmath8 color magnitude diagram for the kormendy - test - consistent , morphologically - selected sample of early type galaxies discussed in section [ sec : ms ] . in particular , black filled circles show the zest type 1 early - type galaxies , and cyan circles represent the bulge - dominated type 2.0 disk galaxies .
the red line in each panel reproduces the final red - sequence calculated at the center of the redshift bin derived in section [ sec : zp ] . at any redshift
, the majority of morphologically - selected early type galaxies is characterized by red colors ( consistent with the final red sequence defined above ) ; however , a fraction of order 40% of these systems has significantly bluer colors than the red sequence ( see tables [ tbl : samples ] and [ tbl : fraction ] ) .
these galaxies would be missed in a pure color - based selection .
this fraction is similar to that of morphologically - selected galaxies at @xmath9 that do not belong to the red - sequence , see table 1 in @xcite .
the morphologically - selected sample of etg progenitors covers the entire observed range of rest
frame @xmath7 colors , although very few of them have @xmath212 .
this is not a consequence of the kormendy - relation selection that has been applied to the sample : this is demonstrated by the black dots in each panel , which show the galaxies with morphological type @xmath146 that are excluded from the final morphological sample on the basis of the kormendy - test .
these `` kormendy - test - rejected '' galaxies are typically faint systems ( of order @xmath213 ) ; there is however no obvious dependence of this galaxy population on the @xmath114 color . versus @xmath8 for the morphologically - selected sample of etgs .
black and cyan filled circles are used for galaxies kept in the sample , after passing the kormendy - relation test described in section [ sec : kormendy ] .
black filled circles are used for the zest type 1 ( early - type morphology ) galaxies , while cyan filled circles are used for the zest type 2.0 ( bulge - dominated disk ) galaxies .
black dots show the galaxies that were excluded from the morphologically - selected sample on the basis of the kormendy test .
the red line in each panel reproduces the final red - sequence derived in section [ sec : zp ] calculated at the center of the redshift bin.,width=340 ] as evident from figure [ fig : colors ] , the color distributions of zest type 1 and type 2.0 galaxies are rather similar ( apart from a small blue `` excess '' at low redshifts of bulge - dominated disks at bright magnitudes ) .
this is further illustrated in figure [ fig : histumvtypes ] , where the rest frame @xmath114 color distributions are shown for both types as solid and dashed histograms , respectively . in each redshift bin , the type 1 and type 2.0 samples are split in two components , one brighter ( left panels ) and the other fainter ( right panels ) than the value of @xmath214 at the center of the given redshift bin ; the @xmath214 values are derived from the schechter fits discussed in section [ sec : lf ] , and are given in the top - right corner of each diagram .
in the left panel of figure [ fig : morphology ] we show the fractions of photometrically - selected early type galaxies ( derived in section [ sec : cmselection ] ) , that belong to different zest morphological classes . to make the comparison among different redshift bins meaningful
, we limit the analysis to @xmath215 at redshift @xmath216 , and we evolve this magnitude limit as a function of redshift according to the evolution of a single - burst stellar population formed at @xmath128 ; according to the relation @xmath217 ( c.f .
section[sec : kormendy ] ) . in figure
[ fig : morphology ] , type 1 early - type galaxies and type 2.0 bulge - dominated galaxies are considered as a single class ( labelled with `` @xmath146 '' ) .
the results for the four redshift bins are represented with different colors ( as described in the figure ) .
galaxies with early - type morphologies , i.e. , the combined sample of type 1@xmath218 galaxies , contribute from @xmath219% at @xmath216 to @xmath2560% at @xmath220 to the sample of photometrically - selected etgs .
this compares to the fraction of @xmath25 58% at @xmath9 @xcite .
the majority of the galaxies that are not classified as early type type 1 or type 2.0 galaxies have the morphology of disk - dominated , small - bulge galaxies ( @xmath13% and @xmath221% of zest type 2.1 and 2.2 , respectively ) ; the remaining few percents are almost equally split between bulgeless disks ( type 2.3 ) and irregular galaxies ( type 3 ) .
@xcite presented the morphological distribution of a sample of combo17 red - sequence selected galaxies with redshifts between @xmath222 .
they find that , down to @xmath223 , 85% of the combined rest - frame @xmath224band luminosity density comes from visually classified e / s0/sa galaxies
. a slightly smaller fraction ( 75% ) was found by @xcite using an automatic morphological classification for early type galaxies based on the sersic index @xmath92 . when limiting the analysis of the cosmos data to the same bright cut in @xmath8 magnitude and considering the same redshift range , we find that a similar fraction of @xmath225% of the rest - frame @xmath224band luminosity density comes from type @xmath146 etgs
. in principle , mergers of old stellar sub - units `` caught in the act '' could have irregular morphologies and not satisfy the kormendy - test described above .
we nonetheless investigate which fraction of the photometrically - selected sample is excluded on the basis of this test , and which are the properties of the rejected objects . .
large black circles identify galaxies with irregular morphological classification ( zest type 3 ) .
red dots are galaxies with seds best fitted by zebra with an elliptical - galaxy template ; blue dots identify objects which are fitted by zebra with a late - type galaxy template.,width=340 ] in figure [ fig : kormcolorselected ] we show the kormendy diagram for the photometrically - selected sample of etgs , split in the four redshift bins ; symbols and lines are as in figure [ fig : kormendy ] . in this figure
we also identify all galaxies that have an irregular ( zest type 3 ) morphological classification by large black circles . indeed , not all of the red sequence selected galaxies are consistent with passively - evolving into the kormendy relation at redshift @xmath9 .
interestingly , however , the vast majority of galaxies that are excluded on the basis of the kormendy criterion have a regular ( zest type 1 or 2 ) morphology .
this is illustrated in the right panel of figure [ fig : morphology ] , where we show the morphological mixture of photometrically - selected galaxies which do not pass the kormendy - test . only up to @xmath252% of these kormendy - test - rejected galaxies
have the type 3 irregular morphology .
this is consistent with the rather short ( @xmath226 gyr ) timescales during which major dissipationless mergers appear as one irregular systems .
at all redshifts , the vast majority ( @xmath2270% ) of kormendy - test - rejected objects in the photometrically - selected sample are disk - dominated galaxies .
therefore , when studying the redshift evolution of the lfs of plausible etg progenitors , it seems appropriate to apply the kormendy - test - rejection criterion also to the photometrically - selected sample , as already done for the morphologically - selected sample .
we thus indicate with _ photometrically - selected _ sample the 2877 galaxies , selected according to the red - sequence criterion described in section [ sec : cmselection ] , that furthermore pass the kormendy - test .
we will nevertheless also show , in our analysis of the lfs , the results obtained when no kormendy - selection is performed on the galaxy samples under study .
there is , obviously , overlap between the morphologically and photometrically selected samples , as shown in figures [ fig : colors ] and figure [ fig : morphology ] .
we therefore also present , in the following , the analysis for the _ combined sample _ of all plausible candidate etg progenitors at each redshifts , i.e. , _ the union of the above - defined morphologically- and photometrically - selected samples_. this combined sample contains a total of 3980 galaxies ; specifically , 1623 of the 2730 galaxies in the final morphologically - classified sample are also members of the final photometrically - classified sample ( see tables [ tbl : samples ] and [ tbl : fraction ] ) .
cccc @xmath227 & 414 & 422 & 574 + @xmath228 & 414 & 407 & 594 + @xmath229 & 862 & 824 & 1214 + @xmath230 & 1040 & 1224 & 1598 + + total & 2730 & 2877 & 3980 + lcc @xmath227 & 63%&61% + @xmath228 & 54%&55% + @xmath229 & 55%&57% + @xmath230 & 65%&54% + for easy reference we summarize in tables [ tbl : samples ] and [ tbl : fraction ] , as a function of redshift , the number of galaxies contained in the morphological , photometric and combined final samples , and the relative fraction of galaxies in common between the morphological and photometric samples .
we derive the @xmath39band lfs using the @xmath231 estimator @xcite : the number of galaxies per unit comoving volume in the range of absolute magnitudes @xmath232 , at redshift @xmath43 , and belonging to the sample @xmath233 can be written as : @xmath234 where the sum is over all galaxies in the sample @xmath233 within the specific range of redshift and absolute magnitude .
@xmath235 is the maximum comoving volume within which a galaxy @xmath153 could still be detected , given the apparent magnitude limits of the survey , i.e. , in our study , @xmath236 ; the volume is appropriately corrected so as to account for the ( small ) fraction of objects that were excluded from the initial acs - selected sample due to the lack of photometric redshifts .
more specifically , @xmath235 is the comoving volume between @xmath237 and @xmath238 , computed , for each galaxy @xmath153 , taking into account the @xmath103correction as a function of redshift .
the @xmath237 and @xmath238 values are , respectively , the maximum between @xmath239 and @xmath240 , and the minimum between @xmath241 and @xmath242 ; the limiting values @xmath241 and @xmath239 are the upper and lower redshift of the considered redshift bin , and @xmath242 and @xmath240 are the redshifts at which the galaxy @xmath153 would have an @xmath48 apparent magnitude of 24 and 16 , respectively , for a given rest frame @xmath39band absolute magnitude and color . the @xmath231 estimator approach does not correct for sources that are missed from the input catalog due to selection effects .
in particular , a catalogue that is defined according to a magnitude cut in a given passband suffers from color dependent selection effects ( e.g. * ? ? ?
* ) . in our @xmath150 sample
, blue objects are detected down to fainter absolute @xmath1 magnitudes for redshifts higher than @xmath243 , i.e. , the redshift at which the observer - frame @xmath107band coincides with the rest - frame @xmath39band ; below @xmath243 , it is the red objects that are instead detected to fainter absolute @xmath1 magnitudes .
correcting for this bias is not straightforward , as it requires an assumption on the color distribution of the galaxy population that is missed from the sample . to avoid this potential source of error , we therefore limit the computation of the lf in each redshift bin to the luminosity range for which we are complete , regardless of galaxy colors . in figure [ fig : lf_ml ] we show , split in four redshift bins , the lfs computed for the different samples of etgs described in the previous sections . specifically , the green , red and black curves represent respectively the _ final morphologically - selected _ sample , the _ final photometrically - selected _ sample ( i.e. , both after the kormendy - test cut ) , and the _ combined _ sample of all plausible progenitors of @xmath9 etgs , which are selected either due to their morphology or to their colors .
the error bars in each luminosity bin take into account poissonian errors only . in appensix [ app : phzerr ] we discuss all other sources of errors , and their impact on the lfs . in particular , we show in the appendix that our main conclusions are largely unaffected by uncertainties in the photometric redshift errors . in the four panels of figure [ fig : lf_ml ] , we highlight the effects of cutting the sample according to the kormendy - test .
in particular we show , with black empty symbols and thin lines , the lfs resulting from the morphologically- ( solid ) and photometrically - selected ( dashed ) samples when the kormendy - relation constraint is not applied . at all redshifts , the inclusion of galaxies that are rejected on the basis of the kormendy - test would lead to a substantial increase in the number of etgs at the faint end ( @xmath244 ) of the lfs .
figure [ fig : lf_ml ] shows that at all redshifts , the shape of the lfs of the final morphologically- and photometrically - selected samples of etgs are very similar .
however , at magnitudes brighter than @xmath245 , morphologically - selected , i.e. , dynamically - relaxed , early type galaxies tend to be more numerous than photometrically - selected , i.e. , old - stellar - population galaxies , by up to a factor of @xmath138 .
in contrast , photometrically
selected etgs are more numerous relatively to the morphologically - selected by about the same factor at fainter magnitudes . from the lowest up to the highest redshift bin , respectively about @xmath246% to @xmath247% of the final sample of photometrically - selected etgs brighter than @xmath248 also belong to the morphologically - selected sample .
these results suggest that , at all redshifts under study , the majority of the most massive etgs are already dynamically - relaxed systems with the appearance of @xmath9 elliptical galaxies and their stellar populations are passively evolving , which is consistent with other features of the lfs that we discuss in section [ sec : disc ] .
cccc + 0.20.4 & @xmath249 & @xmath250 & @xmath251 + 0.40.6 & @xmath252 & @xmath253 & @xmath254 + 0.60.8 & @xmath255 & @xmath256 & @xmath257 + 0.81.0 & @xmath258 & @xmath259 & @xmath260 + + 0.20.4 & @xmath249 & @xmath261 & @xmath262 + 0.40.6 & @xmath252 & @xmath263 & @xmath264 + 0.60.8 & @xmath258 & @xmath265 & @xmath266 + 0.81.0 & @xmath267 & @xmath268 & @xmath269 + + 0.20.4 & @xmath270 & @xmath271 & @xmath272 + 0.40.6 & @xmath273 & @xmath274 & @xmath275 + 0.60.8 & @xmath276 & @xmath277 & @xmath278 + 0.81.0 & @xmath279 & @xmath280 & @xmath281 + we fit a schechter function @xcite to the cosmos lfs , to derive for them an analytical description .
the schecter function best fit parameters , i.e. , the characteristic magnitude @xmath282 , the volume density at @xmath282 , @xmath283 , and the slope of the faint - end , @xmath284 , are derived adopting a @xmath72 procedure .
the best fit parameters for the schechter fits are shown in table [ tbl : schfit ] .
the relatively high normalization factor in the lowest redshift bin is due to the well known overdensity in the central region of the cosmos field ( which is also a reason for using the sdss to determine the cosmic evolution of the lfs in the @xmath35 to @xmath9 window ) . in the fits ,
the faint - end slope @xmath284 is allowed to vary , and @xmath284 values in the range @xmath2850.1 - 0.6 are derived , with _
@xmath286 errors of about @xmath287 ; these values are different from the @xmath288 value adopted by , e.g. , @xcite . however , these values of @xmath284 are derived on the kormendy - consistent samples , and are therefore not directly comparable with those used by other authors .
the faint end slope in the highest redshift bin is however not well defined , given the magnitude cut of our sample ; furthermore , despite the variations in @xmath284 values , all the measurements are consistent within 3@xmath289 with a constant faint end slope .
thus , all in all , the analytical fits quantitatively demonstrate the similarity , discussed above , of the lfs as a function of redshift .
formalism and the zebra maximum likelihood photometric redshifts .
green , red and black show respectively the morphologically- , photometrically - selected , and combined sample of cosmos early - type galaxies .
the black open symbols and lines show the lfs derived without applying the kormendy selection cuts ( solid : morphological sample ; dashed : photometric sample ) .
the error bars indicate poisson errors only .
note that the relatively high normalization factor in the lowest redshift bin is due to the well known overdensity in the central region of the cosmos field .
, width=340 ] the volume of universe covered by the cycle 12 cosmos data in the redshift range @xmath290 is only @xmath291 mpc@xmath292 , and includes less than 50 galaxies ; therefore , cosmic variance and small number statistics make this bin inadequate for comparisons with the higher redshift bins to trace galaxy evolution down to @xmath9 .
furthermore , the @xmath293 redshift bin is affected by a relatively strong overdensity @xcite .
therefore , following the approach described in paper i , in order to study the evolution of the lfs from @xmath19 to @xmath9 , we compare the cosmos lfs with a complete sample of galaxies extracted from the sloan digital sky survey @xcite , appropriately redshifted to @xmath16 so as to provide a direct calibration point for the @xmath294 cosmos data .
the generation of the redshifted sdss images is described in @xcite , and also summarized in paper i and in appendix [ app : sdss ] .
the sdss redshifted images were analyzed following the identical procedure that was adopted for the cosmos galaxies : @xmath5 the sdss galaxies were morphologically classified with zest ; @xmath6 the photometrically - selected sample of sdss early - type galaxies was determined according to the two - step procedure described in section [ sec : cmselection ] ; @xmath295 the kormendy - cut was applied to both the photometrical and morphological samples of etg progenitors ; and @xmath296 finally , the lfs for the morphologically - selected , photometrically - selected and combined samples of redshifted sdss etgs were calculated , using the approach described in section [ sec : lf ] .
we stress that by applying exactly the same procedure to the cosmos and to the sdss samples , the same systematics apply to both the high-@xmath43 and the local comparison sample .
thus , the direct comparison of the two allows us to identify , without biases , any evolutionary effect of the galaxy samples under study from @xmath16 to @xmath9 .
( empty symbols , dashed lines ) , and the @xmath297 cosmos lfs ( filled symbols , solid lines ) , for the morphologically - selected and photometrically - selected samples , and for the combined sample of etgs .
the sdss lfs are shown also brightened by @xmath298 magnitude , i.e. , the amount consistent with the evolution of the red sequence in the rest - frame @xmath114 vs. @xmath8 color - magnitude diagram ( thin dashed curves ) .
, width=340 ] in figure [ fig : lfsdss ] we compare the @xmath229 cosmos lfs ( solid lines ) with the redshifted sdss lfs ( thick dashed lines ) for the morphologically - selected , photometrically - selected and combined samples of etgs .
the thin dashed curves represent the sdss lfs after brightening by @xmath299 magnitude , i.e. , the luminosity evolution from @xmath16 to @xmath9 consistent with the observed evolution of the red - sequence discussed in section [ sec : zp ] . at luminosities higher than @xmath300
( i.e. , @xmath137 mag brighter than @xmath301 , see table [ tbl : schfit ] ) , the @xmath15 cosmos lf of the photometrically - selected sample of etgs well matches the bright end of the corresponding sdss lf , after applying the brightening of @xmath298 magnitudes .
the number density of the brightest and thus most massive photometrically - selected etgs does not evolve from redshift @xmath16 to redshift @xmath9 . at fainter magnitudes
, however , the number density of photometrically - selected early - type galaxies shows an increase of a factor up to three from @xmath15 to @xmath9 .
we stress that the sample of bright photometrically - selected etgs is not contaminated by dusty disks .
indeed , we found that the fraction of type [email protected] in the photometrically - selected sample of etgs at redshift @xmath16 is @xmath303% for galaxies brighter than @xmath304 .
broadly speaking , a similar result is obtained also for the morphologically - selected and combined samples of early type galaxies .
however , adopting a similar evolution as above , the number density of the brightest , @xmath305 morphologically - selected type @xmath146 etgs at @xmath15 in the cosmos field would slightly exceed the number density of the brightened sdss sample .
this suggests a faster luminosity evolution for the morphologically - selected sample of etgs , which is consistent with the latter containing a fraction of galaxies with colors bluer than those predicted for a passively evolving stellar population ( see section [ sec : colors ] ) . these blue colors might arise from a substantial fraction of mass in intermediate - age stars
however , a recent second bursts of star formation , producing only a few percent of the stellar mass , would also make significantly bluer the colors of the stellar population and increase the @xmath39band luminosity of the galaxies . for example , adding a 1 to 5% stellar mass in a second burst of star formation to an underlying old , single - burst population with formation redshift @xmath128 is sufficient to explain the @xmath114 colors of even the bluest morphologically - selected etgs at all redshifts under study .
this suggests that the bulk of the stellar mass in _ all _ morphologically - selected high-@xmath43 etgs is already formed by @xmath19 .
this is consistent with the other studies that we have already mentioned , which find that @xmath306 galaxies with @xmath307 have an integrated specific star formation rate less than the global value .
these studies conclude that the bulk of star formation in massive galaxies occurs at early cosmic epochs and is largely complete by @xmath308 , and that further mass assembly in these galaxies takes place with low specific star formation rates ( e.g. , * ? ? ?
* ; * ? ? ?
furthermore , similar to the case of the photometrically - selected progenitors of etgs , also the morphologically - selected sample and the combined sample of early type galaxy progenitors show a remarkable lack of evolution of the very bright - end of their lf . fainter early type galaxies are instead again about a factor @xmath14 less numerous than at @xmath9 .
more quantitatively , we compute the total density of etgs brighter than @xmath309 in both the combined sample of cosmos and sdss etgs .
we indicate these two densities with @xmath310 and @xmath311 , respectively .
the adopted value of @xmath312 is the one derived from the schechter fit to the @xmath16 lf of the combined sample ( see table [ tbl : schfit ] ) . in order to compute @xmath311
, we integrate the sdss local lf brightened by 0.95 magnitudes .
consistent with the qualitative analysis presented above , we find that @xmath313 the density of @xmath314 etgs remains constant , within the poissonian errors , since @xmath16 to today .
this conclusion might in principle depend on the assumed amount of luminosity evolution ( i.e. , the sfh of the etg population ) , on the cosmic variance , and on the effect of the photometric redshift uncertainties on the lfs as we discuss in appendix [ app : phzerr ] and section [ sec : lf ] , poissonian errors at such bright magnitudes are , however , larger than the uncertainties on the lf induced by the errors on the photometric redshifts
. errors due to the photo@xmath4 can thus be neglected .
the effect of cosmic variance on @xmath310 can be quantified using the prescription of @xcite .
we estimate that , for the given density and cosmos volume , this uncertainty contributes at most @xmath2525% to the error on @xmath310 .
we obtain a similar result by calculating the variance of the total number of @xmath305 , i.e. @xmath315 , galaxies in the 24 mock galaxy catalogs generated by @xcite for the cosmos survey using the millennium run numerical simulations . the sdss density , @xmath311 , is not affected by cosmic variance , since the 1800 sdss galaxies were chosen randomly in a local volume larger than @xmath316 mpc@xmath292 , and the derived sdss lf closely matches the @xcite lf derived from the entire sdss galaxy catalog ( see further discussion in * ? ? ?
what causes the largest uncertainty in @xmath311 is the assumed luminosity evolution for the etg stellar population . in the previous analysis
we have used the evolution of a single stellar population formed at redshift @xmath128 , justified by the fact that this sfh is consistent with the observed evolution of the red - sequence zero point ( see section [ sec : zp ] ) .
however , we can estimate the changes induced on @xmath311 due to a reasonable range of sfhs , and use them as more realistic estimates of the errors on @xmath311 .
although the range of observed @xmath317 reported in the literature is broad ( see , e.g. , table 2 in * ? ? ?
* and references therein ) , a recent study by @xcite shows that stars in massive etgs ( @xmath142 ) , both in clusters and in the field , have a mean luminosity - weighted formation redshift of @xmath143 . for such massive galaxies , @xcite
show that the star - formation time scale is less than @xmath318 gyr , which therefore makes our @xmath144 a reasonable approximation for the measured value .
this formation timescale , together with the error on @xmath144 presented by @xcite , can be used to derive the uncertainty on @xmath311 due to varying the luminosity evolution of the etgs .
we find that an uncertainty in formation redshift of @xmath319 corresponds , roughly , to a @xmath320 at a redshift @xmath16 .
it follows that the total density of @xmath321 etgs in the combined sample is @xmath322 mpc@xmath155 .
it is clear , then , that the largest source of uncertainty on @xmath311 is indeed due to the assumed sfh .
considering together the effects of cosmic variance and varying the sfhs , the ratio between the observed and predicted density of massive etgs at redshift @xmath16 is therefore @xmath323 .
this shows that bright ( @xmath324 ) etgs are already in place at @xmath16 , and that the maximum evolution in the number of bright etgs allowed from @xmath16 to @xmath9 is at most of @xmath13% .
in contrast , there is indeed a dearth of lower - luminosity etgs , at @xmath11 compared with the local universe .
this deficiency of intermediate - to - faint etgs is not an effect of incompleteness , since ( @xmath153 ) both in the morphologically- and photometrically - selected samples , the deficit of early type galaxies is visible at magnitudes @xmath325 , i.e. , well above our magnitude limit ( @xmath326 at @xmath16 ) ; ( @xmath327 ) the central wavelength of f814w filter matches the rest - frame @xmath39band at @xmath30 ; this implies that , down to the faintest magnitude bin considered , there are no color - dependent selection effects in the redshift range @xmath328 ; ( @xmath329 ) the zest classification recovers the morphological class for @xmath330% of the galaxies down to @xmath331 ( paper i ) ; furthermore , possible systematics in the morphological classification are identical for the cosmos and for the local sdss sample used to normalize our results .
arguments for an apparent anti - hierarchical mass assembly of etgs were recently presented based on the redshift evolution of the mass function @xcite and of the luminosity function ( cimatti , daddi & renzini 2006 , brown et al .
2007 , wake et al . 2006 ) . furthermore
, enough star - formation is observed in massive galaxies at higher redshifts to account for all stars that we observe already assembled in massive galaxies by @xmath19 @xcite .
our morphological selection of etgs adds to this picture the key information that not only the full mass assembly of the most massive etgs was completed at an earlier cosmic time compared to the less massive galaxies , but also that the majority of these systems achieved dynamical relaxation by @xmath19 .
a relatively large fraction of the less massive early - type galaxies has either not yet achieved relaxation at the redshift ( epoch ) at which they are observed , or not completed its star - formation .
we highlight our three main results : @xmath332 the shapes of the lfs of morphological , photometrical and combined samples of etg progenitors at all redshifts are remarkably similar ; @xmath333 the vast majority of photometrically - selected massive etgs is already dynamically - relaxed and has thus the morphology of an early - type galaxy at redshifts @xmath19 ; and @xmath334 there is a a deficiency of approximately @xmath335 etgs as opposed to remarkable constancy of the bright end ( @xmath314 ) of the lfs of morphologically- and photometrically - defined etgs .
these findings support a scenario in which the brightest , most massive early type galaxies are already fully assembled @xmath336 gyrs ago , while fainter , less massive early type galaxies keep forming their stars and assembling their mass from @xmath16 to the present .
this trend in the observed evolution of the lfs of early - type galaxies , together with the observed constancy with redshift of the lfs shapes , argues against a significant contribution of dry mergers in the build up of the @xmath9 massive etg population .
furthermore , we suggest that the @xmath9 lower luminosity early - type galaxies are the end product of the conversion of blue irregular and disk galaxies into ` red' , early - type galaxies . this explanation is also supported by the observed increase of about a factor of three in the number density of irregular galaxies from the local universe to @xmath16 ( paper i ) .
abraham , r. g. , van den bergh , s. , & nair , p. 2003
, , 588 , 218 bell , e. f. , et al . 2004a , , 608 , 752b bell , e. f. , et al . 2004b , , 600 , l11 bell , e. f. , et al .
2006a , , 640 , 241 bell , e. f. , phleps , s. , somerville , r. s. , wolf , c. , borch , a. , & meisenheimer , k. 2006b , , 652 , 270 bender , r. , burstein , d. , & faber , s. m. 1993 , , 411 , 153 bernardi , m. , et al .
2003a , , 125 , 1882 bernardi , m. , et al .
2003b , , 125 , 1849 bernardi , m. , et al . 2003c , , 125 , 1866 bertin , e. , & arnouts , s. 1996 , , 117 , 393 blanton , m. r. , et al .
2003 , , 594 , 186 bower , r. g. , lucey , j. r. , & ellis , r. s. 1992 , , 254 , 601 brown , m. j. i. , et al .
, 2007 , , 654 , 858 bruzual , g. & charlot , s. , 2003 , , 344 , 1000 bundy , k. , ellis , r. s. , & conselice , c. j. 2005 , , 625 , 621 carollo , c. m. , danziger , i. j. , & buson , l. 1993 , , 265 , 553 carollo , c. m. , & danziger , i. j. 1994a , , 270 , 743 carollo , c. m. , & danziger , i. j. 1994b , , 270 , 523 capak , p. et al .
2007 , , submitted cimatti , a. , daddi , e. , & renzini , a. 2006 , a&a , in press ( astro - ph/0605353 ) coleman , g. d. , wu , c .- c . , & weedman , d. w. 1980 , apjs , 43 , 393 daddi , e. , et al .
2005 , , 631 , l13 dickinson , m. , papovich , c. , ferguson , h. c. , budavri , t. , 2003 , apj , 587 , 25 djorgovski , s. , & davis , m. 1987 , , 313 , 59 dressler , a. , faber , s. m. , burstein , d. , davies , r. l. , lynden - bell , d. , terlevich , r. j. , & wegner , g. 1987 , , 313 , l37 faber s.m . ,
willmer c. n. a. , wolf c. , koo d. c. , weiner b. j. , et al .
2005 , , submitted ( astro - ph/0506044 ) feldmann , r. et al . , 2006 , 372 , 565 felten , j. e. 1976 , , 207 , 700 ferreras , i. , lisker , t. , carollo , c. m. , lilly , s. j. , & mobasher , b. 2005 , , 635 , 243 hogg d. , 2006 . in _ the fabulous destiny of galaxies : bridging past and present _ , ed .
v le brun , a mazure , s arnouts , & d burgarella .
paris : edition frontieres , in press ( astro - ph/0512029 ) j rgensen , i. , franx , m. & kjrgaard , p. , 1995
, , 273 , 1097 kampczyk , p. et al .
2007 , , accepted , ( astro - ph/0611187 ) kitzbichler et al .
2007 , , submitted kinney , a. l. , calzetti , d. , bohlin , r. c. , mcquade , k. , storchi - bergmann , t. , & schmitt , h. r. 1996 , , 467 , 38 koekemoer , a. et al .
2007 , , submitted kodama , t. , arimoto , n. , barger , a. j. , & aragon - salamanca , a. 1998 , , 334 , 99 kormendy , j. 1977 , , 218 , 333 leauthaud , a. et al .
2007 , , submitted lilly , s. j. , tresse , l. , hammer , f. , crampton , d. , & le fevre , o. 1995 , , 455 , 108 lilly , s. et al .
2007 , , submitted lin , l. , koo , d. c. , willmer , c. n. a. , patton , d. r. , conselice c. j. , et al . 2004 . , 617 , l9 lotz , j. m. , primack , j. & madau , p. 2004
, , 128 , 163 masjedi m , hogg dw , cool rj , eisenstein dj , blanton mr , et al . 2006 . , 644 , 54 mcintosh , d. h. , et al . 2005 , , 632 , 191 oke , j. b. 1974 , , 27 , 21 papovich , c. , et al . 2006 , , 640 , 92 petrosian , v. 1976 , , 209 , l1 renzini , a. 2006 , , 44 rix , h .- w . , et al .
2004 , , 152 , 163 sargent m. , et al .
2007 , , accepted ( astro - ph/0609042 ) scarlata , c. et al .
2007 , , accepted , ( astro - ph/0611644 ) schechter , p. 1976
, , 203 , 297 schmidt , m. 1968 , , 151 , 393 scodeggio , m. m. 2001 , , 121 , 2413 scoville , n. et al . 2007a , , accepted scoville , n. et al .
2007b , , accepted somerville , r. s. , lee , k. , ferguson , h. c. , gardner , j. p. , moustakas , l. a. , & giavalisco , m. 2004 , , 600 , l171 springel , v. , frenk , c.s .
& white , s.d.m .
2006 , , 440 , 1137 strauss , m. a. , et al .
2002 , , 124 , 1810 taniguchi , y. et al .
2006 , , this volume thomas , d. , maraston , c. , bender , r. , & mendes de oliveira , c. 2005 , , 621 , 673 treu , t. , et al .
2005 , , 633 , 174 van dokkum , p. g. , & stanford , s. a. 2003 , , 585 , 78 van dokkum , p. g. , 2005 , , 130 , 2647 van dokkum p. g , & van der marel , r. p. , 2006
, , in press wake , d. a. , et al . 2006 , , 372 , 537 york , d. g. , et al .
2000 , , 120 , 1579
feldmann et al . ( 2006 ) estimated the quality of the photometric redshifts by comparison with spectroscopic redshifts for a sample of @xmath337 galaxies observed within the @xmath64cosmos survey @xcite .
the comparison shows that the zebra photometric redsfhits have an accuracy of @xmath338 for galaxies brighter than @xmath93 , and with redshift @xmath339 . in order to estimate the accuracy at magnitudes fainter than @xmath93
, we simulated a faint version of the spectroscopic redshift catalog by dimming the @xmath66 galaxies with available spectroscopic redshift down to our magnitude limit of @xmath101 .
we then recomputed the zebra photometric redshifts for these artificially - fainted sources .
the accuracy of the zebra photo@xmath4 s down to @xmath101 is @xmath340 . to estimate the effect of the photometric uncertainties on the calculation of the lfs
, we create one hundred versions of the cosmos catalog , by convolving the measured photometric redshift with the photometric redshift error appropriate for the magnitude of each object .
we consider two bins of magnitude : for all galaxies with @xmath67 we used @xmath341 , while we use @xmath342 for all galaxies with @xmath343 .
although the photometric redshift errors should be applied to an `` error free '' catalog , our test provides a conservative estimates of how important the effects of the photometric redshift errors are on the observed lfs . using the simulated cosmos catalogs we extracted 100 versions of the morphologically- and photometrically selected samples of etgs , following the procedures described in sections [ sec : ms ] and [ sec : cmselection ] .
we computed the lfs for these 100 re - simulated samples using the procedure described in section [ sec : lf ] . in figure
[ fig : photozsim ] we present the results of this test in the redshift range @xmath297 .
this is the redshift range we use to derive the main conclusions of our paper . in figure
[ fig : photozsim ] solid circles with errorbars represents the lf computed using the `` true '' photometric redshifts , and the open red circles represent the median of the 100 realizations .
the shaded grey area associated with the simulated volume densities represent the 16@xmath75 and 84@xmath75 percentiles of the simulated distributions within each magnitude bin .
figure [ fig : photozsim ] shows that the dominant effect of the redshift uncertainty is at the faint end of the lf , where the number of galaxies is sistematically higher in the error - convolved lfs .
this effect is stronger in the color selected samples than in the morphologically selected one . at the bright end ,
the only significant effect is for very bright magnitudes ( m@xmath344 , i.e. , @xmath345 ) , where we would overestimate the number of galaxies .
indeed , in the simulations we find @xmath346 galaxies with these magnitudes in 8 out of the 100 mock realizations .
these resultss are consistent with those of brown et al .
( 2007 ) who performed similar simulations and find that the effect of photometric redshfit errors on the lf is only significant for @xmath347 .
we note however , that no such galaxies are observed in the original cosmos catalog .
these tests show that , for the area covered by the present analysis , poissonian errors dominate the contribution in the error budget at the very bright end of the lfs .
the simulations do not include catastrofic redshift failures ( @xmath348 ) , that are found to amount to @xmath137% down to @xmath93 , and @xmath158% down to @xmath101 .
although a small fraction , catastrofic failures could have a significant effect in the poorly populated bright part of the lf .
however , the typical redshift degeneracy observed for the `` training '' sample of galaxies with spectroscopic redshift is between @xmath349 and @xmath350 ( see * ? ? ? * ) , i.e. , outside of the redshift range considered in this study .
we also note that our conclusions are based on galaxies with observed magnitudes brighter than @xmath351 ; at these magnitudes our photometric redshifts are very accurate .
the main conclusion of our paper is based on the number density of galaxies brighter than 2.5@xmath352 , in the redshift range @xmath353 .
the main question to ask is whether the red etgs galaxies brighter than 2.5@xmath352 are really in the redshift range @xmath229 , or have been scattered there because of photometric redshift errors . for 18 etgs brighter than 2.5l * , spectroscopic redsfhit is available from the current version of the zcosmos - bright ( i.e.,@xmath66 ) catalog ( lilly et al . 2007 )
: 18/18 have spectroscopic redshift in the 0.60.8 range .
this , therefore , demonstrate that the our result is robust against photometric redshift errors .
we use the set of artificial images created by @xcite from a selected sample of sdss galaxies in order to calibrate our results with the @xmath9 universe .
the artificial images simulate how the local galaxies would be observed in the cosmos survey at redshift @xmath16 .
details on the generation of these images are given in the original reference . in brief , the sdss galaxies were selected in the redshift range @xmath354 $ ] to be brighter than @xmath355 .
the sdss @xmath356-band images of the selected galaxies were transformed to f814w cosmos acs images at @xmath357 . at this redshift
, the redshifted @xmath356 band well matches the hst f814w passband ; thus , the redshifting of the sdss galaxies needs only to take care of the different pixel scales and point spread functions , and of cosmological surface brightness dimming .
no galaxy size evolution was considered .
the @xmath357 simulated sdss galaxies were then added into the cosmos acs images to reproduce the same circumstances of image crowding , noise , and so on .
we expect the sdss sample to be representative of the local galaxy population down to @xmath358 since , overall , the photometric@xmath302spectroscopic sdss data are mostly complete down to @xmath359 @xcite , i.e. , well below the considered absolute magnitude cut that we applied .
the @xmath16 simulated sdss galaxies were analyzed with the same procedure used for the real cosmos data .
first , we run sextractor to perform the detection and measure the @xmath107band magnitude , position angle and elongation . for the detection we used the same sextractor configuration parameters used to generate the cosmos catalog @xcite . a fraction of 6% of the simulated galaxies
was not detected , while a fraction of 0.1% was found with @xmath360 , and therefore excluded from the sample .
we then classified the galaxies using zest .
rest - frame properties such as @xmath114 colors , @xmath39band absolute magnitudes , radii in kpc , were derived assuming that all galaxies were at redshift @xmath16 ; the @xmath114 colors and @xmath8 magnitudes were derived by logarithmic interpolation between the observed sdss magnitudes ( @xmath361 , @xmath356 , and @xmath362 ) . in figure
[ fig : sdss ] we present the @xmath114-@xmath8 color magnitude diagram ( right panel ) and the kormendy diagram ( left panel ) for the simulated @xmath16 sdss galaxies . in particular , we show on the kormendy diagram all galaxies classified as type 1 or type 2.0 with different colors according to the zebra best fits to their seds . as in figure
[ fig : kormendy ] , red points represent elliptical - galaxy seds , and blue points represent later photometric types .
as expected , the sdss morphologically - classified ellipticals are all consistent with the @xmath9 kormendy relation , as possible small - size faint interlopers are excluded from the sample by the imposed magnitude cut ( see dotted line in the left panel of figure [ fig : sdss ] , which shows the @xmath101 magnitude limit ) .
in the _ right panel _ of figure [ fig : sdss ] we show as gray points all sdss galaxies in the original sample , and , highlighted in red , all galaxies with a zebra elliptical - galaxy sed . the rest
frame colors were derived using the sdss total magnitudes , and therefore represent the global galaxy color . as mentioned in section [ sec : slope ] ,
etgs show color gradients with an average value of @xmath162 @xcite . due to this color gradient
, the galaxy colors depend on the aperture size used for the photometric measurements ; the colors are systematically bluer the larger the aperture . since the average color gradient does not correlate with galaxy luminosity , the global effect of a large aperture is a blueing of the red sequence relative to that derived using a smaller aperture . assuming , for simplicity , that early type galaxies in the local universe have a de vaucouleur profile , and using the color gradients published by @xcite , we estimate that the difference between the @xmath114 colors computed within 30% of the half - light radius and those calculated using total magnitudes amounts to @xmath363 magnitudes . in the figure ,
the colors are corrected to account for the aperture corrections . assuming that the color
magnitude relation of @xmath9 early type galaxies can be interpreted as a pure metallicity sequence @xcite , with less luminous galaxies being more metal poor than more luminous galaxies , in figure [ fig : sdss ] we show the @xmath114 colors derived for single stellar population models of different metallicities formed at @xmath128 @xcite .
the colors of the models , together with the absolute @xmath224band absolute magnitude , metallicity and mass of the galaxy are reported in table [ tbl:1 ] . | we study the evolution since @xmath0 of the rest - frame @xmath1 luminosity function of the early - type galaxies ( etgs ) in @xmath2 degrees@xmath3 in the cosmos field . in order to identify _ all _ plausible progenitors of local etgs
we construct the sample of high@xmath4 galaxies using two complementary criteria : @xmath5 a _ morphological _
selection based on the zurich estimator of structural types , and @xmath6 a _
photometric _ selection based on the galaxy properties in the rest - frame @xmath7-@xmath8 color - magnitude diagram .
we furthermore constrain both samples so as to ensure that the selected high@xmath4 progenitors of etgs are compatible with evolving into systems which obey a fundamental @xmath9 scaling relation for early - type galaxies , i.e. , the @xmath10 kormendy relation . assuming the luminosity evolution derived from studies of the fundamental plane for high@xmath4 etgs
, our analysis shows no evidence for a decrease in the number density of the most massive etgs out to @xmath11 : both the morphologically- and the photometrically - selected sub - samples show no evolution in the number density of bright ( @xmath12 ) etgs . allowing for different star formation histories , and cosmic variance , we estimate a maximum decrease in the number density of massive galaxies at that redshift of @xmath13% .
+ we observe , however , both in the photometrical and in the morphological samples , a deficit of up to @xmath14 of fainter early - type galaxies over the same cosmic period .
our results argue against a significant contribution of recent dissipationless `` dry '' mergers to the formation of the most massive early - type galaxies .
we suggest that the mass growth in low luminosity etgs can be explained with a conversion from @xmath15 to @xmath9 of blue , irregular and disk galaxies into low- and intermediate - mass `` red '' , early - type galaxies , possibly also through gas rich mergers .
this interpretation is consistent with the observed increase of a factor of the order of @xmath14 , from @xmath9 to @xmath16 , of the rest - frame @xmath1-band luminosity function of blue irregular galaxies . |
much of the research regarding graphene , a one - atom thick layer of graphite , has been focused on its outstanding electronic properties and its possible role in future semiconducting devices @xcite , but graphene applications are not restricted to microelectronics . with a mechanical strength of @xmath2100 gpa , a young modulus of @xmath21 tpa @xcite and its impermeability for standard gases as small as he , graphene can be used as an ultra thin protective layer which can e.g. prevent the oxidation of cu @xcite .
graphenes huge potential use as a sensor has already been shown for dna and glucose biosensors @xcite and even the detection of individual gas molecules has been demonstrated @xcite .
having the ability to reconstruct its perfect hexagonal carbon lattice into pentagons and heptagons near defect sites @xcite and the fact that defects can be exploited to manipulate its physical properties , graphene proofs to be an interesting material for ion irradiation experiments @xcite . in previous studies graphene has been irradiated with either low energy ions ( lei ) or light ions and with high ion fluences of ( from 10@xmath3 ions / cm@xmath1 up to 10@xmath5 ions / cm@xmath1 ) normal to the surface .
these reports have shown that it is indeed possible to induce defects with energetic ions even in suspended graphene sheets @xcite .
furthermore , it has been shown that by irradiating single layer graphene ( slg ) with swift heavy ions ( shi ) with @xmath2100 mev kinetic energy under glancing incidence angle , origami like foldings of the graphene sheet are induced by individual ions @xcite . in this paper
, we show that graphene can be used to determine whether an irradiation was successful even in the case where the substrate itself is not radiation sensitive .
we demonstrate how graphene can then be used to assess either the true incidence angle or the true fluence .
furthermore , the fact that graphene folds upon glancing angle shi irradiation can be exploited to mark single ion impacts , rendering them accessible to surface sensitive tools .
finally , we show that the mechanical strength of graphene and its impermeability for small molecules make it a perfect catcher for ejected material .
for our experiments , we prepared graphene on different substrates using the mechanical exfoliation technique of graphite ( hopg - momentive performance materials quartz , zya grade ) in order to obtain graphene of the highest quality @xcite .
graphene was exfoliated in ambient conditions onto a freshly cleaved nacl(100 ) single crystal ( korth kristalle ) , referred to as g / nacl in the following , and srtio@xmath0(100 ) ( crystec ) , referred to as g / sto .
additionally , single layer graphene was prepared under ultra high vacuum ( uhv ) conditions on a si(111)7x7 substrate ( g / si ) . for the latter , a si(111 ) wafer (
n - doped , 10 - 20 ohm@xmath6 cm ) was repeatedly flash heated to 1250@xmath7c and the quality of the reconstruction was confirmed by low energy electron diffraction ( leed ) . for more information regarding this preparation procedure
see @xcite .
shi irradiation took place at the irrsud beamline ( ganil , caen - france ) .
the incidence angles ranged from @xmath81@xmath7 to 3@xmath7 with respect to the sample surface and the ion fluence on the samples were adjusted to around ( 510 ) ions / cm@xmath1 to be able to analyze single , non - overlapping shi impact events . for the irradiations xe@xmath9 and pb@xmath10 ions
were used .
the respective stopping power in the substrate materials was in the range of @xmath11(1219 ) kev / nm , in an amorphous carbon target with the density of graphite it would be @xmath11(1517 ) kev / nm . these values were calculated using the srim 2008 software package @xcite .
after the irradiation experiments , the samples were analysed in ambient with an atomic force microscope ( afm , veeco dimension 3100 system ) .
measurements were performed in tapping mode using nanosensors nchr cantilever with typical first resonant frequencies of @xmath12300 khz .
we begin with the typical image observed when irradiating an exfoliated graphene flake on an insulator , see fig .
[ figure1 ] . here ,
g / sto was irradiated by 91 mev @xmath13 xe@xmath9 shi with a stopping power of @xmath14=19.0 kev / nm under a glancing incidence angle of 2.6@xmath7 . upon impact of the shi
, graphene is folded in a typical pattern .
the exact shape of the pattern may vary depending on different parameters like incidence angle and target material @xcite . at this angle a typical folding consists of two to three folded areas , two of them alongside and symmetrical with respect to the ion beam trajectory , and one of them oriented almost perpedicular to the track ( and usually found downstream ,
see fig .
[ figure1 ] ) .
the mechanism of the folding has been explained in detail elsewhere @xcite , here we will give only a brief description .
shi irradiation of insulators like sio@xmath15 or srtio@xmath16 under glancing incidence is known to create surface tracks of molten material along the ion trajectory @xcite . as this state
is rapidly quenched , a chain of nanosized hillocks remains on the surface .
if the track area is covered by graphene , this hot surface material pushes through the graphene which has been damaged along the track beforehand to some extent by the projectile itself .
the srtio@xmath0 surface track can be seen in the exposed area between the folded graphene ( see fig . [ figure1 ] ) .
after the projectile has reached a certain depth , the surface track in srtio@xmath16 continues underneath an otherwise intact graphene sheet .
note , that virtually every surface track in g / sto observed on single layer graphene has a folding attached to it .
thus , the sensitivity of slg to shi induced damage is 100% .
folding patterns deviating from the regular one , like inverted foldings ( see orange box in fig .
[ figure1 ] ) , can be sometimes observed , too .
these foldings are most likely caused by elastically scattered shi , where the trajectory is not a straight line .
until now , shi irradiation induced foldings in graphene have been studied exclusively on materials which show distinct surface tracks @xcite .
there are however materials , on which no surface tracks or morphological modifications are expected @xcite or detected , like e.g. si @xcite . here
we demonstrate , that on substrates which show no surface tracks in form of protrusions or nano - sized hillocks , graphene foldings can still occur , as can be seen in fig .
[ figure2 ] .
[ figure2](a ) shows an image taken from g / nacl which was irradiated with 104 mev @xmath17pb@xmath10 under an incidence angle of @xmath18 .
the nacl surface at this resolution shows no signs of modification due to the shi impact while the slg flake shows the typical shi induced foldings .
thus , graphene can here be used to detect shi impacts , localize and count them ( fluence control ) .
we determine the zero degree grazing angle by optically aligning the surface parallel to the beam direction by means of a telescope aligned on the optical axis of the beam line .
a stepping motor with a repetability of about 5/100 of a degree is used to turn the target to the nominal angle . in light of the extreme good repeatability of the rotation ,
the main error bar stems from the optical alignment , and we expect an error bar of about 0.2 degree . if the efficiency of damage is known as in the case of g / sto , the true angle of incidence can be calibrated by comparing the experimentally determined fluence with the nominal one : @xmath19 . for g
/ nacl , the efficiency is unknown , but if we assume 100% for the sample shown in fig .
[ figure2 ] we determine @xmath20 which is very close to the nominal one of @xmath21 .
in addition , graphene offers an efficient way to search for possible irradiation damage of the nacl surface .
one could now simply limit the search to the areas marked by foldings and study those areas inside of the graphene and/or foldings located at the edge of the flake in more detail .
this allows for smaller scanning windows with the possibility of enhanced resolution . as the mechanical exfoliation technique of graphene
is generally used under ambient atmosphere , all kind of adsorbates like oxygen , nitrogen or water will be enclosed between graphene and the substrate forming an undefined interface layer ( ifl ) . to study the yet unkown influence of this ifl on the shi irradiation ,
a graphene flake was exfoliated onto a si(111)7x7 substrate in situ . after the in situ preparation
, the sample was exposed to ambient and analysed with afm and raman spectroscopy .
this showed , that while a @xmath21.5 nm native sio@xmath22 layer is formed on top of si , graphene covered si still exhibits terrace steps ( see @xcite and fig.[figure3 ] ) .
the latter indicates that oxidation , and thus the formation of an ifl , might have been successfully prevented .
an afm topography image of this sample after irradiation with 92 mev @xmath23xe@xmath9 shi is shown in fig.[figure3 ] .
on the amorphous sio@xmath22 surface no signs of ion induced surface tracks can be observed .
the native oxide is obviously too thin to give rise to the faint surface tracks typical for bulk sio@xmath15 samples @xcite .
nevertheless , the ion solid interaction can still be detected because on the graphene covered si , distinct surface tracks with an average height of about 0.4 nm are measured .
hence , graphene does not fold with a largely absent ifl despite surface tracks being formed underneath .
this finding shows that in addition to the surface track from the substrate , the ifl must play a role for the folding mechanism .
the exact nature of the tracks detected in g / sio remains unclear .
however , from our data it is obvious that graphene retains material which otherwise would have left the surface . the use of graphene as a protective layer for ion irradiation experiments with gas targets has already been proposed by lehtinen et al .
stolyarova et al .
observed the formation of graphene bubbles after irradiating graphene on sio@xmath24 with 0.4 - 0.7 mev protons normal to the surface @xcite . from fig .
[ figure3 ] it can be clearly seen that the use of graphene as a protective layer does not stop at retaining gaseous materials but can be extended to catching ejected solid material .
in this work we have shown , that ultrathin sheets of graphene can be used to detect and study shi irradiation effects in various ways .
foldings can be used as impact markers enabling calibration of irradiation parameters as well as targeting irradiated areas in materials without distinct track formation .
the ability of graphene to catch ejected material may open a new way to implement foreign material into graphene as well as to study ion induced modifications which are otherwise transient .
this work has been supported by the european community as an integrating activity support of public and industrial research using ion beam technology ( spirit ) under ec contract no .
227012 and by the german science foundation ( spp 1459 : graphene and sfb 616 : energy dissipation at surfaces ) .
the experiments were performed at the irrsud beamline of the grand accelerateur national d@xmath25ions lourds ganil , caen , france . 10 k.i .
bolotin , k.j .
sikes , z. jiang , m. klima , g. fudenberg , j. hone , p. kim , and h.l stormer . , 146(9 - 10):351355 , 2008 .
k. kim , j. choi , t. kim , s. cho , and h. chung .
, 479(7373):338344 , 2011 . c. lee ,
x. wei , j. w. kysar , and j. hone .
321(5887):385388 , 2008 . j. s. bunch , s. s. verbridge , j. s. alden , a. m. van der zande , j. m. parpia , h. g. craighead , and p. l. mceuen .
, 8(8):24582462 , 2008 .
s. chen , l. brown , m. levendorf , w. cai , s. ju , j. edgeworth , x. li , c. w. magnuson , a. velamakanni , r. d. piner , j. kang , j. park , and r. s. ruoff .
, 5(2):13211327 , 2011 .
y. shao , j. wang , h. wu , j. liu , i. a. aksay , and y. lin . , 22(10):10271036 , 2010 .
f. schedin , a. k. geim , s. v. morozov , e. w. hill , p. blake , m. i. katsnelson , and k. s. novoselov . , 6(9):652655 , 2007 .
a. v. krasheninnikov and f. banhart . , 6(10):723733 , 2007 .
v. pereira , j. lopes dos santos , and a. castro neto .
, 77(11):115109 , 2008 .
g. compagnini , f. gianazzo , s. sonde , v. raineri , and e. rimini .
, 201:3201 , 2009 .
j. chen , w. cullen , c. jang , m. fuhrer , and e. williams .
, 102(23):236805 , 2009 .
l. tapaszt , g. dobrik , p. nemes - incze , g. vertesy , ph lambin , and l. bir .
, 78(23):233407 , 2008 .
s. mathew , t. k. chan , d. zhan , k. gopinadhan , a. roy barman , m. b. h. breese , s. dhar , z. x. shen , t. venkatesan , and john t. l. thong .
, 110(8):084309 , 2011 .
s. akcltekin , h. bukowska , t. peters , o. osmani , i. monnet , i. alzaher , b. ban detat , h. lebius , and m. schleberger .
, 98(10):103103 , 2011 .
k. s. novoselov , d. jiang , f. schedin , t. j. booth , v. v. khotkevich , s. v. morozov , and a. k. geim .
, 102(30):1045110453 , 2005 .
o. ochedowski , g. begall , n. scheuschner , m. el kharrazi , j. maultzsch , and m. schleberger . , 23(40):405708 , 2012 .
j. f. ziegler , m. d ziegler , and j. p biersack .
, 268(11 - 12):18181823 , 2010 .
s. akcltekin , h. bukowska , e. akcltekin , h. lebius , and m. schleberger .
, 1354 , 2011 .
e. akcltekin , t. peters , r. meyer , a. duvenbeck , m. klusmann , i. monnet , h. lebius , and m. schleberger . , 2(5):290294 , 2007 .
s. akcltekin , e. akcltekin , t. roll , h. lebius , and m. schleberger . , 267(8 - 9):13861389 , 2009 .
n. itoh , d. m. duffy , s. khaksouri , and a. m. stoneham .
, 21:474205 , 2009 .
s. akcltekin , t. roll , e. akcltekin , m. klusmann , h. lebius , and m. schleberger .
, 267:683 , 2008 .
o. lehtinen , j. kotakoski , a. v. krasheninnikov , a. tolvanen , k. nordlund , and j. keinonen .
, 81(15):153401 , 2010 .
e. stolyarova , d. stolyarov , k. bolotin , s. ryu , l. liu , k. t. rim , m. klima , m. hybertsen , i. pogorelsky , i. pavlishin , k. kusche , j. hone , p. kim , h. l. stormer , v. yakimenko , and g. flynn .
, 9(1):332337 , 2009 .
xe@xmath9 ; @xmath14=19.0 kev / nm ; @xmath26 ) under glancing incidence .
several surface tracks on the srtio@xmath16 substrate are marked by white boxes . folded areas on slg
are connected to surface tracks in srtio@xmath16 underneath . in the orange box , a rare inverted folded structure can be observed . ]
xe@xmath9 ; @xmath14=12.35 kev / nm ; @xmath28 . the sample was prepared by exfoliating graphene in situ onto a si(111)7x7 substrate .
after exposure to ambient bare si is directly oxidized whereas graphene covered si still shows terrace steps . surface tracks on this sample
are observed exclusively on the graphene covered si substrate , marked by white boxes . ] | in this paper we show how single layer graphene can be utilized to study swift heavy ion ( shi ) modifications on various substrates .
the samples were prepared by mechanical exfoliation of bulk graphite onto srtio@xmath0 , nacl and si(111 ) , respectively .
shi irradiations were performed under glancing angles of incidence and the samples were analysed by means of atomic force microscopy in ambient conditions .
we show that graphene can be used to check whether the irradiation was successful or not , to determine the nominal ion fluence and to locally mark shi impacts . in case of samples prepared in situ
, graphene is shown to be able to catch material which would otherwise escape from the surface .
, , @xmath1 , @xmath1 , |
the galaxies in our sample are at @xmath3 and have @xmath4 m@xmath5 . in order to obtain most reliable results ,
our sample was selected to be suitable for image decomposition : galaxies are close to face - on ( @xmath6 ) , are not morphologically disturbed , and have an apparent diameter larger than @xmath7 .
the fits are done with the new version of budda ( @xcite ) and checked individually to avoid wrong results that can arise from automated procedures .
the models include up to four components : bulge , disk , bar and central source .
this allows fitting galaxies with bars without compromising bulge and disk parameters , and to take into account the contribution from type 1 agn .
figure 1 shows an illustrative example . .
at left we show @xmath8 color composites of the original galaxy image , the total model , the models of each separate component and a enhanced residual image . in the latter ,
one clearly sees the spiral arms and the inner ring / lens fragments that surrounds the bar .
the negative residuals also delineate the fainter inter - arm areas of the disk .
this is likewise true for the region between the bar and the inner ring .
the highest residuals reach a level of about 10% .
the radial profiles on the right also demonstrate the fitting ability of the code .
each point in these profiles refer to a single pixel .
only a fraction of the pixels is shown.,title="fig:",width=255 ] . at
left we show @xmath8 color composites of the original galaxy image , the total model , the models of each separate component and a enhanced residual image . in the latter ,
one clearly sees the spiral arms and the inner ring / lens fragments that surrounds the bar .
the negative residuals also delineate the fainter inter - arm areas of the disk .
this is likewise true for the region between the bar and the inner ring .
the highest residuals reach a level of about 10% .
the radial profiles on the right also demonstrate the fitting ability of the code .
each point in these profiles refer to a single pixel .
only a fraction of the pixels is shown.,title="fig:",width=255]0.15 cm
amongst our first main results , with @xmath9 of the sample analysed , we find that : * there is a correlation between the effective radius of the bulge and the 4000 break index d@xmath10(4000 ) , in the sense that smaller bulges have younger stellar populations ( see also @xcite ) ; * the @xmath11 colours of bulges and bars are strongly correlated for galaxies of all bulge - to - total luminosity ratios ( b / t ) . in fact ,
bulges and bars have similar colours .
we are currently investigating whether this is a result of dust attenuation or colour gradients .
we also find a correlation between the colours of disks and bulges that shows the same triangular pattern noted by @xcite using ultraviolet colours .
blue bulges are almost always surrounded by blue disks , but galaxies with red bulges can have either blue or red disks . finally , the correlation between bar and disk colours is weak , even for late - type galaxies , with low values of b / t .
bars are usually redder than their disks ; * there is a gradual increase in b / t and stellar mass from star forming galaxies ( b / t @xmath12 ) to composite systems ( b / t @xmath13 ) and agn hosts ( b / t @xmath14 ) ; * the median _ bar_-to - total luminosity ratio is bar / t @xmath15 regardless of agn or star formation activity and the total stellar mass of the galaxy ; bar / t ranges from @xmath16 0.01 to @xmath16 0.3 ; * there is an anti - correlation between the effective radius of the bulge and the accretion rate onto the supermassive black hole in agn hosts , parameterized as @xmath17}/{\rm m}_\bullet$ ] , in the sense that the hosts of the most powerful agn have smaller bulges ( see also @xcite ) .
in addition , at a fixed bulge size , the hosts of the most powerful agn have bluer disks , in consonance with the results in @xcite ; * the effective radius of the bulge is the _ only _ structural parameter that shows a clear correlation with the accretion rate onto the black hole .
the bulge srsic index and the bulge effective surface brightness show only weak trends .
the accretion rate does not depend on any structural parameter of either the disk or the bar .
this work is supported by the deutsche forschungsgemeinschaft priority program 1177 ( `` witnesses of cosmic history : formation and evolution , of galaxies , black holes and their environment '' ) , and the max planck society . | the structural parameters of bulges , disks and bars of a sample of nearly 1000 nearby galaxies are being determined through sophisticated image decomposition in the @xmath0 , @xmath1 and @xmath2 bands . the sample is carefully drawn from the sloan digital sky survey data release 2 ( sdss dr2 ) , contains 963 galaxies , of which 407 host agn , and
we show that it is representative of the galaxy and agn populations in the local universe .
a large number of other physical properties have also been determined for these galaxies within the sdss collaboration . with these data , we reinforce several recent studies and find a number of results leading to new insights into how the different galaxy components relate to each other and the extent to which the hosts galaxies of agn differ from quiescent galaxies . |
the study of the production of electroweak bosons in association with jets of hadrons in the final state constitutes a fundamental item in the high-@xmath5 physics program at the tevatron .
these events are main backgrounds to many interesting physics processes like , for example , top production , the search for the sm higgs , and supersymmetry . the cdf experiment has recently published precise measurements on jets in events with a z/@xmath6 boson in the final state @xcite , where inclusive jet cross sections as a function of jet transverse momentum and jet multiplicity are measured and compared to pqcd predictions @xcite . at the leading order ( lo ) in pqcd ,
@xmath7+jet events are driven by the processes @xmath8 and @xmath9 , while higher orders contributions , including additional parton radiation , produce mutiple jets in the final state .
next - to - leading order ( nlo ) pqcd predictions for @xmath0+jets production are only available for jet multiplicities @xmath10 up to @xmath11 .
this contribution presents updated results with an increased data sample corresponding to a total integrated luminosity of @xmath3 .
the events are required to have two electrons with @xmath12 gev and a reconstructed invariant mass in the range @xmath13 gev/@xmath14 around the @xmath15 boson mass .
the electron candidates are reconstructed using criteria described in @xcite . in this study
, one electron is required to be central ( @xmath16 ) and fulfill tight selection cuts , while the second electron is required to pass a looser selection and to be either central or forward with @xmath17 .
the events are selected to have a reconstructed primary vertex with @xmath18-position within 60 cm around the nominal interaction point , and at least one jet with transverse momentum @xmath19
gev/@xmath20 , rapidity in the range @xmath21 , and @xmath22 , where @xmath23 denotes the distance between the jet and each of the two electrons in the final state .
the main backgrounds to the @xmath24+jets sample arise from inclusive - jets and @xmath25+jets events , and are estimated from the data .
other background contributions from @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , and @xmath31+jets final states are estimated using monte carlo samples .
the total background in inclusive @xmath24+jets production is about 12@xmath32 for @xmath33 , and increases up to about 17@xmath32 for @xmath34 .
the measured cross sections are corrected for acceptance and smearing effects back to the hadron level using pythia - tune a monte carlo event samples @xcite , cteq5l @xcite parton distribution functions ( pdfs ) for the proton and antiproton , and a bin - by - bin unfolding procedure that also accounts for the efficiency of the @xmath24 selection criteria .
the final results refer to hadron level jets with @xmath19 gev/@xmath20 and @xmath35 , in a limited and well - defined kinematic range for the @xmath0 decay products : @xmath36 gev , @xmath37 , @xmath38 or @xmath39 , @xmath13 gev/@xmath14 , and @xmath22 . in order to avoid any bias on the correction factors due to the particular pdf set used , which translates into slightly different simulated @xmath40 distributions ,
the pythia - tune a monte carlo event sample is re - weighted until it accurately follows the measured @xmath40 spectra .
the unfolding factors @xmath41 are computed separately for the different measurements and vary between 2.0 at low @xmath42 and 2.3 at high @xmath42 .
figure [ fig : pt](top ) shows the measured inclusive jet differential cross sections as a function of @xmath42 in @xmath24+jets production , with @xmath33 and @xmath43 , compared to nlo pqcd predictions .
the cross sections decrease by more than three orders of magnitude as @xmath42 increases from 30 gev / c up to about 300 gev / c .
the nlo pqcd predictions are computed using the mcfm program @xcite with cteq6.1 m pdfs @xcite , with the renormalization and factorization scales set to @xmath44 , and using a midpoint @xcite algorithm with @xmath45 and @xmath46 @xcite to reconstruct jets at the parton level .
the theoretical predictions include parton - to - hadron correction factors @xmath47 that approximately account for non - perturbative contributions from the underlying event and fragmentation into hadrons . in each measurement @xmath48
is estimated using the pythia - tune a monte carlo samples , as the ratio between the nominal @xmath42 distribution and the one obtained by turning off both the interactions between proton and antiproton remnants and the string fragmentation in the monte carlo samples .
the correction decreases as @xmath42 increases from about 1.2 ( 1.26 ) at @xmath42 of 30 gev / c to 1.02 ( 1.01 ) for @xmath49 gev/@xmath20 for @xmath33 ( @xmath43 ) , and is dominated by the underlying event contribution .
the uncertainty on @xmath48 is about @xmath50 ( @xmath51 ) at low @xmath42 and goes down to @xmath52 at high @xmath42 for @xmath33 ( @xmath43 ) .
( black dots ) in @xmath24+jets with @xmath53 compared to nlo pqcd predictions ( open circles ) . for clarity , the measurement for @xmath33 is scaled up by ( @xmath5420 ) .
the shaded bands show the total systematic uncertainty , except for the 5.8@xmath32 luminosity uncertainty .
( middle and bottom ) data / theory ratio as a function of @xmath42 for @xmath33 and @xmath43 , respectively .
the dashed and dotted lines indicate the pdf uncertainty and the variation with @xmath55 of the nlo pqcd predictions , respectively
. , width=264 ] events as a function of @xmath10 compared to lo and nlo pqcd predictions .
the shaded bands show the total systematic uncertainty , except for the 5.8@xmath32 luminosity uncertainty .
( bottom ) ratio of data and nlo to lo pqcd predictions versus @xmath10 .
the dashed and dotted lines indicate the pdf uncertainty and the variation with @xmath55 of the nlo pqcd predictions , respectively . , width=264 ] finally , fig . [
fig : njet ] shows the cross sections @xmath56 for @xmath24+jets events up to .
the data are compared to lo and nlo pqcd predictions .
the parton - to - hadron non - perturbative corrections vary between 1.1 and 1.4 as @xmath10 increases .
the lo pqcd predictions underestimate the measured cross sections by a factor about 1.4 approximately independent of @xmath10 .
good agreement is observed between data and nlo pqcd predictions .
lai _ et al .
j. c * 12 * , 375 ( 2000 ) .
j. pumplin _
_ , j. high energy phys . * 0207 * , 012 ( 2002 ) .
a. abulencia _
( cdf collaboration ) , phys .
d * 74 * , 071103(r ) ( 2006 ) .
ellis and d.e .
soper , phys .
d * 48 * , 3160 ( 1993 ) . | inclusive jet cross sections in @xmath0 events , with @xmath0 decaying into an electron - positron pair , are measured as a function of jet transverse momentum and jet multiplicity in @xmath1 collisions at @xmath2 with the upgraded collider detector at fermilab in run ii , based on an integrated luminosity of @xmath3 .
the measurements cover the rapidity region @xmath4 and the transverse momentum range .
next - to - leading order perturbative qcd predictions are in good agreement with the measured cross sections . |
polars , or am her systems , a subset of magnetic cataclysmic variables ( cvs ) , are a unique class of cvs . as with the typical cv
, polars contain a white dwarf which accretes matter from a roche lobe - filling low - mass secondary star .
however , unlike a typical cv , the white dwarf in a polar has a magnetic field strength on the order of a few tens of megagauss ( mg ) .
these strong magnetic fields cause the disruption of the accreted material that would otherwise form an accretion disk . instead , material transfer is routed on to the white dwarf in the form of an accretion stream which follows the path of one or more of the magnetic field lines to impact the surface at one or both of the magnetic poles . as the material approaches the white dwarf ,
a shock front is created where impact energy is released in the form of extreme ultraviolet ( euv ) and x - ray photons .
surface heating , of up to a few hundred thousand degrees kelvin , and possible penetration of the surface by material blobs completes the accretion region production of high energy flux .
@xcite presents a detailed review of polars .
since the accretion regions in polars are the site of the majority of high energy emission , we would expect that euv and x - ray observations would provide a wealth of information about their accretion geometry .
this is indeed the case as shown , for example , by @xcite .
differentiation and study of eclipses of the accretion region by the secondary , the far and near accretion stream , and the white dwarf ( a self - eclipse ; when the rotation of the cv system causes the accretion pole to pass behind the limb of the white dwarf ) , allows a number of system parameters to be determined .
these include the inclination of the system , the position of the accretion pole on the white dwarf , and if the white dwarf is accreting at one or both of its magnetic poles .
with this type of study in mind , one of us ( sbh ) placed qq vul on the target list of the _ euve _ satellite right angle program @xcite .
serendipitously discovered in a survey of soft x - ray sources @xcite , qq vulpeculae was confirmed as an am her binary through detection of circular and linear polarization @xcite .
qq vul has a relatively long orbital period for a polar , @xmath2 min @xcite , and although the strength of the magnetic field has not been directly measured , it is assumed to be fairly typical , @xmath3 mg @xcite .
blackbody fits to the spectrum of qq vul yield a temperature of @xmath4 k for the x - ray heated accretion regions @xcite .
mass estimates of m@xmath5=0.58m@xmath6 and m@xmath7=0.35m@xmath6 have been determined for qq vul by @xcite .
the initial multi - wavelength observations of qq vul @xcite were able to place some constraints on the system geometry .
these observations suggested a system with a magnetic pole tilted @xmath8 from our line of sight during the linear polarization pulse peak , an orbital inclination of @xmath9 , and a stellar latitude of the accreting magnetic pole in the range @xmath10 .
circular and linear polarization observations @xcite have revealed a weak and diffuse linear polarization pulse centered on maximum light , indicating that the near field accretion column is always in sight , although the pole does graze the limb during a self - eclipse @xcite .
the polarization data also suggest that there is non - radial accretion flow @xcite implying a kink " in the accretion stream which flows to the magnetic pole . throughout its history of observations , qq vul
has repeatedly shown a complex and varying x - ray light curve .
studies undertaken with the use of _ einstein _
@xcite , _ rosat _
@xcite , and _ exosat _ @xcite have all shown the complexities apparent in the x - ray component of qq vul .
@xcite found that the soft x - ray count rate had doubled within a period of two years and that the shape of the light curve they observed was indeed quite different from the initial x - ray light curve of qq vul @xcite .
figure 1 in @xcite provides a comparison of the different x - ray light curves previously obtained for qq vul .
there has been conflicting evidence for whether or not qq vul possesses two accreting poles .
while the double peaked nature of the soft x - ray light curves of @xcite might lead one to interpret it as two - pole accretion , it is noted there that the second pole is not evident in the optical light curve .
other observations @xcite detected soft x - ray spectral variations as a function of orbital phase which could be modeled by an extended multi - temperature accreting region or by two accreting poles with slightly different temperatures
. however , all of these previous observations do agree that if two - pole accretion is taking place , the primary accreting pole is a weaker source in soft x - rays .
recent polarization data seem to strongly suggest that qq vul is undergoing two - pole accretion .
optical polarization data from @xcite can not be explained by one - pole accretion and a second linear polarization peak , seen in the data of @xcite , requires a second accreting region to be in view at certain binary phases .
the euv photometric data were obtained with the _ euve _ satellite using the right - angle pointing scanner telescopes a and b. scanner a imaged qq vul through a lexan / boron filter ( @xmath11 ) , sensitive in the bandpass @xmath12 while scanner b data were obtained in an al / ti / c filter ( @xmath13 ) , sensitive in the bandpass @xmath14 .
details of the photometric properties of the imaging telescopes on board the _ euve _ may be found in @xcite .
our observations of qq vul began on 1996 aug 11 ( gmt ) and continued through 1996 aug 16 ( gmt ) , spanning @xmath15s , or a total of @xmath16 .
the data were passed through _ euve _ standard processing and delivered to us in compressed format on cd - rom .
we then extracted the scanner observations using the standard _
data analysis software packages within iraf .
photometry was performed using an aperture with a seven pixel radius centered on the coordinates of the source and a background annulus having a radius of twenty pixels also centered on the object . due to the large difference in signal - to - noise obtained in the two data sets ,
every 100 data points ( photon events ) were binned together for the raw data from scanner a , and every 30 data points were binned together for the raw data from scanner b. an idl program ( written by m. sirk ) was then used to produce light curve data files that were phased according to the ephemeris of the inferior conjunction of the secondary star in qq vul , hjd 2448446.4710(5 ) + 0.15452011(11)e @xcite .
finally , our resultant light curves , in both scanner a ( lexan / b ) and scanner b ( al / ti / c ) , were rebinned to 0.005 in phase .
figures [ sca_lcurve ] and [ scb_lcurve ] present our obtained _ euve _ light curves phased on the @xcite ephemeris . convolving the mean euv count rate of 0.01 counts / second with the effective area as a function of wavelength for the lexan / b filter @xcite , we find that the observed euv flux is @xmath17 ergs@xmath18s@xmath19@xmath20 .
it is interesting to note that scanner b , viewing qq vul through the al / ti / c filter , detected anything at all . at a wavelength of @xmath21 , and a hydrogen column density of @xmath22 @xcite for a distance to qq vul of 215 pc @xcite , the optical depth is @xmath23 . at such a large optical depth
, the ism transmission of photons from qq vul at this wavelength is essentially zero .
we would therefore not expect to detect photons through this filter and indeed no other polar has been detected by _
euve _ in scanner b at these wavelengths . however , the al / ti / c filter is known to have an x - ray leak peaking near 44@xmath24 @xcite , and we conclude that the data collected here with scanner b is an x - ray light curve for qq vul with a mean effective wavelength of @xmath2544@xmath26 .
we note that this is not the first detection of an x - ray leak with the al / ti / c filter ; x - ray leaks were also reported in
observations of the nova v1974 cygni @xcite .
one flaw in our conclusion concerning the x - ray light curve would be if qq vul were actually quite close by in space , say less than 50 pc .
we therefore independently re - determined the distance to qq vul using bailey s method @xcite and newly obtained infrared observations .
bailey s method relies on the relationship between certain physical parameters of the secondary star in the cv system and the distance to that system .
the relation is : @xmath27 where @xmath28 is the distance , @xmath29 is the @xmath29-band magnitude of the secondary star , @xmath30 is the @xmath29-band surface brightness of the secondary , and @xmath31 is the radius of the secondary star . using data kindly obtained by m. huber with the wyoming infrared observatory on 1998 aug 30 ut ( 9:30 hours ) , we find that qq vul had a @xmath29 magnitude of [email protected] mag .
taking @xmath33 @xcite , and @xmath34 @xcite , the distance to qq vul is determined to be @xmath35 pc ; a value which is in agreement with earlier measurements which suggest a lower limit of @xmath36 pc @xcite . @xmath37 and @xmath38 observations of qq vul by @xcite detected a visual companion star " to qq vul with @xmath39 ( obtained using @xmath40 colors derived for @xmath29 spectral type stars ) .
if possible contamination from this star in the infrared ( i.e. , @xmath29 band ) is taken into consideration , qq vul would be even farther away .
therefore , it seems highly unlikely that the detected signal in scanner b is due to @xmath14 photons but is instead the aforementioned x - ray leak .
the effective bandpass of the x - ray leak in the al / ti / c filter is roughly triangular in shape and covers the range of @xmath41 .
the peak throughput , at 68@xmath26 , is @xmath422% of the normal filter transmission near 171@xmath26 and has zero sensitivity to photons with a wavelength below @xmath43 @xcite . using the effective area ratio of the lexan / b filter to the al / ti / c filter @xcite , and the fact discussed above concerning the total absence of long wavelength photons
, we can determine that the x - rays observed for qq vul have wavelengths from @xmath44 , with a mean effective central wavelength near 44@xmath26 . performing an approximate integration under the triangular bandpass and convolving it with the effective area as a function of wavelength , we find the observed x - ray flux to be @xmath45 ergs@xmath18s@xmath19@xmath20 .
x - ray fluxes ranging from @xmath46 ergs@xmath18s@xmath19@xmath20 @xcite to @xmath47 ergs@xmath18s@xmath19@xmath20 @xcite have been reported in previous studies , which , along with the value determined here , reflect the variability of the source .
we have thus obtained simultaneous time - resolved photometric data in the x - ray ( @xmath48 ) ( figure [ scb_lcurve ] ) and euv ( @xmath49 ) ( figure [ sca_lcurve ] ) wavelength regions which allow us to make a direct comparison of the emitting character of this system in these two wavelength regimes .
the euv and x - ray light curves , figures [ sca_lcurve ] and [ scb_lcurve ] respectively , both show a double peak shape with minima occurring near phase 0.2 and phase 0.85 and maxima at phases 0.0 and 0.45 .
while each light curve reveals similar gross features , we note that there is far less change and detail in the euv data
. this could be due to the fact that qq vul has a broader , more diffuse euv emitting region but a smaller , better defined x - ray emitting region .
the modulations of both light curves are uneven in their minima and maxima . the minima alternate between a deep , essentially complete eclipse at phase 0.85 to a less deep and well - defined dip near phase 0.2 , while the maxima shift between a narrow peaked one near phase 0.0 to a brighter , broader one covering about 0.4 in phase , centered at 0.45 . the two maxima , while showing that the secondary pole is stronger in intensity , are probably nearly equal in phase extent and overall shape , the narrower one being cut - off " around phase 0.9 by an eclipse of the magnetic pole accretion region by the near - field accretion stream @xcite
interpreted as a two - pole accretor , the locations of the two poles would have centers near binary phases 0.1 and 0.55 , that is , almost directly along the line of centers of the binary .
the eclipse by the near - field stream of the accretion region facing the secondary star occurs before phase 0 as is the case in most polars @xcite .
the magnetic pole on the far side of the white dwarf suffers no eclipse , thus , it is visible in phase for approximately one - half of the orbital period , and its shape is consistent with a spot latitude of @xmath50 @xcite given a binary inclination of @xmath51 ( see below ) .
figure [ zoom ] is a close - up of the x - ray light curve minimum near phase 0.8 .
it appears that this minima has two components .
the first half of the broad dip , starting at phase 0.7 , has a slow decline up to phase 0.85 and is likely to be the result of an eclipse by the near - field accretion stream .
the remaining part of this dip shows an abrupt drop to near zero counts and appears to be flat bottomed from phase 0.85 to 0.92 , with the most likely cause being a stellar eclipse of the x - ray emitting region by the secondary star .
if true , this constrains the system inclination to be greater than @xmath52 .
an interesting feature appearing in the x - ray light curve ( figures [ scb_lcurve ] and [ zoom ] ) , but not seen in the euv light curve , is the narrow dip which occurs at phase 0.96 .
using other polar light curves as a guide , this narrow dip feature is likely to correspond to an eclipse of the accretion region by the far - field accretion stream .
it may also be present in the euv data but the noise level precludes its discovery . while unresolved , the short duration of this narrow feature ( 0.015 in phase or 3 min ) provides strong evidence for the compactness of the hard x - ray emitting region in qq vul . translating this time in to a size on the white dwarf surface ( without correction for latitude and assuming r@xmath53 = 7000 km ) we find an emitting region diameter of 660 km or , if circular , @xmath54=0.002 .
taking the gross ratio of the low euv count rate to the higher x - ray count rate ( even as a leaked signal ) one could conclude that the magnetic field strength in qq vul is relatively low , possibly less than 10 - 30 mg @xcite .
however , while in general a large x - ray to euv ratio indicates a lower magnetic field strength in polars , this is not always the case with the difference attributed to the structure and size of the accretion region @xcite .
figure [ hratio ] presents the hardness ratio ( x - ray / euv ) for qq vul . due to the low value of the flux in the euv light curve ,
both the x - ray and euv light curves were re - binned to 0.05 in phase , thereby allowing the hardness ratio to not be dominated by noise spikes due to the low euv flux values .
figure [ hratio ] exhibits an increased hardness near phase 0.25 , with a sharp rise near phase 0.3 .
this peak might indicate the emergence of the second accreting pole .
a comparison of our qq vul x - ray light curve with previous high energy observations shows that the continuous orbital variations and temporal changes , noted by @xcite , appear to be persistent .
some similarities , however , do exist between our x - ray light curve and the 1983 oct and 1985 jun _ exosat _ light curves discussed in @xcite .
for example , the unequal minima and maxima and even the presence of a narrow dip , probably due to an eclipse of the accretion region by the far - field accretion stream .
this narrow dip feature occurs near phase 0.03 in the 1983 oct _
exosat _ light curve and 0.08 in the 1985 jun _ exosat _ light curve , compared with phase 0.96 seen in our light curve [ according to the ephemeris of @xcite ] .
the 1985 sep _ exosat _
light curve @xcite , is very different from our present data as it exhibits a much different shape with nearly equal maxima and yet again a narrow dip but one which appears at yet another phase ( phase 0.71 ) within the light curve .
the fact that the narrow dip is always present but changes phase indicates a movement , within the binary , of the far - field accretion stream similar to that observed in hu aqr @xcite .
keb is supported by a graduate assistantship from the university of wyoming .
sbh acknowledges partial support for this work from nasa cooperative agreement ncc5 - 138 through an
guest observer mini - grant and from nasa adp grant nag5 - 4233 .
the authors wish to thank jennifer cash and david ciardi for their assistance with data reduction , martin sirk for his assistance with data reduction and for extremely useful discussions concerning the performance of the _ euve _ filters , mark huber for providing @xmath29-band photometry for qq vul , and axel schwope for supplying us with his qq vul ephemeris prior to publication .
99 bailey , j. 1982 , , 197 , 31 beardmore , a. p. ramsay , g. , osborne , j. p. , mason , k. o. , nousek , j. a. , & baluta , c. 1995 , , 273 , 742 cropper , m. 1998 , , 295 , 353 cropper , m. 1990 , _ space science reviews _ , 54 , 195 finley , d. s. , jelinsky , p. , bowyer s. , & malina , r. f. 1988 , appl . opt .
27 , 1476 liebert , j. , & stockman , h. s. , 1985 , in _ cataclysmic variables and low mass x - ray binaries _ , eds patterson , j. , & lamb , d. q. , reidel , dordrecht , holland , 151 macdonald , j. 1996 , in _ astrophysics in the euv _ , eds .
bowyer , s. , & malina , r. f. , kluwer , dordrecht , 395 mccarthy , p. , bowyer , s. , & clarke , j. t. 1986 , , 873 , 881 mcdonald , k. , craig , n. , sirk , m. m. , drake , j. j. , fruscione , a. , vallerga , j. v. , & malina , r. f. 1994 , , 108 , 1843 mukai k. , & charles , p. a. 1986 , , 222 , 1 mukai k. , & charles , p. a. 1987 , , 226 , 209 mukai , k. , charles , p. a. , & smale , a. p. 1988
, , 231 , 663 nugent , j. j. , et al .
1983 , , 51 , 1 nousek , j. a. , et al .
1984 , , 277 , 682 nousek , j. , luppino , g. , gajar , s. , bond , h. , grauer , a. , schmidt , g. , hill , g. , & tapia , s. 1982 , _ iau circ . _
3733 osborne , j. p. , et al .
1986 , , 221 , 823 osborne , j. p. , beuermann , k. , charles , p. , maraschi , l. , mukai , k. , & treves , a. 1987 , , 315 , l123 ramsay , g. , mason , k. , cropper , m. , watson , m. g. , & clayton , k. l. , 1994 , , 270 , 692 schwope , a. d. 1991 , phd thesis , berlin university schwope , a. , et al .
1998a , , submitted schwope , a. , et al .
1998b , in _ wild stars in the old west _ ; asp conference series 137 , eds .
howell , s. , kuulkers , e. , & woodward , c. , 44 sirk , m. m. 1999 , euve newsletter , 8 sirk , m. m. , & howell , s. b. 1998 , , 506 , 824 sirk , m. m. , vallerga , j. v. , finley , d. s. , jelinsky , p. , & malina r. f. 1997 , , 110 , 347 stringfellow , g. s. , & bowyer , s. 1996 , in _ astrophysics in the euv _ , eds .
bowyer , s. , & malina , r. f. , kluwer , dordrecht , 401 vallerga , j. , & sirk , m. 2000 , in prep . | we present simultaneous x - ray ( @xmath0 ) and euv ( @xmath1 ) light curves for the magnetic cataclysmic variable qq vulpeculae , obtained with the _ euve _ satellite .
we find that the unique shape of the x - ray light curve is different from previously obtained x - ray light curves of qq vul and provides evidence for two - pole accretion .
detailed examination of the photometric data indicates that qq vul undergoes a stellar eclipse of the x - ray emitting region , indicative of a high binary inclination .
we discuss possible implications for the nature of this system given the observed shape of its euv and x - ray light curves . |
enumeration of finite mathematical structures is an important tool since it allows testing new hypotheses and searching for counterexamples .
additionally , it provides insight into the properties of these structures .
here we concentrate on constructing , up to isomorphism , all modular lattices with a given number of elements .
the algorithm we develop is a modification of the approach of heitzig and reinhold @xcite who enumerated ( up to isomorphism ) all lattices with up to 18 elements .
the number of distributive lattices of size up to 49 were calculated by ern , heitzig and reinhold @xcite . in the online encyclopedia of integer sequences ( ` oeis.org ` ) the relevant sequences are a006981 , a006966 and a006982 , but the sequence for the number of modular lattices was given only up to @xmath2 .
for @xmath3 there are 766 nonisomorphic modular lattices , as was reported in @xcite .
we extend this result to @xmath4 and also count the number of semimodular lattices up to size @xmath5 ( see table [ table1 ] ) .
our algorithm uses an improved method for removing isomorphic copies , which allowed us to recalculate the numbers in @xcite for all lattices up to @xmath6 and go one step further to find the number of nonisomorphic lattices with 19 elements .
the calculations were done on a cluster of 64 processors and took 26 hours for @xmath6 and 19 days for @xmath7 . in the remainder of this section , we define some properties and recall some basic results of ( semi)modular lattices . in section 2 ,
we give an outline of the algorithm used by @xcite to generate finite lattices up to isomorphism .
then , in section 3 , we adapt this algorithm to generate modular lattices up to isomorphism by adding a series of constraints to the algorithm .
section 4 contains an improvement for the algorithm used by @xcite by employing the canonical construction path introduced in @xcite . in section 5 , the algorithm
is adjusted to generate only vertically indecomposable modular lattices . a _ modular lattice _
@xmath8 is a lattice which satisfies the modular law @xmath9 weaker conditions of modularity are semimodularity and lower semimodularity . as usual , we write @xmath10 if @xmath11 is _ covered by _ @xmath12 .
a lattice @xmath8 is _ semimodular _ if for all @xmath13 @xmath14 dually , @xmath8 is _ lower semimodular _ if for all @xmath13 @xmath15 recall that a _ chain _ in a lattice @xmath8 is a subset of @xmath8 such that all elements in the subset are comparable .
we say that a lattice has _ finite length _ if all chains in it have finite cardinality .
the next two well - known results below can be found for example in @xcite .
[ mod - semimod ] a lattice of finite length is modular if and only if it is semimodular and lower semimodular .
a chain @xmath16 in a poset @xmath17 is _ maximal _ if whenever @xmath18 and @xmath19 is a chain in @xmath17 , then @xmath20 . in a finite lattice ,
a maximal chain is a chain from bottom to top such that each element in the chain , other than the top , is covered by some element in the chain .
[ jhcc ] let @xmath8 be a finite semimodular lattice .
then , for any maximal chains @xmath16 and @xmath19 in l , @xmath21 .
there are many ways to represent finite lattices and to construct bigger lattices from smaller lattices .
an algorithm that constructs up to isomorphism all combinatorial objects of a certain kind and of a given size is called an _ orderly algorithm _ if it produces exactly one member of each isomorphism class without testing that this member is nonisomorphic to previously constructed objects .
such algorithms were first introduced by faradzhev @xcite and read @xcite for enumerating finite graphs .
heitzig and reinhold @xcite developed an orderly algorithm to enumerate all finite lattices up to isomorphism and used it to count the number of lattices up to size 18 . since our first algorithm for modular lattices is based on their approach , we recall some of the details here .
let @xmath8 be a lattice .
a nonempty antichain @xmath22 is a _ lattice - antichain _ if @xmath23 for all @xmath24 , where @xmath25 for some @xmath26 .
a finite lattice is called an @xmath1-lattice if its set of elements is @xmath27 , where @xmath28 and @xmath29 are the bottom and top elements . given a lattice antichain @xmath30 and an @xmath1-lattice @xmath8 , a poset @xmath31 with @xmath32 elements is constructed by adding an element @xmath1 to @xmath8 as an atom with @xmath30 as the set of its covers . furthermore , the following lemma states that @xmath31 is a lattice .
a subset @xmath33 of an @xmath1-lattice @xmath8 is a lattice - antichain if and only if @xmath8 is a subposet of an @xmath34-lattice @xmath31 in which the element @xmath1 is an atom and @xmath30 is the set of its covers . in order to generate only one copy of each lattice up to isomorphism
, the weight @xmath35 of an @xmath1-lattice @xmath8 is defined by setting @xmath36 . with this weight , for two @xmath1-lattices @xmath8 and @xmath37
, @xmath38 is said to be ( lexicographically ) smaller than @xmath39 if there is an @xmath40 such that @xmath41 and @xmath42 for all @xmath43 .
an @xmath1-lattice c is called a canonical lattice if there is no @xmath1-lattice isomorphic to @xmath16 that has a smaller weight . in order to check
whether an @xmath1-lattice @xmath8 is canonical , one has to check whether there is a permutation of the elements of @xmath8 that yields an isomorphic copy of @xmath8 with a smaller weight . with these definitions ,
a recursive algorithm is formulated in @xcite which generates exactly all canonical lattices of order less or equal to @xmath1 for a given natural number @xmath44 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ` next_lattice(integer m , canonical m - lattice l ) begin if m < n then for each lattice - antichain a of l do if l^{a } is a canonical lattice then next_lattice(m+1 , l^{a } ) if m = n then output l end ` _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ algorithm 1 the set of all maximal elements in a finite poset @xmath17 is called the _ first level _ of @xmath17 and is denoted by @xmath45 .
the _ @xmath46-th level _ of @xmath17 is recursively defined by @xmath47 following @xcite we define @xmath48 to be the number @xmath49 such that @xmath50 .
although heitzig and reinhold refer to this as the depth of @xmath51 , it is more traditional to consider the depth of @xmath51 to be given by @xmath52 . to avoid confusion , we only use the function @xmath53 rather than the notion of depth .
we say an @xmath1-lattice @xmath8 is _ levelized _ if @xmath54 in other words , the levels form a partition on @xmath55 of the form @xmath56 , where @xmath49 is the number of levels in @xmath8 . throughout the rest of the paper , we consider the bottom level to be @xmath57 , unless indicated otherwise .
the following lemma gives us an important property when generating levelized lattices , since it tells us we only need to consider lattice - antichains that have at least one element in the two bottom levels .
[ l2.2 ] for a levelized @xmath1-lattice @xmath8 and a lattice - antichain @xmath30 , @xmath31 is levelized if and only if @xmath58 .
in order to construct only modular lattices of size @xmath1 using this algorithm , we start by selecting only the lattices that are modular when we get to size @xmath1 . however
, modular lattices constitute a very small fraction of the total number of lattices .
therefore , it is important to add constraints in order to minimize the generation of non - modular lattices . in order to do this ,
we present a series of results to decide when a subtree in a search tree can be cut off , and which lattice antichains must be considered in each step . during this section , we refer to _ descendants _ of a lattice @xmath8 as those lattices generated through the element extension described in @xcite , together with any additional constraints introduced in this section .
[ l3.1 ] for an @xmath1-lattice @xmath8 , and a lattice - antichain @xmath59 , if there exist @xmath60 such that @xmath61 , then all descendants of @xmath31 are non - semimodular . specifically , they are non - modular .
assume without loss of generality that @xmath62 .
let @xmath63 and @xmath64 be the chains of maximal cardinality from 1 to @xmath11 and @xmath12 respectively . for any @xmath65 , @xmath66 is equal to the cardinality of the longest chain from @xmath67 to 1 , hence @xmath68 .
next , in @xmath31 , we have @xmath69 and @xmath70 .
let @xmath37 be a descendant of @xmath31 , and choose any chain @xmath19 from 0 to @xmath1
. then @xmath71 and @xmath72 are maximal chains of different cardinality since @xmath73 . by theorem [ jhcc ]
, it follows that @xmath37 is not semimodular , and therefore is non - modular . from lemmas [ l2.2 ] and [ l3.1 ]
we may conclude the following result .
[ levk ] for the construction of ( semi-)modular lattices using the orderly algorithm of @xcite , it is sufficient to consider lattice - antichains @xmath30 such that @xmath74 or @xmath75
. let @xmath8 be an @xmath1-lattice where @xmath76 is the bottom nonzero level , and let @xmath75 be a lattice - antichain of @xmath8 . if there is an atom of @xmath8 in @xmath77 then all descendants of @xmath31 are non - semimodular , and hence non - modular .
let @xmath78 be an atom of @xmath8 , and choose any @xmath79 . then there exist chains @xmath63 from @xmath11 to 1 and @xmath64 from @xmath12 to 1 of cardinality @xmath49 and @xmath80 respectively .
since the new element @xmath1 is in a new level @xmath81 , and @xmath82 , @xmath12 is contained in the third lowest level of @xmath31 .
therefore , by corollary 3.2 , it is not used in the generation of any descendants , and @xmath12 remains as an atom in all descendants .
hence , the maximal chain @xmath83 is of constant cardinality @xmath84 for any descendant of @xmath31 .
let @xmath37 be a descendant of @xmath31 .
choose any chain @xmath85 from 0 to @xmath1 , then @xmath86 .
therefore , for the maximal chain @xmath87 , @xmath88 by theorem [ jhcc ] , since both @xmath89 and @xmath90 are maximal chains , it follows that @xmath37 is non - semimodular and therefore non - modular .
an observation that significantly decreases the search space is based on the following property of the algorithm : since elements are always added below a lattice antichain , if two elements in the antichain fail semimodularity , then those two elements also fail semimodularity in any of the descendants . therefore , when adding a new element below a lattice antichain , we should check that we are not generating a non - semimodular lattice . for an @xmath1-lattice @xmath8 and a lattice antichain @xmath59 , if there exist @xmath60 which do not have a common cover , then all descendants of @xmath31 are non - semimodular . in @xmath31 , for the new element @xmath1 , @xmath69 and @xmath70 . however , @xmath91 or @xmath92 .
therefore , @xmath31 is not semimodular . furthermore , for any descendant @xmath37 of @xmath31
, it is not possible to add a common cover to @xmath93 .
hence , @xmath37 is not semimodular ( and consequently , not modular ) .
similarly , we can consider when it is not possible to make a non - lower semimodular lattice into a lower semimodular lattice .
let @xmath8 be an @xmath1-lattice , and let @xmath49 be its bottom non - zero level .
if there exist @xmath94 which do not satisfy lower semimodularity , then all descendants of @xmath8 are non - lower semimodular ( and hence non - modular ) . given an @xmath93 such that @xmath95 but @xmath96 or @xmath97 , the algorithm can make @xmath93 satisfy lower semimodularity by adding an element below @xmath93
however , by corollary [ levk ] , we only consider lattice antichains in @xmath57 and @xmath77 .
therefore , if @xmath94 , we can not add a common co - cover , and all descendants @xmath37 of @xmath8 are non - lower semimodular .
this lemma can be incorporated into the algorithm by checking that all elements of @xmath77 satisfy lower semimodularity each time a new level is added .
the preceding results are summarized in the following theorems .
when generating semimodular lattices , for a lattice @xmath8 , we only consider lattice - antichains @xmath30 which satisfy all of the following conditions : * @xmath74 or @xmath75 . * if @xmath75 , there are no atoms in @xmath77 .
* for all @xmath98 , @xmath67 and @xmath99 have a common cover .
when generating modular lattices , for a lattice @xmath8 , we only consider lattice - antichains @xmath30 which satisfy and * if @xmath75 , then @xmath77 satisfies lower semimodularity ( i. e. , for all @xmath100 implies @xmath101 ) .
another improvement can be implemented in the last step when generating lattices of size @xmath1 from those of size @xmath102 , by only considering lattice - antichains @xmath74 and @xmath103 .
this is due to the following result .
[ bottomlev ] for an @xmath1-lattice @xmath8 and a lattice antichain @xmath104 , the @xmath32-lattice @xmath31 is non - modular .
since @xmath104 , there exists @xmath105 such that @xmath106 .
let @xmath107 .
since @xmath108 , there exist chains @xmath63 and @xmath64 from @xmath11 to 1 and @xmath12 to 1 respectively , both of cardinality @xmath49 . in @xmath31 , @xmath69 , but @xmath109 .
thus , for the maximal chains @xmath110 and @xmath83 , @xmath111 thus @xmath31 is non - modular .
when generating finite ( modular ) lattices using algorithm 1 , the majority of the time is spent in testing if the lattice @xmath31 is canonical , an operation of order @xmath112 .
an approach that speeds - up the algorithm significantly , while still generating exactly one isomorphic copy of each ( modular ) lattice is via generation by _ canonical construction path _ , which was introduced by mckay @xcite .
this canonical construction has two components .
the first is to use only one representative of each orbit in the lattice antichains of @xmath8 . in other words , if there is an automorphism @xmath113 on @xmath8 such that @xmath114 for lattice - antichains @xmath115 , only one of these antichains is chosen arbitrarily .
the second is , after the extension of any lattice @xmath8 using @xmath30 , @xmath31 is checked to see if @xmath8 is the inverse through a `` canonical deletion '' .
this uses the canonical labeling of the program ` nauty ` @xcite . in general , a _
canonical labeling _
associates with each @xmath1-lattice @xmath8 a permutation @xmath116 on @xmath117 such that for any @xmath1-lattice @xmath37 we have @xmath118 if and only if @xmath119 i.e. , the permutation maps each lattice to a fixed representative of its isomorphism class .
when a new @xmath34-lattice @xmath31 is generated from @xmath8 and @xmath30 , a canonical labeling @xmath120 of @xmath31 is generated using a partition by levels in ` nauty ` .
let @xmath121 denote the element which maps to @xmath1 under the canonical labeling .
we consider the set @xmath122 , where @xmath123 denotes the lattice obtained by removing @xmath11 from @xmath31 .
note that @xmath124 . if @xmath125 , we say @xmath31 is _ canonical _ and keep it , otherwise it is discarded . using this construction , theorem 1 in @xcite states that starting from any lattice
, exactly one isomorphic copy of each descendant will be output .
thus , starting with the two - element lattice , we can generate exactly one isomorphic copy of each lattice of a given size @xmath1 .
this has an advantage over the construction used in @xcite since it does not require checking all permutations of a lattice , and it uses canonical labeling by ` nauty ` , which is generally considered the most efficient canonical labeling program for small combinatorial structures .
furthermore , this construction is orderly since it only considers the lattices @xmath8 and @xmath31 .
this is beneficial during computations because it does not require storage of previously generated lattices or communication between nodes during parallel computations .
given this , algorithm 1 can be modified : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ` next_lattice2(integer m , canonical m - lattice l ) begin if m < n then lac:=\{a \mid a is a lattice - antichain of l } for each orbit o of the action of aut(l ) on lac select any a\in o c:= canonical labeling of l^a n':=c^{-1}(n ) if f(n)=n ' for some f\in aut(l^a ) then next_lattice2(m+1 , l^{a } ) if m = n then output l end ` _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ algorithm 2
we say a lattice @xmath8 is _ vertically decomposable _ if it contains an element which is neither the greatest nor the least element of @xmath8 but is comparable with every element of @xmath8 . a lattice which is not vertically decomposable is said to be _ vertically indecomposable_. let @xmath126 be the number of unlabeled vertically indecomposable modular lattices .
then the recursive formula @xcite can be used to compute the number of unlabeled modular lattices from the number of unlabeled vertically indecomposable modular lattices .
@xmath127 in order to avoid generating vertically decomposable modular lattices , we only need to avoid using @xmath57 as a lattice antichain of @xmath8 , since then @xmath128 would be comparable to all elements in @xmath31 .
however , not using these lattice antichains would cut off branches of the canonical path that could potentially generate vertically indecomposable canonical lattices .
lemma [ vi ] tells us we can safely avoid using them when @xmath129 .
[ vi ] given an @xmath1-lattice @xmath8 with only one atom @xmath102 , then all descendants of @xmath130 are vertically decomposable . it is clear that @xmath131 and @xmath132 , where @xmath133 . under our construction ,
only lattice - antichains @xmath134 and @xmath135 are considered .
therefore , for any descendant @xmath37 of @xmath31 and any new element @xmath136 such that @xmath137 or @xmath138 , @xmath139 . additionally , for all @xmath140 , @xmath141 .
thus , @xmath37 is vertically decomposable .
this means that we only construct vertically decomposable lattices where the only comparable element is a single atom .
however , these are ignored during the count of vertically indecomposable lattices . note that in the last step , by lemma [ bottomlev ] , we only have to consider lattice - antichains in @xmath77 .
( a0 ) at(0,0 ) ; ( a1 ) at(@xmath142 ) ; at(@xmath143)1 ; at(@xmath144)0 ; ( a0)edge(a1 ) ; ( @xmath145 ) rectangle ( @xmath146 ) ; ( @xmath147 )
+ ( 0,-0.5 ) ; ( b0 ) at(@xmath148 ) ; ( b1 ) at(@xmath149 ) ; ( b2 ) at(@xmath150 ) ; at(@xmath151)0 ; at(@xmath152)1 ; at(@xmath153)2 ; ( b0)edge(b2 ) ; ( b2)edge(b1 ) ; ( @xmath154 ) rectangle ( @xmath155 ) ; ( @xmath156 )
+ ( 0,-0.5 ) ; ( c0 ) at(@xmath157 ) ; ( c1 ) at(@xmath158 ) ; ( c2 ) at(@xmath159 ) ; ( c3 ) at(@xmath160 ) ; at(@xmath161)0 ; at(@xmath162)1 ; at(@xmath163)2 ; at(@xmath164)3 ; ( c0)edge(c2 ) ; ( c0)edge(c3 ) ; ( c3)edge(c1 ) ; ( c2)edge(c1 ) ; ( @xmath165 ) rectangle ( @xmath166 ) ; ( @xmath165 )
+ ( -1,-0.5 ) ; ( d0 ) at(@xmath167 ) ; ( d1 ) at(@xmath168 ) ; ( d2 ) at(@xmath169 ) ; ( d3 ) at(@xmath170 ) ; ( d4 ) at(@xmath171 ) ; at(@xmath172)0 ; at(@xmath173)1 ; at(@xmath174)2 ; at(@xmath175)3 ; at(@xmath176)4 ; ( d0)edge(d2)edge(d3)edge(d4 ) ; ( d2)edge(d1 ) ; ( d3)edge(d1 ) ; ( d4)edge(d1 ) ; ( @xmath177 ) rectangle ( @xmath178 ) ; ( @xmath179 ) + ( 2.7,-0.5 ) ; ( e0 )
at(@xmath180 ) ; ( e1 ) at(@xmath181 ) ; ( e2 ) at(@xmath182 ) ; ( e3 ) at(@xmath183 ) ; ( e4 ) at(@xmath184 ) ; at(@xmath185)0 ; at(@xmath186)1 ; at(@xmath187)2 ; at(@xmath188)3 ; at(@xmath189)4 ; ( e0)edge(e4)edge(e3 ) ; ( e3)edge(e1 ) ; ( e2)edge(e1)edge(e4 ) ; ( @xmath190 ) rectangle ( @xmath191 ) ; ( @xmath177 ) + ( -1.7,-0.5 ) ; ( g0 )
at(@xmath192 ) ; ( g1 ) at(@xmath193 ) ; ( g2 ) at(@xmath194 ) ; ( g3 ) at(@xmath195 ) ; ( g4 ) at(@xmath196 ) ; ( g5 ) at(@xmath197 ) ; at(@xmath198)0 ; at(@xmath199)1 ; at(@xmath200)2 ; at(@xmath201)3 ; at(@xmath202)4 ; at(@xmath203)5 ; ( g0)edge(g2)edge(g3)edge(g4)edge(g5 ) ; ( g1)edge(g2)edge(g3)edge(g4)edge(g5 ) ; ( @xmath204 ) rectangle ( @xmath205 ) ; ( @xmath206 ) + ( 0,-0.5 ) ; ( h0 ) at(@xmath207 ) ; ( h1 ) at(@xmath208 ) ; ( h2 ) at(@xmath209 ) ; ( h3 ) at(@xmath210 ) ; ( h4 ) at(@xmath211 ) ; ( h5 ) at(@xmath212 ) ; at(@xmath213)0 ; at(@xmath214)1 ; at(@xmath215)2 ; at(@xmath216)3 ; at(@xmath217)4 ; at(@xmath218)5 ; ( h0)edge(h3)edge(h4)edge(h5 ) ; ( h1)edge(h2)edge(h3)edge(h4 ) ; ( h2)edge(h5 ) ; ( @xmath219 ) rectangle ( @xmath220 ) ; ( @xmath221 ) +
( 1.5,-0.5 ) ; ( i0 ) at(@xmath222 ) ; ( i1 ) at(@xmath223 ) ; ( i2 ) at(@xmath224 ) ; ( i3 ) at(@xmath225 ) ; ( i4 )
at(@xmath226 ) ; ( i5 ) at(@xmath227 ) ; at(@xmath228)0 ; at(@xmath229)1 ; at(@xmath230)2 ; at(@xmath231)3 ; at(@xmath232)4 ; at(@xmath233)5 ; ( i0)edge(i4)edge(i5 ) ; ( i1)edge(i2)edge(i3)edge(i4 ) ; ( i5)edge(i2)edge(i3 ) ; ( @xmath234 ) rectangle ( @xmath235 ) ; ( @xmath236 ) +
( 1,-0.5 ) ; ( j0 ) at(@xmath237 ) ; ( j1 ) at(@xmath238 ) ; ( j2 ) at(@xmath239 ) ; ( j3 ) at(@xmath240 ) ; ( j4 ) at(@xmath241 ) ; ( j5 ) at(@xmath242 ) ; at(@xmath243)0 ; at(@xmath244)1 ; at(@xmath245)2 ; at(@xmath246)3 ; at(@xmath247)4 ; at(@xmath248)5 ; ( j0)edge(j4)edge(j5 ) ; ( j1)edge(j2)edge(j3 ) ; ( j5)edge(j2)edge(j3 ) ; ( j4)edge(j2 ) ; ( @xmath249 ) rectangle ( @xmath250 ) ; ( @xmath251 ) + ( 0,-0.5 ) ; ( k0 ) at(@xmath252 ) ; ( k1 ) at(@xmath253 ) ; ( k2 ) at(@xmath254 ) ; ( k3 ) at(@xmath255 ) ; ( k5 ) at(@xmath256 ) ; ( k4 ) at(@xmath257 ) ; at(@xmath258)0 ; at(@xmath259)1 ; at(@xmath260)2 ; at(@xmath261)3 ; at(@xmath262)4 ; at(@xmath263)5 ; ( k0)edge(k4)edge(k5)edge(k3 ) ; ( k1)edge(k2)edge(k3 ) ; ( k2)edge(k4)edge(k5 ) ; ( @xmath264 ) rectangle ( @xmath265 ) ; ( @xmath204 ) + ( -0.8,-0.5 ) ; ( l0 )
at(@xmath266 ) ; ( l1 ) at(@xmath267 ) ; ( l2 ) at(@xmath268 ) ; ( l3 ) at(@xmath269 ) ; ( l4 ) at(@xmath270 ) ; ( l5 ) at(@xmath271 ) ; ( l6 ) at(@xmath272 ) ; at(@xmath273)0 ; at(@xmath274)1 ; at(@xmath275)2 ; at(@xmath276)3 ; at(@xmath277)4 ; at(@xmath278)5 ; at(@xmath279)6 ; ( l0)edge(l2)edge(l3)edge(l4)edge(l5)edge(l6 ) ; ( l1)edge(l2)edge(l3)edge(l4)edge(l5)edge(l6 ) ; ( @xmath280 ) rectangle ( @xmath281 ) ; ( @xmath282 ) + ( 0,-0.5 ) ; ( m0 ) at(@xmath283 ) ; ( m1 ) at(@xmath284 ) ; ( m2 ) at(@xmath285 ) ; ( m3 ) at(@xmath286 ) ; ( m4 ) at(@xmath287 ) ; ( m5 ) at(@xmath288 ) ; ( m6 ) at(@xmath289 ) ; at(@xmath290)0 ; at(@xmath291)1 ; at(@xmath292)2 ; at(@xmath293)3 ; at(@xmath294)4 ; at(@xmath295)5 ; at(@xmath296)6 ; ( m0)edge(m6)edge(m3)edge(m4)edge(m5)edge(m6 ) ; ( m1)edge(m2)edge(m3)edge(m4)edge(m5 ) ; ( m2)edge(m6 ) ; ( @xmath297 ) rectangle ( @xmath298 ) ; ( @xmath219 ) + ( -0.2,-0.5 ) ; ( n0 )
at(@xmath299 ) ; ( n1 ) at(@xmath300 ) ; ( n2 ) at(@xmath301 ) ; ( n3 ) at(@xmath302 ) ; ( n4 ) at(@xmath303 ) ; ( n5 ) at(@xmath304 ) ; ( n6 ) at(@xmath305 ) ; at(@xmath306)0 ; at(@xmath307)1 ; at(@xmath308)2 ; at(@xmath309)3 ; at(@xmath310)4 ; at(@xmath311)5 ; at(@xmath312)6 ; ( n0)edge(n3)edge(n4)edge(n5)edge(n6 ) ; ( n1)edge(n2)edge(n3)edge(n4 ) ; ( n2)edge(n5)edge(n6 ) ; ( @xmath313 ) rectangle ( @xmath314 ) ; ( @xmath315 ) + ( 0,-0.5 ) ; ( o0 ) at(@xmath316 ) ; ( o1 ) at(@xmath317 ) ; ( o2 ) at(@xmath318 ) ; ( o3 ) at(@xmath319 ) ; ( o4 ) at(@xmath320 ) ; ( o5 ) at(@xmath321 ) ; ( o6 ) at(@xmath322 ) ; at(@xmath323)0 ; at(@xmath324)1 ; at(@xmath325)2 ; at(@xmath326)3 ; at(@xmath327)4 ; at(@xmath328)5 ; at(@xmath329)6 ; ( o0)edge(o5)edge(o6 ) ; ( o1)edge(o2)edge(o3)edge(o4 ) ; ( o5)edge(o2 ) ; ( o6)edge(o2)edge(o3)edge(o4 ) ; ( @xmath330 ) rectangle ( @xmath331 ) ; ( @xmath332 ) +
( 0,-0.5 ) ; ( p0 ) at(@xmath333 ) ; ( p1 ) at(@xmath334 ) ; ( p2 ) at(@xmath335 ) ; ( p3 ) at(@xmath336 ) ; ( p4 ) at(@xmath337 ) ; ( p5 ) at(@xmath338 ) ; ( p6 ) at(@xmath339 ) ; at(@xmath340)0 ; at(@xmath341)1 ; at(@xmath342)2 ; at(@xmath343)3 ; at(@xmath344)4 ; at(@xmath345)5 ; at(@xmath346)6 ; ( p0)edge(p5)edge(p6 ) ; ( p1)edge(p2)edge(p3)edge(p4 ) ; ( p2)edge(p5 ) ; ( p4)edge(p6 ) ; ( p3)edge(p5)edge(p6 ) ; ( @xmath347 ) rectangle ( @xmath348 ) ; ( @xmath349 ) + (
0,-0.5 ) ; ( q0 ) at(@xmath350 ) ; ( q1 )
at(@xmath351 ) ; ( q2 ) at(@xmath352 ) ; ( q3 ) at(@xmath353 ) ; ( q4 ) at(@xmath354 ) ; ( q5 ) at(@xmath355 ) ; ( q6 ) at(@xmath356 ) ; at(@xmath357)0 ; at(@xmath358)1 ; at(@xmath359)2 ; at(@xmath360)3 ; at(@xmath361)4 ; at(@xmath362)5 ; at(@xmath363)6 ; ( q6)edge(q4)edge(q5)edge(q0 ) ; ( q1)edge(q2)edge(q3 ) ; ( q5)edge(q2)edge(q3 ) ; ( q4)edge(q2 ) ; ( @xmath364 ) rectangle ( @xmath365 ) ; ( @xmath264 ) + ( 0,-0.5 ) ; ( r0 ) at(@xmath366 ) ; ( r1 ) at(@xmath367 ) ; ( r2 ) at(@xmath368 ) ; ( r3 ) at(@xmath369 ) ; ( r6 ) at(@xmath370 ) ; ( r5 ) at(@xmath371 ) ; ( r4 ) at(@xmath372 ) ; at(@xmath373)0 ; at(@xmath374)1 ; at(@xmath375)2 ; at(@xmath376)3 ; at(@xmath377)4 ; at(@xmath378)5 ; at(@xmath379)6 ; ( r0)edge(r4)edge(r5)edge(r3)edge(r6 ) ; ( r1)edge(r2)edge(r3 ) ; ( r2)edge(r4)edge(r5)edge(r6 ) ; ( @xmath380 ) rectangle ( @xmath381 ) ; ( @xmath382 )
+ ( 0.2,-0.5 ) ; ( s0 ) at(@xmath383 ) ; ( s1 ) at(@xmath384 ) ; ( s2 ) at(@xmath385 ) ; ( s3 ) at(@xmath386 ) ; ( s4 ) at(@xmath387 ) ; ( s5 ) at(@xmath388 ) ; ( s6 ) at(@xmath389 ) ; at(@xmath390)0 ; at(@xmath391)1 ; at(@xmath392)2 ; at(@xmath393)3 ; at(@xmath394)4 ; at(@xmath395)5 ; at(@xmath396)6 ; ( s2)edge(s1)edge(s4)edge(s5)edge(s6 ) ; ( s0)edge(s4)edge(s5)edge(s6 ) ; ( s3)edge(s1)edge(s6 ) ; ( @xmath397 ) rectangle ( @xmath398 ) ; ( @xmath399 ) + ( 0,-0.5 ) ; ( t0 ) at(@xmath400 ) ; ( t1 ) at(@xmath401 ) ; ( t2 ) at(@xmath402 ) ; ( t3 ) at(@xmath403 ) ; ( t4 ) at(@xmath404 ) ; ( t5 ) at(@xmath405 ) ; ( t6 ) at(@xmath406 ) ; ( t7 ) at(@xmath407 ) ; at(@xmath408)0 ; at(@xmath409)1 ; at(@xmath410)2 ; at(@xmath411)3 ; at(@xmath412)4 ; at(@xmath413)5 ; at(@xmath414)6 ; at(@xmath415)7 ; ( t0)edge(t2)edge(t3)edge(t4)edge(t5)edge(t6)edge(t7 ) ; ( t1)edge(t2)edge(t3)edge(t4)edge(t5)edge(t6)edge(t7 ) ; ( @xmath416 ) rectangle ( @xmath417 ) ; ( @xmath418 ) + ( 0,-0.5 ) ; ( u0 ) at(@xmath419 ) ; ( u1 ) at(@xmath420 ) ; ( u2 ) at(@xmath421 ) ; ( u3 ) at(@xmath422 ) ; ( u4 ) at(@xmath423 ) ; ( u5 ) at(@xmath424 ) ; ( u6 ) at(@xmath425 ) ; ( u7 ) at(@xmath426 ) ; at(@xmath427)0 ; at(@xmath428)1 ; at(@xmath429)2 ; at(@xmath430)3 ; at(@xmath431)4 ; at(@xmath432)5 ; at(@xmath433)6 ; at(@xmath434)7 ; ( u0)edge(u6)edge(u7 ) ; ( u1)edge(u2)edge(u3)edge(u4)edge(u5 ) ; ( u7)edge(u2)edge(u3)edge(u4)edge(u5 ) ; ( u6)edge(u2 ) ; ( @xmath435 ) rectangle ( @xmath436 ) ; ( @xmath437 )
+ ( 0,-0.5 ) ; ( v0 ) at(@xmath438 ) ; ( v1 ) at(@xmath439 ) ; ( v2 ) at(@xmath440 ) ; ( v3 ) at(@xmath441 ) ; ( v4 ) at(@xmath442 ) ; ( v5 ) at(@xmath443 ) ; ( v6 ) at(@xmath444 ) ; ( v7 ) at(@xmath445 ) ; at(@xmath446)0 ; at(@xmath447)1 ; at(@xmath448)2 ; at(@xmath449)3 ; at(@xmath450)4 ; at(@xmath451)5 ; at(@xmath452)6 ; at(@xmath453)7 ; ( v0)edge(v5)edge(v6)edge(v7 ) ; ( v1)edge(v2)edge(v3)edge(v4 ) ; ( v7)edge(v2)edge(v3)edge(v4 ) ; ( v2)edge(v5)edge(v6 ) ; ( @xmath454 ) rectangle ( @xmath455 ) ; ( @xmath456 ) + ( 0,-0.5 ) ; ( w0 ) at(@xmath457 ) ; ( w1 ) at(@xmath458 ) ; ( w2 ) at(@xmath459 ) ; ( w3 ) at(@xmath460 ) ; ( w4 ) at(@xmath461 ) ; ( w5 ) at(@xmath462 ) ; ( w6 ) at(@xmath463 ) ; ( w7 ) at(@xmath464 ) ; at(@xmath465)0 ; at(@xmath466)1 ; at(@xmath467)2 ; at(@xmath468)3 ; at(@xmath469)4 ; at(@xmath470)5 ; at(@xmath471)6 ; at(@xmath472)7 ; ( w0)edge(w5)edge(w6 ) ; ( w1)edge(w2)edge(w3)edge(w4 ) ; ( w2)edge(w5 ) ; ( w4)edge(w6 ) ; ( w3)edge(w5)edge(w6 ) ; ( w7)edge(w0)edge(w2)edge(w4 ) ; ( @xmath473 ) rectangle ( @xmath474 ) ; ( @xmath475 )
+ ( 0.1,-0.5 ) ; ( x0 ) at(@xmath476 ) ; ( x1 ) at(@xmath477 ) ; ( x2 ) at(@xmath478 ) ; ( x3 ) at(@xmath479 ) ; ( x4 ) at(@xmath480 ) ; ( x5 ) at(@xmath481 ) ; ( x6 ) at(@xmath482 ) ; ( x7 ) at(@xmath483 ) ; at(@xmath484)0 ; at(@xmath485)1 ; at(@xmath486)2 ; at(@xmath487)3 ; at(@xmath488)4 ; at(@xmath489)5 ; at(@xmath490)6 ; at(@xmath491)7 ; ( x6)edge(x4)edge(x5)edge(x0 ) ; ( x1)edge(x2)edge(x3 ) ; ( x5)edge(x2)edge(x3 ) ; ( x4)edge(x2 ) ; ( x7)edge(x0)edge(x4 ) ; ( @xmath492 ) rectangle ( @xmath493 ) ; ( @xmath364 ) + ( -0.1,-0.5 ) ; ( z0 ) at(@xmath494 ) ; ( z1 ) at(@xmath495 ) ; ( z2 ) at(@xmath496 ) ; ( z3 ) at(@xmath497 ) ; ( z4 ) at(@xmath498 ) ; ( z5 ) at(@xmath499 ) ; ( z6 ) at(@xmath500 ) ; ( z7 ) at(@xmath501 ) ; at(@xmath502)0 ; at(@xmath503)1 ; at(@xmath504)2 ; at(@xmath505)3 ; at(@xmath506)4 ; at(@xmath507)5 ; at(@xmath508)6 ; at(@xmath509)7 ; ( z6)edge(z4)edge(z5)edge(z0 ) ; ( z1)edge(z2)edge(z3 ) ; ( z5)edge(z2)edge(z3)edge(z6)edge(z7 ) ; ( z4)edge(z2 ) ; ( z7)edge(z0 ) ; ( @xmath510 ) rectangle ( @xmath511 ) ; ( @xmath512 )
+ ( 0,-0.5 ) ; ( y0 ) at(@xmath513 ) ; ( y1 ) at(@xmath514 ) ; ( y2 ) at(@xmath515 ) ; ( y3 ) at(@xmath516 ) ; ( y4 ) at(@xmath517 ) ; ( y5 ) at(@xmath518 ) ; ( y6 )
at(@xmath519 ) ; ( y7 ) at(@xmath520 ) ; at(@xmath521)0 ; at(@xmath522)1 ; at(@xmath523)2 ; at(@xmath524)3 ; at(@xmath525)4 ; at(@xmath526)5 ; at(@xmath527)6 ; at(@xmath528)7 ; ( y2)edge(y1)edge(y4)edge(y5)edge(y6)edge(y7 ) ; ( y0)edge(y4)edge(y5)edge(y6)edge(y7 ) ; ( y3)edge(y1)edge(y7 ) ; ( @xmath529 ) rectangle ( @xmath530 ) ; .number of lattices and ( vertically indecomposable = v.i . ) ( semi)modular lattices up to isomorphism .
new numbers are in bold . [ cols= " > , > , > , > , > , > " , ]
let @xmath531 denote the number of modular lattices of size @xmath1 ( up to isomorphism ) . in this section
we give a simple argument for a lower bound of this sequence .
let @xmath533 be the three element lattice with 0 and 1 as bottom and top respectively .
consider the following two extensions of an @xmath1-lattice @xmath8 : @xmath534 @xmath535 we declare an @xmath1-lattice @xmath8 to be an @xmath536-lattice or a @xmath537-lattice if it is obtained through the @xmath536 or @xmath537 construction respectively . for @xmath539
, we consider the cases where @xmath8 is a modular @xmath536-lattice and a modular @xmath537-lattice obtained through this construction .
if @xmath8 is an @xmath536-lattice , then there are two atoms @xmath1 and @xmath102 in @xmath539 , both of which are covered by @xmath12 , independently of the choice of @xmath12 .
therefore , @xmath1 and @xmath102 satisfy ( lower ) semimodularity .
since the new element @xmath1 is not the common cover or co - cover of any two elements in @xmath539 and @xmath8 is modular , it follows that @xmath539 is modular .
if @xmath8 is a @xmath537-lattice , then there is only one choice of @xmath12 ( the element used in the previous step ) since there is only one cover for @xmath102 .
notice that the first @xmath537 step starting from an @xmath536-lattice ( or @xmath533 ) will generate an @xmath540 sublattice formed by the bottom element @xmath28 , the two atoms @xmath1 and @xmath102 and the cover @xmath12 used .
after @xmath49 successive @xmath537 steps , there will be an @xmath541 sublattice formed by @xmath28 , @xmath12 , and the atoms @xmath542, ,@xmath1 .
therefore , after any @xmath537 step , the new element will share a common cover ( @xmath12 ) with all the atoms it shares a common co - cover ( @xmath28 ) with , and vice versa .
since all other elements of @xmath539 satisfy ( lower ) semimodularity by modularity of @xmath8 , it follows that @xmath539 is modular .
next , we want to show that @xmath538 and @xmath539 are not isomorphic to any other lattice obtained from this construction .
consider two modular lattices @xmath8 and @xmath37 generated by the @xmath536-@xmath537 construction . since @xmath538 has a unique atom and @xmath543 has more than one atom , it follows that @xmath544 .
if @xmath547 , there is more than one choice of atom .
let @xmath548 be the number of atoms in both , then they both have been obtained through @xmath49 @xmath537 steps , and the @xmath49 atoms added through @xmath537 steps are automorphic
. therefore removal of any of these atoms in @xmath539 and @xmath543 will generate two isomorphic modular lattices , hence @xmath549 .
consequently no two non - isomorphic lattices can generate isomorphic lattices through the @xmath536-@xmath537 construction .
we conclude the proof by induction . for @xmath550 , there are @xmath551 modular lattices in the @xmath552 construction ( the initial @xmath533 lattice ) .
assume there are @xmath0 non - isomorphic modular lattices of size @xmath1 constructed via the @xmath536-@xmath537 construction .
then each of these lattices will produce 2 new modular lattices which are not isomorphic to any of the lattices produced by any other non - isomorphic lattice . as a result , there are @xmath553 non - isomorphic modular lattice of size @xmath32 , thus completing the induction .
( a0 ) at(0,0)0 ; ( a2 ) at(@xmath142)2 ; ( a1 ) at(@xmath554)1 ; ( a2)edge(a1)edge(a0 ) ; ( @xmath555 ) + ( 0.6,0 ) ; ( b0 ) at(@xmath556)0 ; ( b2 ) at(@xmath557)2 ; ( b3 ) at(@xmath558)3 ; ( b1 ) at(@xmath149)1 ; at(@xmath559)@xmath537 ; ( b2)edge(b1)edge(b0 ) ; ( b3)edge(b1)edge(b0 ) ; ( @xmath560 ) + ( 0.6,0 ) ; ( c0 ) at(@xmath561)0 ; ( c2 ) at(@xmath562)2 ; ( c3 ) at(@xmath563)3 ; ( c4 ) at(@xmath564)4 ; ( c1 ) at(@xmath565)1 ; at(@xmath566)@xmath536 ; ( c2)edge(c1)edge(c4 ) ; ( c3)edge(c1)edge(c4 ) ; ( c4)edge(c0 ) ; ( @xmath567 ) + ( 0.6,0 ) ; ( d0 ) at(@xmath568)0 ; ( d2 ) at(@xmath569)2 ; ( d3 ) at(@xmath168)3 ; ( d4 ) at(@xmath171)4 ; ( d5 ) at(@xmath169)5 ; ( d1 ) at(@xmath570)1 ; at(@xmath160)@xmath537 ; ( d2)edge(d1)edge(d4 ) ; ( d3)edge(d1)edge(d4 ) ; ( d4)edge(d0 ) ; ( d5)edge(d0)edge(d3 ) ; ( @xmath571 ) + ( 0.6,0 ) ; ( e0 ) at(@xmath572)0 ; ( e2 ) at(@xmath573)2 ; ( e3 ) at(@xmath181)3 ; ( e4 ) at(@xmath183)4 ; ( e5 ) at(@xmath574)6 ; ( e6 ) at(@xmath575)5 ; ( e1 ) at(@xmath576)1 ; at(@xmath577)@xmath537 ; ( e2)edge(e1)edge(e4 ) ; ( e3)edge(e1)edge(e4 ) ; ( e4)edge(e0 ) ; ( e5)edge(e0)edge(e3 ) ; ( e6)edge(e0)edge(e3 ) ; ( @xmath578 ) + ( 0.3,0 ) ; at(@xmath579 ) ; ( @xmath580 )
+ ( 0.3,0 ) ; ( f0 ) at(@xmath581)0 ; ( f2 ) at(@xmath582)2 ; ( f3 ) at(@xmath583)3 ; ( f4 ) at(@xmath584)4 ;
( f5 ) at(@xmath585 ) ; ( f6 ) at(@xmath586)5 ;
( f7 ) at(@xmath587 ) ; ( f8 ) at(@xmath588 ) ; ( f1 ) at(@xmath589)1 ; at(@xmath590)@xmath537 ; at(@xmath591)@xmath537 ; ( f2)edge(f1)edge(f4 ) ; ( f3)edge(f1)edge(f4 ) ; ( f4)edge(f0 ) ; ( f6)edge(f0)edge(f3 ) ; let @xmath592 and @xmath593 be the number of ( nonisomorphic ) lattices and distributive lattices of size @xmath1 . lower and upper bounds for these sequences are given in @xcite and @xcite respectively : @xmath594 @xmath595 the lower bound for modular lattices obtained in the preceding theorem can be improved slightly for @xmath596 by counting a larger class of planar modular lattices .
however , it seems that currently the best known upper bound for ( semi)modular lattices is the same as the one for all lattices . as suggested by one of the referees , we conclude with some observations about possible future research .
there are several alternative representations for ( finite ) modular lattices , based on partial order geometries ( see e.g. @xcite ) or join - covers or the incidence of join and meet irreducibles .
it is possible that enumeration algorithms using these representations would be more efficient , but this has not ( yet ) been explored .
the algorithm we use can also be fairly easily adapted to other classes of lattices , such as 2-distributive lattices or lattices of breadth @xmath597 , either with or without adding modularity .
however , any enumeration algorithm similar to the one presented here that builds lattices by adding elements one - by - one can not build only modular lattices if it is supposed to build all modular lattices ( this can be seen for example by removing any element from the subspace lattice of a finite projective plane , see figure [ fano ] ) . ( -3,1.5)(-2,3)(2,1.5)(3,3)(0,1.5)(1,3)(-2,1.5)(-1,3)(3,1.5)(-3,3)(1,1.5)(2,3)(-1,1.5)(0,3)(-3,1.5 ) ; ( 0,0)node(-3,1.5)node(-3,3)node(0,4.5)node ( -2,3)node(-2,1.5)node(0,0)(-1,1.5)node(-1,3)node(0,4.5) ( 0,3)node(0,1.5)node(0,0)(1,1.5)node(1,3)node(0,4.5) ( 2,3)node(2,1.5)node(0,0)(3,1.5)node(3,3)node(0,4.5 ) ; we would like to thank michael fahy and nikos hatzopoulus for help with operating the computing clusters .
we also thank the summer undergraduate research fellowship at chapman university for financial support .
r. belohlavek and v. vychodil : residuated lattices of size @xmath598 12 , order 27 ( 2010 ) , 147161 .
d. j. benson and j. h. conway : diagrams for modular lattices , j. pure and applied algebra 37 ( 1985 ) 111116 .
r. dedekind : ber die von drei moduln erzeugte dualgruppe , math .
ann . 53 ( 1900 ) , 371403 .
m. ern , j. heitzig and j. reinhold : on the number of distributive lattices , electron .
j. combin . 9 ( 2002 ) , 23 pp .
u. faigle and c. herrmann : projective geometry on partially ordered sets , transactions a.m.s .
266 ( 1981 ) , 267291 .
i. a. faradzhev , constructive enumeration of combinatorial objects , internat .
cnrs no.260 , combinatoire et theorie des graphes , paris ( 1976 ) , 131135 .
g. grtzer , `` general lattice theory '' , 2nd edition , birkhuser , 1998 .
j. heitzig and j. reinhold : counting finite lattices , algebra univers .
48 ( 2002 ) 4353 .
d. j. kleitman and k. j. winston : the asymptotic number of lattices , annals of discrete math . 6
( 1980 ) 243249 .
b. d. mckay : isomorph - free exhaustive generation , j. algorithms 26 ( 1998 ) 306324 .
b. d. mckay and a. piperno : practical graph isomorphism , ii , http://arxiv.org/abs/1301.1493 , ( 2013 ) , 22 pp .
r. c. read : every one a winner , or how to avoid isomorphism search when cataloguing combinatorial configurations , annals of discrete math . 2 ( 1978 ) 107120 . | modular lattices , introduced by r. dedekind , are an important subvariety of lattices that includes all distributive lattices .
heitzig and reinhold @xcite developed an algorithm to enumerate , up to isomorphism , all finite lattices up to size 18 . here
we adapt and improve this algorithm to construct and count modular lattices up to size 24 , semimodular lattices up to size 22 , and lattices of size 19 .
we also show that @xmath0 is a lower bound for the number of nonisomorphic modular lattices of size @xmath1 . |
some random phenomena occur at discrete times or locations , with the individual events largely identical , such as a sequence of neural action potentials .
a stochastic point process @xcite is a mathematical construction which represents these events as random points in a space .
fractal stochastic point processes exhibit scaling in all of the statistics considered in this paper ; fractal - rate stochastic point processes do so in some of them . in this work
we consider the simulation and estimation of fractal and fractal - rate stochastic point processes on a line , which model a variety of observed phenomena in the physical and biological sciences .
this work provides an extension and generalization of an earlier paper along these lines @xcite .
figure [ basics ] shows several representations that are useful in the analysis of point processes .
figure [ basics](a ) demonstrates a sample function of a point process as a series of impulses occurring at specified times @xmath1 . since these impulses have vanishing width , they are most rigorously defined as the derivative of a well - defined counting process @xmath2 [ fig .
[ basics](b ) ] , a monotonically increasing function of @xmath3 , that augments by unity when an event occurs . accordingly
, the point process itself is properly written as @xmath4 , since it is only strictly defined within the context of an integral .
the point process is completely described by the set of event times @xmath5 , or equivalently by the set of interevent intervals .
however , the sequence of counts depicted in fig . [ basics](c ) also contains much information about the process . here the time axis is divided into equally spaced contiguous counting windows of duration @xmath6 sec to produce a sequence of counts @xmath7 , where @xmath8 - n[kt]$ ] denotes the number of events in the @xmath9th window . as illustrated in fig .
[ basics](d ) , this sequence forms a discrete - time random process of nonnegative integers .
in general , information is lost in forming the sequence of counts , although for a regular point process the amount lost can be made arbitrarily small by reducing the size of the counting window @xmath6 .
an attractive feature of this representation is that it preserves the correspondence between the discrete time axis of the counting process @xmath7 and the absolute `` real '' time axis of the underlying point process . within the process of counts @xmath7 , the elements @xmath10 and
@xmath11 refer to the number of counts in windows separated by precisely @xmath12 sec , so that correlation in the process @xmath7 is readily associated with correlation in the underlying point process @xmath4 .
the complete characterization of a stochastic process involves a description of all possible joint probabilities of the various events occurring in the process .
different statistics provide complementary views of the process ; no single statistic can in general describe a stochastic process completely .
fractal stochastic processes exhibit scaling in their statistics .
such scaling leads naturally to power - law behavior , as demonstrated in the following . consider a statistic @xmath13 which depends continuously on the scale @xmath14 over which measurements are taken .
suppose changing the scale by any factor @xmath15 effectively scales the statistic by some other factor @xmath16 , related to the factor but independent of the original scale : @xmath17 the only nontrivial solution of this scaling equation for real functions and arguments , that is independent of @xmath15 and @xmath14 is @xmath18 for some constants @xmath19 and @xmath20 @xcite .
thus statistics with power - law forms are closely related with the concept of a fractal .
the particular case of fixed @xmath15 admits a more general solution @xcite : @xmath21 \ , .\ ] ] consider , for example , a commonly encountered first - order statistic for a stochastic point process , the interevent interval histogram ( iih ) .
this estimates the interevent - interval probability density function ( iipdf ) @xmath22 by computing the relative frequency of occurrence of interevent intervals as a function of interval size .
this measure highlights the behavior of the times between adjacent events , but reveals none of the information contained in the relationships among these times , such as correlation between adjacent time intervals . for a fractal stochastic point process ( fspp )
the iipdf follows the form of eq .
( [ scale0 ] ) , so that @xmath23 over a certain range of @xmath3 , where @xmath24 .
a number of statistics may be used to describe an fspp , and each statistic which scales will in general have a different scaling exponent @xmath20 .
each of these exponents can be simply related to a more general parameter @xmath0 , the fractal exponent , where the exact relation between these two exponents will depend upon the statistic in question .
for example , the exponent @xmath20 of the iipdf defined above is related to the fractal exponent @xmath0 by @xmath25 .
sample functions of the fractal renewal point process ( frp ; see sec .
[ frp_def ] ) are true fractals ; the expected value of their generalized dimensions ( see sec .
[ othermeas ] ) assumes a nonintegral value between the topological dimension ( zero ) and the euclidean dimension ( unity ) @xcite . however , the sequence of unitary events observed in many biological and physical systems do not exhibit power - law - distributed iihs but nevertheless exhibit scaling in other statistics .
these processes therefore have integral generalized dimensions ( see sec . [ othermeas ] ) and are consequently not true fractals .
they are nevertheless endowed with rate functions that are indeed either fractals or their increments : fractal gaussian noise , fractal brownian motion , or closely related processes .
therefore , such point processes are more properly termed fractal-_rate _ stochastic point processes ( frspps ) .
the fractal exponent @xmath0 defined above is related to the more commonly encountered hurst exponent @xmath26 @xcite .
the relationship is ambiguous , however , since some authors @xcite use the formula @xmath27 for all values of @xmath0 , while others @xcite use @xmath28 for @xmath29 to restrict @xmath26 to the range @xmath30 . in this paper
, we avoid this confusion by considering @xmath0 directly instead of @xmath26 .
many phenomena are readily represented by fspps and frspps .
we provide several examples drawn from the physical @xcite and biological @xcite sciences .
a multiple trapping model has been used to show how traps that are exponentially distributed over a large range of energies lead to a power - law decay of current in an amorphous semiconductor @xcite .
if a pulse of light strikes such a semiconductor , the many carriers excited out of their traps will be available to carry current until they are recaptured , which happens relatively quickly . at some point each carrier will be released from its trap by thermal excitation and become mobile for a time , and then be recaptured by another trap . for exponentially distributed energy states with identical capture cross sections , the electrons tend to be trapped in shallow states at first , but the probability of being caught in a deep trap increases as time progresses .
this leads to a current that decreases as a power - law function of time .
the multiple trapping model may be usefully recast in terms of a standard fractal renewal process ( see sec .
[ frp_def ] ) @xcite .
consider an amorphous semiconductor with localized states ( traps ) with energies which are exponentially distributed with parameter @xmath31 between a minimum energy @xmath32 of the order of @xmath33 , where @xmath34 is boltzmann s constant and @xmath35 is the temperature in degrees kelvin ; and a maximum energy @xmath36 determined by the bandgap of the material .
for a particular trap with energy @xmath37 , the corresponding mean waiting time to release is @xmath38 where @xmath39 is the average vibrational period of the atoms in the semiconductor .
if we define characteristic time cutoffs @xmath40 and @xmath41 , and the power - law exponent @xmath42 , then the mean waiting time @xmath43 has a density which decays as a power law with this exponent between those two cutoffs .
each trap holds carriers for times that are exponentially distributed given the conditional parameter @xmath43 , and averaging this exponential density over all possible values of @xmath43 yields the unconditional trapping time density , which is itself approximately power law : @xmath44 for @xmath45 .
thus each carrier will be trapped for a period that is essentially power - law distributed . upon escaping from a trap ,
the carrier can conduct current for a short time until it is again captured by another trap .
thus each carrier executes a series of current - carrying jumps well described by a standard frp . assuming that each carrier acts independently of the others , the action of the carriers as a whole can be modeled as the superposition of a collection of such component processes , which converges to the fractal - gaussian - noise driven poisson process ( fgndp ; see secs . [ fgn ] and [ fdspp ] ) in the limit of a large number of carriers .
again , both experimental @xcite and theoretical @xcite results point to a power - law or fractal decay in the power spectral density , while several other statistics also show scaling behavior .
burst noise occurs in many communications systems and is characterized by relatively brief noise events which cluster together , separated by relatively longer periods of quiet .
berger and mandelbrot @xcite long ago showed that burst errors in communication systems are well modeled by a version of a fractal renewal process ( see sec .
[ frp_def ] ) , and in particular that the interevent times were essentially independent of each other for time scales determined by the resolution and the duration of the observation .
moreover , the rate of traffic flow itself displays fractal fluctuations on a variety of high - speed packet - switching networks conducting different types of traffic @xcite .
this has been demonstrated for time scales greater than about one sec in both the power spectral density and the fano factor ( see secs .
[ psdest ] and [ ffest ] ) . over these time scales , the fractal - shot - noise driven poisson process ( fsndp ; see secs
. [ fsn ] and [ fdspp ] ) has been successfully employed to model the traffic @xcite .
ion channels reside in biological cell membranes , permitting ions to diffuse in or out @xcite .
these channels are usually specific to a particular ion , or group of related ions , and block the passage of other kinds of ions .
further , most channels have gates , and thus the channels may be either open or closed . in many instances ,
intermediate conduction states are not observed .
some ion channels may be modeled by a two - state markov process @xcite , with one state representing the open channel , and the other representing the closed channel .
this model generates exponentially distributed dwell times in both states , which are , in fact , sometimes observed .
however , many ion channels exhibit independent power - law distributed closed times between open times of negligible duration @xcite , and are well described by a fractal renewal point process ( see sec . [ frp_def ] )
@xcite .
communication in the nervous system is usually mediated by the exocytosis of multiple vesicular packets ( quanta ) of neurotransmitter molecules at the synapse between two cells , triggered by an action potential ( information - bearing signal ) at the presynaptic cell @xcite . even in the absence of an action potential ,
however , many neurons spontaneously release individual packets of neurotransmitter @xcite .
on arrival at the postsynaptic membrane , each neurotransmitter packet induces a miniature endplate current ( mepc ) ; superpositions of mepc - like events comprise the postsynaptic endplate currents elicited by nerve impulses arriving at the presynaptic cell @xcite .
analysis of the sequence of mepc and mepc - like events in a variety of preparations reveals the presence of memory over a range of times and frequencies @xcite .
the fractal - lognormal - noise driven poisson process model ( flndp ; see secs .
[ fln ] and [ fdspp ] ) provides an excellent model for this process , for a variety of statistical measures @xcite .
many biological neurons carry information in the form of sequences of action potentials , which are localized regions of depolarization traveling down the length of an axon .
action potentials or neural spikes are brief and largely identical events and are well represented by a point process .
the firing patterns of an isolated neuron and an isolated axons have been shown to have fractal characteristics .
musha and colleagues generated a synthetic spike train consisting of electrical impulses separated by independent intervals drawn from the same gaussian distribution @xcite .
this nonfractal stimulus as applied to one end of an excised squid giant axon ; the membrane voltage at the far end was considered to be the response spike train . the resulting power spectral density followed a @xmath46-type decay .
spontaneous firing from an excised giant snail neuron yielded similar results @xcite , although for this preparation the data was selected to reduce fluctuations , and analyzed by interevent number rather than in real time .
more recently , fractal behavior has been observed in the sequence of action potentials recorded from several _ in vivo _ vertebrate preparations .
real - time recordings were made under both spontaneous and driven firing conditions . over short time
scales , nonfractal stochastic point processes prove adequate for representing such nerve spikes , but over long time scales ( typically greater than one sec ) the fractal nature becomes evident @xcite .
estimators of the rate of the process converge more slowly than for nonfractal processes , displaying fluctuations which decrease as a power - law function of the time used to estimate the rate @xcite . with the inclusion of the refractory effects of nerve fibers , frspp models have been shown to properly characterize the statistical properties of certain potentials in peripheral primary auditory fibers in several species , over all time scales and for a broad variety of statistical measures @xcite ; only four parameters are required .
this process possibly arises from superpositions of fractal ion - channel transitions @xcite or by fractal vesicular exocytosis in inner - ear sensory cells , as described briefly in secs .
[ ionch ] and [ exocytosis ] , respectively . many neurons in the mammalian visual system
transmit information by means of action potentials as well , and frspps also provide suitable models for describing the behavior of these neurons @xcite .
the gamma renewal process , which is nonfractal , has proven to be a useful model for some of these processes over short time scales @xcite . however , nerve spike trains recorded from both cat retinal - ganglion - cell and lateral - geniculate - nucleus neurons , like those recorded from primary auditory neurons , exhibit fractal behavior over time scales greater than about @xmath47 sec , thus necessitating the use of an frspp description @xcite .
long - duration temporal correlation has also been observed in spike trains from cat striate - cortex neurons @xcite and from an insect visual interneuron @xcite .
action - potential data collected from rabbit somatosensory cortical neurons and cat respiratory control neurons show experimental iih plots which decay as power laws @xcite .
in addition , for the data lengths considered , serial correlation coefficients among interevent intervals from these data sets do not differ from zero in a statistically significant way .
( however , differences may well emerge for longer data sets , as has been seen , for example , in auditory - nerve - fiber recordings @xcite . ) these two interevent - interval characteristics independence and a power - law distribution define the fractal renewal point process ( see sec .
[ frp_def ] ) , which is an fspp . on the other hand , action potentials in the cat mesencephalic reticular formation ( mrf ) exhibit activity that can be modeled by an frspp .
during rem phases of sleep , these neurons exhibit @xmath46-type spectra @xcite , as do certain hippocampal and thalamic neurons @xcite .
the cluster poisson point process ( see sec . [ cluster ] ) has been used successfully to model mrf neural activity @xcite .
furthermore , there is some evidence that these neurons have iihs whose tails decay in power - law fashion @xcite , so that a type of fspp may prove suitable for describing these spike trains .
the sequence of human heartbeats exhibits considerable variability over time , both in the short - term and in the long - term patterns of the times between beats @xcite .
a point process of the heartbeats is formed by focusing on the times between maximum contractions ( r - r intervals ) .
a particular frspp with an integrate - and - fire substrate has been constructed and shown to successfully describe these events @xcite . as with auditory and visual - system spike trains , over short time scales nonfractal point processes provide suitable models for the pattern of times between contractions ; for times longer than roughly @xmath48 sec , only fractal models suffice .
interestingly , parameters of the frspp used to model the data indicate promise for the diagnosis of various disease states of the heart @xcite .
as the examples described above illustrate , many phenomena are amenable to modeling by frspps .
the value of the fractal exponent @xmath0 can often provide important information regarding the nature of an underlying process , and can also serve as a useful classification tool in some cases , as with human heartbeat data as mentioned in section [ heartbeat ] .
accordingly , it is desirable to estimate @xmath0 reliably , although this task is often confounded by a variety of issues @xcite .
here we briefly review some of the techniques used for estimating @xmath0 . by way of illustration
, we apply these methods to a train of action potentials recorded from a lateral - geniculate - nucleus ( lgn ) relay neuron in the cat visual system .
there are 24285 events in this particular spike train , with an average interevent interval of 0.133 sec , comprising a total duration of 3225 sec @xcite .
the interevent - interval histogram ( iih ) , an estimate of the interval probability function , and the generalized dimension ( gd ) @xcite , are useful measures for fractal stochastic point processes , and in particular for the fractal renewal process ( see sec .
[ frp_def ] ) .
the iih is simply the relative frequency of interevent - interval occurrence in the data set , ignoring all correlations among the intervals . since realizations of fspps form true fractals , all measures considered in this paper , including the iih , exhibit scaling for these processes .
thus the iih proves useful in elucidating the fractal nature of fspps . for frspps ,
on the other hand , the fractal structure is not manifested in first - order interevent - time statistics , so that the iih does not scale over a significant range of interevent times .
rather , the fractal character of these processes lie in the dependencies among the interevent intervals , which are captured by the measures considered in secs .
[ crsec][afest ] .
the generalized dimension is of interest for fractal stochastic point processes . if a data segment of length @xmath49 is divided into intervals of length @xmath6 , with @xmath50 representing the number of points in the @xmath9th interval ( see fig .
1(c ) ) , then the generalized dimension @xmath51 of a point process is defined as @xmath52 where the sum extends over all non - empty intervals .
particular cases are the capacity or box - counting dimension @xmath53 , the information dimension @xmath54 , and the correlation dimension @xmath55 .
the sum is a form of time averaging ; for a stochastic point process , it is convenient to replace the sum by the product of @xmath56 and the expected value of @xmath57 . for a fractal - rate point process , however , analytical values of the @xmath51 will not in general equal analytical values of the dimensions obtained from the pg or af , and in fact the @xmath51 will often assume integer values ( see sec .
[ frsppdef ] ) .
the cr measures the correlations between pairs of events with a specified time delay between them , regardless of intervening events , and is related to the autocorrelation function used with continuous processes .
the cr is defined as @xmath58 \mbox { and } { \cal e}[\tau,\tau+\delta ] \right\ } } { \delta^{2}},\ ] ] where @xmath59 $ ] denotes the occurrence of at least one event of the point process in the interval @xmath60 and @xmath43 is the delay time . for an ideal fractal or fractal - rate stochastic point process with a fractal exponent @xmath0 in the range @xmath61 , the coincidence rate assumes the form @xcite @xmath62 , \label{fractalcr}\ ] ] where @xmath63 is the mean rate of the process , @xmath64 denotes the dirac delta function , and @xmath65 is a constant representing the fractal onset time .
a stationary , regular point process with a cr following this form for _ all _ delay times @xmath43 exhibits infinite mean .
further , statistics of fractal data sets collected from experiments exhibit scaling only over a finite range of times and frequencies , as determined by the resolution of the experimental apparatus and the duration of the experiment . nevertheless , in much of the following we employ eq .
( [ fractalcr ] ) without cutoffs since we find that the cutoffs do not significantly affect the mathematical results in many cases .
we employ similar ideal forms for other second - order measures defined later in this paper for the same reasons .
the coincidence rate can be directly estimated from its definition . however , in practice the cr is a noisy measure , since its definition essentially involves a double derivative .
furthermore , for fspps and frspps typical of physical and biological systems , the cr exceeds its asymptotic value @xmath66 by only a small fraction at any large but practical value of @xmath43 , so that determining the fractal exponent with this small excess presents serious difficulties .
therefore we do not specifically apply this measure to the lgn data , although the formal definition of coincidence rate plays a useful role in developing other , more reliable measures .
the power spectral density ( psd ) is a familiar and well - established measure for continuous - time processes . for point
processes the psd and the cr introduced above form a fourier transform pair , much like the psd and the autocorrelation function do for continuous - time processes .
the psd provides a measure of how the power in a process is concentrated in various frequency bands . for an ideal fractal point process
, the psd assumes the form @xmath67 , \label{fractalpsd}\ ] ] for relevant time and frequency ranges , where @xmath63 is the mean rate of events and @xmath68 is a cutoff radian frequency . the psd of a point process can be estimated using the periodogram ( pg ) @xmath69 of the sequence of counts , rather than from the point process itself @xcite .
this method introduces a bias at higher frequencies , since the fine time resolution information is lost as a result of the minimum count - window size . nevertheless ,
since estimating the fractal exponent principally involves lower frequencies where this bias is negligible , and employing the sequence of counts permits the use of vastly more efficient fast fourier transform methods , we prefer this technique . alternate definitions of the psd for point processes ( and thus for the pg used to estimate them ) exist ; for example , a different psd may be obtained from the real - valued discrete - time sequence of the interevent intervals .
however , features in this psd can not be interpreted in terms of temporal frequency @xcite .
figure [ vispp ] displays the pg for the visual system lgn data calculated using the count - based approach . throughout the text of this paper
we employ radian frequency @xmath70 ( radians per unit time ) to simplify the analysis , while figures are plotted in common frequency @xmath71 ( cycles per unit time ) in accordance with common usage . for low frequencies ,
the pg decays as @xmath72 , as expected for a fractal point process . fitting a straight line ( shown as dotted ) to the doubly logarithmic plot of the pg , over the range from 0.002 hz to 1 hz , provides an estimate @xmath73 .
another measure of correlation over different time scales is provided by the fano factor ( ff ) , which is the variance of the number of events in a specified counting time @xmath6 divided by the mean number of events in that counting time .
this measure is sometimes called the index of dispersion of counts @xcite . in terms of the sequence of counts illustrated in fig .
[ basics](c ) , the fano factor is simply the variance of @xmath7 divided by the mean of @xmath7 , i.e. , @xmath74 -{\mbox{\rm e}}^{2}[z_{k}]}{{\mbox{\rm e}}[z_{k}]},\ ] ] where @xmath75 $ ] represents the expectation operation . the ff may be obtained from the coincidence rate by an integral transform @xmath76\ , d\tau,\ ] ] where @xmath63 is the average rate of events of the point process , and the lower limit of the integral approaches zero from the positive side @xcite .
the ff varies as a function of counting time @xmath6 .
the exception is the homogeneous poisson point process ( hpp ) , which is important as a benchmark in point - process theory just as the gaussian is in the theory of continuous stochastic processes . for an hpp ,
the variance - to - mean ratio is always unity for any counting time @xmath6 .
any deviation from unity in the value of @xmath77 therefore indicates that the point process in question is not poisson in nature .
an excess above unity reveals that a sequence is less ordered than an hpp , while values below unity signify sequences which are more ordered . for an ideal fspp or frspp with @xmath78 ,
the ff assumes the form @xmath79 where @xmath80 is a fractal onset time that differs from @xmath39 in eq .
( [ fractalcr ] ) [ see eq .
( [ relati ] ) ] .
therefore a straight - line fit to an estimate of @xmath77 vs. @xmath6 on a doubly logarithmic plot can also be used to estimate the fractal exponent . however , the estimated slope of the ff saturates at unity , so that this measure finds its principal applicability for processes with @xmath29 @xcite .
figure [ visrff ] shows the estimated ff curve for the same data set as shown in fig .
[ vispp ] . for counting times
@xmath6 greater than approximately 0.3 sec , this curve behaves essentially as @xmath81 .
the estimated value is @xmath82 ( dotted line ) , closely agreeing with the value obtained from the pg .
fair agreement between these two measures exists in other sensory spike - train and heartbeat sequence data sets @xcite and in simulated frspps @xcite .
the allan variance , as opposed to the ordinary variance , is defined in terms of the variability of _ successive _ counts @xcite . in analogy with the fano factor ( ff )
, we define the allan factor ( af ) for the sequence of counts shown in fig . [ basics ] as @xmath83}{2{\mbox{\rm e}}[z_{k}]}. \label{afdef}\ ] ] in terms of quantities defined earlier
, we have @xmath84\ , d\tau . \label{afcalc}\ ] ] as for the ff , the value of the allan factor for the hpp is unity .
for an ideal frspp with @xmath85 , the af also assumes a power - law - varying form @xmath86 where @xmath87 is another fractal onset time , different from @xmath39 and @xmath80 [ see eq .
( [ relati ] ) ] ; therefore a straight line fit of an estimate of @xmath88 vs. @xmath6 on a doubly logarithmic plot yields yet another estimate of the fractal exponent .
figure [ visaff ] shows the estimated allan factor curve for the same data set as shown in figs .
[ vispp ] and [ visrff ] . from visual inspection
, this measure appears to be considerably `` rougher '' than the ff , which is typical of the data sets we have analyzed .
nevertheless it is clear that for counting times @xmath6 greater than approximately 0.3 sec , its behavior can also be approximated as @xmath81 . to estimate the value of @xmath0 , a straight line fit to the doubly logarithmic plot of the estimate of @xmath88 vs. counting time
@xmath6 was provided .
the value of @xmath89 obtained agrees well with the values calculated using the pg and ff .
excellent agreement between the af and the pg exists in other sensory spike - train and heartbeat sequence data sets @xcite and simulations @xcite .
furthermore , the af appears to yield superior estimates of @xmath0 than the ff in all cases @xcite . in particular , instead of saturating at a power - law exponent of unity , estimates of @xmath0 obtained from the af can range up to a value of three @xcite .
thus we do not employ the ff in the remaining sections of this paper .
the allan variance , @xmath90 $ ] may be recast as the variance of the integral of the point process under study multiplied by the following function : @xmath91 equation ( [ haar ] ) defines a scaled wavelet function , specifically the haar wavelet .
this can be generalized to any admissible wavelet @xmath92 ; when suitably normalized , the result is a wavelet allan factor ( waf ) @xcite .
this generalization enables an increased range of fractal exponents to be estimated , at the cost of reducing the range of times over which the waf varies as @xmath93 .
in particular , for a particular wavelet with regularity ( number of vanishing moments ) @xmath94 , fractal exponents @xmath95 can be reliably estimated @xcite .
for the haar basis , @xmath96 , while all other wavelet bases have @xmath97 .
thus the waf employing bases other than the haar should prove useful in connection with frspps for which @xmath98 ; for processes with @xmath99 , however , the af appears to be the best choice @xcite .
for an ideal fspp or frspp with @xmath78 , any one of eqs .
( [ fractalcr ] ) , ( [ fractalpsd ] ) , ( [ fractalff ] ) , and ( [ fractalaf ] ) implies the other three , with @xcite @xmath100 for larger values of @xmath0 ,
the ff and cr do not scale , although the psd and af do ; thus eqs .
( [ fractalpsd ] ) , and ( [ fractalaf ] ) imply each other for these frspps .
therefore , over the range of @xmath0 for which the af exhibits scaling @xcite @xmath101 in principle , any of the these statistics could be solved for @xmath0 when it is known that @xmath29 , although best results obtain from the pg and the af . since the gd of a fractal can not exceed the euclidean dimension , which assumes a value of unity for one - dimensional point processes
, fspps can not exist for @xmath102 ; only frspps are possible at these values of the fractal exponent .
in previous work we defined an fspp and several frspps and derived a number of their statistics @xcite .
we include brief summaries of these here , as well as of the clustered poisson point - process model of grneis and colleagues @xcite .
we also introduce two new fractal - rate processes ( fractal lognormal noise and fractal exponential noise ) and define two new methods for generating a point processes from a rate function : integrate - and - fire and controlled variability integrate - and - fire .
these prove useful in isolating the effects of fractal components from those of nonfractal noise introduced through the mechanics of generating the point process .
for all of the processes considered , the scaling relation in eq .
( [ scale0 ] ) holds for a number of statistical measures .
the one - dimensional homogeneous poisson point process ( hpp ) is perhaps the simplest nonfractal stochastic point process @xcite .
the hpp is characterized by a single constant quantity , its rate , which is the number of events expected to occur in a unit interval .
a fundamental property of the hpp is that it is memoryless ; given this rate , knowledge of the entire history and future of a given realization of a hpp yields no information about the behavior of the process at the present .
the hpp belongs to the class of renewal point processes ; times between events are independent and identically distributed .
the hpp is the particular renewal point process for which this distribution follows a decaying exponential form .
we now turn to a renewal process that is fractal : the standard fractal renewal process ( frp ) @xcite . in the standard frp ,
the times between adjacent events are independent random variables drawn from the same fractal probability distribution .
in particular , the interevent - interval probability density function @xmath22 decays essentially as a power law function @xmath103 where @xmath0 is the fractal exponent , @xmath104 and @xmath105 are cutoff parameters , and @xmath9 is a normalization constant
. the frp exhibits fractal behavior over time scales lying between @xmath104 and @xmath105 .
this process is fully fractal for @xmath61 : the power spectral density , coincidence rate , fano factor , allan factor , and even the interevent - time survivor function all exhibit scaling as in eq .
( [ scale0 ] ) with the same power - law exponent @xmath0 or one simply related to it .
further , for this process the capacity or box - counting dimension @xmath53 assumes the value @xmath0 @xcite ; since the frp is ergodic , the generalized fractal dimension @xmath51 becomes independent of the index @xmath106 @xcite , so @xmath107 for all @xmath106 , and all fractal dimensions coincide .
a different ( nonrenewal ) point process results from the superposition of a number @xmath108 of independent frps ; however , for this combined process as @xmath108 becomes large , and indeed for any frspp , the interevent - time probability density function no longer scales , and the generalized dimensions @xmath51 no longer equal @xmath0 , although the pg and af ( and the ff and cr , for @xmath78 ) retain their scaling behavior .
as the number of such processes increases , and for certain ranges of parameters , this superposition ultimately converges to the fractal gaussian - noise driven poisson process @xcite ( see also secs . [ fgn ] and [ fdspp ] ) .
the standard frp described above is a point process , consisting of a set of points or marks on the time axis as shown in fig .
[ frp_diag](a ) ; however , it may be recast as a real - valued process which alternates between two values , for example zero and unity .
this alternating frp would then start at a value of zero ( for example ) , and then switch to a value of unity at a time corresponding to the first event in the standard frp . at the second such event in the standard frp , the alternating frp would switch back to zero , and would proceed to switch back and forth at every successive event of the standard frp .
thus the alternating frp is a bernoulli process , with times between transitions given by the same interevent - interval probability density as in the standard frp , as portrayed in fig .
[ frp_diag](b ) .
many point processes derive from a continuous - time function @xmath109 which serves as the stochastically varying rate of the point process .
these are known as fractal - rate stochastic processes ( frspps ) .
we confine our discussion to rate processes ( and therefore frspps ) which are ergodic .
perhaps the simplest means for generating a point process from a rate is the integrate - and - fire ( if ) method @xcite . in this model ,
the rate function is integrated until it reaches a fixed threshold @xmath110 , whereupon a point event is generated and the integrator is reset to zero .
thus the occurrence time of the @xmath111st event is implicitly obtained from the first occurrence of @xmath112 with such a direct conversion from the rate process to the point process , any measure applied to these point processes will return results closely related to the fractal structure of the underlying rate process @xmath109 . in particular , over small frequencies , the theoretical psds of the rate process and of the resulting point process coincide .
we now turn to methods for generating several different kinds of fractal rate functions .
more complex methods for point - process generation , as delineated in secs .
[ fjif ] and [ fdspp ] , can also be applied to these same rate functions .
gaussian processes are ubiquitous in nature and are mathematically tractable .
the values of these processes at any number of fixed times form a jointly gaussian random vector ; this property , the mean , and the spectrum completely define the process . for use as a rate process
, we choose a stationary process with a mean equal to the expected rate of events , and a spectrum of the form of eq .
( [ fractalpsd ] ) with cutoffs .
fgn properly applies only for @xmath29 ; the range @xmath113 is generated by fractal brownian motion @xcite .
however , in the interest of simplicity , we employ the term fractal gaussian noise for all values of @xmath0 .
a number of methods exist for generating fgn @xcite : typically the result is a sampled version of fgn with equal spacing between samples .
interpolation between these samples is usually achieved by selecting the most recently defined value . with a rate process described by fgn serving as the input to an if process , the fractal - gaussian - noise driven integrate - and - fire process ( fgnif ) results .
generally , we require that the mean of the rate process be much larger than its standard deviation , so that the times when the rate is negative ( during which no events may be generated ) remain small .
the fgnif has successfully been used to model the human heartbeat @xcite ( see sec . [ heartbeat ] ) .
all of the rate processes considered in the following subsections converge to fgn under appropriate circumstances , and thus the point processes they yield through the if construct converge to the fgnif . a related process results from passing fgn through a memoryless exponential transform .
since the exponential of a gaussian is lognormal , we call this process fractal lognormal noise ( fln ) .
if @xmath114 denotes an fgn process with mean @xmath115 $ ] , variance @xmath116 $ ] , and autocovariance function @xmath117 , then @xmath118 $ ] is an fln process with moments @xmath119 = \exp\{n{\rm e}[x ] + n^2{\rm var}[x]/2\}$ ] and autocovariance function @xmath120\{\exp[k_x(\tau ) ] - 1\}$ ] . when this rate serves as the input to a doubly stochastic poisson point process ( see sec . [ fdspp ] )
, the result provides an excellent model for vesicular exocytosis ( see sec .
[ exocytosis ] ) . by virtue of the exponential transform ,
the autocorrelation functions of the gaussian process @xmath114 and the lognormal process @xmath109 differ ; thus their spectra differ .
in particular , the psd of the resulting point process does not follow the exact form of eq .
( [ fractalpsd ] ) , although for small values of @xmath116 $ ] , as encountered in modeling exocytosis ( for which @xmath116\le 0.6 $ ] ) @xcite , the experimental pg closely resembles this ideal form . in the limit of very small
@xmath116 $ ] , the exponential operation approaches a linear transform , and the rate process becomes fgn .
other nonlinear transforms may be applied to fgn , yielding other fractal rate processes ; we consider one which generates an exponentially distributed process . if @xmath121 and @xmath122 denote two independent and identically distributed fgn processes with zero mean , variance @xmath116 $ ] , and autocovariance function @xmath117 , then @xmath123 is a fractal exponential noise ( fen ) process with moments @xmath119 = 2^n n ! \left({\rm var}[x]\right)^n$ ] and autocovariance function @xmath124 .
the rate @xmath109 turns out to be exponentially distributed .
if @xmath117 scales as in eq .
( [ fractalcr ] ) with an exponent @xmath125 in the range @xmath126 , then so will @xmath127 , but with an exponent @xmath128 .
this process may prove useful in the study of certain kinds of thermal light @xcite .
such light has an electric field which comprises two independent components , with gaussian amplitude distributions .
the light intensity is the sum of the squares of the two fields , and therefore has an exponential amplitude distribution .
a number of independent , identical alternating fractal renewal processes ( see sec . [ frp_def ] ) may be added together , yielding a binomial process with the same fractal exponent as each individual alternating frp @xcite .
this binomial process can serve as a rate process for an integrate - and - fire process ; the fractal - binomial - noise driven integrate - and - fire ( fbnif ) results .
it is schematized in fig .
[ fbndp_dg ] .
as the number of constituent processes increases , the fbn rate converges to fgn ( see sec .
[ fgn ] ) with the same fractal exponent . though the hpp itself is not fractal , linearly filtered versions of it , denoted shot noise , can exhibit fractal characteristics . in particular , if the impulse response function of the filter has a decaying power - law ( fractal ) form , the result is fractal shot noise ( fsn ) @xcite . if fsn serves as the rate for an if process , the fractal - shot - noise driven integrate - and - fire process ( fsnif ) results ; fig .
[ fsndp_dg ] schematically illustrates the fsndp as a two - stage process .
the first stage produces a hpp with constant rate @xmath129 .
its output @xmath130 becomes the input to a linear filter with a power - law decaying impulse response function @xmath131 where @xmath0 is a fractal exponent between zero and two , @xmath104 and @xmath105 are cutoff parameters , and @xmath9 is a normalization constant .
this filter produces fractal shot noise @xmath132 at its output , which then becomes the time - varying rate for the last stage , an integrate - and - fire process .
the resulting point process @xmath2 reflects the variations of the fractal - shot - noise driving process . under suitable conditions
, fsn converges to fgn as provided by the central limit theorem @xcite ; the result is then the fgnif .
frspps based on an integrate - and - fire substrate have only one source of randomness , which is the rate process @xmath109 . those based on a poisson - process substrate ( see sec .
[ fdspp ] ) have a second source which depends explicitly on the rate process .
we now generalize the family of frspps to include those which have a second source of randomness that may be specified independently of the first .
this new process differs from the simple if process based on @xmath109 , by the imposition of a specific form of jitter on the interevent times . after generating the time of the @xmath133st event @xmath134 in accordance with eq .
( [ ifdef ] ) , the @xmath133st interevent time @xmath135 is multiplied by a gaussian - distributed random variable with unit mean and variance @xmath136 .
thus @xmath134 is replaced by @xmath137 where @xmath138 is a zero - mean , unity - variance gaussian random variable .
the result is the jittered - integrate - and - fire ( jif ) family of processes .
employing the fractal - rate functions delineated in sec .
[ fif ] yields the fractal - gaussian - noise driven jittered - integrate - and - fire process ( fgnjif ) ( sec .
[ fgn ] ) , the fractal - lognormal - noise driven jittered - integrate - and - fire process ( flnjif ) ( sec .
[ fln ] ) , the fractal - exponential - noise driven jittered - integrate - and - fire process ( fenjif ) ( sec . [ fen ] ) , the fractal - binomial - noise driven jittered - integrate - and - fire process ( fbnjif ) ( sec .
[ fbn ] ) , and the fractal - shot - noise driven jittered - integrate - and - fire process ( fsnjif ) ( sec . [ fsn ] ) .
in these point processes , the standard deviation @xmath139 of the gaussian jitter is a free parameter that controls the strength of the second source of randomness .
two limiting situations exist . in the limit @xmath140 the second source of randomness is absent and the jif processes reduce to the simple if process considered in sec .
[ fif ] . the opposite limit , @xmath141 leads to a homogeneous poisson process ; none of the fractal behavior in the rate process appears in the point process , as if the rate were constant . between these two limits , as @xmath139 increases from zero , the fractal onset times of the resulting point processes
increase , and the fractal characteristics of the point processes are progressively lost , first at higher frequencies ( shorter times ) and subsequently at lower frequencies ( longer times ) . finally , we note that another generalization of the integrate - and - fire formalism is possible @xcite . the threshold for firing @xmath110 need
not remain constant , but can itself be a random variable or stochastic process .
randomness imparted to the threshold can then be used to represent variability in other elements of a system , such as the amount of neurotransmitter released per exocytic event or the duration of ion channel openings .
the homogeneous poisson process ( hpp ) with rate @xmath142 would then form a special case within this construct , generated when the rate is fixed and constant at @xmath143 , and @xmath110 is an independent , exponentially distributed unit mean random variable associated with each interevent interval . if the rate process @xmath109 is also random , a doubly stochastic poisson point process results instead of the hpp .
we have seen that an integrate - and - fire process ( with or without jitter ) converts a fractal rate process into an frspp .
a poisson process substrate may be employed instead of the if substrate , in which case the outcome is the family of doubly stochastic poisson point processes ( dspps ) @xcite .
dspps display two forms of randomness : one associated with the stochastically varying rate , and the other associated with the underlying poisson nature of the process even if the rate were fixed and constant . with fractal rate processes
, a fractal - rate dspp ( frdspp ) results @xcite .
again , simple relations exist between measures of the rate @xmath109 and the point process @xmath144 .
in particular , for the theoretical psds we have @xmath145 + s_{\lambda}(\omega),\ ] ] where @xmath146 is the psd of the point process and @xmath147 is the psd of the rate . as with the fractal if processes considered above , the amount of randomness introduced by the second source in fractal dspps ( the poisson transform ) may also be adjusted .
however , in this case the amount introduced depends explicitly on the rate @xmath109 , rather than being independent of it as with the jif family of frspps .
for example , for a particular rate function @xmath109 suppose that its integral between zero and @xmath3 assumes the value @xmath148 : @xmath149 then given @xmath148 , the number of events @xmath2 generated in the interval @xmath150 $ ] has a poisson distribution , and in particular the mean and variance of @xmath2 both assume a value of @xmath148 .
therefore , if @xmath151 is multiplied by a constant value , then for a fixed integration time @xmath3 , the mean and variance of @xmath2 will change by the same factor . as this
constant value increases , the coefficient of variation ( the standard deviation divided by the mean ) of @xmath2 decreases , and the contribution of the poisson transform to the total randomness of the point process decreases , ultimately becoming negligible as the multiplication constant increases without limit .
however , the dspp formalism is less flexible than the jif formalism for introducing randomness , since in the former case it explicitly depends on the mean number of events , while in the latter case the variance may be independently adjusted with the parameter @xmath139 .
again , a variety of fractal point processes exist in the frdspp class , depending on the rate process chosen .
examples include the fractal - gaussian - noise driven poisson point process ( fgndp ) ( sec .
[ fgn ] , @xcite ) , the fractal - lognormal - noise driven poisson point process ( flndp ) ( sec .
[ fln ] , @xcite ) , the fractal - exponential - noise driven poisson point process ( fendp ) ( sec .
[ fen ] ) , the fractal - binomial - noise driven poisson point process ( fbndp ) ( sec .
[ fbn ] , @xcite ) , and the fractal - shot - noise driven poisson point process ( fsndp ) ( sec .
[ fsn ] , @xcite ) . for the fgndp
a slight modification is necessary since a negative rate has no meaning for a poisson point process .
therefore , rather than using a gaussian process directly we instead set the rate to zero for negative values of this process .
we require that the mean rate be much larger than its standard deviation , so that the effect of this nonlinear limiting can essentially be ignored .
again , all forms of frspps with positive finite mean rates converge to the fgndp under appropriate conditions @xcite .
other important formulations for frspps exist .
grneis and colleagues defined a clustered poisson point process in which each member of a sequence of primary events from an hpp gives rise to a train of secondary events ( as in the fsndp ) , but where the events in the train form a segment of a renewal process ( usually also an hpp , but often with a different rate ) , with a fractal ( power - law distributed ) duration @xcite .
the resulting process indeed exhibits power - law scaling in the same statistics as other frspps @xcite .
the cluster poisson point process is therefore a bartlett - lewis - type process , whereas the fsndp is a process of the neyman - scott type @xcite .
this process has been used successfully to model mesencephalic reticular formation neural spike trains @xcite ( see sec .
[ cns ] ) .
given a segment of an frspp , we seek an estimate @xmath152 of the true fractal exponent @xmath0 of the entire process from which the segment was derived .
several effects contribute to estimation error for finite - length segments , regardless of the method used .
the fractal exponent provides a measure of the relative strengths of fluctuations over various time scales ; for an frspp with a relatively large fractal exponent , for example , relatively more energy is concentrated in longer time scale fluctuations than for an frspp with a smaller fractal exponent .
variance stems from the inherent randomness of the strengths of these fluctuations .
a collection of finite realizations of an frspp with the same parameters will exhibit fractal fluctuations of varying strengths , leading to a distribution of estimated fractal exponents .
bias appears to arise from cutoffs in the time ( frequency ) domain which give rise to oscillations in the frequency ( time ) domain , confounding the pure power - law behavior of the fractal pg ( af ) . in addition , the physical limitations of the measurement process itself impose practical limits on the range of time scales available .
although algorithms exist for accurately compensating for these cutoffs , they presuppose a detailed knowledge of the process _ a priori _ , which is not in the spirit of estimating an unknown signal . consequently , in this paper we do not attempt to compensate for these cutoffs in this manner . in previous work
@xcite we determined the theoretical expected fano factor values for an frspp with finite duration , and the corresponding expected bias for the corresponding fractal exponent .
bias for the power spectral density was also considered , but not for a finite data length .
the af and the pg highlight the fractal behavior of frspps particularly well , and thus prove to be most useful for estimating fractal exponents , as indicated in sec . [ analysis ] .
we proceed to investigate the bias in these measures , and the effects of this bias on fractal - exponent estimation .
the variance of pg - based estimators was examined previously @xcite ; that for the af appears not to be readily amenable to analytical study . unlike the fano factor
, estimates of the af do not suffer from bias due to finite data length ; thus fractal exponent estimates based on the af do not either .
given a particular data set , the estimated af @xmath153 at a particular counting time @xmath6 is given by @xmath154 where @xmath155 is the number of samples , @xmath7 represents the sequence of counts as shown in fig .
[ basics](c ) , and the functional dependence of @xmath50 upon @xmath6 is suppressed for notational clarity .
this estimate of the af is simply the estimate of the allan variance divided by twice the estimate of the mean ; however , computing the expected value of this estimate is not straightforward .
we therefore employ the true mean rather than its estimate in eq .
( [ meana ] ) .
this does not appreciably affect the result , since the error so introduced remains a constant factor for all counting times , and so cancels in power - law slope calculations where logarithms are used .
in any case , the estimate of the allan variance exhibits far larger variations than the estimate of the mean , so the fluctuations in the af estimate are dominated by the former . therefore , the expected value of the af estimate becomes @xmath156 & = & { \rm e}\left\ { \left .
\widehat{{\rm e}}\left[(z_k - z_{k+1})^2\right ] \right/ 2 \widehat{{\rm e}}[z_k ] \right\ } \nonumber\\ & \approx & \left .
{ \rm e}\left[(z_k - z_{k+1})^2\right ] \right/ 2 { \rm e}[z_k ] \nonumber\\ & = & \left .
{ \rm e}\left[(z_0 - z_1)^2\right ] \right/ 2 { \rm e}[z_0 ] \nonumber\\ & = & ( \lambda t)^{-1 } { \rm e}\left[z_0 ^ 2 - z_0 z_1\right ] , \label{afbiaseq}\end{aligned}\ ] ] which is independent of the number of samples , @xmath155 , and therefore of the duration of the recording @xmath49 . in the limit @xmath157 , the expected estimated af for an frspp with @xmath85 follows the form of eq .
( [ fractalaf ] ) precisely , by definition of an ergodic process . since eq .
( [ afbiaseq ] ) does not depend on @xmath49 , the quantity @xmath158 $ ] must assume a constant value independent of @xmath49 ; thus the af does not exhibit bias caused by finite data duration .
finally , since the expected value of the af estimate is unbiased , the expected value of the estimate of @xmath0 itself is expected to have negligible bias .
this is a simple and important result .
estimates of the ordinary variance and fano factor , in contrast , do depend on the duration . in this case
we have @xmath159 & = & ( n-1)^{-1}\sum_{k=0}^{n-1}(z_{k}-\widehat{\rm e}[z])^2 \nonumber\\ & = & ( n-1)^{-1}\sum_{k=0}^{n-1 } \left(z_{k}-n^{-1}\sum_{l=0}^{n-1}z_{l}\right)^2 \nonumber\\ & = & ( n-1)^{-1}\left[\sum_{k=0}^{n-1}z_k^2 - 2n^{-1}\sum_{k=0}^{n-1}\sum_{l=0}^{n-1 } z_k z_l + n^{-2}\sum_{k=0}^{n-1}\sum_{l=0}^{n-1}\sum_{m=0}^{n-1 } z_l z_m\right ] \nonumber\\ & = & ( n-1)^{-1}\sum_{k=0}^{n-1}z_k^2 - n^{-1}(n-1)^{-1}\sum_{k=0}^{n-1}\sum_{l=0}^{n-1 } z_k z_l \nonumber\\ & = & n^{-1}\sum_{k=0}^{n-1}z_k^2 - n^{-1}(n-1)^{-1}\sum_k^{n-1}\sum_{l\neq k } z_k z_l.\end{aligned}\ ] ] these cross terms , which do depend on the number of samples , lead to an estimated fano factor with a confounding linear term @xmath160 \approx 1 + ( t / t_0)^\alpha - t\left/\left(t_0^{\alpha } l^{1-\alpha}\right)\right . \label{ffbias}\ ] ] for @xmath78 @xcite . the last term on the right - hand - side of eq .
( [ ffbias ] ) leads to bias in estimating the fractal exponent ; for this and other reasons , we do not employ the ff in fractal - exponent estimation .
computation of the periodogram ( estimated psd ) proves tractable only when obtained from the series of counts @xmath161 , rather than from the entire point process @xmath2 .
this , and more importantly the finite length of the data set , introduce a bias in the estimated fractal exponent as shown below .
we begin by obtaining the discrete - time fourier transform of the series of counts @xmath161 , where @xmath162 : @xmath163 the periodogram then becomes @xmath164 with an expected value @xmath165 & = & m^{-1 } \sum_{k=0}^{m-1 } \sum_{m=0}^{m-1 } e^{j2\pi ( k - m)n / m } { \mbox{\rm e}}\left[z_k z_m\right ] \nonumber\\ & = & m^{-1 } \sum_{k=0}^{m-1 } \sum_{m=0}^{m-1 } e^{j2\pi ( k - m)n / m } { \mbox{\rm e}}\left[\int_{s=0}^t\int_{t=0}^t\ , dn(s+kt)dn(t+mt)\right ] \nonumber\\ & = & m^{-1 } \sum_{k=0}^{m-1 } \sum_{m=0}^{m-1 } e^{j2\pi ( k - m)n / m } \int_{s=0}^t\int_{t=0}^t g_n[s - t+(k - m)t]\ , ds\ , dt \nonumber\\ & = & m^{-1 } \sum_{k=0}^{m-1 } \sum_{m=0}^{m-1 } e^{j2\pi ( k - m)n / m } \int_{u =- t}^t\int_{v=|u|}^{2t-|u| } g_n[u+(k - m)t]\ , { \frac{\displaystyle { du\ , dv}}{\displaystyle { 2 } } } \nonumber\\ & = & m^{-1 } \sum_{k=0}^{m-1 } \sum_{m=0}^{m-1 } e^{j2\pi ( k - m)n / m } \int_{u =- t}^t(t-|u|)\ , g_n[u+(k - m)t]\ , du \nonumber\\ & = & m^{-1 } \sum_{k=0}^{m-1 } \sum_{m=0}^{m-1 } e^{j2\pi ( k - m)n / m } \int_{u =- t}^t(t-|u| ) \nonumber\\ & & \qquad \times \int_{\omega=-\infty}^\infty s_n(\omega)\ , e^{j\omega [ u + ( k - m)t]}\ , { \frac{\displaystyle { d\omega}}{\displaystyle { 2 \pi}}}\ , du \nonumber\\ & = & ( 2\pi m)^{-1 } \int_{\omega=-\infty}^\infty s_n(\omega ) \left|\sum_{k=0}^{m-1 } e^{jk(2\pi n / m + \omega t)}\right|^2 \int_{u =- t}^t(t-|u|)\ , e^{j\omega u } \ , du\ , d\omega \nonumber\\ & = & ( 2\pi m)^{-1 } \int_{\omega=-\infty}^\infty s_n(\omega ) { \frac{\displaystyle { \sin^2(\pi n + m\omega t/2)}}{\displaystyle { \sin^2(\pi n / m + \omega t/2 ) } } } { \frac{\displaystyle { 4\sin^2(\omega t/2)}}{\displaystyle { \omega^2}}}\ , d\omega \nonumber\\ & = & { \frac{\displaystyle { t}}{\displaystyle { \pi m } } } \int_{-\infty}^\infty s_n(2x / t ) { \frac{\displaystyle { \sin^2(mx)\sin^2(x)}}{\displaystyle { x^2\sin^2(x + \pi n / m)}}}\ , dx.\end{aligned}\ ] ] for a general fractal stochastic point process , where the psd follows the form of eq .
( [ fractalpsd ] ) , we therefore have @xmath165 & = & { \frac{\displaystyle { \lambda t}}{\displaystyle { \pi m } } } \int_{-\infty}^\infty \left[1 + ( \omega_0 t/2)^\alpha |x|^{-\alpha}\right ] { \frac{\displaystyle { \sin^2(mx)\sin^2(x)}}{\displaystyle { x^2\sin^2(x + \pi n / m)}}}\ , dx . \label{fractalpgex}\end{aligned}\ ] ] focusing on the smaller values of @xmath166 useful in estimating fractal exponents permits the use of two approximations in eq .
( [ fractalpgex ] ) .
for the values of @xmath0 used in this paper ( @xmath167 ) , the integrand of eq .
( [ fractalpgex ] ) will only be significant near @xmath168 , yielding @xmath165 & \approx & { \frac{\displaystyle { \lambda t}}{\displaystyle { \pi m } } } \int_{-\infty}^\infty \left[1 + ( \omega_0 t/2)^\alpha |x|^{-\alpha}\right ] { \frac{\displaystyle { m\pi\delta(x + \pi n / m)\sin^2(x)}}{\displaystyle { x^2}}}\ , dx \nonumber\\ & = & \lambda t \left[1 + ( 2\pi n / \omega_0 m t)^{-\alpha}\right ]
{ \frac{\displaystyle { \sin^2(\pi n / m)}}{\displaystyle { ( \pi n / m)^2 } } } \nonumber\\ & \approx & \lambda t \left[1 + ( 2\pi n / \omega_0 m t)^{-\alpha}\right],\end{aligned}\ ] ] which is of the same form as eq . ( [ fractalpsd ] ) . improvement of this estimation procedure appears to require numerical integration of eq .
( [ fractalpgex ] ) , which proves nontrivial since the integrand exhibits oscillations with a small period @xmath169 .
fortuitously , for the parameter values employed in this paper the integrand appears peaked near @xmath168 , so that not too many ( @xmath170 ) oscillations need be included in the calculations . indeed , numerical results employing this method , and with @xmath171 for which eq .
( [ fractalpgex ] ) is known to assume the simple value @xmath172 , agree within @xmath173 .
these results form an extension of earlier approaches @xcite which ignored the effects of imposing periodic boundary conditions on the fourier transform , and of binning the events .
numerical integration of eq .
( [ fractalpgex ] ) followed by a least - squares fit on a doubly logarithmic plot leads to results for the expected bias of the psd - based estimate of the fractal exponent , shown in table [ tabgau02 ] .
other methods exist for compensating for finite data length in power spectral estimation , such as maximizing the entropy subject to the correlation function over the available range of times ( see , e.g. , @xcite ) .
having considered various possible theoretical sources of error in simulating frspps with a desired fractal exponent , we now proceed to investigate their effects upon fractal exponent estimates based on the power spectral density and allan factor measures . to this end
we employ simulations of three of the frspps outlined above : fgn driving a deterministic integrate - and - fire process ( fgnif ) , fgn driving an if process followed by gaussian jitter of the interevent intervals ( fgnjif ) , and fgn driving a poisson process ( fgndp ) .
we choose these three processes from the collection presented in sec .
[ mathform ] for a number of reasons .
first , the fgnif , fgnjif , and fgndp derive from a common continuous - time rate process , fractal gaussian noise ( fgn ) , and thus differences among these point processes must lie in the point - process generation mechanism rather than in the fractal rate processes .
second , fgn admits fractal exponent of any value , while the fractal exponents can not exceed two for fractal binomial noise and fractal shot noise .
third , of all the frspps examined in sec .
[ mathform ] , those based on the fgn appear to suffer the least from the effects of cutoffs , so that expected values of the pg and the af most closely follow the pure power - law forms of eqs .
( [ fractalpsd ] ) and ( [ fractalaf ] ) respectively . to generate the rate function with @xmath174 spectral properties , we computed discrete - time approximations to fractal gaussian noise ( fgn ) by forming a conjugate - symmetric sequence @xmath175 $ ] of @xmath176 samples which satisfies @xmath177\right|=\left\ { \begin{array}{ll } m \rho & \mbox{for $ k=0$}\\ c k^{-\alpha/2 } & \mbox{for $ 1\leq k \leq m/2 $ } , \end{array } \right.\ ] ] where
@xmath178 $ ] is the desired average value of the fgn process , and @xmath20 is a constant determining the relative strength of the fgn .
the phases of @xmath175 $ ] were independent and uniformly distributed in the interval @xmath179 for @xmath180 , while both @xmath181 $ ] and @xmath182 $ ] were strictly real .
a discrete - time sequence @xmath183 $ ] was obtained by taking the inverse discrete fourier transform of @xmath175 $ ] .
the resulting fgn process is periodic with period @xmath176 , so to reduce the effects of this periodicity we employ only the first half of the sequence .
specifically , we used an original sequence @xmath183 $ ] of length @xmath184 points and kept the first @xmath185 samples as @xmath186 $ ] . further , since rate functions must not attain negative values , we chose the mean value of the fgn , @xmath187 , sufficiently large compared to the relative strength of the fgn @xmath20 to ensure that @xmath186 $ ] remained positive for all @xmath188 different values of @xmath186 $ ] in all simulations .
in particular for @xmath29 this was adjusted by choosing the ff intercept time @xmath80 to be 10 times the average interevent time , while for @xmath102 the psd fractal onset frequency @xmath68 was chosen to be @xmath189 times the average rate @xmath187 .
the resultant sequence @xmath186 $ ] was taken to represent equally spaced samples of a continuous - time process ; to simplify subsequent analysis , we chose the duration of each sample to be one sec , without loss of generality .
varying the mean event production rate @xmath187 while keeping all other parameters fixed ( or in fixed relation to @xmath187 , as with the ff intercept time and psd corner frequency ) resulted in point processes which contained varying expected numbers of events .
in particular , we employed values of @xmath190 and @xmath191 events per sample of @xmath186 $ ] , leading to a range of roughly @xmath192 to @xmath193 events in the resulting point process .
the following procedures were employed to compute the estimates described in sec .
[ analysis ] . for the pg ,
the absolute event times were quantized into @xmath194 bins , which therefore formed a rate estimate of the process with counting windows having a duration of @xmath195 sec .
then a discrete fourier transform was performed , followed by replacing the fourier components with their square magnitudes .
finally , a least - squares fit was performed on the logarithm of these quantities vs. the logarithm of the frequencies @xmath13 for various selected ranges of frequencies .
these slopes form pg - based estimates of the fractal exponent .
estimates of the af were calculated over a range of counting times @xmath6 in a geometric sequence with ten counting times per decade .
the slopes of the af estimate were again obtained from linear least - squares fits to the logarithm of the af values vs. the logarithm of the size of the counting time @xmath6 ; these slopes became the af - based estimates of the fractal exponent .
as indicated in sec .
[ af : bias ] , finite data lengths do not impart extraneous bias to the af - based estimates .
for each type of frspp and value of @xmath187 employed , 100 simulations were performed for each of three different design values of the fractal exponent : @xmath196 , 0.8 , and 1.5 ; af and pg curves were computed for each simulation .
estimates of the fractal exponent were obtained by two methods for each of these two measures . in the first
, all 100 af or pg curves were averaged together , after which the doubly logarithmic slope of this average curve was computed , yielding a single estimate of the fractal exponent . the second method consisted of first obtaining the doubly logarithmic slopes of the individual af or pg curves , and then averaging these slopes .
this second method yields the mean fractal exponent as well as the standard deviation ( and other possible statistics of the fractal exponent ) . as shown in sec .
[ fif ] , the fgn - driven integrate - and - fire process ( fgnif ) is obtained by identifying the fgn sequence @xmath186 $ ] as samples of a rate function @xmath109 , taken to be constant over each sample of @xmath186 $ ] ( 1 sec ) .
this rate function is integrated until the result attains a value of unity , whereupon an output point is generated and the integrator is reset to zero . the generation of the continuous rate function from the piecewise - constant fgn does not significantly change the observed fractal structure of the resulting point process . for frequencies less than the inverse of the sampling time @xmath6 ( 1 hz in this example ) , the psds of the piecewise - constant and exact versions of fgn differ by a factor of @xmath197 ; fractal behavior depends on low frequencies ( @xmath198 ) where the above factor is essentially unity . since the fgnif process has just one source of randomness , namely the underlying fgn , the estimators for the fractal exponent should therefore display only the ( fractal ) behavior of the fgn itself , along with finite - size effects .
bias or variance due to a second source of randomness , such as the poisson transform in the fgndp , will not be present .
thus for our purposes the fgnif process serves as a benchmark for the accuracy of fractal exponent estimators , with a bias and variance depending only on finite data size effects and not on the rate @xmath187 .
the results for the fgnif process with a rate @xmath199 and three different fractal exponents ( @xmath0=0.2 , 0.8 and 1.5 ) are summarized for the af in table [ tabgau01 ] and for the pg in table [ tabgau02 ] . .
af - based fractal exponent estimates @xmath152 for simulated fgnif processes with a rate @xmath200 , for different time ranges .
the governing rate processes have theoretical fractal exponents of @xmath196 , @xmath201 , and @xmath202 .
the estimated fractal exponents were obtained from straight - line fits on doubly logarithmic coordinates to an average curve of 100 independent simulations ( fit of avg . ) and from averages of the slopes of the individual curves ( avg . of fits ) .
the deviations are the standard deviations obtained from the second averaging procedure . [
cols="^,^,<,<,<",options="header " , ] for both the fgnjif and the fgndp processes , estimating fractal exponents over larger ranges of time or frequency leads to increased bias , particularly for large values of @xmath139 and small values of @xmath187 . since the results depart from those obtained with the fgnif , the additional randomness as @xmath187 decreases , or as @xmath139 increases , leads to a dilution of the power - law behavior of the af and pg , changing its onset to at larger times and smaller frequencies . in the limits
@xmath203 and @xmath141 , the result is essentially white noise and the estimated fractal exponent will be zero .
best results for the fgnif were obtained with the af - based fractal - exponent estimator using counting times in the range @xmath204 ; this yielded root - mean - square error results which were accurate to within @xmath205 for all design values of @xmath0 ( tables [ tabgau01 ] and [ rmse ] ) .
for both the af- and pg - based estimators , performance degraded at longer times and smaller frequencies ( tables [ tabgau01 ] , [ tabgau02 ] , and [ rmse ] ) .
finite - duration effects in the pg and relatively few counting windows in the af produce significant error over these scales .
the rms error remained near @xmath206 for both estimators , over all design values of the fractal exponent , and over all time and frequency ranges examined .
the fgnjif and fgndp processes impose a second source of randomness on the fractal behavior of the fgn kernel .
this additional randomness is essentially uncorrelated , and thus may be modeled as the linear superposition of white noise on the original fractal signal ; this affects the shapes of both the af and the pg . as this white - noise level increases
, it dilutes the fractal structure present in the fgn process , beginning at the shortest time scales and moving progressively towards longer ones .
this effect reduces the range of times and frequencies over which fractal behavior appears , resulting in a decreased estimated fractal exponent .
this proved most significant for larger design values of @xmath0 , and for the estimator based on the pg .
indeed , magnitudes of the bias exceeded @xmath201 in one case .
the af - based estimator fared significantly better , although in this case it employed counting times which were five times the inverse of the frequencies used for the pg .
these results compare favorably to earlier studies using the same number of events @xcite , for which inaccuracies of @xmath206 or larger were obtained for a variety of fractal stochastic point processes , all of which included a second source of randomness .
this study considered data sets with an average of @xmath207 points ; experimental data may contain far fewer points , and thus yield poorer fractal - exponent estimation statistics .
in particular , finite - size effects , which lead to increased bias for the pg - based estimator and increased variance for the af - based estimator will reduce the ranges of times and frequencies over which reliable estimation can be achieved .
similarly , data sets with larger numbers of points ( computer network traffic , for example ) would yield superior performance .
in addition , if several independent data sets are available , known to be from identical experiments , then the af plots or the resulting estimates could be averaged , reducing the effective variance . in this case , the reduced bias in the af - based estimator would render it far superior to that based on the pg .
finally we mention that there exist several methods for reducing the bias of the fractal exponent estimates if one has some _ a priori _ knowledge of the process of interest .
such an approach however does not follow the philosophy of estimating a completely unknown signal and we therefore did not attempt to compensate for the bias by the use of such methods .
we have investigated the properties of fractal and fractal - rate stochastic point processes ( fspps and frspps ) , focusing on the estimation of measures that reveal their fractal exponents .
the fractal - gaussian - noise driven integrate - and - fire process ( fgnif ) is unique as a point process in that it exhibits no short - term randomness ; we have developed a generalization which includes jitter , the fgnjif , for which the short - term randomness is fully adjustable .
in addition to the randomness contributed by the fractal nature of the rate , all other frspps appear to exhibit additional randomness associated with point generation , which confounds the accuracy of fractal - exponent estimation .
the fgnjif proved crucial in elucidating the role that such randomness plays in the estimation of fractal exponents in other frspps .
we presented analytical results for the expected biases of the pg- and af - based fractional exponent estimators due to finite data length .
we followed this theoretical work with a series of simulations of frspps for three representative fractal exponents : @xmath196 , @xmath201 , and @xmath202 . using these simulations together with the analytical predictions , we delineated the sources of error involved in estimating fractal exponents . in particular
, using the fgnjif we were able to separate those factors intrinsic to estimation over finite frspp data sets from those due to the particular form of frspp involved .
we conclude that the af - based estimate proves more reliable in estimating the fractal exponent than the pg - based method , yielding rms errors of @xmath205 for data segments of frspps with @xmath207 points over all values of @xmath0 examined .
finally , we note that wavelet generalizations of the af appear to yield comparable results @xcite , suggesting that the af estimate may be optimal in some applications .
we are grateful to m. shlesinger for valuable suggestions .
this work was supported by the austrian science foundation under grant no .
s-70001-mat ( st , mf , hgf ) , by the office of naval research under grant no .
n00014 - 92-j-1251 ( sbl , mct ) , by the whitaker foundation ( sbl ) , and by the joint services electronics program through the columbia radiation laboratory ( mct ) .
equation ( [ fractalpgex ] ) yields predictions for the pg - based fractal exponent estimation bias for a general fractal - rate stochastic point process ( frspp ) .
more accurate results may be obtained for particular frspps , taking into account the structure of these individual processes .
we consider the case of the fgnif process , with the further specialization that the original fgn array is obtained by direct fourier synthesis with half of the array discarded , and that the analyzing periodogram subsequently doubles the number of points , yielding the same number as in the original array .
we begin with a discrete - time power spectral density @xmath208 which is periodic with period @xmath176 , and with @xmath209 . for simulation purposes
we define a conjugate symmetric sequence @xmath210 based on the square root of @xmath208 : @xmath211 where @xmath212 are independent and uniformly distributed in @xmath213 , @xmath214 represents the signum function , and @xmath215 denotes complex conjugation .
the corresponding time - domain signal may be obtained by inverse fourier transformation : @xmath216 we next define a subsampled sequence @xmath50 , also of length @xmath176 , by @xmath217 for which the corresponding sequence in the frequency domain becomes @xmath218 \nonumber\\ & = & 2 e^{-j\pi
n / m } \cos(\pi n / m ) \sum_{l=0}^{m/2 - 1 } x_l e^{-j4\pi ln / m } \nonumber\\ & = & 2 m^{-1 } e^{-j\pi n / m } \cos(\pi n / m ) \sum_{l=0}^{m/2 - 1 } e^{-j4\pi ln / m } \sum_{q=1}^{m-1 } e^{j2\pi ql / m } \tilde x_q.\end{aligned}\ ] ] finally , we compute the expected value of the periodogram of @xmath219 @xmath220 & \equiv & m^{-1 } { \mbox{\rm e}}\left[\tilde z_n \right]^2 \nonumber\\ & = & 4 m^{-3 } \cos^2(\pi n / m ) \sum_{l=0}^{m/2 - 1 } e^{-j4\pi ln / m } \sum_{k=0}^{m/2 - 1 } e^{j4\pi kn / m } \nonumber\\ & & \qquad \times \sum_{q=1}^{m-1 } e^{j2\pi ql / m }
\sum_{r=1}^{m-1 } e^{-j2\pi rk / m } { \mbox{\rm e}}\left[\tilde x_q \tilde x^*_r \right ] \nonumber\\ & = & 4 m^{-2 } \cos^2(\pi n / m ) \sum_{l=0}^{m/2 - 1 } e^{-j4\pi ln / m } \sum_{k=0}^{m/2 - 1 } e^{j4\pi kn / m } \sum_{q=1}^{m-1 } e^{j2\pi ql / m } e^{-j2\pi qk / m } s_x(q ) \nonumber\\ & = & 4 m^{-2 } \cos^2(\pi n / m ) \sum_{q=1}^{m-1 } s_x(q ) \left| \sum_{l=0}^{m/2 - 1 } e^{-j2\pi ( 2n - q)l / m}\right|^2 \nonumber\\ & = & \cos^2(\pi n / m ) \left\{s_x(2n ) + 4 m^{-2 } \sum_{\stackrel{\scriptstyle q=1 } { q \mbox{\scriptsize \hspace{2pt } odd}}}^{m-1 } \csc^2\left[\pi(2n - q)/m\right ] s_x(q ) \right\ } .
\label{app_psd_sum}\end{aligned}\ ] ] numerical computation based on eq .
( [ app_psd_sum ] ) yields results substantially in agreement with those of eq .
( [ fractalpgex ] ) .
99 d. r. cox and v. isham , _ point processes _ ( chapman and hall , london , 1980 ) . s. b. lowen and m. c. teich , `` estimation and simulation of fractal stochastic point processes , '' _ fractals _ * 3 * , 183 - 210 ( 1995 ) .
w. rudin , _ principles of mathematical analysis , _
( mcgraw hill , new york , 1976 ) , p. 197 .
m. f. shlesinger and b. j. west , `` complex fractal dimension of the bronchial tree , '' _ phys .
lett . _ * 67 * , 2106 - 2108 ( 1991 ) .
b. b. mandelbrot , _ the fractal geometry of nature _ ( w. h. freeman , new york , 1983 )
. h. e. hurst , `` long term storage capacity of reservoirs , '' _ trans .
. civil eng . _ * 116 * , 770808 ( 1951 ) .
p. flandrin , `` on the spectrum of fractional brownian motions , '' _ ieee trans .
theory _ * 35 * , 197199 ( 1989 ) .
p. flandrin , `` wavelet analysis and synthesis of fractional brownian motion , '' _ ieee trans .
theory _ * 38 * , 910917 ( 1992 ) .
p. flandrin , `` time scale analyses and self - similar stochastic processes , '' _ proc .
nato advanced study institute on wavelets and their applications _
( il ciocco , italy , 1992 ) . g. w. wornell and a. v. oppenheim , `` estimation of fractal signals from noisy measurements using wavelets , '' _ ieee trans . sig .
_ * 40 * , 611623 ( 1992 )
. r. j. barton and h. v. poor , `` signal detection in fractional gaussian noise , '' _ ieee trans . inform .
theory _ * 34 * , 943959 ( 1988 ) .
w. h. press , `` flicker noises in astronomy and elsewhere , '' _ comm .
astrophys . _ * 7 * ( no . 4 ) , 103119 ( 1978 ) .
m. j. buckingham , _ noise in electronic devices and systems _
( wiley - halsted , new york , 1983 ) , ch .
m. b. weissman , `` @xmath46 noise and other slow , nonexponential kinetics in condensed matter , '' _ rev .
_ * 60 * , 537571 ( 1988 )
. j. b. bassingthwaighte , l. s. liebovitch , and b. j. west , _ fractal physiology _
( american physioloical society , new york , 1994 ) . b. j. west and w. deering , `` fractal physiology for physicists : lvy statistics , '' _ phys . rep .
_ * 246 * , 1100 ( 1994 ) . c. m. anderson and a. j. mandell , `` fractal time and the foundations of consciousness : vertical convergence of @xmath46 phenomena from ion channels to behavioral states , '' in _ fractals of brain , fractals of mind _ ( advances in consciousness research * 7 * ) , eds .
e. maccormac and m. stamenov ( john benjamin , amsterdam , 1996 ) , pp .
. g. pfister and h. scher , `` dispersive ( non - gaussian ) transient transport in disordered solids , '' _ adv .
_ * 27 * , 747798 ( 1978 ) . j. orenstein , m. a. kastner , and v. vaninov , `` transient photoconductivity and photo - induced optical absorption in amorphous semiconductors , '' _ phil . mag .
b _ * 46 * , 2362 ( 1982 ) .
m. a. kastner , `` the peculiar motion of electrons in amorphous semiconductors , '' in _ physical properties of amorphous materials , _ eds .
d. alser , b. b. schwartz , and m. c. steele , ( plenum , new york , 1985 ) , pp .
. t. tiedje and a. rose , `` a physical interpretation of dispersive transport in disordered semiconductors , '' _ solid state commun . _
, 4952 ( 1980 ) . m. f. shlesinger , `` fractal time and @xmath46 noise in complex systems , '' _ ann . new york acad .
* 504 * , 214228 ( 1987 ) .
s. b. lowen and m. c. teich , `` fractal renewal processes generate @xmath46 noise , '' _ phys .
e _ * 47 * , 9921001 ( 1993 ) .
s. b. lowen and m. c. teich , `` fractal renewal processes as a model of charge transport in amorphous semiconductors , '' _ phys .
b _ * 46 * , 18161819 ( 1992 ) . v. k. bhatnagar and k. l. bhatia , `` frequency dependent electrical transport in bismuth - modified amorphous germanium sulfide semiconductors , ''
_ j. non - cryst .
sol . _ * 119 * , 214231 ( 1990 ) .
w. tomaszewicz , `` multiple - trapping carrier transport due to modulated photogeneration , '' _ phil . mag .
* 61 * , 237243 ( 1990 ) .
j. m. berger and b. b. mandelbrot , `` a new model for the clustering of errors on telephone circuits , '' _
ibm j. res .
dev . _ * 7 * , 224236 ( 1963 ) . b. b. mandelbrot , `` self - similar error clusters in communication systems and the concept of conditional stationarity , '' _ ieee trans .
comm . tech .
* 13 * , 7190 ( 1965 ) .
w. e. leland , m. s. taqqu , w. willinger , and d. v. wilson , `` on the self - similar nature of ethernet traffic , '' in _ proc .
acm sigcomm 1993 _ , ( 1993 ) , pp
. 183193 .
a. erramilli and w. willinger , `` fractal properties in packet traffic measurements , '' in _ proc .
petersburg reg .
_ , ( st . petersburg , russia , 1993 ) , pp . 144 - 158 . b. k. ryu and h. e. meadows , `` performance analysis and traffic behavior of xphone videoconferencing application on an ethernet , '' in _ proc .
3rd int .
conf . comp .
w. liu , ( 1994 ) , pp .
w. e. leland , m. s. taqqu , w. willinger , and d. v. wilson , `` on the self - similar nature of ethernet traffic ( extended version ) , '' _ ieee / acm trans .
* 2 * , 115 ( 1994 ) .
j. beran , r. sherman , m. s. taqqu , and w. willinger , `` long - range dependence in variable - bit - rate video traffic , '' _ ieee trans .
comm . _ * 43 * , 15661579 ( 1995 ) .
b. k. ryu and s. b. lowen , `` modeling self - similar traffic with the fractal - shot - noise - driven poisson process , '' cent . for telecomm .
res . , tech . rep.392 - 94 - 39 , ( columbia university , new york , 1994 ) .
b. sakmann and e. neher , _ single - channel recording _
( plenum , new york , 1983 ) .
l. j. defelice and a. isaac , `` chaotic states in a random world : relationship between the nonlinear differential equations of excitability and the stochastic properties of ion channels , '' _ j. stat .
phys . _ * 70 * , 339354 ( 1993 ) .
p. luger , `` internal motions in proteins and gating kinetics of ionic channels , '' _ biophys .
j. _ * 53 * , 877884 ( 1988 ) .
g. l. millhauser , e. e. salpeter , and r. e. oswald , `` diffusion models of ion - channel gating and the origin of power - law distributions from single - channel recording , '' _ proc .
( usa ) * 85 * , 15031507 ( 1988 ) .
l. s. liebovitch and t. i. tth , `` using fractals to understand the opening and closing of ion channels , '' _ ann
. biomed .
eng . _ * 18 * , 177194 ( 1990 ) .
m. c. teich , `` fractal character of the auditory neural spike train , '' _ ieee trans
. biomed .
eng . _ * 36 * , 150160 ( 1989 ) .
s. b. lowen and m. c. teich , `` fractal renewal processes , '' _ ieee trans .
theory _ * 39 * , 16691671 ( 1993 ) .
s. b. lowen and m. c. teich , `` fractal auditory - nerve firing patterns may derive from fractal switching in sensory hair - cell ion channels , '' in _ noise in physical systems and @xmath46 fluctuations _
( aip conference proceedings * 285 * ) , eds .
h. handel and a. l. chung , ( american institute of physics , new york , 1993 ) , pp .
. b. katz , _ nerve , muscle , and synapse _
( mcgraw - hill , new york , 1966 ) .
p. fatt and b. katz , `` spontaneous subthreshold activity at motor nerve endings , ''
_ j. physiol .
_ ( london ) * 117 * , 109128 ( 1952 )
. j. del castillo and b. katz , `` quantal components of the end - plate potential , '' _ j. physiol . _ ( london ) * 124 * , 560573 ( 1954 ) .
s. b. lowen , s. s. cash , m - m .
poo , and m. c. teich , `` neuronal exocytosis exhibits fractal behavior , '' in _ computational neuroscience , _ ed .
j. m. bower ( plenum , new york , 1997 ) , in press .
s. b. lowen , s. s. cash , m - m .
poo , and m. c. teich ,
quantal neurotransmitter secretion rate exhibits fractal behavior , " _ j. neurosci .
_ , in press .
t. musha , y. kosugi , g. matsumoto , and m. suzuki , `` modulation of the time relation of action potential impulses propagating along an axon , '' _ ieee trans .
eng . _ * bme-28 * , 616623 ( 1981 ) .
t. musha , h. takeuchi , and t. inoue , `` @xmath46 fluctuations in the spontaneous spike discharge intervals of a giant snail neuron , '' _ ieee trans
. biomed .
eng . _ * bme-30 * , 194197 ( 1983 ) .
s. b. lowen and m. c. teich , `` auditory - nerve action potentials form a non - renewal point process over short as well as long time scales , '' _
j. acoust .
* 92 * , 803806 ( 1992 ) .
m. c. teich , d. h. johnson , a. r. kumar , and r. g. turcott , `` rate fluctuations and fractional power - law noise recorded from cells in the lower auditory pathway of the cat , '' _ hear .
res . _ * 46 * , 4152 ( 1990 ) .
m. c. teich , r. g. turcott , and s. b. lowen , `` the fractal doubly stochastic poisson point process as a model for the cochlear neural spike train , '' in _ the mechanics and biophysics of hearing ( lecture notes in biomathematics , vol .
87 ) , _ eds .
p. dallos , c. d. geisler , j. w. matthews , m. a. ruggero , and c. r. steele ( springer - verlag , new york , 1990 ) , pp .
354361 . n. l. powers , r. j. salvi , and s. s. saunders , `` discharge rate fluctuations in the auditory nerve of the chinchilla , '' in _ abstracts of the fourteenth midwinter research meeting of the association for research in otolaryngology , _ ed . d. j. lim ( association for research in otolaryngology , des moines , ia ) , abstract 411 , p. 129 .
n. l. powers and r. j. salvi , `` comparison of discharge rate fluctuations in the auditory nerve of chickens and chinchillas , '' in _ abstracts of the fifteenth midwinter research meeting of the association for research in otolaryngology , _ ed .
d. j. lim ( association for research in otolaryngology , des moines , ia , 1992 ) , abstract 292 , p. 101 .
m. c. teich , `` fractal neuronal firing patterns , '' in _ single neuron computation , _ eds .
t. mckenna , j. davis , and s. zornetzer ( academic , boston , 1992 ) , pp .
a. h. kumar and d. h. johnson , `` analyzing and modeling fractal intensity point processes , '' _ j. acoust .
* 93 * , 33653373 ( 1993 ) .
m. c. teich and s. b. lowen , `` fractal patterns in auditory nerve - spike trains , '' _ ieee eng . med .
biol . mag .
_ * 13 * ( no . 2 ) , 197202 ( 1994 ) . o. e. kelly , `` analysis of long - range dependence in auditory - nerve fiber recordings , '' master s thesis , rice university ( 1994 ) . o. e. kelly , d. h. johnson , b. delgutte , and p. cariani , `` fractal noise strength in auditory - nerve fiber recordings , '' _ j. acoust
. soc .
_ * 99 * , 22102220 ( 1996 )
. s. b. lowen and m. c. teich , `` the periodogram and allan variance reveal fractal exponents greater than unity in auditory - nerve spike trains , '' _ j. acoust .
soc . am _ * 99 * , 35853591 ( 1996 ) . s. b. lowen and m. c. teich , `` refractoriness - modified fractal stochastic point processes for modeling sensory - system spike trains , '' in _ computational neuroscience , _ ed .
j. m. bower ( academic , new york , 1996 ) , pp
. 447 - 452 .
s. b. lowen and m. c. teich , `` estimating scaling exponents in auditory - nerve spike trains using fractal models incorporating refractoriness , '' in _ diversity in auditory mechanics , _ eds .
e. r lewis , g. r. long , r. f. lyon , p. m. narins , c. r. steele , and e. hecht - pointar ( world scientific , singapore , 1997 ) , pp .
197 - 204 .
m. c. teich , c. heneghan , s. b. lowen , t. ozaki , and e. kaplan , `` fractal character of the neural spike train in the visual system of the cat , '' _ j. opt
. soc .
a _ * 14 * , 529546 ( 1997 ) . j. b. troy and j. g. robson , `` steady discharges of x and y retinal ganglion cells of cat under photopic illuminance , '' _ visual neuroscience _ * 9 * , 535553 ( 1992 ) . m. c. teich , r. g. turcott , and r. m. siegel , `` temporal correlation in cat striate - cortex neural spike trains , '' _ ieee eng .
biol . mag .
_ * 15 * ( no . 5 ) , 7987 ( 1996 ) .
r. g. turcott , p. d. r. barker , and m. c. teich , `` long - duration correlation in the sequence of action potentials in an insect visual interneuron , '' _
j. statist .
simul . _ * 52 * , 253271 ( 1995 ) .
m. e. wise , `` spike interval distributions for neurons and random walks with drift to a fluctuating threshold , '' in _ statistical distributions in scientific work , vol .
6 _ , eds . c. taillie , g. p. patil , and b. a. baldessari ( d. reidel , hingham , ma , 1981 ) , pp
. 211231
. m. yamamoto , h. nakahama , k. shima , t. kodama , and h. mushiake , `` markov - dependency and spectral analyses on spike - counts in mesencephalic reticular formation during sleep and attentive states , '' _ brain research _ * 366 * , 279289 ( 1986 ) .
f. grneis , m. nakao , m. yamamoto , t. musha , and h. nakahama , `` an interpretation of @xmath46 fluctuations in neuronal spike trains during dream sleep , '' _ biol .
cybern _ * 60 * , 161169 ( 1989 ) f. grneis , m. nakao , y. mizutani , m. yamamoto , m. meesman , and t. musha , `` further study on @xmath46 fluctuations observed in central single neurons during rem sleep , '' _ biol .
cybern . _ * 68 * 193198 ( 1993 ) .
t. kodama , h. mushiake , k. shima , h. nakahama , and m. yamamoto , `` slow fluctuations of single unit activities of hippocampal and thalamic neurons in cats . i. relation to natural sleep and alert states , '' _ brain research _ * 487 * , 2634 ( 1989 ) . m. yamamoto and h. nakahama , `` stochastic properties of spontaneous unit discharges in somatosensory cortex and mesencephalic reticular formation during sleep waking states , '' _ j. neurophysiology _ * 49 * , 11821198 ( 1983 ) .
m. kobayashi and t. musha , `` @xmath46 fluctuations of heartbeat period , '' _ ieee trans . biomed .
eng . _ * bme-29 * , 456457 ( 1982 ) .
r. d. berger , s. akselrod , d. gordon , and r. j. cohen , `` an efficient algorithm for spectral analysis of heart rate variability , '' _ ieee trans .
* bme-33 * , 900904 ( 1986 ) .
r. g. turcott and m. c. teich , `` long - duration correlation and attractor topology of the heartbeat rate differ for healthy patients and those with heart failure , '' _ proc .
spie _ * 2036 * ( chaos in biology and medicine ) , 2239 ( 1993 ) . r. g. turcott and m. c. teich , `` interevent - interval and counting statistics of the human heartbeat recorded from normal subjects and patients with heart failure , '' _ ann
. biomed .
* 24 * , 269293 ( 1996 ) .
h. e. schepers , j. h. g. m. van beek , and j. b. bassingthwaighte , `` four methods to estimate the fractal dimension from self - affine signals , '' _ ieee eng . med .
biol . mag .
_ * 11 * , 5764 ( 1992 )
. j. beran , `` statistical methods for data with long - range dependence , '' _ statistical science _ * 7 * , 404427 ( 1992 ) .
h. g. e. hentschel and i. procaccia , `` the infinite number of generalized dimensions of fractals and strange attractors , '' _ physica d _ * 8 * , 435444 ( 1983 ) .
p. grassberger , `` generalized dimensions of strange attractors , '' _ phys .
lett . a _ * 97 * , 227 - 230 ( 1983 ) . j. theiler , `` estimating fractal dimension , '' _ j. opt
. soc . am . a _ * 7 * , 10551073 ( 1990 ) . k. matsuo , b. e. a. saleh , and m. c. teich , `` cascaded poisson processes , '' _ j. math .
* 23 * , 23532364 ( 1982 ) .
s. b. lowen and m. c. teich , `` estimating the dimension of a fractal point processes , '' _ proc .
spie _ * 2036 * ( chaos in biology and medicine ) , 6476 ( 1993 ) .
d. w. allan , `` statistics of atomic frequency standards , '' _ proc .
ieee_@xmath221 * 54 * , 221 - 230 ( 1966 ) .
j. a. barnes and d. w. allan , `` a statistical model of flicker noise , '' _ proc .
ieee _ * 54 * , 176178 ( 1966 ) .
m. c. teich , c. heneghan , s. b. lowen , and r. g. turcott , `` estimating the fractal exponent of point processes in biological systems using wavelet- and fourier - transform methods , '' in _ wavelets in medicine and biology _ , eds .
a. aldroubi and m. unser ( crc press , boca raton , fl , 1996 ) , pp .
383412 . c. heneghan , s. b. lowen , and m. c. teich , `` wavelet analysis for estimating the fractal properties of neural firing patterns , '' in _ computational neuroscience _
j. m. bower , ( academic press , san diego , 1996 ) , pp .
. s. b. lowen , `` refractoriness - modified doubly stochastic poisson point process , '' _ center for telecommunications research , technical report _ * 449 - 96 - 15 * ( columbia university , new york , 1996 ) .
s. b. lowen and m. c. teich , `` doubly stochastic poisson point process driven by fractal shot noise , '' _ phys .
a _ * 43 * , 41924215 ( 1991 ) . f. grneis and h .- j .
baiter , `` more detailed explication of a number fluctuation model generating @xmath46 pattern , '' _ physica _ * 136a * , 432452 ( 1986 ) .
f. a. haight , _ handbook of the poisson distribution _
( wiley , new york , 1967 ) .
b. b. mandelbrot , `` a fast fractional gaussian noise generator , '' _ water resources res . _
* 7 * , 543553 ( 1971 ) .
t. lundahl , w. j. ohley , s. m. kay , and r. siffert , `` fractional brownian motion : a maximum likelihood estimator and its application to image texture , '' _ ieee trans . med .
imag . _ * 5 * , 152161 ( 1986 ) .
m. a. stoksik , r. g. lane , and d. t. nguyen , `` accurate synthesis of fractional brownian motion using wavelets , '' _ electron .
lett . _ * 30 * , 383384 ( 1994 ) .
b. e. a. saleh , _ photoelectron statistics _ ( springer verlag , berlin , 1978 ) .
s. b. lowen and m. c. teich , `` generalised @xmath46 shot noise , '' _ electron .
lett . _ * 25 * , 10721074 ( 1989 ) . s. b. lowen and m. c. teich , `` fractal shot noise , '' _ phys .
lett . _ * 63 * , 17551759 ( 1989 ) . s. b. lowen and m. c. teich , `` power - law shot noise , '' _ ieee trans .
theory _ * 36 * , 13021318 ( 1990 ) .
d. r. cox , `` some statistical methods connected with series of events , '' _ j. roy .
b _ * 17 * , 129164 ( 1955 ) .
f. grneis , `` a number fluctuation model generating @xmath46 pattern , '' _ physica _ * 123a * , 149160 ( 1984 ) . f. grneis and t. musha , `` clustering poisson process and @xmath46 noise , '' _ jpn .
j. appl .
phys . _ * 25 * , 15041509 ( 1986 ) .
j. skilling , ed . , _ maximum entropy and bayesian methods _ ( kluwer , boston , 1988 ) . | fractal and fractal - rate stochastic point processes ( fspps and frspps ) provide useful models for describing a broad range of diverse phenomena , including electron transport in amorphous semiconductors , computer - network traffic , and sequences of neuronal action potentials .
a particularly useful statistic of these processes is the fractal exponent @xmath0 , which may be estimated for any fspp or frspp by using a variety of statistical methods .
simulated fspps and frspps consistently exhibit bias in this fractal exponent , however , rendering the study and analysis of these processes non - trivial . in this paper , we examine the synthesis and estimation of frspps by carrying out a systematic series of simulations for several different types of frspp over a range of design values for @xmath0 .
the discrepancy between the desired and achieved values of @xmath0 is shown to arise from finite data size and from the character of the point - process generation mechanism . in the context of point - process simulation ,
reduction of this discrepancy requires generating data sets with either a large number of points , or with low jitter in the generation of the points . in the context of fractal data analysis , the results presented here suggest caution when interpreting fractal exponents estimated from experimental data sets |
the recent wilkinson microwave anisotropy probe ( _ wmap _ ) first - year data release has shown that , despite advances in our ability to compute the power spectrum of microwave background anisotropies and increases in computer power , cosmological parameter estimation is still relatively slow and computationally expensive . for example , for the simplest parameter space of flat , lambda cold dark matter ( lcdm ) models , the _ wmap _ team needed about a week of supercomputer time ( 32 cpus on an sgi origin ) to determine the two parameters joint 2-@xmath0 confidence contours in the parameter space @xcite . allowing a non - flat geometry
slows down the analysis significantly and more than quadruples the time required to draw the 2-@xmath0 contours .
the main bottleneck in the parameter estimation process is the theoretical computation of the power spectrum anisotropy for a very large number of models , scattered throughout the parameter space .
this is mostly done using publicly available codes such as cmbfast ( @xcite ) or camb ( @xcite ) . to accelerate the calculations , ( dash)@xcite introduced a shortcut in the calculation of the angular power spectra , based on analytical and semi - analytical approximations .
rigorous analysis of future microwave background and other complementary data sets require recomputing the likelihood contours for cosmological parameters each time new data becomes available .
additionally for each new experiment , not one but many likelihood analyses are required to explore the impacts of potential systematic errors . not every researcher has access to months of supercomputer time for these computations ; a fast and reasonably accurate method to compute confidence regions for cosmological parameters is needed .
motivated by the enormous computational burden that a proper 3 to 4-@xmath0 exploration of the likelihood hypersurface represents and by the related need to understand more clearly the physical parameters that control the shape of the power spectrum , kosowsky , milosavljevic and jimenez ( kmj ) @xcite presented a new set of `` physical '' parameters formed from the usual cosmological parameters which have nearly - orthogonal effects on the microwave background temperature power spectrum .
thus , with this set of parameters , the efficiency of monte carlo techniques for evaluating the likelihood region increases significantly ; these parameters have been used in the wmap analysis @xcite and then incorporated into lewis suite of microwave background analysis codes @xcite .
subsequent work in a similar vein constructed a set of normal parameters that locally transformed the likelihood in parameter space into gaussian form@xcite . in the new kjm `` basis '' , it is also possible to obtain an efficient approximation of the temperature power spectrum in a region around the fiducial model as a simple linear extrapolation from a fiducial model power spectrum .
this new technique makes possible to compute approximate power spectra , several orders of magnitude faster than any previous methods , and was extensively used by the _ wmap _
collaboration @xcite in the preliminary analysis of the data .
despite a significant increase in computational speed and physical insight afforded by the kmj parameters , some remaining technical issues prevented a direct application of the kmj power spectrum fits to real data .
in particular : ( a ) the parameters and the linear fits were only computed for the temperature power spectrum and not for the polarization ; ( b ) the fiducial model used by kmj had an optical depth to the last scattering surface which was too small ( @xmath2 ) compared with the _ wmap _ best - fit model ( @xmath3 ) ; ( c ) for a fiducial model with higher @xmath4 some covariance remained between the slope of the power spectrum @xmath5 , the amplitude of the fluctuations @xmath6 and the optical depth parameter @xmath7 which suggested that the kmj parameters @xmath8 , @xmath9 and @xmath5 were not completely orthogonal ; ( d ) errors in the power spectrum estimation became larger than 1% in some relevant regions of the parameter space . here
we present a refined version of the kmj algorithm , cmbwarp , that solves these problems .
we extend the method to include polarization spectra , make all parameters nearly orthogonal , use a fiducial model given by the best fit model to _ wmapext _ data@xcite , and obtain the power spectra using polynomial fits for improved accuracy over linear extrapolations . throughout the region of parameter space within _ wmap
_ s first - year data 3-@xmath0 confidence region for each parameter , marginalized over the other n-1 parameters . ] , the accuracy is now comparable to the difference between cmbfast and other independent boltzmann codes .
we make cmbwarp publicly available from http:/www.physics.upenn.edu/@xmath1raulj / cmbwarp ; in addition , we also release accompanying markov chain code that can be used to perform parameter estimation .
we first review the kmj choice of physical parameters and describe the slightly modified set of parameters presented here which yield an orthogonal basis for temperature and polarization spectra .
our model space is the standard class of inflation - like cosmological models , specified by five parameters determining the background homogeneous spacetime ( matter density @xmath10 , radiation density @xmath11 , vacuum energy density @xmath12 , baryon density @xmath13 , and hubble parameter @xmath14 ) , two parameters determining the spectrum of primordial scalar perturbations ( scalar amplitude @xmath9 and power law index @xmath5 ) , and a single parameter @xmath4 describing the total optical depth since reionization . for the present
we postpone additional complications such as tensors , massive neutrinos or a varying vacuum equation of state . throughout this paper
, we use a fiducial cosmological model corresponding to wmap s best fit lcdm model with parameters : @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 . as it will be clear below
, the choice for @xmath9 is convention - dependent . in our convention @xmath22 which corresponds to @xmath23 in the fiducial model .
we will construct fast and accurate approximations to the microwave background power spectra in some region of parameter space surrounding this fiducial model .
the five parameters describing the background cosmology induce complex dependencies in the microwave background power spectra through multiple physical effects ; therefore their effects on the power spectra are not orthogonal .
a characteristic scale in the power spectrum is the angular scale of the first acoustic peak .
it is advantageous to choose this angular scale as a parameter which can be varied independently of any other parameters .
this angular scale is in turn determined by the ratio of the sound horizon at last scattering ( which determines the physical wavelength of the acoustic waves ) to the angular diameter distance to the surface of last scattering ( which determines the apparent angular size of this yardstick ) .
kmj used this quantity as the first of the following set of orthogonal _ physical _ parameters . the analytic theory underlying any such choice of physical parameters
has been worked out in detail ( see @xcite , and also @xcite ) .
the characteristic angular scale is @xmath24 where @xmath25 is the scale factor at recombination , the sound horizon is given by @xmath26 \nonumber \\
\!\!\!\!\!\!\int_0^a \!\!{dx \over \left\{\!\left(1\ ! + \ ! { 3\omega_b\over 4\omega_{\rm rad}}x\right)\ !
\left[(1-\omega)x^2 + \omega_\lambda x^4 \!+\ ! \omega_{\rm m } x\ ! + \ !
\omega_{\rm rad}\right]\right\ } ^{1/2 } } \label{rsa - integral}\end{aligned}\ ] ] and the angular diameter distance in a standard frw spacetime by @xmath27 with @xmath28 see , e.g. , @xcite .
the other physical parameters employed by kmj are : @xmath29 @xmath30 is the physical baryon density determining the baryon driving effect on the acoustic oscillations @xcite ( which modifies the relative odd and even acoustic peak heights ) ; @xmath31 is the matter - radiation density ratio at recombination ; @xmath32 determines the late - time integrated sachs - wolfe effect arising from a late vacuum - dominated phase , but otherwise varying @xmath33 while holding the other parameters fixed represents a nearly exact degeneracy ( sometimes called the `` geometrical degeneracy '' ) ; @xmath34 couples only to other small physical effects and is an approximate degeneracy direction .
this choice of parameters is not unique , but this specific set is particularly useful and easy to interpret . given values for @xmath35 , @xmath30 , @xmath32 , @xmath31 , and @xmath34 , they can be inverted to the corresponding cosmological parameters by rewriting the definition of @xmath35 in terms of @xmath30 , @xmath32 , @xmath31 , @xmath34 , and @xmath14 , then searching in @xmath14 until the desired value for @xmath35 is obtained ; @xmath13 and @xmath12 then follow immediately , while @xmath10 and @xmath11 can be obtained with a few iterations to determine a precise value for @xmath25 .
the other cosmological parameters which affect the microwave background power spectrum are the optical depth to reionization and the amplitude and power law index of the scalar perturbation power spectrum .
these turn out to be crucial parameters in order to make an approximate method like the one described here to work . for reionization
, kmj used the physical parameter @xmath36 the factor by which the microwave background anisotropies on small scales are damped due to compton scattering by free electrons after the universe is reionized .
this damping occurs for all scales smaller than the horizon size at reionization .
the primordial power spectrum of density perturbations is generally taken to be a power law , @xmath37 , with @xmath38 corresponding to the scale - invariant harrison - zeldovich spectrum . making the approximation that @xmath39 and @xmath40 have a direct correspondence , the effect of @xmath5 on the microwave background power spectrum can be modeled as @xmath41 which is a good approximation for power law power spectra . a first - order deviation from a power law , characterized by @xmath42 @xcite ,
can be represented in a similar way .
note that since the dependence is exponential , a linear extrapolation is never a good approximation over the entire interesting range @xmath43 .
the choice of @xmath44 is somewhat arbitrary : changing @xmath44 simply gives a different overall normalization . in this work
we choose @xmath45 as the best pivot point to reproduce the cmbfast power spectra .
these parameters have nearly orthogonal effects on the power spectra , so the effect of varying a particular parameter can be considered independently from the other parameters , within a reasonably large region of parameter space .
this property can be exploited to create simple and efficient approximations to the microwave background power spectra , based on a few computed power spectra .
the kmj paper chose a fiducial cosmological model described by a set of parameters @xmath46 and computed its power spectrum @xmath47 using available codes .
the effects of parameter @xmath48 were modeled as a simple rescaling of the angular multipole @xmath49 and the effects of @xmath50 as a overall normalization factor , the effects of @xmath5 as in equation [ n_effect ] above .
then for each of the remaining parameters @xmath51 , a linear fit was used to describe the variation of @xmath52 as @xmath51 moved away from its value in the fiducial model .
this set of numerical fits gives approximations to the power spectrum which are remarkably good in some region of parameter space surrounding the fiducial model .
however , the linear extrapolations are not good enough to compute the power spectrum to the accuracy of the cmbfast code used to compute the fiducial model and the dependencies on the individual parameters . here
we improve on a simple linear extrapolation by performing polynomial fits ( to the 4th order ) to the computed power spectra dependencies .
models extracted randomly from a mcmc for _ wmap _ first year data .
the solid line is for flat models while the dashed line is for universes with arbitrary geometry .
the top panel shows the percent mean difference averaged from @xmath53 to @xmath54 , while the bottom panel shows the percent rms difference . note that the agreement is at the 0.5% level or better . ]
but for the te power spectrum .
since te crosses zero at several points , the vertical axis is plotted in @xmath55 .
the dotted line in the bottom panel shows the cosmic variance error for the fiducial model . ] unfortunately , the kmj choice of physical parameters for @xmath56 and @xmath5 does not work if the fiducial model has a high @xmath4 .
even small deviations from the fiducial model result in strong covariances between the two parameters ( see further discussion in sec .
fortunately , a simple technique breaks the degeneracy between these two parameters : we use @xmath56 as a multiplicative factor to the whole power spectrum parameter for @xmath57 . for @xmath58 the shape of the spectrum does not depend on @xmath59 but only on @xmath5 as follows : @xmath60 where the @xmath61 subscript denotes the fiducial value .
we then fix the normalization for @xmath62 by imposing the condition that the spectrum must be continuous at @xmath63 . for the amplitude @xmath6 we use the same definition as in kmj . in the rest of the paper
, we refer to the parameters @xmath35 , @xmath64 , @xmath32 , @xmath31 , @xmath34 , @xmath6 , and @xmath56 as `` physical parameters '' , as opposed to the usual `` cosmological parameters ''
@xmath13 , @xmath65 , @xmath11 , @xmath66 , @xmath14 , @xmath67 , and a quadrupole - based normalization .
the set of approximation techniques for the physical parameters , comprising polynomial fits for @xmath48 , @xmath68 , @xmath33 , @xmath69 , and @xmath70 , a multiplicative @xmath8 dependence , and an approximate @xmath5 dependence given by eqs .
( [ n_effect ] ) and ( [ n_corr ] ) provide an extremely simple , fast , and accurate method of computing the temperature power spectrum of the microwave background .
slight additional modifications at low @xmath49 values described below provide excellent fits to the e - polarization and the te cross - power .
all of these approximations taken as a whole comprise the cmbwarp algorithm .
for the cross power spectrum between the temperature and the e - mode polarization , no modification is done for @xmath71 and the whole spectrum is modeled using the the same prescription for the temperature for @xmath72 but instead of the multiplicative factor @xmath8 we use the corresponding polynomial fit for @xmath8 .
this models accurately the te power spectrum neglecting secondary effects
. however , reionization of the neutral intergalactic medium by the first generation of stars re - scatters the cmb photons , producing the so called `` reionization bump '' at large angular scales .
we model this effect using the following fitting formula for @xmath73 , inspired by @xcite : @xmath74 where @xmath75 is the te @xmath76 value of the fiducial model for each @xmath77 , evaluated at @xmath78.\ ] ] the subscript @xmath61 denotes the fiducial model and @xmath79 denotes the integer part of @xmath80 ; we smoothly interpolate in the transition region .
this approximation is accurate to a few percent . for the e - polarization power spectra
, the power spectrum for all @xmath81 s is modeled using the same prescription as for the temperature power spectrum for @xmath72 but instead of the multiplicative factor @xmath8 we use the corresponding polynomial fit for @xmath8 .
the perturbative approach of cmbwarp greatly speeds up the calculation of cmb power spectra since it requires only a few floating point operations per multipole moment calculated .
computation time for a single non - flat lcdm model is on the order of _ 0.005 seconds _ for cmbwarp ; about 100 seconds for cmbfast on an athlon 2400 + workstation . in other words
, cmbwarp is about 20 thousand times faster than cmbfast .
we have computed our numerical fits up to @xmath81 of 1500 and tested thoroughly the performance of the fit in the same @xmath81 range .
we have checked that the same type of approximation also holds up to @xmath81 of 3500 . here
we present a direct application to data using _ wmap _ first year data and finer angular scale experiments which probe scales up to @xmath81 of 1500 . for cmbwarp to be useful in the analysis of present and forthcoming high - precision data sets
, it must be not only fast but also accurate .
two key questions are : ( a ) can the approximate approach reproduce , with reasonable accuracy , cmbfast power spectra ? and ( b ) can it recover unbiased and accurate cosmological parameters and confidence regions in a likelihood analysis for a realistic data set ? we demonstrate below that the answer is yes .
the first test we perform is to compare power spectra from cmbfast and cmbwarp for a large set of models . for the comparison set , we select models from converged markov chain monte carlos for the _ wmap _ one - year data for both flat and non - flat lcdm models .
the chains are run using cmbfast ( version 4.4 using the high precision option ) and the @xmath82 are saved ; we select models with parameters within ( roughly ) the @xmath83 marginalized confidence level region for each parameter .
we then calculate power spectra for cmbwarp and compare the two sets of @xmath82 . for tt , te and ee power spectra we present both the mean difference and the root - mean - square between cmbfast and cmbwarp , as a function of multipole @xmath49 .
the first quantity is an estimate of the possible bias introduced by the perturbative approximation , while the latter is an estimate of the scatter .
a comparison between cmbfast and boltzmann codes @xcite for one cosmological model ( the standard lcdm model ) reveals differences at the 0.1% level in the temperature power spectrum and the 1% level in the polarization power spectrum .
figure [ fig : ttrms ] shows this comparison for @xmath84 flat ( solid line ) and @xmath84 open ( dashed line ) models for the temperature power spectrum .
the average difference is typically 0.3% : of the same order of magnitude of the difference between cmbfast and boltzmann codes .
the rms plot shows not only that cmbwarp is unbiased , but that the rms scatter is only at the 0.5% level .
as we demonstrate below , this level of accuracy is sufficient for accurate recovery of cosmological parameters from current and future cmb temperature experiments
. the te power spectrum crosses zero at several values of @xmath49 , so a fractional comparison is not possible .
instead we compute the residual difference in @xmath85k@xmath86 between the two models . to be more specific we compute @xmath87 ( rms ) and @xmath88 ( bias ) .
figure [ fig : terms ] shows the results .
the discrepancy is largest at large @xmath40 , and is of the order of a 1 - 2% percent .
these discrepancies are always small compared to the cosmic variance @xmath89 , as shown in the figure . and @xmath4 using _ wmap _ tt + te
. the dashed line is for cmbfast while the solid line is for cmbwarp . ] finally , fig .
[ fig : eerms ] shows the comparison between cmbfast and cmbwarp for the ee power spectrum . for @xmath90
the ee power spectrum is very close to zero , thus for @xmath91 we do a residual comparison as for te .
the discrepancy with cmbfast is only few % of a @xmath92 and significantly below cosmic variance . for @xmath93
the ee power spectrum is not close to zero and we show a fractional comparison .
the agreement with cmbfast is at the 1 - 2% level , similar to that between cmbfast and boltzmann codes .
we now turn our attention to the determination of cosmological parameters and their errors . given the satisfactory comparison between our approximate power spectra and those from cmbfast
, we need to demonstrate that when we determine cosmological parameters from real measurements , cmbwarp yields the same parameter values and confidence intervals as cmbfast . for this purpose
we have calculated likelihood regions for two data sets : first the _ wmap _ one - year temperature power spectrum data @xcite , second the _ wmap _ one - year temperature and te power spectra @xcite complemented by the acbar data points @xcite extending to @xmath94 .
the above tests are performed for flat and open cosmologies . in the likelihood analysis we use the full covariance matrix @xcite and we explore the likelihood surface with a markov chain ( see @xcite for more details ) .
the procedure is effectively equivalent to that illustrated in @xcite , but with cmbwarp replacing cmbfast .
a markov chain with @xmath95 models is now computed in 1.5 hours on a single intel 3.2 ghz cpu , the computational bottleneck in the calculation now being the evaluation of the likelihoods and not the computation of the power spectra .
figure [ fig : wmaptt1d ] shows the marginalized one - dimensional likelihoods obtained from chains computed with cmbfast ( dashed line ) and cmbwarp ( solid line ) , for _ wmap _ 1yr tt data only and for a flat model .
the recovered best - fit parameters display no significant bias ( the parameter shift is always less that @xmath96 ) , the 1-@xmath0 error is recovered to within @xmath97 accuracy or better , and the 2-@xmath0 error is recovered to 20% .
note that despite the fact that @xmath4 is effectively unconstrained when only temperature data are used , cmbwarp still recovers the same shape of the likelihood for @xmath4 .
in the second test including the small - scale temperature measurements and the _ wmap _ one - year te data , the extra polarization data help to greatly constrain the integrated optical depth to reionization .
once again , chains of flat models generated with cmbfast and cmbwarp return virtually the same cosmological parameters as can be seen in fig .
[ fig : wmapttte1d ] . finally , fig .
[ fig : wmapopen ] compares the recovered values of @xmath4 and @xmath14 when the model is not constrained to be flat .
the set of approximations described here , cmbwarp , comprise a highly accurate and extremely fast method of approximating the microwave background temperature and polarization power spectra in standard inflation - like cosmological models for the range of cosmological parameters which encompass the likelihood regions allowed by the _ wmap _ first - year data . for 8 parameters , cmbwarp computes each multipole moment with @xmath1 50 floating - point operations ; it is unlikely that any significantly more efficient computational scheme can be constructed .
note that the computational effort is the same for all multipole moments independent of @xmath49 ; conventional codes require large increases in computational cost as larger-@xmath49 multipoles are computed .
this fact will become more important in the coming years as experiments such as planck , act @xcite and the south pole telescope map the microwave background at high resolution .
cmbwarp is particularly well - suited to explorations of cosmological parameter space using markov chain techniques . as cosmological data continues to improve , joint analysis combining
many different data sets will continue .
the technique allows easy incorporation of the true microwave background constraints into joint analysis with a minimum of computational expense . a useful extension of cmbwarp would be a comparably fast and accurate computation of the transfer function and matter power spectrum .
the optimal set of `` physical '' parameters for describing the matter power spectrum will not necessarily be the same as for the microwave background ; the crucial property of a set of parameters is that they must have orthogonal effects on the matter power spectrum .
this will be addressed elsewhere .
the speed and flexibility of cmbwarp make possible analysis with many more parameters which are not computationally feasible using standard techniques : one example is estimation of the primordial power spectrum using a number of parameters to describe it ( see , e.g. , @xcite ) , taking into account the effects of the standard cosmological parameters .
markov chain monte carlo explorations of cosmological parameter space have now become common .
the need for very fast , accurate , and portable microwave background power spectrum computations in this context is readily apparent .
we hope cmbwarp proves to be a useful tool for these analyses .
the work of rj is partially supported by nsf grant ast-0206031 .
we thank p. g. castro and andy taylor for useful discussions .
rj and lv thank the `` centro de investigaciones de astronomia '' ( cida ) and the university of edinburgh for hospitality while part of this work was carried out .
ak is a cottrell scholar of the research corporation .
we acknowledge the use of the legacy archive for microwave background data analysis ( lambda ) .
support for lambda is provided by the nasa office of space science . | we improve the algorithm of kosowsky , milosavljevic , and jimenez ( 2002 ) for computing power spectra of the cosmic microwave background .
the present algorithm computes not only the temperature power spectrum but also the e - mode polarization and the temperature - polarization cross power spectra , providing the accuracy required for current cosmological parameter estimation .
we refine the optimum set of cosmological parameters for computing the power spectra as perturbations around a fiducial model , leading to an accuracy better than 0.5% for the temperature power spectrum throughout the region of parameter space within _ wmap _ s first - year 3@xmath0 confidence region .
this accuracy is comparable to the difference between the widely - used cmbfast code ( seljak and zaldarriaga 1996 ) and boltzmann codes .
our algorithm ( cmbwarp ) makes possible a full exploration of the likelihood region for eight cosmological parameters _ in about one hour on a laptop computer_. we provide the code to compute power spectra as well as the markov chain monte carlo algorithm for cosmological parameters estimation at http://www.physics.upenn.edu/@xmath1raulj/cmbwarp . |
radiative corrections to the @xmath4 ( @xmath5 ) vertex have long received considerable attention . apart from its sensitivity to new physics effects , this vertex has theoretical interest by its own as it may serve as a probe of the gauge sector of the standard model ( sm ) . in this context , the one - loop contributions to the @xmath6 vertex , which defines the static electromagnetic properties of the @xmath7 boson have been calculated in the sm @xcite and several of is extensions such as two - higgs doublet models ( thdms ) @xcite , supersymmetric models @xcite , 331 models @xcite , etc .
similar attention has been paid to the study of the @xmath8 vertex . despite the attention
has focused mainly on the static properties of the @xmath7 boson , a more comprehensive study of the @xmath7 boson properties requires the analysis of the off - shell @xmath4 vertex , particularly when the neutral @xmath9 boson is off - shell and the @xmath7 bosons are on - shell as the resulting vertex could be tested with high precision at the cern large hadron collider ( lhc ) or the planned future linear colliders .
it is well known that although the on - shell @xmath4 vertex renders a gauge independent amplitude by itself , gauge independence is lost once any external particle goes off - shell , thereby requiring the use of a nonconventional calculation scheme .
the difficulties to obtain a gauge independent green function for the @xmath0 vertex have long been known . in particular
, a careless calculation of the radiative corrections to the @xmath0 vertex via a conventional gauge - fixing procedure irremediably leads to an ill - behaved gauge dependent green function .
for instance , one - loop corrections to the @xmath0 vertex were first calculated in the sm by the authors of ref .
@xcite using the feynman-t hooft gauge via a conventional gauge - fixing scheme .
the result was an infrared divergent green function that violates unitarity .
aware of the fact that an off - shell green function is generally gauge dependent , authors of ref .
@xcite invoked a diagrammatic method known as the pinch technique ( pt ) @xcite to obtain a gauge independent @xmath0 green function satisfying the requirements of gauge invariance and infrared finiteness .
the pt exploits the fact that although the off - shell green functions are gauge dependent , the @xmath10-matrix elements to which they contribute are gauge independent . in this way ,
order by order in perturbation theory , one can construct a gauge independent green function for an off - shell vertex by combining its contribution with all other feynman diagrams that also contribute to a particular physical process , thereby getting rid of any @xmath11 dependent terms . in this respect , while the @xmath0 vertex does not represent a physical process by its own , it can contribute to a gauge invariant physical process such as @xmath12 scattering . by this method , a gauge invariant and gauge independent @xmath0 green function can be constructed .
the pt was also used in ref .
@xcite to calculate the one - loop contributions to the @xmath0 vertex from the unconstrained minimal supersymmetric standard model ( mssm ) .
the pt is thus a diagrammatic method that allows one to remove any unphysical gauge dependent term from an off - shell @xmath13-point function at the level of the feynman graphs contributing to a certain physical process , thereby yielding a well - behaved gauge - independent and gauge - invariant green function .
the pt may turn itself , however , into a somewhat cumbersome task .
there is also the alternative of removing any gauge dependent term at the level of the generating functional . in this respect , although the background field method ( bfm ) @xcite renders a gauge invariant quantum action , no mechanism has been found yet that allows one to obtain both gauge - invariant and gauge - independent green functions from the generating functional .
eventually , a connection between the pt and the bfm has been found @xcite : it turns out that the gauge - invariant and gauge - independent green function obtained via the pt is exactly reproduced if it is calculated through the bfm using the feynman rules given in the feynman-t hooft gauge ( @xmath14 ) .
although such a correspondence , which so far remains a puzzle , was first established at the one - loop level @xcite , it has been shown that it persists at all orders of perturbation theory @xcite . therefore , instead of using the pt
, one can use the bfm feynman-t hooft gauge ( bfmfg ) to calculate gauge - independent off - shell amplitudes .
along these lines , in ref .
@xcite we calculated the one - loop contributions to the @xmath0 vertex from the gauge sector of the so called minimal 331 model @xcite . in order to construct a well - behaved gauge independent green function , rather than using the pt , we invoked an alternative method inspired in the bfm and brst symmetry @xcite , which is well suited to study the sensitivity of the @xmath0 vertex to the virtual effects of the new gauge bosons predicted by 331 models .
we will follow a similar approach here to calculate the one - loop contributions from the 331 model with right - handed neutrinos @xcite .
although we will shortly describe the calculation scheme below , for more details we refer the reader to ref .
@xcite .
331 models have been the source of great interest lately .
recent studies within the framework of this class of theories have focused on neutrino physics @xcite , @xmath15 physics @xcite , higgs boson physics @xcite , bilepton physics @xcite , supersymmetric extensions @xcite , dark matter @xcite , cp violation @xcite , and theoretical aspects @xcite .
models of this kind are based on the @xmath16 gauge symmetry @xcite and are unique in the sense that they require that all the 3 fermion families be summed up in order to cancel anomalies , in contrast with other models in which anomaly cancelation is achieved family by family . as a consequence ,
331 models require that the number of fermion families be a multiple of 3 , the quark color number .
this may suggest a solution to the family replication problem .
apart from this feature , 331 models are interesting as they predict a pair of massive gauge bosons arranged in a doublet of the electroweak group .
while the minimal 331 model predicts a pair of singly - charged @xmath17 and a pair of doubly - charged @xmath18 gauge bosons , the model with right - handed neutrinos predicts a pair of neutral no self - conjugate gauge bosons @xmath19 instead of the doubly - charged ones .
these new gauge bosons are called bileptons since they carry two units of lepton number , thereby being responsible for lepton - number violating interactions @xcite .
the neutral bilepton has been deemed a promising candidate in accelerator experiments since it may be the source of neutrino oscillations @xcite .
the reason why the effects of the bileptons on the @xmath0 vertex are worth studying is because their couplings to the sm gauge bosons are rather similar to the sm gauge boson couplings .
current bounds put the bilepton masses in the range of a few hundred gevs @xcite , which means that the bileptons may show up via their virtual effects in low - energy observables .
this is an important reason to investigate the effect of these particles on the @xmath0 vertex .
furthermore , due to the spontaneous symmetry breaking ( ssb ) hierarchy , the splitting between the charged and neutral bilepton masses is bounded by @xmath20 , so the bilepton masses might be almost degenerate since they are expected to be heavier than @xmath20 .
we will see below that this is particularly suited for our calculation method .
the rest of the paper is organized as follows . in sec .
ii we present a brief introduction to the 331 model with right - handed neutrinos .
a survey of the quantization method is presented in sec .
iii along with a detailed discussion on the gauge fixing procedure for the bileptons , whereas sec .
iv is devoted to the analytical results and the analysis .
finally , the conclusions are presented in sec . v.
the 331 model with right - handed neutrinos was first introduced in ref .
@xcite . in a previous paper , we worked in the context of this model and calculated the static electromagnetic properties of the @xmath7 boson @xcite . also , the one - loop contributions to the static electromagnetic properties of the neutral no self - conjugate @xmath19 boson were calculated in ref .
it was shown that the main contributions to the @xmath6 vertex arise from the gauge sector , i.e. from the bileptons , as the fermion sector does not contribute and the higgs sector gives a negligibly contribution quite similar to that arisen in a thdm @xcite . therefore , although for completeness we will present a short description of the main features of the model , the following discussion will be mainly focused on the gauge sector as it is the one which is expected to give the dominant contributions to the @xmath0 vertex .
furthermore , we will see that our calculation scheme is suited to analyze the effects of the bileptons on the @xmath0 green function . in the fermion sector of the 331 model with right - handed neutrinos ,
leptons of the same generation are arranged in left - handed triplets and right - handed singlets , the same being true for each quark generation .
apart from the introduction of right - handed neutrinos , there are three new quarks @xmath21 , @xmath22 , and @xmath23 , which have the following electric charge @xmath24 and @xmath25 . in order to cancel the @xmath26 anomaly ,
two quark families transform as @xmath26 antitriplets and the remaining one as a triplet . at the first stage of ssb , when the @xmath27 group is spontaneously broken into @xmath3 ,
the new quarks get their masses and emerge as singlets of the electroweak group
. consequently , they can not interact with the @xmath7 boson at the tree level .
it follows that there is no contribution from the new quarks to the @xmath0 vertex at the one - loop level .
as far as the scalar sector is concerned , several higgs sectors have been proposed in the literature to achieve the ssb in 331 models @xcite . as for the 331 model with right - handed neutrinos
, it requires the simplest higgs sector of this class of theories @xcite , i.e. only three triplets of @xmath26 are required to reproduce the sm physics at the fermi scale : @xmath28 where @xmath29 , @xmath30 and @xmath31 are @xmath3 doublets with hypercharge -1 , 1 and -1 respectively .
the vacuum expectation values ( vev ) are @xmath32 , @xmath33 , and @xmath34 .
the triplet @xmath35 breaks the @xmath27 group into @xmath3 at the @xmath36 scale , whereas @xmath37 and @xmath38 are meant to break @xmath3 into @xmath39 . in the gauge sector
, the model predicts the existence of five new gauge bosons : two singly charged @xmath17 , two neutral no self - conjugate @xmath19 , and a neutral self - conjugate @xmath15 .
all these new gauge bosons and the new quarks as well acquire their masses at the @xmath36 scale . at this stage of ssb ,
the @xmath15 boson emerges as a singlet of the electroweak group and so no interaction between the @xmath15 boson and the @xmath7 boson arises .
however , in the following stage of ssb , at the fermi scale , the @xmath15 boson couples with the @xmath7 boson via the @xmath40 mixing angle . as a consequence , the @xmath15 contribution to the @xmath0 vertex will be proportional to the square of the @xmath40 mixing angle , which is highly suppressed according to the most recent experimental bounds @xcite .
we will thus ignore this contribution in this work .
in sharp contrast , the bileptons arise as a doublet of the electroweak group at the @xmath36 scale , which means that they couple to the sm gauge bosons in a rather peculiar way : due to the fact that the @xmath41 group is completely embedded in @xmath26 , the bileptons couplings to the sm gauge bosons have similar strength and lorentz structure than those of the sm gauge boson couplings themselves . these new interactions , which arise solely from the yang - mills sector , do not involve any mixing angle and are entirely determined by the @xmath41 coupling constant and the weak angle @xmath42 . in the first stage of ssb ,
the bilepton masses are degenerate : @xmath43 .
however , when @xmath3 is broken down to @xmath39 , the bilepton masses degeneration is also broken .
once the higgs kinetic - energy sector is diagonalized , there emerge the following mass - eigenstate fields : @xmath44 with masses @xmath45 , @xmath46 , and @xmath47 . the remaining three gauge fields @xmath48 , @xmath49,and @xmath50 define the self - conjugate mass eigenstates . from the above expressions ,
it is straightforward to obtain the following upper bound on the bilepton masses splitting @xcite : @xmath51 therefore , the bilepton masses become nearly degenerate when they are much larger than @xmath52 . while in the minimal 331 model an upper bound on the bilepton masses of the order of 1 tev can be derived from the theoretical constraint @xmath53 @xcite , which is obtained from matching the gauge couplings constants at the @xmath54 breaking scale , in the version with right - handed neutrinos the corresponding bound is highly relaxed as the theoretical constraint is @xmath55 @xcite .
the yang - mills sector induces all the couplings we need to compute the bilepton contribution to the @xmath0 vertex , so we will devote particular attention to it .
the respective lagrangian can be written as @xmath56 where @xmath57 and @xmath58 , being @xmath59 the structure constants associated with @xmath26 .
when the mass eigenstate fields are introduced , the above lagrangian can be split into three @xmath3-invariant terms @xcite : @xmath60 while @xmath61 stands for the usual sm yang - mills lagrangian , @xmath62 comprises the interactions between the new gauge bosons and the sm ones .
the latter lagrangian is the only one relevant for the present discussion and we will get back to it later . as for the last term , @xmath63 , it induces the couplings of the @xmath15 boson to the bileptons . at the one - loop level
there are no contributions to the @xmath0 vertex induced by this lagrangian . in order to calculate the one - loop correction to the @xmath0 vertex
, we will introduce a @xmath3-covariant gauge - fixing procedure for the bileptons .
this will be discussed in the following section .
we now turn to present an overview of the quantization method used to obtain a gauge invariant and gauge independent green function for the @xmath0 vertex in the 331 model with right - handed neutrinos .
this method , which is inspired in the bfm and brst formalism , was comprehensively discussed in our previous work @xcite and we refrain from presenting a detailed discussion here . the bfm @xcite is a powerful tool that allows one to construct a gauge invariant quantum action out of which gauge invariant green functions can be obtained that are free of pathologies and satisfy simple ward identities . in a conventional quantization formalism , all the fields appearing in the action are quantized . in the bfm
, the gauge fields are decomposed into a quantum field and a classical ( background ) field .
while the quantum fields are integrated out , the background fields are treated as sources . as a result
, the quantum fields can only appear as internal lines in loop diagrams , whereas the background fields appear as external lines . in principle , it is necessary to gauge - fix both the quantum and the classical fields in order to define @xmath10-matrix elements , but it is enough to gauge - fix the former to obtain off - shell green functions . in this respect , while gauge invariance with respect to the quantum fields is broken when they are gauge - fixed , this process leaves unaltered the gauge invariance with respect to the classical fields .
the resulting green functions will be gauge invariant but they will still maintain the dependence on the gauge parameter that characterizes the gauge - fixing scheme used for the quantum fields .
however , we can exploit the connection between the bfm and the pt to obtain a gauge invariant and gauge independent green function .
we just need to use the bfm feynman rules given in the feynman-t hooft gauge .
we will see below that our quantization method incorporates some features of the bfm .
although gauge invariance with respect to the full theory is broken when a conventional quantization scheme is applied , one can still preserve the gauge invariance under a subgroup of the theory .
this approach is well suited when the virtual effects of the gauge fields associated with the subgroup are expected to be considerably small . in this context ,
following the philosophy of the effective lagrangian approach , where a @xmath3 effective lagrangian is constructed out of the light ( sm ) fields to assess the effects of the heavy fields on a low - energy observable , it would be convenient to analyze the virtual effects of the bileptons on the @xmath0 green function in a @xmath3-covariant manner .
a quantization scheme for the bileptons would only be required since the sm gauge fields would only appear as external lines .
thus , a @xmath3-invariant effective lagrangian can be constructed by introducing a @xmath3-covariant gauge - fixing procedure for the bileptons , which then can be integrated out .
the gauge - fixing procedure must involve the @xmath3-covariant derivative given in the representation in which the heavy fields transform under this group , which is the reason why such gauges are known as nonlinear or covariant .
this class of gauges have proved to be very useful for the calculation of radiative corrections in the sm and beyond @xcite .
since we want to introduce a nonlinear gauge fixing procedure for the bileptons , we need to be careful as the difficulties to implement the faddeev - popov method ( fpm ) @xcite in such a case have long been known .
more specifically , it is known that renormalizability becomes ruined when the fpm is applied to a nonlinear gauge .
we can invoke instead the brst formalism @xcite to construct the most general renormalizable nonlinear gauge - fixing term @xcite .
we will thus introduce a @xmath3-covariant gauge - fixing term for the bileptons , which will lead to an invariant quantum action out of which a gauge invariant @xmath0 green function will be obtained . invoking the connection between the pt and the bfm
, the use of the feynman-t hooft feynman rules will render a gauge independent green function .
finally , we would like to point out that both our quantization method and the bfm are meant to construct gauge - invariant quantum actions .
the main difference resides in the fact that while the bfm allows one to analyze any new physics effects regardless the energy scale , ours is only appropriate to study heavy physics effects on low - energy ( sm ) green functions .
this stems from the fact that while in the bfm gauge - invariance would be preserved with respect to the gauge group of a complete theory , in our quantization method there is only invariance under a subgroup of such a theory . in the case of the present paper ,
our quantization method preserves gauge invariance under the @xmath3 group rather than @xmath27 .
while the calculation will be greatly simplified because of electroweak invariance , the price to be paid is that our result will only be approximate as it will only account for the bilepton effects on the @xmath0 vertex at the @xmath36 scale , when the bilepton masses are still degenerate .
we will see below that this is indeed a good approximation .
the procedure to gauge - fix the yang - mills lagrangian of the 331 model with right - handed neutrinos is similar to that described in the case of the minimal 331 model .
so , we refrain from presenting a detailed discussion here and refer the reader to ref .
@xcite . after introducing the most general action for a yang - mills system consistent with brst symmetry and renormalization theory @xcite
, we integrate out the auxiliary fields to obtain an action defined by the following gauge - fixed @xmath64-invariant lagrangian @xmath65 where @xmath66 and @xmath67 are the gauge - fixing term and the ghost sector lagrangian , respectively . as for the yang - mills lagrangian @xmath68 , which is given in eq .
( [ l_ym ] ) , for our calculation we only need the @xmath62 term as it is the only one that induces the interactions between the bileptons and the sm gauge bosons .
it can be expressed as @xmath69 where @xmath70 is a doublet of the electroweak group with hypercharge @xmath71 , and @xmath72 is the electroweak covariant derivative .
we have also introduced the definitions @xmath73 , @xmath74 , and @xmath75 , with @xmath76 the pauli matrices . as for the gauge - fixing term @xmath66 and the ghost sector lagrangian @xmath67 , they can be written as @xmath77 and @xmath78 where @xmath79 are the gauge - fixing functions and @xmath80 stands for the antighost fields . in addition , @xmath81 are the ghost fields associated with the @xmath82 fields , @xmath59 are the @xmath26 structure constants , and @xmath11 is the gauge parameter .
according to ref .
@xcite , the most general @xmath3-covariant gauge - fixing functions @xmath79 consistent with renormalization theory is given by : @xmath83 we can now insert this expression into the gauge - fixed lagrangian . apart from analyzing the dynamics induced by each term of the gauge - fixed lagrangian , we would like to put special emphasis on the covariance under @xmath3 .
the covariant structure of the gauge - fixing term becomes manifest when the mass eigenstates @xmath84 and @xmath85 are introduced in an @xmath3 doublet @xmath86 we will now decompose the gauge - fixing lagrangian into three terms in order to analyze the dynamics it induces : @xmath87 where each term is separately @xmath3 invariant : @xmath88\nonumber \\ % & & -\frac{g^2\sqrt{3}\sqrt{1 - 4s^2_w}}{2\sqrt{2}c_w}z'_\mu(\chi^{'0*}y^{\mu % \dag}\phi_y+\chi^{'0 } \phi^\dag_y y^\mu)\nonumber \\ % & & -\frac{3g^2(1 - 4s^2_w)}{4c^2_w\xi}z'_\mu z'_\nu y^{\mu % \dag}y^\nu.\end{aligned}\ ] ] whereas the third term , denoted by @xmath89 , involves the @xmath15 boson and is irrelevant here .
we note that the first term of @xmath90 defines the bilepton propagators and gives new contributions to the trilinear and quartic couplings involving the bileptons and the sm gauge bosons , which originally arise from the @xmath62 lagrangian .
after some algebra , we are lead to the feynman rules for these modified couplings , which are shown in fig .
the lorentz structure associated with the trilinear couplings is given by @xmath91 whereas those of the quartic couplings are @xmath92 and @xmath93g_{\alpha\mu}g_{\beta\nu } + \left[\left(1+\xi^{-1}\right)q^v_{y^-}- 2q^v_{y^{0}}\right]g_{\alpha\nu}g_{\beta\mu}.\ ] ] the reason why the trilinear couplings @xmath94 and @xmath95 have the same lorentz structure is a consequence of @xmath3 invariance .
this is also reflected in the fact that a simple ward identity is satisfied by these vertices : @xmath96 where @xmath97 is the two - point vertex function @xmath98 feynman rules for the trilinear and quartic vertices involving the bileptons and sm gauge fields in the @xmath3-covariant @xmath99-gauge .
@xmath100 , @xmath101 , and @xmath102 for @xmath1 , whereas @xmath103 , @xmath104 , and @xmath105 for @xmath106 .
in addition , @xmath107 .
( [ gamn])-([gwvyy ] ) for the lorentz structures.,width=566 ] note that the unphysical masses for the pseudo - goldstone bosons associated with the bileptons are determined by the scalar part of @xmath90 , which also modifies various unphysical couplings arising from the higgs potential . as for @xmath108
, it allows one to remove various unphysical vertices that arise from the @xmath35 kinetic - energy sector @xmath109 , which can be written as : @xmath110 where each term is @xmath3-invariant by itself : @xmath111 @xmath112 once again , the terms denoted by @xmath89 are irrelevant for our calculation as they involve the @xmath15 boson . from these expressions
it is clear that @xmath113 induces the couplings of the pseudo - goldstone bosons to the sm gauge bosons .
the respective feynman rules for these coupling can be extracted straightforwardly and are presented in fig .
[ sfr ] . as for @xmath114
, it induces the bilinear terms @xmath115 and @xmath116 , together with the unphysical trilinear couplings @xmath117 , @xmath118 , and some quartic vertices irrelevant for our calculation .
it turns out that all these couplings exactly cancel when @xmath119 and @xmath120 are combined : @xmath121 where @xmath89 stands for irrelevant surface terms .
this is the reason why this gauge - fixing procedure will simplify considerably our calculation as we can get rid of several feynman diagrams involving the couplings @xmath117 and @xmath118 .
finally , we would like to show the covariant structure of the ghost sector and extract the feynman rules necessary for our calculation .
we introduce the mass eigenstates in @xmath3 doublets : @xmath122 where the mass eigenstates are defined as @xmath123 and @xmath124 , and similar expressions for the antighost field @xmath125 .
the @xmath67 lagrangian can thus be written as follows : @xmath126\nonumber \\ & & + \frac{ig}{\sqrt{2}}y^\dag_\mu m_cd^\mu \bar{c}_y+\frac{ig}{2}y^\dag_\mu { \cal m}_c\bar{c}_y-\frac{\xi g}{2}\big[\chi^{'0*}\chi^{'0}c^\dag_y\bar{c}_y+\chi^{'0}\phi^\dag_ym_c\bar{c}_y -(c^\dag_y\phi_y)(\phi^\dag_y\bar{c}_y)\big]\nonumber \\ & & + \frac{i\sqrt{2}}{\xi}\big[(\bar{m}_cc_y+m_c\bar{c}_y)^\dag ( d_\mu y^\mu)-(d_\mu y^\mu)(\bar{m}_cc_y+m_c\bar{c}_y)\big]\nonumber \\ & & -g\big[\phi^\dag_y(\bar{m}_cc_y+m_c\bar{c}_y)\chi^{'0}+\chi^{'0*}(\bar{m}_cc_y+m_c\bar{c}_y)^\dag \phi_y\big]+{\rm h.c.}\nonumber \\ & & -\frac{1}{2}f^{\bar{a}bc}f^{cde}\bar{c}^{\bar{a}}\bar{c}^bc^dc^e,\end{aligned}\ ] ] where @xmath127
@xmath128 where @xmath129 , @xmath130 is the covariant derivative given in the adjoint representation of @xmath26 , and @xmath131 is obtained from @xmath132 after replacing the ghost fields by the antighost fields . under a @xmath64 unitary transformation @xmath133 ,
@xmath132 transforms as @xmath134 .
the same is true for @xmath131 and @xmath135 .
it is thus clear that @xmath67 is @xmath3 invariant . from @xmath67 ,
it is straightforward to show that the feynman rules for the trilinear and quartic couplings involving the ghost fields and the sm gauge bosons have the same lorentz structure than those involving the pseudo - goldstone bosons and the sm gauge bosons , which stems from the fact that each sector is @xmath64 invariant by itself .
the feynman rules for the ghost ( antighost ) fields are summarized in fig .
[ sfr ] together with the feynman rules for the pseudo - goldstone bosons .
we see that the trilinear vertices @xmath136 and @xmath137 , with @xmath10 standing for a commutative ( pseudo - goldstone boson ) or anticommutative ( ghost ) charged scalar , satisfy simple ward identities : @xmath138 where @xmath139 and @xmath140 stands for the two - point vertex functions @xmath141 .
feynman rules for the trilinear and quartic vertices involving sm gauge fields and scalar unphysical particles ( pseudo - goldstone bosons and ghosts ) in the @xmath64-covariant @xmath99-gauge . in this gauge , the @xmath7 and @xmath9 couplings to pseudo - goldstone bosons and ghosts coincide.,width=528,height=336 ]
the most general transverse cp - even vertex function for the @xmath142 coupling has the form @xcite @xmath143 + 2\delta \kappa_v ( q_\beta g_{\alpha \mu}-q_\alpha g_{\beta \mu } ) + \frac{4\delta q_v}{m^2_w}\big(p_\mu q_\alpha q_\beta-\frac{1}{2}q^2p_\mu g_{\alpha \beta}\big)\big\}.\end{aligned}\ ] ] the longitudinal terms were ignored as they become negligibly small when @xmath144 couples to a light fermion current , as in @xmath145 scattering .
the sm tree - level values are @xmath146 and @xmath147 .
while @xmath148 , and @xmath149 are ultraviolet finite at the one - loop level , @xmath150 is divergent and requires renormalization .
the coefficients @xmath151 and @xmath152 , obtained from the @xmath6 vertex function , determine the static electromagnetic properties of the @xmath7 boson , namely , its magnetic dipole moment @xmath153 and its electric quadrupole moment @xmath154 . from the feynman rules presented in figs . 1 and 2
, we can construct all the feynman diagrams contributing to the @xmath4 vertex at the one - loop level , which we show in fig . 3 , and extract the @xmath155 and @xmath156 coefficients .
although there is a similar set of feynman diagrams for the pseudo - goldstone bosons and the ghost fields , the unphysical particles only contribute to eq .
( [ g_wwv ] ) via the triangle diagram .
moreover , it turns out that the ghost ( antighost ) contribution is minus twice that of the pseudo - goldstone bosons , which is due to the separate @xmath64 invariance of these sectors .
we have evaluated the loop amplitudes in the feynman-t hooft gauge via the passarino - veltman reduction method @xcite .
we have verified that the bilepton contribution can be written exactly as eq .
( [ g_wwv ] ) , which means the @xmath0 green function satisfies the ward identity @xmath157 .
it turns out that the form factors are the same regardless of @xmath9 , which means that , just as occurs at the tree - level , @xmath158 and @xmath159 only differ by the @xmath160 coefficient .
thus @xmath3 invariance is preserved at one - loop level .
the @xmath160 coefficient is given by @xmath161 , which , after inserting the charge values , gives @xmath162 for @xmath1 and @xmath163 for @xmath106 .
we now introduce the following short - hand notation @xmath164 , @xmath165 , @xmath166 , @xmath167 , @xmath168 , and express the @xmath169 and @xmath156 coefficients in terms of two- and three - point passarino - veltman scalar functions : @xmath170\end{aligned}\ ] ] @xmath171,\end{aligned}\ ] ] where @xmath172 and @xmath173 , with the scalar two - point @xmath174 and three - point @xmath175 functions given in the usual notation . from here , it is clear that ultraviolet divergences cancel up in the @xmath7 form factors . before the numerical evaluation
, it is interesting to analyze the static electromagnetic form factors , i.e. the scenario when @xmath176 .
note that @xmath177 and @xmath178 vanish when @xmath176 , whereas @xmath179 @xcite .
after taking the @xmath180 limit in the last two terms of eq .
( [ q_v ] ) , @xmath152 can be written as @xmath181 where lhpital rule has been used for the indeterminate limit .
since any three- and two - point scalar function and their derivatives can be written as a combination of two - point functions @xcite , we can write @xmath182 and @xmath183 we thus obtain @xmath184 whereas @xmath185 can be obtained immediately from ( [ k_v ] ) @xmath186 in the limit of very large @xmath187 ( @xmath188 ) we obtain @xmath189 which can be used to show that both @xmath152 and @xmath190 vanish in the large bilepton mass limit , i.e. these coefficients are of decoupling nature . furthermore , it is evident that @xmath152 decreases more rapidly than @xmath190 as @xmath191 increases .
feynman diagrams for the @xmath4 vertex in the @xmath3-covariant gauge . the pseudo - goldstone bosons and ghosts contribute through an identical set of diagrams , but only the triangle ones give a nonvanishing contribution to @xmath155 and @xmath156.,width=384 ] we emphasize that the above expressions account for the bilepton effects on the @xmath0 vertex at the @xmath36 scale , when the bilepton masses are degenerate and equal to @xmath191 .
these results are thus approximate but can give a realistic estimate of this class of effects .
we now proceed to the numerical evaluation and analysis .
we now would like to analyze the behavior of @xmath156 and @xmath169 as functions of @xmath192 and the bilepton mass , which are the only free parameters which they depend on .
as far as @xmath192 is concerned , we will focus on the values that can be at the reach of the future planned linear colliders as this vertex has the chance of being studied through @xmath193 scattering .
we thus consider the range 200 gev @xmath194 2 tev @xcite . as for the bilepton mass , it is convenient to give a short account on the existing bounds from both the theoretical and experimental sides
. first of all , we already mentioned that in the 331 model with right - handed neutrinos the matching of the gauge couplings constants at the @xmath16 breaking scale requires that @xmath195 @xcite , from which an upper bound on the bilepton mass can be derived after considering radiative corrections to @xmath196 and the most recent experimental data @xcite .
the respective bounds is somewhat weak and contrasts with the case of the minimal 331 model , which requires that @xmath197 tev @xcite for consistency with the theoretical bound @xmath198 .
therefore , whereas the bileptons predicted by the minimal 331 model could be searched for via collider experiments in a near future , thereby confirming or ruling out the model , the 331 model with right - handed neutrinos would still leave open the door . we have also mentioned that because of the symmetry breaking hierarchy , the bilepton mass splitting is bounded by @xmath199 .
this means that @xmath200 and @xmath201 can not be arbitrarily different .
in fact , the heavier the bilepton masses , the closer its degeneracy .
our approximation of mass degeneracy has thus much sense .
we now would like to examine the lower bounds on the bilepton masses . in ref .
@xcite it was argued that the data from neutrino neutral current elastic scattering give a lower bound on the mass of the new neutral gauge boson @xmath202 in the range of 300 gev , which along with the symmetry - breaking hierarchy yield @xmath203 gev .
a similar bound was obtained in ref .
@xcite from the observed limit on the wrong muon decay @xmath204 , which leads to @xmath205 gev at 90% c.l .
these lower bounds on @xmath206 are in agreement with that obtained from the latest bnl measurement on the muon anomaly @xcite .
it is then reasonable to consider the range 200 gev @xmath207 2000 gev for our numerical analysis
. this will allow us to assess the behavior of the form factors and get an estimate of their order of magnitude .
a word of caution is in order here : it has been pointed out that some of the existing bounds on the bilepton masses are too model dependent and thus can be relaxed by considering extra assumptions @xcite , so a relatively light bilepton can not be ruled out yet . in fig .
[ figffqd ] we show the @xmath7 form factors @xmath169 and @xmath156 vs the center - of - mass energy @xmath208 , i.e. the momentum carried by the virtual @xmath9 boson , for the bilepton mass values @xmath209 , @xmath210 , and @xmath211 gev . on the other hand ,
the dependence of the @xmath7 form factors on the bilepton mass is shown in fig .
[ figffmd ] , where we have plotted @xmath169 and @xmath156 vs the bilepton mass for @xmath212 and @xmath211 gev .
the @xmath7 form factors @xmath156 ( upper plots ) and @xmath169 ( lower plots ) vs the center - of - mass energy @xmath208 for various values of the bilepton mass .
the form factors are in units of @xmath213 and @xmath214 is in units of gev .
the solid ( dashed ) lines represent the real ( imaginary ) part of the form factors.,width=480 ] the @xmath7 form factors @xmath156 ( upper plots ) and @xmath169 ( lower plots ) vs the bilepton mass @xmath191 for @xmath212 , and @xmath211 gev .
the form factors are in units of @xmath213 and @xmath191 is in units of gev .
the solid ( dashed ) lines represent the real ( imaginary ) part of the form factors.,width=480 ] as far as the energy dependence of the form factors , from fig .
[ figffqd ] it is clear that both @xmath156 and @xmath155 are of the order of about @xmath215 for a relatively light bilepton with a mass of 200 gev and a center - of - mass energy around the energy threshold @xmath216 , where the form factors reach their extremum values and develop an imaginary part . below this threshold ,
the form factors are purely real , which reflects the fact that the bileptons that couple to the @xmath9 boson are necessarily virtual , whereas the direct production of a bilepton pair becomes possible for energies @xmath217 .
although the direct production of a bilepton pair would be preferred over the search for their virtual effects , the latter would be useful for a high precision test of the @xmath4 vertex .
we can also see that the form factors decrease by one order of magnitude when @xmath218 gev and by two orders of magnitude when @xmath219 gev , which is depicted in fig .
[ figffmd ] .
for very high energies , the form factors have a good behavior as they approach zero asymptotically after reaching an extremum value above @xmath216 .
therefore , unitarity is respected , which stems from the fact that the @xmath4 green function that we have obtained is gauge invariant . in fig .
[ figffqd ] we can also observe that , except for the reversed sign , the curves for @xmath156 and @xmath169 are very similar . for a relatively light bilepton both form factors are of about the same order of magnitude , but the magnitude of @xmath156 decreases more quickly as @xmath191 increases .
it is interesting to compare our results with those obtained in the sm and some of its extensions .
as far as the sm is concerned , the gauge boson contribution to the @xmath156 form factor is of the order of @xmath220 for @xmath214 in the range @xmath221 gev , whereas @xmath156 is about one order of magnitude below @xcite .
these contributions are of the same order of magnitude than those of the fermion and scalar sectors of the theory @xcite .
as far as supersymmetric models are concerned , the total contribution from the unconstrained mssm is about the same order of magnitude or larger than the sm total contribution for some set of the values of the parameters of the model @xcite .
our results in the 331 model with right - handed neutrinos and relatively light bileptons , with a mass around 200 gev , is of the same order of magnitude than the sm contribution .
these results are similar to those obtained in the minimal 331 model @xcite and other weakly coupled theories .
unless a very high precision is achieved in future particle colliders , it would be very hard to discriminate the contributions form different models to the @xmath4 vertex .
however , testing this vertex would still be useful for probing some particular model once it has been confirmed by experimental data .
we have used a nonconventional quantization method inspired on the bfm and brst symmetry to analyze the effects of the new gauge bosons ( bileptons ) predicted by the 331 model with right - handed neutrinos on the of - shell @xmath0 vertex .
hopefully , this class of effects might be searched for at a future linear collider through @xmath222 scattering . in particular , it has been pointed out that clic would reach a sensitivity of about @xmath215 in the measurement of the so called form factors , @xmath156 and @xmath169 , characterizing the @xmath0 vertex @xcite . following the philosophy of the bfm
, our method considers the bileptons as quantum fields and the sm gauge bosons as classical fields .
a nonlinear @xmath3 invariant gauge - fixing is then introduced for the bileptons , which are integrated out in the generating functional .
the result is a quantum action defined by a @xmath3 invariant lagrangian effective lagrangian is constructed out of the sm fields to analyze the virtual effects of the heavy fields , which have been integrated out . ] out of which a gauge invariant and gauge independent green function can be obtained . to this end , we made use of the link between the diagrammatic method known as the pt and the bfm , which establishes that the gauge invariant and gauge independent green function obtained via the pt coincides with that obtained via the bfm as long as the feynman-t hooft gauge feynman rules are used when calculating the latter .
we emphasize that our method is only approximate as the quantum action is only invariant under @xmath3 rather than @xmath27 .
this method is suited to analyze the bilepton effects at the @xmath27 breaking scale , when their masses are still degenerate .
the advantages of our calculation scheme are twofold : the introduced nonlinear gauge - fixing term for the bileptons allows one to remove several unphysical vertices , which in turn allows one to get rid of several feynman diagrams ; on the other hand , preserving the electroweak invariance turn the calculation into a simple task as each sector of the theory gives a gauge invariant contribution satisfying simple ward identities by its own .
once our method was applied , we obtained the form factors @xmath156 and @xmath169 , which were analyzed for several values of the bilepton mass and convenient values of the momentum carried by the virtual @xmath9 boson .
it was found that the bilepton effects on the @xmath0 vertex are of the same order of magnitude than the sm and the minimal 331 model contributions , provided that the bilepton mass is of the order of a few hundred of gevs .
for very heavy bileptons , the respective contribution to the @xmath0 is negligibly small .
this indicates that it will be very hard to discriminate between different class of effects on the @xmath0 vertex arising from distinct models .
however , if a high precision is achieved in future linear colliders , the @xmath0 vertex might serve as a probe to some particular model by then confirmed by other experimental data .
the good behavior of the form factors at high energies was also discussed , for it is a reflect of the gauge invariance and gauge independence of the @xmath0 green function obtained via our quantization method .
j. montano , f. ramrez - zavaleta , g. tavares - velasco , and j. j. toscano , phys .
d * 72 * , 055023 ( 2005 ) .
w. a. bardeen , r. gastmans and b. lautrup , nucl .
b * 46 * , 319 ( 1972 ) .
e. n. argyres , g. katsilieris , a. b. lahanas , c. g. papadopoulos and v. c. spanos , nucl .
b * 391 * , 23 ( 1993 ) .
g. couture , j. n. ng , j. l. hewett and t. g. rizzo , phys .
d * 36 * , 859 ( 1987 ) . c. l. bilchak , r. gastmans and a. van proeyen , nucl .
b * 273 * , 46 ( 1986 ) ; g. couture , j. n. ng , j. l. hewett and t. g. rizzo , phys .
d * 38 * , 860 ( 1988 ) ; a. b. lahanas and v. c. spanos , phys .
b * 334 * , 378 ( 1994 ) ; t. m. aliev , _ ibid .
_ * 155 * , 364 ( 1985 ) .
j. l. garcia - luna , g. tavares - velasco and j. j. toscano , phys .
d * 69 * , 093005 ( 2004 ) .
j. papavassiliou and k. philippides , phys .
d * 48 * , 4255 ( 1993 ) .
j. m. cornwall , phys .
d * 26 * , 1453 ( 1982 ) ; j. m. cornwall and j. papavassiliou , phys .
d * 40 * , 3474 ( 1989 ) ; j. papavassiliou , phys .
d * 41 * , 3179 ( 1990 ) .
a. arhrib , j. l. kneur and g. moultaka , phys .
b * 376 * , 127 ( 1996 ) .
b. s. dewitt , phys .
rev . * 162 * , 1195 ( 1967 ) ; _ dynamical theory of groups and fields _ , gordon and breach , new york ( 1965 ) ; c. j. isham , r. penrose , and d. w. sciama , _ quantum gravity 2 _ , oxford university press , oxford ( 1981 ) ; g. t hooft , acta universitatis wratislavensis * 368 * , 345 ( 1976 ) ; m. t. grisaru , p. van nieuwenhuizen , and c. c. wu , phys . rev .
d * 12 * , 3203 ( 1975 ) ; h. kluberg - stern and j. b. zuber , phys .
d * 12 * , 482 3159 ( 1975 ) ; d. g. boulware , phys .
d * 23 * , 389 ( 1981 ) ; c. f. hart , phys .
d * 28 * , 1993 ( 1983 ) ; l. f. abbott , nucl .
* b185 * , 189 ( 1981 ) ; acta phys . pol .
b * 13 * , 33 ( 1982 ) ; l. f. abbott , m. t. grisaru , and r. k. schaefer , nucl . phys .
* b229 * , 372 ( 1983 ) ; for the bfm formulation of the electroweak theory see : a. denner , g. weglein , and s. dittmaier , nucl . phys . * b440 * , 95 ( 1995 ) ; acta phys .
b * 27 * , 3645 ( 1996 ) .
j. papavassiliou , phys . rev .
d * 51 * , 856 ( 1995 ) .
d. binosi and j. papavassiliou , phys .
d * 66 * , 025024 ( 2002 ) ; d. binosi and j. papavassiliou , phys .
d * 66 * , 076010 ( 2002 ) ; d. binosi and j. papavassiliou , phys .
d * 66 * , 111901 ( 2002 ) ; d. binosi and j. papavassiliou , j. phys .
g * 30 * , 203 ( 2004 ) .
f. pisano and v. pleitez , phys .
d * 46 * , 410 ( 1992 ) .
p. h. frampton , phys .
lett . * 69 * , 2889 ( 1992 ) . c. becchi , a. rouet and r. stora , commun .
phys . * 42 * , 127 ( 1975 ) ; c. becchi , a. rouet and r. stora , annals phys . * 98 * , 287 ( 1976 ) ; i. v. tyutin , lebedev report no .
fian 39 ( 1975 ) ( unpublished ) .
r. foot , h. n. long and t. a. tran , phys .
d * 50 * , 34 ( 1994 ) ; h. n. long , phys . rev .
d * 54 * , 4691 ( 1996 ) ; w. a. ponce and o. zapata , phys .
d * 74 * , 093007 ( 2006 ) ; p. v. dong , h. n. long and d. v. soa , arxiv : hep - ph/0610381 ; p. v. dong , d. t. huong , m. c. rodriguez and h. n. long , eur .
j. c * 48 * , 229 ( 2006 ) ; a. palcu , mod .
a * 21 * , 2027 ( 2006 ) ; a. palcu , mod .
a * 21 * , 1203 ( 2006 ) ; d. chang and h. n. long , phys .
d * 73 * , 053006 ( 2006 ) ; n. v. .
cortez and m. d. tonasse , phys .
d * 72 * , 073005 ( 2005 ) ; e. r. barreto , y. a. coutinho and j. s. borges , arxiv : hep - ph/0703099 ; e. ramirez barreto , y. a. coutinho and j. sa borges , arxiv : hep - ph/0605098 ; p. v. dong , h. n. long and d. t. nhung , phys .
b * 639 * , 527 ( 2006 ) ; a. carcamo , r. martinez and f. ochoa , phys .
d * 73 * , 035007 ( 2006 ) ; m. a. perez , g. tavares - velasco and j. j. toscano , phys .
d * 69 * , 115004 ( 2004 ) [ arxiv : hep - ph/0402156 ] .
y. b. liu , y. h. cao and h. m. han , arxiv : hep - ph/0701169 ; d. van soa and d. le thuy , arxiv : hep - ph/0610297 ; j. e. cieza montalvo , n. v. .
cortez and m. d. tonasse , arxiv : hep - ph/0610276 ; j. e. cieza montalvo , n. v. .
cortez , j. sa borges and m. d. tonasse , nucl .
b * 756 * , 1 ( 2006 ) ; p. van dong , h. n. long and d. van soa , phys .
d * 73 * , 075005 ( 2006 ) ; p. v. dong , h. n. long , d. t. nhung and d. v. soa , phys .
d * 73 * , 035004 ( 2006 ) ; n. a. ky and n. t. h. van , phys .
d * 72 * , 115017 ( 2005 ) ; d. a. gutierrez , w. a. ponce and l. a. sanchez , int .
j. mod .
a * 21 * , 2217 ( 2006 ) ; g. tavares - velasco and j. j. toscano , phys .
d * 70 * , 053006 ( 2004 ) .
p. v. dong ,
d. t. huong , m. c. rodriguez and h. n. long , arxiv : hep - ph/0701137 ; s. sen and a. dixit , arxiv : hep - ph/0510393 . h. n. long , n. q. lan , d. v. soa and l. n. thuc , arxiv : hep - ph/0611276 . c. promberger , s. schatt and f. schwab , arxiv : hep - ph/0702169 .
a. g. dias , j. c. montero and v. pleitez , phys .
d * 73 * , 113004 ( 2006 ) ; a. g. dias , j. c. montero and v. pleitez , phys .
b * 637 * , 85 ( 2006 ) ; d. l. anderson and m. sher , phys . rev .
d * 72 * , 095014 ( 2005 ) ; p. v. dong , d. t. huong , t. t. huong and h. n. long , phys .
d * 74 * , 053003 ( 2006 ) .
h. n. long and t. inami , phys .
d * 61 * , 075002 ( 2000 ) .
n. a. ky , h. n. long , d. v. soa , l. p. trung and v. t. van , arxiv : hep - ph/0009187 ; n. a. ky and h. n. long , arxiv : hep - ph/0103247 ; h. n. long and l. p. trung , phys .
b * 502 * , 63 ( 2001 ) .
m. d. tonasse , phys .
b * 381 * , 191 ( 1996 ) ; m. b. tully and g. c. joshi , int .
j. mod .
a * 18 * , 1573 ( 2003 )
; h. n. long , mod . phys .
a * 13 * , 1865 ( 1998 ) ; w. a. ponce , y. giraldo and l. a. sanchez , phys .
d * 67 * , 075001 ( 2003 ) ; r. a. diaz , r. martinez and f. ochoa , phys .
d * 69 * , 095009 ( 2004 ) ; n. a. ky and n. t. h. van , phys .
d * 72 * , 115017 ( 2005 ) .
j. t. liu and d. ng , z. phys .
c * 62 * , 693 ( 1994 ) .
d. ng , phys . rev .
d * 49 * , 4805 ( 1994 ) .
a. g. dias , r. martinez and v. pleitez , eur .
j. c * 39 * , 101 ( 2005 ) ; a. g. dias , phys .
d * 71 * , 015009 ( 2005 ) .
m. bace and n. d. hari dass , ann .
phys . * 94 * , 349 ( 1975 ) ; m. b. gavela , g. girardi , c. malleville , and p. sorba , nucl . phys . *
b193 * , 257 ( 1981 ) ; n. m. monyonko , j. h. reid , and a. sen , phys .
b * 136 * , 265 ( 1984 ) ; n. m. monyonko and j. h. reid , phys .
d * 32 * , 962 ( 1985 ) ; g. jikia and a. tkabladze , phys .
b * 323 * , 453 ( 1994 ) ; _ ibid .
_ * 332 * , 441 ( 1994 ) ; g. jikia , phys . lett .
b * 298 * , 224 ( 1993 ) ; nucl .
b405 * , 24 ( 1993 ) ; _ ibid . _ * 412 * , 57 ( 1994 ) ; d. a. dicus and c kao , phys .
d * 49 * , 1265 ( 1994 ) ; a. abbasabadi , d. bowser - chao , d. a. dicus , and w. w. repko , phys .
d * 52 * , 3919 ( 1995 ) ; j. m. hern ' andez , m. a. p ' erez , g. tavares - velasco , and j. j. toscano , phys .
d * 60 * , 013004 ( 1999 ) ; u. cotti , j. l. daz - cruz and j. j. toscano , phys .
d * 62 * , 035009 ( 2000 ) ; j. l. daz - cruz , j. hernndez - snchez , and j. j. toscano , phys .
b * 512 * , 339 ( 2001 ) ; j. hernndez - snchez , m. a. prez , g. tavares - velasco , and j. j. toscano , phys .
d * 69 * , 095008 ( 2004 ) .
l. d. faddeev and v. n. popov , phys .
b * 25 * , 29 ( 1967 ) .
the most general renormalizable gauge - fixing functions of @xmath99 type , nonlinear in both the vector and scalar sectors , along with a discussion on the difficulties to implement the fpm , have been introduced from the brst standpoint in u. cotti , j. l. daz - cruz , and j. j. toscano , phys .
b * 404 * , 308 ( 1997 ) ; for the implications on the ghost sector see : j. g. mndez and j. j. toscano , rev .
. fs . * 50 * , 346 ( 2004 ) .
i. a. batalin and g. a. vilkovisky , phys .
d * 28 * , 2567 ( 1983 ) ; _ ibid . _ d * 30 * , 508 ( 1984 ) ( e ) ; j. gomis , j. paris and s. samuel , phys .
rept . * 259 * , 1 ( 1995 ) .
k. j. f. gaemers and g. j. gounaris , z. phys .
c * 1 * , 259 ( 1979 ) ; k. hagiwara , r. d. peccei , d. zeppenfeld , and k. hikasa , nucl .
b282 * , 253 ( 1987 ) ; u. baur and d. zeppenfeld , _ ibid . _ * b308 * , 127 ( 1988 ) ; * b325 * , 253 ( 1989 ) .
g. passarino and m. j. g. veltman , nucl .
b * 160 * , 151 ( 1979 ) .
r. g. stuart , comput .
commun . * 48 * , 367 ( 1988 ) ; g. devaraj and r. g. stuart , nucl .
b * 519 * , 483 ( 1998 ) .
e. accomando _ et al .
_ [ clic physics working group ] , arxiv : hep - ph/0412251 . w. m. yao _ et al . _
[ particle data group ] , j. phys .
g * 33 * , 1 ( 2006 ) .
v. pleitez , phys .
d * 61 * , 057903 ( 2000 ) . | in a recent paper @xcite , we investigated the effects of the massive charged gauge bosons ( bileptons ) predicted by the minimal 331 model on the off - shell vertex @xmath0 ( @xmath1 , @xmath2 ) using a @xmath3 covariant gauge - fixing term for the bileptons .
we proceed along the same lines and calculate the effects of the gauge bosons predicted by the 331 model with right - handed neutrinos .
it is found that the bilepton effects on the @xmath0 vertex are of the same order of magnitude than those arising from the sm and several of its extensions , provided that the bilepton mass is of the order of a few hundred of gevs . for heavier bileptons ,
their effects on the @xmath0 vertex are negligible .
the behavior of the form factors at high energies is also discussed as it is a reflect of the gauge invariance and gauge independence of the @xmath0 green function obtained via our quantization method . |
recently , a hierarchy of hubbard - like hamiltonians has been proposed to describe the behavior of ultracold fermions in one - dimensional optical lattices @xcite .
these lattices can be realized with a pair of lasers propagating at a given angle @xmath1 ( @xmath2 represents the familiar counterpropagating case ) , with global confinement ensured by a magnetic trap ( see fig .
[ fig : setup ] , for a detailed description of this setup see @xcite ) .
the pair of lasers give rise to a interference pattern needed to obtain a periodic potential by ac stark effect .
the lattice constant can be adjusted tuning the angle @xmath1 according to the relation @xmath3 where @xmath4 is the laser wavelength .
in addition , it is possible to control both the barrier height of the periodic potential ( as a function of the laser intensity ) and the interaction between fermions via an external magnetic field ( feshbach resonance , see , e.g. @xcite ) .
these simple considerations allow one to understand how ultracold - atoms physics offers the possibility to explore experimentally a wide range of parameters set that would be unattainable in other contexts , such as the hubbard model in condensed matter physics . as a first step towards the description of the cited models , we propose here a mean - field algebraic approach based on coherent - states procedure @xcite for a fermionic one - dimensional array of harmonic wells .
although the analytical approach followed here may be regarded as completely general , future numerical analysis will concentrate on a dimer with a six - level structure per well as depicted in fig .
[ fig : levels ] ) .
the approach followed here allows one a straightforward reformulation of the the usual mean - field approach for quantum system ( based on the ` linearization ' of the hamiltonian and the subsequent solving appropriate self - consistency equations ) in terms of a corresponding classical dynamical system . while , for fermions , the interpretation of the aforementioned classical dynamical system as a semiclassical approximation seems not beyond need of justification , it is possible to give a precise physical interpretation to the dynamical variables of the classical problem in terms of expectation values of quantum operators . in general
, it is possible to consider a mean - field approach to a given problem as the constrained minimization of the hamiltonian @xmath5 over a algebra @xmath6 .
a different choice of @xmath6 will lead to different mean - field solutions ( @xcite ) .
in particular , we will focus on the so(2r ) coherent states that , as it will be shown , will lead to the hartree - fock - bogoliubov @xcite mean - field approximation , whose effectiveness has been proven for a single spherical harmonic trap in @xcite . the paper is organized as follows . in section [ sec : moddesc ] a brief discussion of the general model considered will be given , along with some possible approximations in different physical situations .
as we already mentioned , the fully - analytical control over the physical parameters allows to conceive various hamiltonians that may have direct experimental relevance . in section [ sec : cohst ] so(2r ) coherent states and the relevant algebra will be defined .
the end of this rather technical section will be devoted to the physical interpretation of the choice of so(2r ) as the algebra for the mean - field procedure . in section [ sec : clh ] the classical hamiltonian @xmath7 will be deduced and the functional dependence in terms of quantum operators expectation values will be investigated . finally , in section
[ sec : evzeta ] the analysis of the classical dynamical system whose hamiltonian is @xmath7 is performed : lie - poisson brackets ( namely the ` classical ' commutators ) and , consequently , the evolution equations for the dynamical variables are given .
in @xcite , along the lines introduced in @xcite , a generalized hubbard hamiltonian has been introduced to describe the behavior of alkali - metal fermionic atoms in a one - dimensional optical lattice of oblate three dimensional ( 2 + 1d ) harmonic oscillators ( pancakes ) @xmath8 in eq .
( [ eq : rhh ] ) , @xmath9 must be considered as a multiple index @xmath10 whose origin can be traced back to the space(+local)-modes approximation . in this picture @xmath11 and @xmath12
are the local 2 + 1 d harmonic oscillator quantum numbers , @xmath13 is the site quantum number and @xmath14 is the spin quantum number . in the following
we will confine our analysis to situation where radial modes only are involved in the system dynamics ( i.e. we will `` freeze '' the axial quantum number @xmath15 to zero ) .
the validity of this assumption is guaranteed as long as the radial trapping frequency @xmath16 is much smaller than the axial trapping frequency @xmath17 , i.e. @xmath18 . in this case
the tunneling coefficient assumes the following form : @xmath19 if we allow nearest - neighbor hopping only @xmath20 where @xmath21 is a known function of the external parameters .
another assumption concerns the two - body interaction term @xmath22 , which is treated within the pseudopotential approximation , leading to a delta - like spatial dependence , thus excluding nearest - neighbor two - body interaction terms .
if the fermionic nature of the interacting particles is taken into account we have @xmath23 finally , we give the expression for the one - particle energy term which is essentially given by the single - particle energy of the 3d harmonic oscillator @xmath24 \ , , \ ] ] where @xmath25 represents a `` hopping correction '' to the single particle energy term .
for the case considered in fig .
[ fig : levels ] the selection rules imposed on the two - body interaction term select three possible values on @xmath22 that can be classified as : i ) lowest - level/ lowest - level interaction terms , ii ) lowest - level / highest - level interaction terms , iii ) highest - level / highest - level interaction terms .
since in our mean - field analysis we would like to keep trace of the ( possible ) atom pairing , the most appropriate coherent - states algebra seems to be , according to @xcite , the algebra spanned by the @xmath26 operators @xmath27 , i.e. so(2r ) .
its commutation relations can be written as @xmath28=\left [ y^2_{\alpha\beta},y^2_{\gamma\delta}\right]=0 \nonumber\\ & & \left [ y^1_{\alpha\beta},y^2_{\gamma\delta}\right]= y^3_{\gamma\beta}\delta_{\alpha\delta}+y^3_{\delta\alpha}\delta_{\beta\gamma } -y^3_{\gamma\alpha}\delta_{\beta\delta}-y^3_{\delta\beta}\delta_{\alpha\gamma}\nonumber\\ & & \left[y^1_{\alpha\beta},y^3_{\gamma\delta}\right]= y^1_{\alpha\delta}\delta_{\beta\gamma}-y^1_{\beta\delta}\delta_{\alpha\gamma}\nonumber\\ & & \left[y^2_{\alpha\beta},y^3_{\gamma\delta}\right]=y^2_{\beta\gamma}\delta_{\alpha\delta}-y^2_{\alpha\gamma}\delta_{\beta\delta}\ , , \end{aligned}\ ] ] having defined @xmath29 with the above definitions the coherent states can be expressed as @xmath30 |0>\,.\ ] ] to evaluate the expectation value of the hamiltonian @xmath5 defined by equation ( [ eq : rhh ] ) over the coherent state of the form ( [ eq : cohst ] ) , it is necessary to evaluate the action of the operator @xmath31\ ] ] over the fermionic raising and lowering operators .
namely @xmath32 \hat{c}^\dagger_\alpha \nonumber \\ & & \exp \left [ -\sum_{1\leq i \neq
j \leq r } \left ( \eta_{i , j } \hat{c}^\dagger_i \hat{c}^\dagger_j - \eta^*_{i , j } \hat{c}_j\hat{c}_i \right ) \right ] \end{aligned}\ ] ] which , exploiting the bch formula @xcite can be written as @xmath33_m \,.\ ] ] it can be shown that in the last summation the two first terms only survive , leading to to the following expression for @xmath34 and @xmath35 respectively @xmath36 with @xmath37 .
we are now in the position to evaluate @xmath38 where the term @xmath39 has been added to take into account the particle number constraint . with eq .
( [ eq : cohst ] ) , eq .
( [ eq : expval ] ) becomes @xmath40\omega^\dagger|0\rangle \ , , \ ] ] where @xmath41 with @xmath42 since @xmath43 is a unitary operator , we can write @xmath44 the following expectation values must then be evaluated . for the one - body term @xmath45 \cdot \left
[ \hat{c}_\beta + \sum_k \zeta_{k\beta } \hat{c}^\dagger_k \right]\ ] ] and for the interaction term @xmath46 \cdot \left [ \hat{c}^\dagger_\beta + \sum_j \zeta^*_{j\beta } \hat{c}_j \right ] \cdot
\nonumber\\ & & \hspace{-1 cm } \left [ \hat{c}_\delta + \sum_k \zeta_{k\delta } \hat{c}^\dagger_k \right ] \cdot \left [ \hat{c}_\gamma + \sum_l \zeta_{l\gamma } \hat{c}^\dagger_l \right ] \ , .
\end{aligned}\ ] ] as it can be directly verified , in the calculation of the expectation values over the vacuum state @xmath47 only the following terms survive @xmath48 @xmath49 the two expectation values over the vacuum state give @xmath50 hence @xmath51\ ] ]
hamiltonian ( [ eq : hcl ] ) can be shown to represent the effective hamiltonian associated with @xmath52 within the time - dependent variational principle procedure @xcite .
the latter is based on approximating the quantum states of the system by a trial state @xmath53 satisfying the weak form of the schrdinger equation @xmath54 . here
, we assume that @xmath55 , up to an irrelevant phase factor , is the coherent state defined in equation ( [ eq : cohst ] ) .
the variational procedure allows one to derive the effective lagrangian @xmath56 , depending on dynamical variables @xmath57 , which in turn supplies the effective hamiltonian ( [ eq : hcl ] ) .
such a procedure provides as well the dynamical equations pertaining to hamiltonian ( [ eq : hcl ] ) and the relevant lie - poisson brackets .
the latter exhibit the same algebraic structure of commutators ( [ eq : commy ] ) and will be defined below . a quite direct physical insight about coherent - state parameters @xmath57
is achieved when considering the expectation values for the elements of the lie algebra so(2r ) over the coherent states @xmath58 .
we have @xmath59 showing how parameters @xmath57 are related to microscopic physical processes of creation / destruction of lattice fermions .
moreover eq . ( [ eq : hcl ] ) can be written as @xmath60\ , .\ ] ] the three terms represent the direct , the exchange and the pairing term in the hfb mean - field approximation . according to @xcite the variables @xmath61 , @xmath62 and @xmath63 represent the canonical variable for the classical hamiltonian @xmath7 . with a well - known procedure
@xcite it is possible to describe the time evolution of those canonical variables in terms of their poisson brackets with @xmath7 . to write the lie - poisson brackets for the given dynamical system we can make explicit the structure constants for the so(2r ) algebra @xmath64 thus the poisson brackets have the following form @xmath65 remembering that @xmath66 it is possible to write @xmath67 \nonumber \\ & & \sum_{\beta\gamma\eta}\left[u_{\rho\beta\gamma\eta}\left(\zeta_{\theta\eta}\xi_{\beta\gamma}-\zeta_{\theta\gamma}\xi_{\beta\eta}\right ) + u_{\theta\beta\gamma\eta}\left(\zeta_{\rho\eta}\xi_{\beta\gamma}-\zeta_{\rho\gamma}\xi_{\beta\eta}\right)\right]+ \nonumber \\ & & \sum_{\alpha\gamma\eta } \zeta_{\eta\gamma}\left(u_{\alpha\theta\gamma\eta}\xi_{\alpha\rho}-u_{\alpha\rho\gamma\eta}\xi_{\alpha\theta}\right)+ \nonumber \\ & & \sum_{\beta\gamma\eta } \zeta_{\eta\gamma}\left(u_{\rho\beta\gamma\eta}\xi_{\beta\theta}-u_{\theta\rho\gamma\eta}\xi_{\beta\rho}\right)\end{aligned}\ ] ] that provide the set of dynamical equations governing the evolution of the coherent state that approximates the system quantum state . in particular , they allow to find the mean - field ground the state for the system and to perform a weakly - excited state analysis
in this paper we have formulated an hfb mean - field approximation for a one - dimensional array of oblate harmonic oscillators loaded with neutral fermionic atoms . as already pointed out by grasso et al .
@xcite , the numerical solution to eq .
( [ eq : evzeta2 ] ) appears to be rather demanding from a computational point of view .
it seems then appropriate for future work to concentrate on the simplest situation beyond known models like , as already mentioned , a dimer with a six - level local structure . in this case
the evaluation of ground - state properties in this mean - field picture as a function of the relevant parameters(i.e @xmath68 , @xmath69 , @xmath70 ) reduces to the fixed - point analysis of eq.([eq : evzeta2 ] ) .
moreover , an extension to finite - temperature properties does not seem beyond the possibilities of the analytical techniques here outlined and may represent one of the future lines of research .
in the present section we will explicitly calculate the terms obtained form eqs .
( [ cs2],[cs2.1 ] ) leading to eqs .
( [ eq : cs3 ] ) . to evaluate ( [ cs2 ] ) we need to perform the following product calculation @xmath71 \cdot \left [ \hat{c}_\beta + \sum_j \zeta_{j\beta}\hat{c}^\dagger_j \right]\ ] ] which is equal to the evaluation of eq .
( [ eq : app2 ] ) over the vacuum state @xmath47 leads to vanishing contributions for all non number - conseving terms and for all the terms with a lowering operator on the right - hand side ( or a raising operator on the left - hand side ) .
namely ( see eq .
( [ eq : cs3 ] ) @xmath73 with an analogous procedure it is possible to evaluate the expression given by eq .
( [ cs2.1 ] ) @xmath74 & & \cdot \left [ \hat{c}^\dagger_\beta + \sum_j \zeta^*_{j\beta } \hat{c}_j \right ]
\cdot \nonumber\\ & & \hspace{-1 cm } \left [ \hat{c}_\delta + \sum_k \zeta_{k\delta } \hat{c}^\dagger_k \right ]
\cdot \left [ \hat{c}_\gamma + \sum_l \zeta_{l\gamma } \hat{c}^\dagger_l \right ] \end{aligned}\ ] ] leading to @xmath75 \,.\end{aligned}\ ] ] with the same argument needed to obtain eq .
( [ eq : app4 ] ) we can write the expectation value of the operator defined by eq .
( [ eq : app6 ] ) over the vacuum state @xmath47 as @xmath76 which is the expression given by eq .
( [ eq : cs3.1 ] ) .
s. peil , j. v. porto , b. laburthe tolra , j. m. obrecht , b. e. king , 2 s. l. rolston m. subbotin , and w. d. phillips .
patterned loading of a bose - einstein condensate into an optical lattice . , 67:051603(r ) , 2003 . | we consider a one - dimensional optical lattice of three - dimensional harmonic oscillators which are loaded with neutral fermionic atoms trapped into two hyperfine states . by means of a standard variational coherent - state procedure
, we derive an effective hamiltonian for this quantum model and the hamiltonian equations describing its evolution . to this end , we identify the algebra @xmath0 of two - fermion operators describing the relevant microscopic quantum processes of our model whereby the natural choice for the trial state appears to be a so(2r ) coherent state .
the coherent - state parameters , playing the role of dynamical variables for the effective hamiltonian , are shown to identify with the @xmath0-operator expectation values thus providing a clear physical interpretation of this algebraic mean - field picture . |
quantum field theory is the mathematical language of nature . viewed in this light ,
much of the last century of theoretical physics can be seen as quantum field theory calculations performed in a variety of approximations .
the most fruitful scheme so far is clearly perturbation theory in small couplings , which has seen wide use and great success .
nevertheless there are important physical situations where naive perturbation theory fails and must be enhanced by partial resummation if it can be used at all .
for example , massless particles in thermal plasmas produce large loop corrections which must be resummed , causing thermal mass generation and non - analyticities in thermodynamic functions .
similarly , large logarithms can invalidate naive perturbation theory in problems involving disparate length or energy scales .
resummation of these logarithms leads to the renormalization group running of coupling constants .
although the need to go beyond naive perturbation theory is clear in many cases , _ ad hoc _ resummations are problematic because perturbation series are asymptotic in nature @xcite .
systematic resummation schemes are required to guarantee consistency with the original non - perturbative theory .
there are three such schemes which can claim to be widely studied and successful : the renormalization group ( rg ) @xcite , large @xmath1 expansion @xcite , and @xmath2-particle irreducible effective actions ( @xmath2pieas ) @xcite .
this work focuses on the latter .
@xmath2pieas are a functional technique that combine the advantages ( and some disadvantages ) of perturbation theory and variational methods . based on a legendre transform procedure ,
@xmath2pieas are guaranteed to be equivalent to the original theory and can capture analytic features of the exact theory that are invisible to perturbation theory , and so side - step potential issues with _ ad hoc _
resummations @xcite . the 1piea , developed by goldstone , salam and weinberg @xcite and jona - lasinio @xcite , effects a resummation of tadpoles by defining an effective action functional @xmath3 $ ] which depends on the expectation value @xmath4 of the quantum field .
the physical value of @xmath5 is determined by the variational equation of motion @xmath6 and , when this equation is satisfied , one can drop all one particle reducible diagrams from the diagrammatic perturbation expansion .
the 2piea technique was originally developed by lee and yang @xcite , luttinger and ward @xcite , baym @xcite and others in the context of many - body theory , then extended by cornwall , jackiw and tomboulis @xcite to relativistic field theory in the functional formalism . in the 2piea formalism
one defines an action functional @xmath7 $ ] depending not just on the mean field , but also its correlation function @xmath8 , and solves for @xmath9 using @xmath10 .
this effects a resummation of all self - energy insertions and removes two particle reducible diagrams from the diagrammatic expansion .
apparently de dominicis and martin @xcite were the first to realize that the @xmath11 and @xmath12 formalisms are special cases of a general construction giving higher order effective actions for @xmath13 .
early work on higher effective actions was carried on by vasiliev @xcite , whose book was unfortunately not available in english for more than twenty years , though there were some reviews suggesting that the english literature was at least aware of these developments ( e.g. @xcite ) . the recent resurgence of interest in higher @xmath2pieas has largely been driven by their advantages for non - equilibrium problems and can likely be credited to the reviews by berges @xcite and advances in computer power .
one of the disadvantages of @xmath2pieas is that finite order truncations of the effective actions generically do not respect the symmetry properties one expects from the exact theory .
this is because the correlation functions of the 1piea obey different ward identities ( wis ) than the 2piea etc .
this can also be understood in terms of the patterns of resummations effected by @xmath2pieas which , if truncated , do not preserve the order by order cancellations required to maintain all symmetries .
similar remarks apply for global symmetries and gauge invariances . the most notable physical consequence of this is a violation of goldstone s theorem for theories with spontaneous symmetry breaking : would - be goldstone bosons gain an unphysical mass in the hartree - fock approximation , leading also to unphysical reaction thresholds and decay rates .
more subtle effects appear in higher order approximations .
similarly , in gauge theories a residual gauge dependence appears in quantities that should be physical .
this work focuses on the particular scheme devised by pilaftsis and teresi @xcite for improving the symmetry properties of 2pieas .
symmetry improvement enforces 1pi style wis on the 2piea by the use of lagrange multipliers .
the resulting constraint forces the goldstone boson correlation functions to obey goldstone s theorem , giving physically correct massless goldstone bosons at all orders and even at finite temperatures below the critical temperature of the symmetry breaking phase transition .
so far symmetry improvement has only been applied to non - gauged scalar field theories in equilibrium .
it restores goldstone s theorem and produces physically reasonable absorptive parts in propagators @xcite and has been shown to restore the second order phase transition of the @xmath14 linear sigma model in the hartree - fock approximation @xcite .
it has been used to study pion strings evolving in the thermal bath of a heavy ion collision @xcite ( though note that the symmetry improved 2piea was only used to calculate an equilibrium finite temperature effective potential ; this work did not constitute what we would call a true non - equilibrium calculation using symmetry improved effective actions ) .
symmetry improvement has also been demonstrated to improve the evaluation of the effective potential of the standard model by taming the infrared divergences of the higgs sector , treated as an @xmath14 scalar field theory with gauge interactions turned off @xcite .
the symmetry improvement idea can also be extended to 3pieas @xcite , which requires enforcing an additional set of wis for the three point vertex function and the fixing of an ambiguity in the constraint procedure . during review of this manuscript
we became aware of the recent preprint @xcite which shows that symmetry improved 2pi equations of motion can lose solutions even in equilibrium as the result of a truncation artifact .
this loss of solutions is distinct from the one we find below and shows that more work is required to understand the domain of applicability of symmetry improved effective actions .
there is a strong motivation to extend symmetry improvement beyond equilibrium since one of the major reasons for using @xmath2pieas in the first place is their ability to handle non - equilibrium situations .
@xmath2pieas give an entirely mechanical way to set up the generic initial value problem as a _ closed _ system of integro - differential equations directly for the mean fields and low order correlation functions , which are simply related to the handful of physical quantities ( densities , conserved currents , etc . )
one is most often interested in .
apart from a truncation to some finite loop order these equations need not be subject to any further approximation .
hence , apart from the symmetry issue and issues involved in the renormalization process , @xmath2pieas give potentially the most general and accurate framework available for the computation of real time properties in quantum field theory .
motivated by this , we seek to extend the symmetry improvement technique to non - equilibrium situations .
the ultimate goal of this program would be a tractable and manifestly gauge invariant set of equations of motion for highly excited yang - mills - higgs theories with chiral fermion matter based on the self - consistently complete 4piea . in the meantime we content ourselves with an analysis of the symmetry improved 2piea for scalar fields in the linear response regime .
we investigate the linear response approximation rather than a generic non - equilibrium situation for several reasons .
first , the linear response approximation is simply far more tractable than the general non - equilibrium situation as the response functions only depend on the equilibrium properties of the theory .
second , linear response is widely applicable in the real world : many systems are `` close enough '' to equilibrium for practical purposes .
third , the linear response approximation is a nice laboratory to isolate the novel features of symmetry improvement constraints in non - equilibrium settings .
indeed , we can find results that are independent of any truncation of the effective action .
finally , we expect any physically reasonable formalism to give a well formed linear response approximation , though this depends on the assumption that the exact behavior is an analytic ( or at least not too singular ) function of the external perturbation within some neighborhood of zero perturbation .
this is true of all quantum mechanical systems ( so long as the hamiltonian remains bounded below ) , but for field theories the infinite number of degrees of freedom may complicate the situation .
the outline of the remainder of this work is as follows . in section [ sec : linear - response - theory ]
we review linear response theory and 2pieas .
then in section [ sec : symmetry - improvement ] we review the symmetry improvement method , re - deriving the wi constraints in the presence of external sources which were neglected in previous work . in section [ sec : implications - of - constraints ] we derive the consequences of the constraints for the linear response functions , noting that a careful treatment of the constraint procedure requires that not just the wi , but also its derivatives , must vanish . in section [ sec : discussion ]
we reach our conclusions about the feasibility of symmetry improvement within the linear response approximation and sketch some ideas for future work . in appendix
[ sec : the - alternate - constraint ] we discuss an alternative symmetry improvement procedure that relaxes the over - determining wis , but still leads to difficulties due to the necessity of the derivatives of the constraint to also vanish .
finally , appendix [ sec : mechanical - analogy ] includes a mechanical analogy which illustrates some of the subtle points about the constraint procedure in a simpler setting .
our conventions follow our previous paper @xcite . in particular @xmath15 and @xmath16 .
loop counting factors of @xmath17 will be kept .
repeated indices are summed , and spacetime arguments going along with @xmath0 indices are implicitly integrated over ( `` dewitt notation '' ) . where explicitly written spacetime integrals are @xmath18 and momentum integrals are @xmath19 .
@xmath20 $ ] represents the ( anti-)time ordered product of the factors in @xmath21 $ ] .
it is not necessary here to distinguish between real time , matsubara and schwinger - keldysh time contours . for a physical quantity @xmath22
we denote its equilibrium value by @xmath23 and its shift under linear response by @xmath24 so that @xmath25 plus higher order terms .
linear response theory studies the effect of small externally applied perturbations on a system initially in equilibrium .
consider a quantum system which is subjected to an external driving potential @xmath26 where @xmath27 is a c - number function of time representing the strength of the driving and @xmath28 is the interaction hamiltonian ( the reason for the name will become apparent ) .
if the initial state of the system is described by a density matrix @xmath29 at time @xmath30 , with @xmath31 for @xmath32 , then at time @xmath33 the expectation of an operator @xmath34 ( in the interaction picture with respect to the external perturbation ) is : @xmath35\rho_{0 } \right .
\nonumber \\ & \left .
\times \tilde{\mathrm{t}}\left[\mathrm{e}^{-i\int_{t_{0}}^{t}j\left(\tau\right)\hat{b}\left(\tau\right)\mathrm{d}\tau}\right]\hat{a}\left(t\right)\right\ } \nonumber \\ & = \left\langle \tilde{\hat{a}}\left(t\right)\right\rangle+i\int_{t_{0}}^{t}\left\langle \left[\hat{a}\left(t\right),\hat{b}\left(\tau\right)\right]\right\rangle j\left(\tau\right)\mathrm{d}\tau \nonumber \\ & + \mathcal{o}\left(j^{2}\right),\end{aligned}\ ] ] where @xmath36 denotes the expected value in the absence of perturbation .
this leads us to define the response function @xmath37\right\rangle \theta\left(t-\tau\right)$ ] ( which only depends on the time difference due to the equilibrium assumption about @xmath29 ) such that @xmath38 ( the limits can be pushed to @xmath39 thanks to the step function in @xmath40 , and the equation becomes trivial in the fourier domain . )
the goal of linear response theory is to compute @xmath41 for perturbations @xmath42 and observables @xmath34 of interest .
the condition for validity of the approximation is that the quadratic term , which is @xmath43,\hat{b}\left(\tau_{2}\right)\right]\right\rangle \mathrm{d}\tau_{2}\mathrm{d}\tau_{1},\ ] ] is much smaller than the linear term , which occurs for sufficiently small sources @xmath44 and times @xmath45 .
we now specialize to the scalar @xmath0 field theory defined by the action @xmath46=\int_{x}\mathcal{l}\left[\phi\right]=\int_{x}\frac{1}{2}\partial_{\mu}\phi_{a}\partial^{\mu}\phi_{a}-\frac{1}{2}m^{2}\phi_{a}\phi_{a}-\frac{\lambda}{4!}\left(\phi_{a}\phi_{a}\right)^{2},\ ] ] where @xmath47 and we choose to operate in the spontaneous symmetry breaking regime with @xmath48 .
the vacuum expectation value @xmath49 is given by @xmath50 and the tree level mass of the radial ( `` higgs '' ) mode is @xmath51 .
we take the vacuum expectation value to be in the last component @xmath52 .
the symmetry transformation is @xmath53 where @xmath54 with @xmath55 are the generators of rotations .
when we have cause to use specific generators we will write @xmath56 with @xmath57 in @xmath58 to denote the plane of rotation , and have @xmath59 .
considering this theory in a particle physics context , the most likely external perturbations will be linear or quadratic in the fields .
for example , we could be using this theory to describe mesons coupling to hadrons through yukawa interactions @xmath60 , or higgs fields coupling to a yang - mills sector through ( possibly some subset of ) the conserved currents with terms like @xmath61 and @xmath62 .
the theory can also represent an extended higgs or dark matter sector coupled via portal terms to the standard model higgs with @xmath63 , or coupled to a standard model singlet scalar sector @xmath64 through @xmath65 .
finally the fields could represent a multi - field inflaton , moduli or an extended gravitational sector , all of whose interactions to the standard model sector will be mediated by terms of the preceding forms to leading order in an effective field theory expansion .
thus , on very general grounds we expect that the external perturbation can be taken as a linear or quadratic function of the fields .
the only notable exception is the coupling of the field theory to gravity , which includes the quartic interaction @xmath66 .
we now consider the notable observables in the theory .
chiefly we will be interested in the field expectation values @xmath67 , the conserved @xmath0 currents @xmath68 and the energy - momentum tensor @xmath69 again we see that , apart from the @xmath70 term in the energy , the observables are also linear or quadratic in the fields .
note that if desired @xmath71 can be approximated in a mean field approximation by terms of the form @xmath72 , @xmath73 and @xmath74 plus corrections that can be found diagrammatically .
this means that many of the local quantities one might want can be determined by a method yielding @xmath67 and @xmath75\right\rangle -\varphi_{a}\varphi_{b}$ ] subject to generic quadratic perturbations of the form @xmath76 .
this problem seems tailor made for the 2piea .
the 2piea is defined as the double legendre transform of the connected generating functional @xmath77=-i\hbar\ln z\left[j , k\right],\ ] ] where the partition function is @xmath78=\int\mathcal{d}\left[\phi\right]\exp\frac{i}{\hbar}\left(s\left[\phi\right]+j_{a}\phi_{a}+\frac{1}{2}\phi_{a}k_{ab}\phi_{b}\right).\ ] ] here @xmath79 and @xmath80 are the externally applied sources coupling linearly and quadratically to the field respectively . performing the double legendre transform
gives the 2piea functional @xmath81=w - j_{a}\frac{\delta w}{\delta j_{a}}-k_{ab}\frac{\delta w}{\delta k_{ab}},\ ] ] where on the right hand side @xmath79 and @xmath82 are eliminated in terms of @xmath67 and @xmath83 by inverting @xmath84 the equations of motion for @xmath7 $ ] are then @xmath85 the result of performing the legendre transform is the expression @xmath81=s\left[\varphi\right]+\frac{i\hbar}{2}\mathrm{tr}\ln\left(\delta^{-1}\right)+\frac{i\hbar}{2}\mathrm{tr}\left(\delta_{0}^{-1}\delta\right)+\gamma_{2}\left[\varphi,\delta\right],\label{eq:2piea}\ ] ] where the free propagator is @xmath86\nonumber \\ & = \left[\left(-\square_{x}-m^{2}-\frac{\lambda}{6}\varphi_{c}\varphi^{c}\right)\delta_{ab}-\frac{\lambda}{3}\varphi_{a}\varphi_{b}\right ] \nonumber \\ & \times\delta\left(x - y\right),\end{aligned}\ ] ] and @xmath87 $ ] is the sum of all two particle irreducible vacuum feynman diagrams with propagators @xmath9 and vertices obtained from the cubic and quartic parts of the shifted action @xmath88 $ ] .
the equation of motion for @xmath9 ( in the absence of sources ) is simply the dyson equation : @xmath89 where the self - energy is @xmath90\delta\left(x - y\right)\nonumber \\ & + i\frac{\hbar}{2}\int_{rsuv}v_{acd}\left(x , r , s\right)v_{bef}\left(y , u , v\right ) \nonumber \\ & \times \delta_{ce}\left(r , u\right)\delta_{df}\left(s , v\right ) + \mathcal{o}\left(\hbar^{2}\right),\end{aligned}\ ] ] where the second line is the expansion to one loop order and we have introduced the three point vertex function @xmath91\nonumber \\ & \times \delta\left(x - y\right)\delta\left(x - z\right)\end{aligned}\ ] ] for convenience .
the _ hartree - fock _ approximation is obtained by retaining only the @xmath92 term in @xmath93 .
note that none of our results actually require any truncation of @xmath93 . for the equilibrium solution
we make the spontaneous symmetry breaking ansatz @xmath94 where @xmath95 are the goldstone / higgs propagators respectively .
we define the masses @xmath96 from the corresponding exact propagators .
the free propagator is @xmath97 so that @xmath98 and @xmath99 to lowest order .
however , truncations of the vev equation of motion @xmath100 _ do not _ generically obey @xmath101 so that @xmath102 .
symmetry improvement replaces the vev equation of motion by the constraint @xmath103 so that goldstone s theorem is exactly satisfied whenever @xmath104 .
the value of @xmath49 is then determined by the self - consistent solution of ( [ eq : dyson - eq ] ) . to connect this formalism to linear response theory we must expand @xmath105 and @xmath106 about their source - free equilibrium values @xmath107 and @xmath108 in ( [ eq : unmodified - vev - eom])-([eq : unmodified - delta - eom ] ) and match terms order by order in the sources , treating the responses @xmath109 and @xmath110 as first order , as typical of a perturbation theory analysis .
at lowest order we find ( [ eq : unmodified - vev - eom])-([eq : unmodified - delta - eom ] ) with no sources for the equilibrium solutions and at first order we find @xmath111 where all derivatives on the left hand sides are evaluated at the equilibrium values .
it is possible to eliminate the fluctuations from these equations by introducing the linear response functions @xmath112 , @xmath113 , @xmath114 and @xmath115 : @xmath116 and demanding that the resulting equations hold for any value of the sources @xmath44 , @xmath117 . doing
this leads to the system @xmath118 where note that in the last two equations we have to symmetrize the right hand sides before removing the source @xmath117 ( since by the symmetry of @xmath117 only the symmetric part contributes ) .
these equations determine the linear response functions entirely in terms of the equilibrium properties of the theory ( in particular , the second derivatives of the effective action evaluated at the equilibrium solution ) .
note that the last equation can be recast as a bethe - salpeter equation for the @xmath119 by using ( [ eq:2piea ] ) and ( [ eq : dyson - eq ] ) to write the left hand side of ( [ eq : unmodified - delta - eom ] ) as @xmath120-\sigma_{ab}\left[\varphi,\delta\right ] & = -\delta_{ab}^{-1}+\left(\delta_{0ab}^{-1}\left[\varphi\right]-\delta_{0ab}^{-1}\left[\tilde{\varphi}\right]\right ) \nonumber\\ & + \left(\delta_{0ab}^{-1}\left[\tilde{\varphi}\right]-\sigma_{ab}\left[\tilde{\varphi},\tilde{\delta}\right]\right)+\left(\sigma_{ab}\left[\tilde{\varphi},\tilde{\delta}\right]-\sigma_{ab}\left[\varphi,\delta\right]\right ) \nonumber\\ & = -\delta\delta_{ab}^{-1}+\frac{\delta\delta_{0ab}^{-1}}{\delta\varphi_{c}}\delta\varphi_{c}-\left(\frac{\delta\sigma_{ab}}{\delta\varphi_{c}}\delta\varphi_{c}+\frac{\delta\sigma_{ab}}{\delta\delta_{cd}}\delta\delta_{cd}\right),\end{aligned}\ ] ] then using the identity @xmath121 and the definitions of the linear response functions followed by some rearrangement to give @xmath122 this is an equation which determines the four point kernel @xmath119 iteratively , i.e. a bethe - salpeter equation with the last quantity in braces being the bethe - salpeter kernel .
while this is the general formalism connecting effective actions to linear response theory , we do not need to use it because symmetry improvement constraints change the picture significantly .
in particular , we can derive the implications of the constraints for linear response theory without explicitly computing any equilibrium solutions , taking any derivatives of the effective action , performing any truncation of the self - energies , or worrying about renormalization of the bethe - salpeter equation .
our results are independent of all of these details .
here we review the symmetry improvement procedure since the derivation of the constraint and the limiting procedure are affected by the presence of sources .
the idea is to force the 2pi propagator @xmath123 to mimic the behavior of the corresponding 1pi function @xmath124/\delta\varphi\delta\varphi$ ] under a symmetry transformation . to derive the required
wi we perform a symmetry transformation of the 1piea @xmath125 ( note that our `` dewitt '' integration convention can be maintained if we define the generators to be spacetime delta functions : @xmath126 . )
this is the `` master '' wi governing all 1pi correlation functions of the theory . to find the identity governing the propagator
we take a field derivative then apply the equations of motion , giving @xmath127 which holds for all rotations @xmath128 so we define @xmath129 the symmetry improvement program then interprets the @xmath123 as the _ 2pi _ propagator .
one could also motivate the replacement @xmath130 based on the form of the right hand side of ( [ eq : unmodified - vev - eom ] ) .
we do not make this replacement . the only affect at the linear response level would be to renormalize @xmath131 in our results below which , since @xmath79 are freely chosen functions , gives no new physics . to enforce the wi
we add a lagrange multiplier term to the 2piea @xmath132 with @xmath133 where @xmath134 are the lagrange multiplier fields . in the previous paper @xcite
we included a transverse projector @xmath135 : @xmath136 to ensure that only goldstone modes are involved in the constraint .
this turns out to make no difference in equilibrium .
however , this constraint gives a different scheme beyond the equilibrium approximation .
( note that pilaftsis and teresi @xcite did not write the constraint in an @xmath0 covariant form , hence both of the above are valid generalizations of their procedure . ) using @xmath137 instead of @xmath138 merely seems to shuffle around the pathologies we derive below rather than remove them .
we shall henceforth use @xmath138 since dropping the projector greatly simplifies the following algebra . see appendix [ sec : the - alternate - constraint ] for discussion of the scheme using @xmath137 .
the equations of motion following from the symmetry improved effective action are @xmath139 where the last equation simplifies to @xmath140 on using the identity @xmath141 .
we now recall what happens to the right hand sides of these equations in equilibrium with @xmath142 ( this is reviewing @xcite ) .
the only non - trivial wis are @xmath143 , so that the constraint enforces @xmath103 , i.e. the goldstone mass vanishes if @xmath104 as expected . using homogeneity @xmath144 , and equations ( [ eq : eq - ssb - ansatz - vev])-([eq : eq - ssb - ansatz - delta ] )
the other equations of motion become @xmath145 where @xmath146 is the volume of spacetime .
now , since @xmath147 the right hand sides vanish unless @xmath148 .
we can consistently set @xmath149 and all other components of @xmath134 to zero .
thus the constraint is a singular one .
we regulate the divergence by setting @xmath150 and take the limit @xmath151 such that @xmath152 is a constant .
this gives @xmath153 thus the propagator equations of motion are unmodified and the vev equation is modified by the presence of a homogeneous force that acts to push @xmath49 away from the minimum of the effective potential to the point where @xmath154 .
in the linear response approximation @xmath79 and @xmath82 no longer vanish , and we must solve equations ( [ eq : wi ] ) , ( [ eq : si - eom - vev ] ) and ( [ eq : si - eom - delta ] ) to first order in the sources without any assumption of homogeneity . to do this
we expand all quantities @xmath105 , @xmath106 and @xmath155 and match terms order by order , considering the @xmath109 etc . as first order .
working first on ( [ eq : wi ] ) gives the pair of equations : @xmath156 the first equation is simply the equilibrium constraint as expected since the analysis of the last section holds with all quantities decorated with tildes .
the second equation is new . using the identity ( [ eq : variation - of - inverse - delta ] ) gives to first order @xmath157 the first term vanishes by virtue of ( [ eq : wi - zeroth - order ] ) , thus @xmath158 the nature of this equation is remarkable .
the _ constraint _ yields a _ wave equation _ for the fluctuations . working out the component equations
gives : @xmath159 the desired linear response functions are then just the green functions for these equations , i.e. @xmath160 etc .
there are two major flaws with these equations .
the first is a physical misprediction : @xmath161 propagates masslessly due to the equilibrium goldstone propagator even though the higgs propagator has a non - zero mass .
the second is that @xmath162 is overdetermined , i.e. its initial value problem is ill posed . in the fourier domain equations ( [ eq : phig - eom - w - higgs - propagator ] ) and ( [ eq : phig - eom - w - goldstone - propagator ] )
read @xmath163 which obviously only allows solutions if there are modes satisfying @xmath164 and if @xmath165 is only supported on these modes . if any other perturbation is applied or a solution of @xmath164 does not exist the system is inconsistent .
this is clearly not the expected behavior physically
. one can relax ( [ eq : phig - eom - w - higgs - propagator ] ) by using the projected constraint ( [ eq : projected - constraint ] ) instead of ( [ eq : simple - constraint ] ) , however this leads to further difficulties as discussed in appendix [ sec : the - alternate - constraint ] .
there is further danger lurking in the right hand sides of equations ( [ eq : si - eom - vev ] ) and ( [ eq : si - eom - delta ] ) .
since the constraint procedure involves a limit @xmath166 there is a danger that terms on the right hand sides can diverge .
consider the expansion of the right hand side of ( [ eq : si - eom - vev ] ) to first order : @xmath167 using ( [ eq : lagrange - multipliers ] ) the term proportional to @xmath168 can be written @xmath169 which , for the @xmath170 limit to exist , requires that the term in braces vanishes , i.e. @xmath171 a similar analysis for ( [ eq : si - eom - delta ] ) gives @xmath172 we call ( [ eq : secondary - constraint - phi - eom ] ) and ( [ eq : secondary - constraint - delta - eom ] ) the _ secondary _ constraints of the scheme and , by contrast , equation ( [ eq : wi - first - order ] ) the _ primary _ constraint .
the secondary constraints must be enforced so that no divergences appear in the @xmath166 limit of the equations of motion .
note that one can take @xmath147 without any problems in ( [ eq : secondary - constraint - delta - eom ] ) , so that constraint is satisfied identically .
similarly , using ( [ eq : variation - of - inverse - delta ] ) , ( [ eq : secondary - constraint - phi - eom ] ) becomes @xmath173 which is also automatically satisfied in the @xmath147 limit .
the fact that the secondary constraints automatically vanish is a consequence of using the unprojected constraint ( [ eq : simple - constraint ] ) . had we used ( [ eq : projected - constraint ] ) instead , the troublesome equation of motion ( [ eq : phig - eom - w - higgs - propagator ] )
would be gone , but the secondary constraints become non - trivial and lead again to pathologies ( see appendix [ sec : the - alternate - constraint ] ) .
we have shown that the imposition of symmetry improvement constraints is incompatible with the linear response approximation . since the original symmetry improvement scheme of pilaftsis and teresi
was not formulated @xmath0-covariantly there are actually two natural generalizations : a scheme we have used previously @xcite based on ( [ eq : projected - constraint ] ) and a new one based on the simpler constraint ( [ eq : simple - constraint ] ) .
both schemes are equivalent in equilibrium and both lead , in different ways , to pathologies in the linear response approximation .
there are two types of pathology appearing in our results .
the first is that the higgs field fluctuations propagate masslessly .
the second and more serious pathology is that the goldstone field fluctuations are over - determined , and this over - determination happens in both constraint schemes we consider ( though in different ways ) .
there is no simple modification of the constraint which could possibly fix these problems since we can understand them as a consequence of treating @xmath5 and @xmath9 as independent variables in the 2piea and the inability of symmetry improved 2piea to enforce also the three point _ vertex _
wi . to see these aspects of the problem consider the first order ward identity for the 1piea , which is the same in form as ( [ eq : wi - first - order ] ) , @xmath174 only now @xmath9 is not independent and @xmath175 is to be understood as arising purely from the variation @xmath109 .
to lowest order in @xmath17 : @xmath176 substituting this into the wi gives @xmath177 and working out the components we get the set of equations @xmath178\delta\varphi_{n}\left(y\right ) & = -j_{n},\\ \int_{y}\left[\tilde{\delta}_{h}^{-1}\left(x , y\right)+\frac{\lambda v^{2}}{3}\delta\left(x - y\right)\right]\delta\varphi_{g}\left(y\right ) & = -j_{g},\\ \int_{y}\tilde{\delta}_{g}^{-1}\left(x , y\right)\delta\varphi_{g}\left(y\right ) & = -j_{g},\end{aligned}\ ] ] which is to be contrasted with ( [ eq : phin - eom - massless])-([eq : phig - eom - w - goldstone - propagator ] ) .
this system is consistent if @xmath179 which , of course , is true .
thus , for the 1piea the wi propagates higgs and goldstone fluctuations with the correct masses and source terms .
this is expected because the wi was constructed to be satisfied by the 1pi correlation functions .
there is no longer any reason for this to work if @xmath110 is independent of @xmath109 .
we can extend this analysis by considering ( [ eq : wi - first - order ] ) again . from the definition in terms of the 1piea , @xmath180 , we can write @xmath181 and note that @xmath182 is the three point vertex function .
these relations no longer hold identically for the 2pi correlation functions , but they do hold numerically for the _ exact _ solutions of the untruncated 2pi equations of motion .
we can now extend this to 2piea by writing @xmath183 where @xmath184 encapsulates the additional variations of @xmath9 in the truncated 2piea formalism .
there is a wi for the three point vertex function ( c.f . equation ( 3.6 ) of @xcite ) , @xmath185 which is unaffected by the presence of sources . as with ( [ eq : wi ] ) , @xmath186 in the exact theory , but not automatically in truncations . combining these relations with ( [ eq : wi - first - order ] ) gives : @xmath187 which can be rearranged as @xmath188 if the right hand side vanishes one has ( using the symmetry of @xmath189 ) @xmath190 which is the correct equation of motion for fluctuations .
the failure of these equations to be satisfied is measured by the terms on the right hand side of ( [ eq : fluctuation - eom - vertex - wi ] ) , which have two distinct meanings . the first , @xmath191 ,
measures the failure of the _ vertex _ wi to be satisfied by 2pi approximations , while the second term , @xmath192 , measures the failure of the variations in the 2pi functions @xmath5 and @xmath9 to be linked according to the 1pi relations .
one could try to eliminate the second error term by constraining the variation @xmath110 to be related to @xmath109 in a appropriate way .
this would no longer be working strictly within the 2piea formalism .
rather , it would define a hybrid 2pi-1pi scheme where one computes the equilibrium properties using the symmetry improved 2piea , then defines a resummed 1piea by eliminating @xmath9 from @xmath193 in an appropriate way .
this is similar to the usual resummed 1piea method , only now symmetry improvement is applied self - consistently at the 2pi level .
the first error term is more formidable .
it comes down to the failure of the master 1pi wi in @xmath2piea truncations .
the master wi encodes relations between _ all _ correlation functions in the exact theory
. however , a symmetry improved @xmath2piea scheme only has the power to enforce constraints between the lowest @xmath2 of them . violations of wis involving higher order correlation functions are inevitable in any scheme with fixed finite @xmath2 .
this is not a major issue in equilibrium because one can solve for the @xmath2-point correlation functions in a self - consistently complete @xmath2-loop truncation , and so long as one does not care about the behavior of higher order correlation functions their problems remain invisible .
however , once external sources are turned on and the system departs from equilibrium , variations of the correlation functions appear and these can be related to higher order correlation functions .
violations of the higher order ward identities then feedback into the equations of motion for the fluctuations , leading to an inconsistent system overall .
this leads to our general conclusion : _ all symmetry improved @xmath2piea schemes are incompatible with the linear response approximation_. it is interesting that the same number of non - trivial constraints is found in both schemes , despite the attempt to reduce their number in the projected scheme .
this is because the projection operator depends on the fields @xmath5 , so it has non - trivial derivatives
. this will be the case for any @xmath0-covariant projection scheme .
this motivates a study of whether it is possible to _ non_-covariantly project out the troublesome constraints , i.e. explicitly break the symmetry in order to save it .
it is worth mentioning the logical possibility that the true behavior of the solutions to the truncated si-2piea equations of motion , when solutions exist , is non - analytic as a function of the sources in the neighborhood of equilibrium .
this would invalidate any attempted taylor series expansion in the sources and so we should expect problems at the linear response level .
we do not know at this time how to further analyze this possibility
. it would also be interesting to examine whether alternatives to symmetry improvement , such as the method for enforcing constraints using external sources @xcite , can be extended to non - equilibrium situations .
we plan in future work to investigate a scheme with _
softy _ imposed symmetry improvement as a potential workaround for the issues found here and in @xcite .
here we consider using the constraint ( [ eq : projected - constraint ] ) instead of ( [ eq : simple - constraint ] ) for the symmetry improvement . the projection operator in ( [ eq : projected - constraint ] )
forces only goldstone modes to be involved in the constraint .
note that the constraints which have been projected out are valid wis which are obeyed by the 1piea .
this alternate scheme consists of picking a subset of the wis to enforce , chosen in the only @xmath0-covariant way available .
the new constraint is @xmath194 at lowest order this becomes @xmath195 which is the same as before .
however , at first order there are new terms due to the variation of @xmath196 . for reference
we give the first and second derivatives of @xmath196 evaluated at @xmath107 : @xmath197 & = -\frac{1}{v}\left(\delta_{de}\delta_{cn}+\delta_{dn}\delta_{ce}-2\delta_{dn}\delta_{cn}\delta_{en}\right)\delta\left(x - z\right),\\ \frac{\delta^{2}p_{dc}^{\perp}\left(x\right)}{\delta\varphi_{f}\left(w\right)\delta\varphi_{e}\left(z\right)}\left[\tilde{\varphi}\right ] & = \frac{1}{v^{2}}f_{dcfe}\delta\left(z - w\right)\delta\left(x - z\right),\end{aligned}\ ] ] where @xmath198 which is symmetric in @xmath199 and in @xmath200 . in the linear response approximation
we need both @xmath201\delta\varphi_{e}\left(z\right)=-\frac{1}{v}\left(\delta_{cn}\delta\varphi_{d}\left(x\right)+\delta_{dn}\delta\varphi_{c}\left(x\right)-2\delta_{dn}\delta_{cn}\delta\varphi_{n}\left(x\right)\right),\\ \delta\left[\frac{\delta p_{cd}^{\perp}\left(x\right)}{\delta\varphi_{e}\left(z\right)}\right ] & \to\int_{w}\frac{\delta^{2}p_{dc}^{\perp}\left(x\right)}{\delta\varphi_{f}\left(w\right)\delta\varphi_{e}\left(z\right)}\left[\tilde{\varphi}\right]\delta\varphi_{f}\left(w\right)=\frac{1}{v^{2}}f_{dcfe}\delta\left(x - z\right)\delta\varphi_{f}\left(x\right).\end{aligned}\ ] ] the first order constraint is @xmath202 the second term vanishes due to ( [ eq : wi - zeroth - order ] ) and the first term is just ( [ eq : first - order - constraint - simple - scheme ] ) with only the @xmath203 equations picked out .
this gives the set of wave equations : @xmath204 which would be the end of the story if not for the secondary constraints .
however , do note that the higgs field still propagates masslessly . in order to find the secondary constraints we write the equations of motion @xmath205 the secondary constraints are that the variation of the terms multiplying @xmath206 vanish , since if they did not divergences would arise as @xmath166 .
we start work on the right hand side of ( [ eq : si - eom - delta - projected - scheme ] ) by demanding @xmath207\nonumber \\ & = -i\frac{1}{n-1}\tilde{\ell}_{hn}^{h}\tilde{p}_{gf}^{\perp}\int_{x}\delta\left[p_{fc}^{\perp}\left(x\right)\delta_{cd}^{-1}\left(x , z\right)\int_{y}\delta_{ea}^{-1}\left(w , y\right)t_{ab}^{gn}\varphi_{b}\left(y\right)\right],\end{aligned}\ ] ] which gives the constraint @xmath208\nonumber \\ & = \tilde{p}_{ef}^{\perp}\int_{x}\delta p_{fc}^{\perp}\left(x\right)\tilde{\delta}_{cd}^{-1}\left(x , z\right)\int_{y}\tilde{\delta}_{g}^{-1}\left(w , y\right)v+\tilde{p}_{ec}^{\perp}\int_{x}\delta\delta_{cd}^{-1}\left(x , z\right)\int_{y}\tilde{\delta}_{g}^{-1}\left(w , y\right)v\nonumber \\ & + \tilde{p}_{ad}^{\perp}\int_{x}\tilde{\delta}_{g}^{-1}\left(x , z\right)\int_{y}\delta\delta_{ea}^{-1}\left(w , y\right)v-\delta_{en}\tilde{p}_{gd}^{\perp}\int_{x}\tilde{\delta}_{g}^{-1}\left(x , z\right)\int_{y}\tilde{\delta}_{h}^{-1}\left(w , y\right)\delta\varphi_{g}\left(y\right)\nonumber \\ & + \tilde{p}_{ed}^{\perp}\int_{x}\tilde{\delta}_{g}^{-1}\left(x , z\right)\int_{y}\tilde{\delta}_{g}^{-1}\left(w , y\right)\delta\varphi_{n}\left(y\right),\end{aligned}\ ] ] we find again that every term is proportional to @xmath147 so the constraint is satisfied automatically . now working on the right hand side of ( [ eq : si - eom - phi - projected - scheme ] )
gives the secondary constraint @xmath209\nonumber \\ & = i\tilde{p}_{gf}^{\perp}\int_{x}\delta\left[\frac{\delta p_{fc}^{\perp}\left(x\right)}{\delta\varphi_{d}\left(z\right)}\right]\int_{y}\tilde{\delta}_{ca}^{-1}\left(x , y\right)t_{ab}^{gn}\tilde{\varphi}_{b}\left(y\right ) + i\tilde{p}_{gf}^{\perp}\int_{x}\frac{\delta p_{fc}^{\perp}\left(x\right)}{\delta\varphi_{d}\left(z\right)}\left[\tilde{\varphi}\right]\int_{y}\delta\delta_{ca}^{-1}\left(x , y\right)t_{ab}^{gn}\tilde{\varphi}_{b}\left(y\right)\nonumber \\ & + i\tilde{p}_{gf}^{\perp}\int_{x}\frac{\delta p_{fc}^{\perp}\left(x\right)}{\delta\varphi_{d}\left(z\right)}\left[\tilde{\varphi}\right]\int_{y}\tilde{\delta}_{ca}^{-1}\left(x , y\right)t_{ab}^{gn}\delta\varphi_{b}\left(y\right ) + i\tilde{p}_{gf}^{\perp}\int_{x}\frac{\delta p_{fc}^{\perp}\left(x\right)}{\delta\varphi_{d}\left(z\right)}\left[\tilde{\varphi}\right]t_{ca}^{gn}j_{a}\left(x\right)\nonumber \\ & + i\tilde{p}_{gf}^{\perp}\int_{x}\delta p_{fc}^{\perp}\left(x\right)\tilde{\delta}_{ca}^{-1}\left(x , z\right)t_{ad}^{gn } + i\tilde{p}_{gf}^{\perp}\int_{x}\tilde{p}_{fc}^{\perp}\left(x\right)\delta\delta_{ca}^{-1}\left(x , z\right)t_{ad}^{gn}.\end{aligned}\ ] ] ( note that the @xmath79 term in the third line is present because in the first line the @xmath210 $ ] truly means `` linear piece of @xmath21 $ ] '' , not `` variation of @xmath21 $ ] . '' ) plugging in the expressions for @xmath211 , @xmath212 and @xmath213 $ ] and simplifying gives @xmath214 the first term vanishes as @xmath147 .
the second term also vanishes as @xmath215 the remaining terms become , in the case @xmath216 : @xmath217 which is satisfied identically .
the remaining constraint is for @xmath218 : @xmath219 again the last term vanishes and @xmath220 the second line follows by the hermiticity of @xmath221 .
finally we obtain another wave equation @xmath222 again @xmath223 is over - determined , but by a _ different _
equation this time .
here we investigate a very simple mechanical system which illustrates several of the unusual features of the constraint procedure and linear response formulation we have used .
it shows : ( a ) why the lagrange multiplier diverges , ( b ) why constraints must be imposed in the linear response approximation to begin with , ( c ) why secondary constraints arise . consider a unit mass classical particle constrained to move without friction on a circular hoop of radius @xmath224 in the @xmath225 plane .
its lagrangian is @xmath226 where @xmath227 is the lagrange multiplier and the form of the constraint @xmath228 is chosen to mimic the singular constraint procedure .
the equations of motion are @xmath229 we now consider the source free case @xmath230 . the constraint terms vanish unless @xmath231 as @xmath232 .
we set @xmath233 and @xmath234 and take the limit such that @xmath235 is a constant .
note that @xmath236 .
then the equations of motion become @xmath237 with the solutions @xmath238,\\ y & = r\sin\left[\omega\left(t - t_{0}\right)\right],\end{aligned}\ ] ] where @xmath30 and @xmath239 are determined by the initial conditions .
we take as the static solution @xmath240 and @xmath241 , which determines @xmath242 and @xmath243 .
now we turn on the sources @xmath244 and @xmath245 and investigate the linear response by setting @xmath246 , @xmath247 , @xmath248 where the tilde variables are the source free solutions .
the variation of the constraint is @xmath249 regardless of the behavior of @xmath250 and @xmath251 , so long as they are non - singular in the @xmath151 limit .
however , the first order equations of motion become @xmath252 where we introduce the radial force @xmath253\left(\tilde{x},\tilde{y}\right)$ ] , whose physical function is to balance the applied force normal to the constraint surface , resulting in the net transverse source @xmath254 .
now notice the terms proportional to @xmath255 in the equations of motion . in order for these terms to be well behaved in the limit @xmath256 we must have @xmath257 , i.e. the response remains within the constraint surface ( to first order ) .
thus the vanishing of these terms in addition to the vanishing of @xmath258 is required to fully enforce that the response be tangential to the constraint surface .
we also note that by examining the @xmath259 terms in the equation of motion one can identify which component of the applied force acts normal to the constraint surface ( and hence produce no physical response ) . applying the static solution we find @xmath260\left(r,0\right)$ ] . for this
to be well behaved as @xmath256 requires @xmath261 , which also determines @xmath262 via the @xmath250 equation of motion .
the @xmath251 equation of motion is @xmath263 with the solution ( taking into account the initial conditions @xmath264 ) : @xmath265 now we can compare this to the exact solution . substituting the ansatz @xmath266 , @xmath267 ,
the lagrangian and equation of motion become @xmath268 where @xmath269 is the tangential component of the force .
the solution satisfying the initial conditions @xmath270 is @xmath271 to linear order in @xmath272 the @xmath273 and @xmath274 components are just @xmath275 which , on putting @xmath276 , gives @xmath277 which is just the solution obtained previously .
if in contrast we never imposed the constraints on the linear response solution we would have the equations of motion @xmath278 with the solutions @xmath279 the @xmath251 component is correct but @xmath250 is in error already at linear order .
in fact , the solution should not even depend on @xmath244 until second order .
25ifxundefined [ 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/physrev.85.631 [ * * , ( ) ] @noop _ _ , oxford graduate texts ( , , ) link:\doibase 10.1016/s0370 - 1573(03)00263 - 1 [ * * , ( ) ] in link:\doibase 10.1063/1.1843591 [ _ _ ] , vol .
( , ) pp . link:\doibase 10.1016/j.nuclphysb.2015.09.021 [ * * , ( ) ] link:\doibase 10.1103/physrev.127.965 [ * * , ( ) ] link:\doibase 10.1007/bf02750573 [ * * , ( ) ] link:\doibase 10.1103/physrev.117.22 [ * * , ( ) ] link:\doibase 10.1103/physrev.118.1417 [ * * , ( ) ] link:\doibase 10.1103/physrev.127.1391 [ * * , ( ) ] link:\doibase 10.1103/physrevd.10.2428 [ * * , ( ) ] link:\doibase 10.1063/1.1704064 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1002/prop.19820300402 [ * * , ( ) ] link:\doibase 10.1007/bf02811226 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1016/j.nuclphysb.2013.06.004 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2014.02.011 [ * * , ( ) ] http://journals.aps.org/prd/abstract/10.1103/physrevd.92.085036 [ * * , ( ) ] link:\doibase 10.1088/1742 - 6596/631/1/012008 [ * * , ( ) ] @noop ( ) , link:\doibase 10.1103/physrevd.91.085020 [ * * ( ) , 10.1103/physrevd.91.085020 ] @noop ( ) , link:\doibase 10.1016/j.nuclphysb.2016.02.022 [ * * , ( ) ] | we investigate the linear response of an @xmath0 scalar quantum field theory subject to external perturbations using the symmetry improved two particle irreducible effective action ( si-2piea ) formalism [ a. pilaftsis and d. teresi , nucl .
phys .
b874 , 594 ( 2013 ) ] . despite satisfactory equilibrium behavior
, we find a number of unphysical effects at the linear response level .
goldstone boson field fluctuations are over - determined , with the only consistent solution being to set the fluctuations and their driving sources to zero , except for momentum modes where the higgs and goldstone self - energies obey a particular relationship .
also higgs field fluctuations propagate masslessly , despite the higgs propagator having the correct mass .
these pathologies are independent of any truncation of the effective action and still exist even if we relax the over - determining ward identities , so long as the constraint is formulated @xmath0-covariantly .
we discuss possible reasons for the apparent incompatibility of the constraints and linear response approximation and possible ways forward . |
relatively uniform light curves and spectral evolution of type ia supernovae ( sne ia ) have led to the use of sne ia as a `` standard candle '' to determine cosmological parameters . whether a statistically significant value of the cosmological constant can be obtained depends on whether the peak luminosities of sne ia are sufficiently free from the effects of cosmic and galactic evolutions @xcite .
sne ia have been widely believed to be a thermonuclear explosion of a mass - accreting white dwarf ( wd ) .
however , the immediate progenitor binary systems have not been clearly identified yet @xcite . in order to address the above questions regarding the nature of high - redshift sne ia
, we need to identify the progenitors systems and examine the `` evolutionary '' effects ( or environmental effects ) on those systems @xcite . here
we review several issues such as chandra vs. sub - chandra , double degenerates vs. single degenerate ( aic
vs. sn ia ) , c / o ratio , and rotation . for complete discussion , see earlier reviews on sn ia progenitors @xcite .
there exist two models proposed as progenitors of sne ia : 1 ) the chandrasekhar mass model , in which a mass - accreting carbon - oxygen ( c+o ) wd grows its mass @xmath1 up to the critical mass @xmath2 near the chandrasekhar mass and explodes as an sn ia ( e.g. , @xcite ) , and 2 ) the sub - chandrasekhar mass model , in which an accreted layer of helium atop a c+o wd ignites off - center for a wd mass well below the chandrasekhar mass ( e.g. , @xcite ) .
the early time spectra of the majority of sne ia are in excellent agreement with the synthetic spectra of the chandrasekhar mass models , while the spectra of the sub - chandrasekhar mass models are too blue to be comparable with the observations @xcite .
however , the peculiar sn ia 2000cx might be the sub - chandrasekhar mass explosion as will be discussed in 6 .
, scaledwidth=85.0% ] for the evolution of accreting wds toward the chandrasekhar mass , two scenarios have been proposed : 1 ) a double degenerate ( dd ) scenario , i.e. , merging of double c+o wds @xcite , and 2 ) a single degenerate ( sd ) scenario , i.e. , accretion of hydrogen - rich matter via mass transfer from a binary companion ( e.g. , @xcite ) .
the issue of dd vs. sd is still debated , although theoretical modeling has indicated that the merging of wds leads to the accretion - induced collapse ( aic ) rather than sn ia explosion @xcite . whether the effect of rotation on accretion changes the conclusion on aic vs. sn ia
will be discussed in 5.2 . in the sd chandrasekhar mass model
, a wd explodes as a sn ia only when its rate of the mass accretion ( @xmath3 ) is in a certain narrow range ( fig .
[ nomotof1 ] @xcite ) .
in particular , if @xmath3 exceeds the critical rate @xmath4 , the accreted matter extends to form a common envelope @xcite . here
@xmath4 is the rate at which steady burning can process the accreted hydrogen into helium as @xmath5 . for @xmath3@xmath6 @xmath4 , the strong peak of fe opacity @xcite drives the radiation - driven wind from the wd @xcite .
if the wind is sufficiently strong , the wd can avoid the formation of a common envelope and steady hydrogen burning increases its mass continuously at a rate @xmath4 by blowing the extra mass away in a wind . for @xmath3@xmath7 0.5 @xmath4 ,
hydrogen shell burning becomes unstable to trigger weak shell flashes but still burns a large fraction of accreted hydrogen .
recurrent novae appear in the upper - right region of fig .
[ nomotof1 ] . in this way ,
strong winds from the accreting wd play a key role to increase the wd mass to @xmath8 . for the actual binary systems which grow @xmath1 to @xmath8 ,
the following two systems are appropriate .
this is a symbiotic binary system consisting of a wd and a low mass red - giant ( rg ) @xcite .
the immediate progenitor binaries in this symbiotic channel to sne ia may be observed as symbiotic stars , luminous supersoft x - ray sources , or recurrent novae like t crb or rs oph , depending on the wind status @xcite .
this is a system consisting of a mass - accreting wd and a lobe - filling , more massive , slightly evolved main - sequence or sub - giant star . in this scenario ,
a c+o wd is originated , not from an agb star with a c+o core , but from a red - giant star with a helium core of @xmath9 .
the helium star , which is formed after the first common envelope evolution , evolves to form a c+o wd of @xmath10 with transferring a part of the helium envelope onto the secondary main - sequence star @xcite .
a part of the progenitor systems are identified as the luminous supersoft x - ray sources @xcite during steady h - burning ( but without wind to avoid extinction ) , or the recurrent novae like u sco if h - burning is weakly unstable @xcite
. actually these objects are characterized by the accretion of helium - rich matter .
the rate of sne ia originating from these channels in our galaxy is estimated with equation ( 1 ) of @xcite .
the realization frequencies of sne ia through the wd+rg and wd+ms channels are estimated as @xmath11 0.0017 yr@xmath12 ( wd+rg ) and @xmath11 0.001 yr@xmath12 ( wd+ms ) , respectively .
the total sn ia rate of the wd+ms / wd+rg systems becomes @xmath11 0.003 yr@xmath12 , which is close enough to the inferred rate of our galaxy .
the optically thick winds are driven by a strong peak of opal opacity due to iron lines .
thus the wind velocity @xmath13 is lower for lower [ fe / h ] .
the sn ia regions are much smaller for lower metallicity , and very few sn ia occurs at [ fe / h ] @xmath14 in this model .
it is possible to test such metallicity effects on sne ia with the chemical evolution of galaxies . in the one - zone uniform model for the chemical evolution of the solar neighborhood ,
the heavy elements in the metal - poor stars originate from the mixture of the sn ii ejecta of various progenitor masses .
the abundances averaged over the progenitor masses of sne ii predicts [ o / fe ] @xmath15 ( e.g. , @xcite ) .
later sne ia start ejecting mostly fe , so that [ o / fe ] decreases to @xmath16 around [ fe / h ] @xmath17 .
the low - metallicity inhibition of sne ia predicts that the decrease in [ o / fe ] starts at [ fe / h ] @xmath18
. such an evolution of [ o / fe ] well explains the observations @xcite .
we should note that some anomalous stars have [ o / fe ] @xmath11 0 at [ fe / h ] @xmath19 .
the presence of such stars , however , is not in conflict with the metallicity dependence of sne ia , but can be understood as follows : the formation of such anomalous stars ( and the diversity of [ o / fe ] in general ) indicates that the interstellar materials were not uniformly mixed but contaminated by only a few sne ii ( or even single sn ii ) ejecta .
the fe and o abundances produced by a single sn ii vary depending mainly on the mass of the progenitor .
relatively smaller mass sne ii ( @xmath20 ) and higher explosion energies tend to produce [ o / fe ] @xmath16 @xcite .
those metal - poor stars with [ o / fe ] @xmath16 may be born from the interstellar medium polluted by such sne ii .
alternatively , such stars were born in nearby dwarf spheroidal galaxies and captured by our galaxy ( see below ) .
the chemical abundances of individual stars in local dwarf spheroidal galaxies ( dsph ) have been measured ( e.g. , @xcite for recent observations and references therein ) .
these stars have [ fe / h ] @xmath21 but the ratios between the @xmath22 elements and fe are found to be as low as [ @xmath22/fe ] @xmath11 0 being significantly lower than the metal - poor halo - stars in our galaxy .
if the low [ @xmath22/fe ] was due to the fe - enrichment by sne ia , it implies that sne ia appeared in low metallicity environment , which may be difficult to explain with the metallicity dependent sn ia model
. however , it is possible to produce such low [ @xmath22/fe ] with core - collapse sne . as discussed above , yields from 13 - 15 @xmath23 stars have low ( even sub - solar ) [ @xmath22/fe ] .
thus if imf in those dwarf spheroidal galaxies is steep enough or the upper limit mass of core - collapse sne is truncated around 20 @xmath23 @xcite , the contribution of 13 - 15 @xmath23 stars could be significantly larger than in our galaxy .
also it is possible that some fraction of agb stars suffers less mass loss because of low metallicity , thus growing degenerate c+o cores to the chandrasekhar mass before losing entire envelopes .
if this is the case , those metal - poor agb stars undergo sn ia - like explosions ( sn i+1/2 ) .
if the mass range is very narrow , say @xmath11 7.5 - 8.0 @xmath23 , sne i+1/2 do not affect [ o / fe ] for our galaxy if integrated over the salpeter imf up to 50 @xmath23 @xcite . however , if imf is steeper or the upper - mass limit is lower , contributions of sne i+1/2 and 13 - 15 @xmath23 stars are larger , thus reducing [ @xmath22/fe ] .
the abundance patterns in dsph as well as some damped lyman-@xmath22 systems can provide important constraint on the progenitor mass ranges of sne ia , sne i+1/2 , and sne ii .
also the appearance of sne ia in more metal - rich regions of dsph may not be impossible in view of spatial inhomogeneity .
there are some observational indications that sne ia depend on the type of the host galaxies .
the most luminous sne ia seem to occur only in spiral galaxies , while both spiral and elliptical galaxies are hosts for dimmer sne ia . thus the mean peak brightness is dimmer in ellipticals than in spiral galaxies @xcite .
also the sne ia rate per unit luminosity at the present epoch is almost twice as high in spirals as in ellipticals @xcite .
umeda et al .
@xcite suggested that the variation of the c / o ratio is the main cause of the variation of sne ia brightness , with larger c / o ratio yielding brighter sne ia .
here we will show that the c / o ratio depends indeed on the metallicity and age of the companion of the wd , and that the model can explain most of the observational trends discussed above .
we then make some predictions about the brightness of sn ia at higher redshift .
the c / o ratio in c+o wds depends primarily on the main - sequence mass of the wd progenitor and on metallicity .
the most important metallicity effect is that the radiative opacity is smaller for lower @xmath24 .
therefore , a star with lower @xmath24 is brighter , thus having a shorter lifetime than a star with the same mass but higher @xmath24 . in this sense ,
the effect of reducing metallicity for these stars is almost equivalent to increasing a stellar mass . for stars with larger masses and/or smaller @xmath24 , the luminosity is higher at the same evolutionary phase . with a higher nuclear energy generation rate , these stars have larger convective cores during h and he burning , thus forming larger he and c - o cores .
c mass fraction included in the convective core of mass , @xmath25 , just before the sn ia explosion as a function of the c+o core mass before the onset of mass accretion , @xmath26 .
the lower bounds of @xmath26 obtained from the age effects and the conditions for strong wind to blow are also shown by arrows @xcite .
[ nomotof2],scaledwidth=90.0% ] according to the evolutionary calculations for 3@xmath279 @xmath28 stars @xcite , the c / o ratio and its distribution are determined in the following evolutionary stages of the close binary .
\(1 ) at the end of central he burning in the 3@xmath279 @xmath28 primary star , c / o@xmath29 in the convective core .
the mass of the core is larger for more massive stars .
\(2 ) after central he exhaustion , the outer c+o layer grows via he shell burning , where c / o@xmath30 @xcite .
( 3a ) if the primary star becomes a red giant ( case c evolution ) , it then undergoes the second dredge - up , forming a thin he layer , and enters the agb phase .
the c+o core mass , @xmath26 , at this phase is larger for more massive stars . for a larger @xmath26
the total carbon mass fraction is smaller .
( 3b ) when it enters the agb phase , the star greatly expands and is assumed here to undergo roche lobe overflow ( or a super - wind phase ) and to form a c+o wd .
thus the initial mass of the wd , @xmath31 , in the close binary at the beginning of mass accretion is approximately equal to @xmath26 .
( 4a ) if the primary star becomes a he star ( case bb evolution ) , the second dredge - up in ( 3a ) corresponds to the expansion of the he envelope .
( 4b ) the ensuing roche lobe overflow again leads to a wd of @xmath31 = @xmath26 .
\(5 ) after the onset of mass accretion , the wd mass grows through steady h burning and weak he shell flashes , as described in the wd wind model .
the composition of the growing c+o layer is assumed to be c / o=1 .
\(6 ) the wd grows in mass and ignites carbon when its mass reaches @xmath32 , as in the model c6 @xcite .
carbon burning grows into a deflagration for a central temperature of @xmath33 k and a central density of @xmath34 g @xmath35 . at this stage ,
the convective core extends to @xmath36 and the material is mixed almost uniformly ( model c6 ) .
figure [ nomotof2 ] shows the carbon mass fraction @xmath0(c ) in the convective core of this pre - explosive wd , as a function of metallicity ( @xmath24 ) and initial mass of the wd before the onset of mass accretion , @xmath26 .
we note : \1 ) @xmath0(c ) is smaller for larger @xmath37 .
\2 ) the dependence of @xmath0(c ) on metallicity is small when plotted against @xmath26 , even though the relation between @xmath26 and the initial stellar mass depends sensitively on @xmath24 @xcite . in the chandrasekhar mass models for sne ia , brightness of sne ia
is determined mainly by the mass of @xmath38ni synthesized ( @xmath39 ) .
observational data suggest that @xmath40 for most sne ia lies in the range @xmath41 ( e.g. , @xcite ) . here
we postulate that @xmath40 and consequently brightness of a sn ia increase as the progenitors c / o ratio increases ( and thus @xmath42 decreases ) . as illustrated in figure
[ nomotof2 ] , the range of @xmath43 is the result of an @xmath0(c ) range @xmath44 , which is the range of @xmath0(c ) values of our progenitor models . the @xmath0(c ) @xmath40
@xmath42 relation we adopt is still only a working hypothesis , which needs to be proved from studies of the turbulent flame during explosion ( e.g. , @xcite ) . assuming the relation between @xmath40 and @xmath45(c ) given in figure [ nomotof2 ] , the model predicts the absence of brighter sne ia in lower metallicity environment . in this model
, the age of the progenitor system also constrains the range of @xmath0(c ) in sne ia . in the sd scenario ,
the lifetime of the binary system is essentially the main - sequence lifetime of the companion star , which depends on its initial mass @xmath46 . in order for the wd mass to reach @xmath8 ,
the donor star should transfer enough material at the appropriate accretion rates .
the donors of successful sn ia cases are divided into two categories : one is composed of slightly evolved main - sequence stars with @xmath47 ( for @xmath24=0.02 ) , and the other of red - giant stars with @xmath48 ( for @xmath24=0.02 ) @xcite . if the progenitor system is older than 2 gyr , it should be a system with a donor star of @xmath49 in the red - giant branch .
systems with @xmath50 become sne ia in a time shorter than 2 gyr .
likewise , for a given age of the progenitor system , @xmath46 must be smaller than a limiting mass .
this constraint on @xmath46 can be translated into the presence of a minimum @xmath26 for a given age , as follows : for a smaller @xmath46 , i.e. for the older system , the total mass which can be transferred from the donor to the wd is smaller . in order
for @xmath1 to reach @xmath8 , therefore , the initial mass of the wd , @xmath51 , should be larger .
this implies that the older system should have larger minimum @xmath26 as indicated in figure [ nomotof2 ] . using the @xmath0(c)-@xmath26 and @xmath40-@xmath0(c ) relations ( fig .
[ nomotof2 ] ) , we conclude that wds in older progenitor systems have a smaller @xmath0(c ) , and thus produce dimmer sne ia
. among the observational indications which can be compared with our model is the possible dependence of the sn brightness on the morphology of the host galaxies .
hamuy et al .
@xcite found that the most luminous sne ia occur in spiral galaxies , while both spiral and elliptical galaxies are hosts to dimmer sne ia .
hence , the mean peak brightness is lower in elliptical than in spiral galaxies . in our model ,
this property is simply understood as the effect of the different age of the companion . in spiral galaxies ,
star formation occurs continuously up to the present time .
hence , both wd+ms and wd+rg systems can produce sne ia . in elliptical galaxies ,
on the other hand , star formation has long ended , typically more than 10 gyr ago .
hence , wd+ms systems can no longer produce sne ia . since a wd with smaller @xmath26 is assumed to produce a brighter sn ia ( larger @xmath52 ) , our model predicts that dimmer sne ia occur both in spirals and in ellipticals , while brighter ones occur only in spirals .
the mean brightness is smaller for ellipticals and the total sn ia rate per unit luminosity is larger in spirals than in ellipticals .
these properties are consistent with observations .
our model predicts that when the progenitors belong to an old population , or to a low metal environment , the number of very bright sne ia is small , so that the variation in brightness is also smaller , which is shown in figure [ nomotof3 ] . in spiral galaxies ,
the metallicity is significantly smaller at redshifts @xmath53 , and thus both the mean brightness of sne ia and its range tend to be smaller ( fig .
[ nomotof3 ] ) . at @xmath54 sne
ia would not occur in spirals at all because the metallicity is too low .
in elliptical galaxies , on the other hand , the metallicity at redshifts @xmath55 is not very different from the present value .
however , the age of the galaxies at @xmath56 is only about 5 gyr , so that the mean brightness of sne ia and its range tend to be larger at @xmath53 than in the present ellipticals because of the age effect .
we note that the variation of @xmath0(c ) is larger in metal - rich nearby spirals than in high redshift galaxies .
therefore , if @xmath0(c ) is the main parameter responsible for the diversity of sne ia , and if the light curve shape ( lcs ) method is confirmed by the nearby sne ia data @xcite , the lcs method can also be used to determine the absolute magnitude of high redshift sne ia . ]
here we consider the metallicity effects only on the c / o ratio ; this is just to shift the main - sequence mass - @xmath42 relation , thus resulting in no important evolutionary effect .
however , some other metallicity effects could give rise to evolution of sne ia between high and low redshifts ( i.e. , between low and high metallicities ) ( see , e.g. , @xcite ) .
the accreting wd gains angular momentum from the rotating disk in addition to the mass , thus rotating faster and faster .
rapid rotation of a wd affects its limiting mass and final structure .
the effects of rotation on the evolution of accreting wds have been studied mostly with 1d approximate models ( e.g. , @xcite ) .
uenishi et al .
@xcite have calculated the axisymmetric structure of _ uniformly _ rotating wds and followed the evolutionary sequence in the total angular momentum @xmath57 and the wd mass @xmath58 ( fig . [ nomotof4 ] ) .
here accreting gas is supposed to obey keplerian rotation law and gives its angular momentum to the wd .
the almost straight lines show the evolutionary tracks of the accreting wds starting from the initial masses of 0.6 @xmath23 , 0.8 @xmath23 , and 1.07 @xmath23 , respectively . after gaining @xmath11 0.1 @xmath23
, the wds reach the critical rotation at the upper edge of the @xmath59 diagram where the equatorial centrifugal force is equal to the gravity .
for wds which have reached the critical rotation , paczyski @xcite and pooham & narayan @xcite found a solution that permits the wd to accrete without becoming secularly unstable .
in such a solution , the angular momentum is transported backwards from the wd to the disk while the mass of the wd increases .
then the wd evolves along the upper envelope in the @xmath57-@xmath58 plane and eventually explodes at the upper - right corner .
( along the right edge of the diagram , the central density of wds is 2 @xmath60 10@xmath61 g @xmath35 where carbon is ignited . ) then all the rapidly rotating wds explode with the same @xmath62 .
in other words , _ uniform _ rotation leads to the uniformity rather than the diversity of sne ia .
( 10@xmath63 erg sec ) and the wd mass @xmath58 ( @xmath23 ) of uniformly rotating wds starting from the initial masses of 0.6 @xmath23 , 0.8 @xmath23 , and 1.07 @xmath23 @xcite .
[ nomotof4],scaledwidth=90.0% ] however , rotation could produce the diversity of sne ia in the following way .
the uniform rotation is realized if the timescale of angular momentum transport in the wds is much shorter than the accretion time scale .
if the angular momentum transport is much slower , then the initially non - rotating part of the wd may remain non - rotating , while the accreted outer part of the wd rotates fast and reaches the critical rotation .
when the wd reaches the carbon ignition at the limiting mass , the angular momentum brought into the wd is larger ( smaller ) for smaller ( larger ) @xmath64 .
the wd with smaller @xmath64 grows to larger @xmath65 at the explosion .
although @xmath65 @xmath11 1.45 @xmath23 are not so different from each other , larger @xmath65 could produce brighter sne ia .
also the distribution of the angular velocity is different , i.e. , the non - rotating core is smaller ( larger ) for smaller ( larger ) @xmath64
. such a difference might affect the flame propagation and lead to different amount of @xmath38ni .
if the wd with smaller @xmath64 produce a brighter sn ia with the effect of rotation , the upper limit of sne ia brightness is higher in spirals than in ellipticals .
sne ia in spirals can originate from wds with a wide range of @xmath64 , thus having a large dispersion of brightness than those in ellipticals .
this could explain the observation that the most luminous sne ia appear to be observed only in spirals , while dimmer sne ia are observed in both spirals and ellipticals .
nomoto & iben @xcite and saio & nomoto @xcite have simulated the merging of double wds in 1d and shown that the rapidly accreting wds undergo off - center carbon ignition if @xmath3 @xmath6 2 @xmath60 10@xmath66 @xmath28 yr@xmath12 because of compressional heating . afterwards
carbon flame propagates inward through the center and converts c+o into o+ne+mg .
then the final outcome is most likely accretion - induced collapse rather than sne ia .
recently piersanti et al .
@xcite calculated the evolution of wds with rotation in 1d approximation and argued that the lifting effect of rotation reduces compressional heating , and the wd could avoid off - center carbon ignition and reach the central carbon ignition to produce sne ia .
however , they did not follow the accretion after the critical rotation is reached because the backward transport of angular momentum to the disk was not taken into account .
saio & nomoto @xcite have also calculated the accretion of c+o onto the c+o wd with rotation for various timescale of angular momentum transport in 1d approximation .
the outermost layer of the accreting wd quickly reaches the critical rotation as in figure [ nomotof4 ] .
afterwards , the angular momentum is transported backward to disk and accretion continues . for @xmath3@xmath6 4 @xmath60 10@xmath66 @xmath28 yr@xmath12 ,
off - center carbon burning is ignited prior to the central c - ignition .
thus the lifting effect of rotation increases the critical accretion rate for the occurrence of off - center c - ignition by a factor of @xmath11 2 compared with the non - rotating case , but the basic conclusion is the same as non - rotating case , i.e. , the accretion - induced collapse is the most likely outcome in the dd scenario @xcite .
, scaledwidth=90.0% ] supernova 2000cx was a very well observed sn ia that in certain respects resembled the peculiar `` powerful '' sn 1991 t , but also showed some photometric and spectroscopic characteristics that are unprecedented among well observed sne ia ( li et al .
@xcite ; hereafter l01 ) .
rudy et al .
( @xcite ; hereafter r02 ) obtained infrared spectra covering the range 0.8 to 2.5 @xmath67 m , 6 and 5 days before optical maximum .
they attributed an absorption feature near 1.0 @xmath67 m to mg ii @xmath6810926 .
hatano et al .
@xcite have used spectrum code synow and found that the absorption feature is most likely due to he i @xmath6810830 .
the presence of helium would have important implications for models . in figure [ nomotof5 ] the infrared spectrum @xcite
is compared with four synthetic spectra , each containing lines of only one ion .
the synthetic spectra have @xmath69 k and @xmath70 km s@xmath12 ( mg ii ) , 16,000 km s@xmath12 ( c i ) , and 20,000 km s@xmath12(si i and he i ) . mg ii @xmath6810926 has been considered as a possible identification for the absorption feature by r02 .
however , the upper left panel of figure [ nomotof5 ] shows that when @xmath6810926 is strong enough to account for the 1.0 @xmath67 m absorption , other mg ii lines produce unwanted absorptions near 8600 ( @xmath689226 ) and 9400 ( @xmath689632 ) . in lte @xmath689226
has a larger optical depth than @xmath6810926 for any reasonable excitation temperature .
it is unlikely that mg ii is responsible for the 1.0 @xmath67 m feature .
the lower left panel of figure [ nomotof5 ] shows that c i , with @xmath6810695 producing the 1.0 @xmath67 m absorption , has a similar problem of unwanted features , due to @xmath689055 and @xmath6811755 .
the upper right panel shows that a group of si i lines ( mainly a multiplet at @xmath6810790 ) fit the 1.0 @xmath67 m feature without severe problems ; a feature due to @xmath6812047 is from the same lower level and is stronger , but one might argue that it is possibly present , blended with other lines , in the observed spectrum . however , unless the level of ionization in sn 2000cx was much lower at the epoch of the infrared spectrum than it was at the time of the first optical spectrum , just a few days later , si i lines would not be expected .
finally , the lower right panel of figure [ nomotof5 ] shows that he i @xmath6810830 , with a optical depth at the photosphere of 0.9 , can account for the 1.0 @xmath67 m feature without causing any obvious problems . by this process of elimination , hatano et al .
@xcite find that the most likely identification of the 1.0 @xmath67 m feature in sn 2000cx is he i @xmath6810830 .
the evidence for helium in sn 2000cx , although not conclusive , is strong enough to be taken seriously . both l01 and r02 suggest a delayed detonation as a favorable model for sn 2000cx .
in particular , r02 argued that the 1.0 @xmath67 m feature is due to mg ii and the presence of such high velocity mg as @xmath6 20,000 km s@xmath12 supports the delayed detonation model .
however , this argument does not hold if the 1.0 @xmath67 m feature is due to he i. rather the delayed detonation models do not contain such high velocity he ( e.g. , @xcite ) . the presence of he , instead , may suggest the sub - chandrasekhar model for sn 2000cx . in the sub - chandrasekhar mass models ,
the he detonation in the outer he layer synthesize mostly @xmath38ni but significant amount of he is also present due to strong @xmath22-rich freezeout @xcite .
the velocities of he are @xmath11 11,000 - 14,000 km s@xmath12 near the bottom of the he layer and @xmath11 25,000 - 30,000 km s@xmath12 near the surface , depending on @xmath1 and the mass of the he layer ( @xmath71 ) .
the he velocities and the @xmath38ni mass of sn 2000cx could be consistent with the model with relatively large @xmath1 and @xmath71 . in
the outer detonated layers , little si is produced , which is also consistent with the weak si features of sn 2000cx .
the presence of high velocity he with @xmath38ni ( and little si ) is also seen in the late detonation model w7dhe of the chandrasekhar mass white dwarf @xcite .
the explosion of this type would occur for slower accretion than the sub - chandrasekhar mass explosion @xcite . | a key question for supernova cosmology is whether the peak luminosities of type ia supernovae ( sne ia ) are sufficiently free from the effects of cosmic and galactic evolution . to answer this question , we review the currently popular scenario of sn ia progenitors , i.e. , the single degenerate scenario for the chandrasekhar mass white dwarf ( wd ) models .
we identify the progenitor s evolution with two channels : ( 1 ) the wd+rg ( red - giant ) and ( 2 ) the wd+ms ( near main - sequence he - rich star ) channels . the strong wind from accreting wds plays a key role , which yields important age and metallicity effects on the evolution .
we suggest that the variation of the carbon mass fraction @xmath0(c ) in the c+o wd ( or the variation of the initial wd mass ) causes the diversity of sn ia brightness .
this model can explain the observed dependence of sne ia brightness on the galaxy types .
we then predict how sn ia brightness evolves along the redshift ( with changing metallicity and age ) for elliptical and spiral galaxies .
such evolutionary effects along the redshift can be corrected as has been made for local sne ia .
we also touch on several related issues : ( 1 ) the abundance pattern of stars in dwarf spheroidal galaxies in relation to the metallicity effect on sne ia , ( 2 ) effects of angular momentum brought into the wd in relation to the diversities and the fate of double degenerates , and ( 3 ) possible presence of helium in the peculiar sn ia 2000cx in relation to the sub - chandrasekhar mass model . invited review published in `` from twilight to highlight : the physics of supernovae , '' eds . w. hillebrandt & b. leibundgut , eso / springer series `` eso astrophysics symposia '' ( berlin : springer ) p.115127(2003 ) . |
calculations from a simple hadronic rescattering model have been shown to agree reasonably well with experimental results obtained from the relativistic heavy ion collider ( rhic ) with @xmath0 = 130 gev and 200 gev au + au collisions @xcite .
more specifically , these calculations provide good representations of the data for 1 ) the particle mass dependence of the @xmath1 distribution slope parameters ( i.e. radial flow ) , 2 ) the @xmath2 , particle mass , and pseudorapidity dependences of the elliptic flow , and 3 ) the @xmath2 , centrality , and azimuthal angle dependences of two - pion hanbury - brown - twiss ( hbt ) measurements . although the agreement between the model and data is sometimes more qualitative than quantitative , this agreement with such a broad range of rhic experimental observables is still a noteworthy accomplishment for a single , simple , model .
many hadrons are initially produced in a relatively small volume in rhic - type collisions , particularly at early times in the interaction .
it is a challenging task for a monte carlo calculation of this type to deal accurately with binary collisions between particles in such a large particle density environment without introducing numerical artifacts which may affect the results .
one such undesirable artifact that can occur when the interaction range between two particles is much greater than the scattering mean - free - path is non - causality of the collisions resulting in superluminal particle velocities@xcite .
the method of subdivision can be used to minimize these artifacts associated with high particle density@xcite .
the effects of such artifacts on observables calculated from the rescattering model are studied in the present work in two ways : 1 ) varying the time step and 2 ) using the subdivision method .
the goal will be to determine whether non - causal artifacts are present in the rescattering model and , if so , whether they have appreciable effects on the calculated observables .
this will provide a test of whether or not one can have confidence in the results of this rescattering model for rhic energies .
section ii will describe the calculational methods used and section iii with give the results of the study .
a brief description of the rescattering model calculational method is given below .
the method used is similar to that used in previous calculations for lower cern super proton synchrotron ( sps ) energies @xcite . rescattering is simulated with a semi - classical monte carlo calculation which assumes strong binary collisions between hadrons .
the monte carlo calculation is carried out in three stages : 1 ) initialization and hadronization , 2 ) rescattering and freeze out , and 3 ) calculation of experimental observables .
relativistic kinematics is used throughout .
all calculations are made to simulate rhic - energy au+au collisions .
the hadronization model employs simple parameterizations to describe the initial momenta and space - time of the hadrons similar to that used by herrmann and bertsch @xcite .
the initial momenta are assumed to follow a thermal transverse ( perpendicular to the beam direction ) momentum distribution for all particles , @xmath3 where @xmath4 is the transverse mass , @xmath2 is the transverse momentum , @xmath5 is the particle rest mass , @xmath6 is a normalization constant , and @xmath7 is the initial `` temperature parameter '' of the system , and a gaussian rapidity distribution for mesons , @xmath8}\ ] ] where @xmath9}$ ] is the rapidity , @xmath10 is the particle energy , @xmath11 is the longitudinal ( along the beam direction ) momentum , @xmath12 is a normalization constant , @xmath13 is the central rapidity value ( mid - rapidity ) , and @xmath14 is the rapidity width .
two rapidity distributions for baryons have been tried : 1 ) flat and then falling off near beam rapidity and 2 ) peaked at central rapidity and falling off until beam rapidity .
both baryon distributions give about the same results .
the initial space - time of the hadrons for @xmath15 fm ( i.e. zero impact parameter or central collisions ) is parameterized as having cylindrical symmetry with respect to the beam axis .
the transverse particle density dependence is assumed to be that of a projected uniform sphere of radius equal to the projectile radius , @xmath16 ( @xmath17 , where @xmath18 fm and @xmath19 is the atomic mass number of the projectile ) . for @xmath20 ( non - central collisions )
the transverse particle density is that of overlapping projected spheres whose centers are separated by a distance @xmath21 .
the particle multiplicities for @xmath20 are scaled from the @xmath15 values by the ratio of the overlap volume to the volume of the projectile . the longitudinal particle hadronization position ( @xmath22 ) and time ( @xmath23 )
are determined by the relativistic equations @xcite , @xmath24 where @xmath25 is the particle rapidity and @xmath26 is the hadronization proper time .
thus , apart from particle multiplicities , the hadronization model has three free parameters to extract from experiment : @xmath14 , @xmath7 and @xmath26 .
the hadrons included in the calculation are pions , kaons , nucleons and lambdas ( @xmath27 , k , n , and @xmath28 ) , and the @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , and @xmath35 resonances . for simplicity
, the calculation is isospin averaged ( e.g. no distinction is made among a @xmath36 , @xmath37 , and @xmath38 ) .
resonances are present at hadronization and also can be produced as a result of rescattering .
initial resonance multiplicity fractions are taken from herrmann and bertsch @xcite , who extracted results from the helios experiment @xcite .
the initial resonance fractions used in the present calculations are : @xmath39 , @xmath40 , @xmath41 , @xmath42 , @xmath43 and , for simplicity , @xmath44 .
the second stage in the calculation is rescattering which finishes with the freeze out and decay of all particles . starting from the initial stage ( @xmath45 fm / c ) , the positions of all particles are allowed to evolve in time in small time steps ( normally @xmath46 fm / c is used ) according to their initial momenta . at each time
step each particle is checked to see a ) if it decays , and b ) if it is sufficiently close to another particle to scatter with it .
isospin - averaged s - wave and p - wave cross sections for meson scattering are obtained from prakash et al . @xcite .
the calculation is carried out to 100 fm / c , although most of the rescattering finishes by about 30 fm / c .
the rescattering calculation is described in more detail elsewhere @xcite .
calculations are carried out assuming initial parameter values and particle multiplicities for each type of particle . in the last stage of the calculation ,
the freeze - out and decay momenta and space - times are used to produce observables such as pion , kaon , and nucleon multiplicities and transverse momentum and rapidity distributions .
the values of the initial parameters of the calculation and multiplicities are constrained to give observables which agree with available measured hadronic observables . as a cross - check on this ,
the total kinetic energy from the calculation is determined and compared with the rhic center of mass energy to see that they are in reasonable agreement .
particle multiplicities were estimated from the charged hadron multiplicity measurements of the rhic phobos experiment @xcite .
calculations were carried out using isospin - summed events containing at freezeout for central collisions ( @xmath15 fm ) about 5000 pions , 500 kaons , and 650 nucleons ( @xmath28 s were decayed ) .
the hadronization model parameters used were @xmath47 mev , @xmath14=2.4 , and @xmath26=1 fm / c .
it is interesing to note that the same value of @xmath26 was required in a previous rescattering calculation to successfully describe results from sps pb+pb collisions @xcite . as an indicator of the presence of superluminal artifacts , the transverse signal propagation velocity , @xmath48 ,
is calculated for all particle pairs in the rescattering calculation for each time step from@xcite @xmath49 where @xmath50 is the signal propagation velocity vector , @xmath51 is the position of particle 1 for the present collision with particle 2 at time @xmath52 , @xmath53 is the position of particle 2 for its previous collision at time @xmath54 , and @xmath55 and @xmath56 are the perpendicular components of @xmath57 in the transverse plane .
note that even though the rescattering calculations are carried out with relativistic kinematics , it is still possible for @xmath48 to be greater than @xmath58 for some particle pairs due to the details of how the two - particle collisions are implemented .
the method of subdivision is based on the invariance of monte carlo particle - scattering calculations for a simultaneous decrease of the scattering cross sections by some factor , @xmath59 , and increase of the particle density by @xmath59 , where @xmath59 is called the subdivision@xcite . as @xmath59 becomes sufficiently large , non - causal artifacts
become insignificant .
the present rescattering calculation will be tested comparing pion observables from the `` nominal '' calculation , i.e. @xmath60 , with subdivisions of @xmath61 and @xmath62 .
note that a subdivision of @xmath61 has been shown to significantly suppress non - causal artifacts present in some monte carlo calculations@xcite . since the particle density increase is accomplished by increasing the particle number by a factor @xmath59 , the computer cpu time taken per event increases by a factor @xmath63 .
for the present study , 1640 , 124 and 120 events were generated for the @xmath60 , @xmath61 , and @xmath62 samples , respectively .
fortunately , the statistical value of the @xmath64 events is @xmath59 times greater per event than for @xmath60 . as an example , the cpu time taken to generate the @xmath62 sample of events was 600 cpu - hours on 2.8 ghz pc processors at the ohio supercomputing center .
figures 1 - 8 present results from the hadronic rescattering model simulating hadrons produced in rhic - energy collisions for various time steps and subdivisions .
quantities plotted are @xmath48 histograms including all particles in the calculation and several hadronic observables
. descriptions of how the observables are extracted from the rescattering calculation are given elsewhere@xcite .
all calculations are carried out for a mid - rapidity bin , @xmath65 , and for an impact parameter of 8 fm to simulate a medium non - central collision which should result in significant elliptic flow for the purposes of the present test .
statistical errors are shown either as error bars or are of the order of the marker size when error bars are not shown .
figures 1 - 4 show comparisons of various quantities for three different time steps in the rescattering calculations , @xmath66 and @xmath67 fm / c .
the time step used in calculating observables which have been compared with rhic data is @xmath46 fm / c@xcite , so the other time steps used represent half of the usual step and twice the usual step .
figure 1 shows @xmath48 histograms .
transverse signal velocity distributions for all particles for various time steps.,width=529 ] it is seen that @xmath48 extends beyond @xmath58 with an exponential tail for all time steps , indicating the presence of superluminal artifacts . for @xmath68 all time steps
lie on top of each other whereas for @xmath69 they differ somewhat according to what would be expected for the different time steps , i.e. smaller and thus more frequent time steps should enhance the superluminal artifacts , whereas larger and thus less frequent time steps should reduce the artifacts . for @xmath70 fm / c an enhancement of @xmath71 in the artifacts over the nominal case
is seen , while in the case of @xmath72 fm / c a reduction in the artifacts by about @xmath73 over the nominal time step is seen .
the effects of these time steps on various hadronic observables are shown in figures 2 - 4 .
pion , kaon , and nucleon @xmath1 distributions for various time steps.,width=529 ] pion elliptic flow vs. @xmath2 for various time steps.,width=529 ] two - pion hbt source parameters vs. @xmath2 for various time steps .
the ordinate scale for @xmath74 is shown to the right.,width=529 ] figure 2 shows @xmath1 distributions for pions , kaons , and nucleons , figure 3 shows the elliptic flow , @xmath75 , vs. @xmath2 for pions and figure 4 shows two - pion hbt source parameters@xcite vs. @xmath2 . as seen in all three figures , even though the different time steps have noticeable effects on the artifacts ( as seen in figure 1 ) , they do not significantly effect the hadronic observables , the observables for all three time steps agreeing within the statistical errors of the calculations . since it could be argued that the time step study presented above may not reduce the superluminal artifacts enough to see significant effects on the hadronic observables , the subdivision test is also employed since it provides a greater ability to reduce these artifacts .
this is seen in figure 5 which shows @xmath48 histograms for the three subdivision used : @xmath76 and @xmath77 .
transverse signal velocity distributions for all particles for @xmath78 and @xmath77.,width=529 ] while all three subdivisions mostly lie on top of each other for @xmath68 , one sees factors of 5 - 10 reductions in the superluminal artifacts for @xmath61 and @xmath77 .
if these artifacts have a large influence on the values of the hadronic observables extracted from the rescattering model , it should surely be evident for the @xmath61 and @xmath77 cases .
figures 6 - 8 show the effects of these different subdivisions on pion @xmath2 distributions , elliptic flow and hbt , respectively .
pion @xmath2 distributions for @xmath78 and @xmath77.,width=529 ] pion elliptic flow vs. @xmath2 for @xmath78 and @xmath77.,width=529 ] two - pion hbt source parameters vs. @xmath2 for @xmath78 and @xmath77 .
the ordinate scale for @xmath74 is shown to the right.,width=529 ] as was the case in the time step study , the hadronic observables shown in figures 6 - 8 are not significantly effected by the different subdivisions used in spite of the fact that superluminal artifacts are greatly suppressed in the subdivision study .
it can be concluded from figures 1 - 8 that non - causal ( superluminal ) artifacts while being present in the hadronic rescattering code used in this study do not significantly effect the hadronic observables calculated from this code .
it is possible to speculate on why this is the case .
there are three main features of the code which might tend to reduce the effects of these artifacts : 1 ) individual particles are allowed to scatter only once per time step , 2 ) a `` scattering time '' of two nominal time steps is defined during which particles that have scattered are not allowed to rescatter , and 3 ) once two particles have scattered with each other , they are not allowed to scatter with each other again in the calculation .
the present study was performed for non - central collisions ( @xmath79 fm ) , but it is expected that the same conclusion , i.e. that superluminal artifacts do not significantly effect the results of calculations with the rescattering code , would also be obtained for central collisions . this is because the particle density is not very different for central collisions in the model than for mid - peripheral collisions since the particle multiplicities are scaled by the overlap volume for @xmath20 , as mentioned earlier . in summary , two methods have been used to test the validity of the present rescattering model calculations in the presence of non - causal ( superluminal ) artifacts .
it is found that the results of this model are not appreciably effected by such artifacts , thus strengthing the confidence in the results presented previously from this rescattering model for rhic .
the author wishes to thank tams csrg , dnes molnr and derek teaney for helpful discussions relating to this paper .
this work was supported by the u. s. national science foundation under grant phy-0355007 . | it has been shown that calculations based on a hadronic rescattering model agree reasonably well with experimental results from rhic au+au collisions .
because of the large particle densities intrinsically present at the early time steps of monte carlo calculations attempting to model rhic collisions undesirable artifacts resulting in non - causality may be present .
the effects of such artifacts on observables calculated from the rescattering model are studied in the present work in two ways : 1 ) varying the time step and 2 ) using the subdivision method .
it is shown that although non - causal artifacts are present in the rescattering model they have no appreciable effects on the calculated observables , thus strengthing the confidence in the results of this rescattering model for rhic energies . |
the analysis of the phases of strong interactions presents many fascinating aspects mechanisms of confinement , different realisations of the chiral symmetry , the nature of the symmetric phase , the emergence of conformality , and many others .
all these topics are under active scrutiny both theoretically and experimentally @xcite .
while strong interactions spontaneously break chiral symmetry in ordinary qcd at zero temperature , the chiral symmetry is realised either at high temperatures in the so - called quark - gluon plasma ( qgp ) phase and at a large number of flavours @xmath14 ( even at zero temperature ) @xcite . in the latter case
, the theory is expected to become not only chirally but also conformally invariant .
this is due to the emergence of an infra - red fixed point ( irfp ) for @xmath14 at a coupling which is not strong enough to break chiral symmetry .
both physics intuition and phenomenological analysis based on functional renormalisation group @xcite and finite temperature holographic qcd @xcite indicate that the conformal phase of cold , many flavour qcd and the high temperature chirally symmetric phase are continuously connected . in particular , the onset of the conformal window coincides with the vanishing of the transition temperature , and the conformal window appears as a zero temperature limit of a possibly strongly interacting qgp .
the analysis of the finite temperature phase transition is a well - established line of research within the lattice community , and our approach will be completely conventional here .
according to the pisarski - wilczek scenario @xcite , the most likely possibility for @xmath15 is a first order chiral transition in the chiral limit , turning into a crossover above a critical mass endpoint , and/or on lattices which are not large enough
. we will identify such crossover with confidence for a number of flavours ranging from four to eight , and we will complement these results with those of the deconfinement transition in the quenched model . then , we study the approach to the conformal phase in the light of the chiral phase transition at finite temperature with variable number of flavours . one problem of this approach is the setting of a common scale among theories which are essentially different .
we will propose two alternative possibilities to handle this problem , one evolving from our previous work @xcite , and the other from a recent analysis @xcite .
interestingly , this latter approach analyses the dependence of the confinement parameters on the matter content , and proposes microscopic mechanisms for confinement motivated by such @xmath4 dependence .
further , we will argue that even results in the bare lattice parameters can be used directly to locate the critical number of flavours , thus generalising to finite temperature the miransky - yamawaki phase diagram , ref .
@xcite .
a second zero of the two - loop beta - function of a non - abelian gauge theory implies , at least perturbatively , the appearance of irfp conformal symmetry @xcite . in colour su(@xmath16 ) gauge theory with @xmath4 massless fundamental fermions , the second zero appears at @xmath17 , before the loss of asymptotic freedom ( laf ) at @xmath18 .
analytic studies of the conformal transition of strong interactions have produced a variety of predictions for the conformal threshold : the schwinger - dyson approach with rainbow resummations @xcite or the functional renormalisation group method @xcite suggest the onset of conformal window around @xmath19 .
an all - order perturbative beta - function @xcite inspired by the nsvz beta - function of sqcd @xcite leads to a bound @xmath20 .
instanton studies at large @xmath4 @xcite claimed a qualitative change of behaviour at @xmath21 .
@xmath22 has also been estimated for different fermion representations @xcite . the sub
critical region , when @xmath4 gets closer and closer to @xmath5 , is interesting per se : the question is whether the chiral dynamics there shows any difference with the standard qcd dynamics .
significant differences with respect to the qcd dynamics might offer a basis to model builders interested in beyond - the standard model theories .
the recent discovery of a 125 gev boson at the lhc poses the question as to whether there are light composite scalars which might be identified with such boson , in alternative to a standard model higgs boson .
conformal dynamics might well help these studies @xcite . in our study
, such pre - conformal dynamics could manifest itself either with a clear observation of a separation of scales , or with a manifestation of a critical behaviour when approaching @xmath5 .
one possibility is to observe the miransky - yamawaki essential singularity @xcite .
alternatively , in an frg approach @xcite , the pseudo - critical line is almost linear with @xmath4 for small @xmath4 , and displays a singular behaviour when approaching @xmath5 , which could be the only observable effects , beyond miransky scaling .
a _ jumping _
scenario in which the change from a qcd dynamics to the conformal window is abrupt is also a distinct possibility @xcite .
clearly , as in any system undergoing a phase transition , the nature and extent of the critical window are purely dynamical questions whose answer can not be guessed a priori . since
the underlying dynamics is completely non - perturbative , lattice calculations are the only tool to perform an ab initio , rigorous study of these phenomena , and many lattice studies have recently appeared @xcite .
this paper is one step of our ongoing program @xcite@xcite which aims at elucidating the phase diagram of qcd with fundamental fermions on the lattice , and in the continuum .
further studies either with fundamental fermions @xcite @xcite or other representations @xcite have contributed to our current understanding of this challenging field . however , only a subset of these studies has addressed issues related with pre conformal dynamics @xcite @xcite , @xcite which are the main theme of this paper .
the direct inspection of theories at fixed @xmath4 is often inconclusive , especially close to the expected threshold @xmath5 . also because of this ,
we feel it is a useful approach to try to observe directly the approach to conformality by monitoring the evolution of the pre
conformal results as a function of @xmath4 . in this paper
, we investigate the thermal chiral phase transition for @xmath23 colour su@xmath0 qcd by using lattice qcd monte carlo simulations with staggered fermions . here
, @xmath21 and @xmath3 is expected to be in the important regime as suggested by the results in refs .
we combine our findings with those of our early work for @xmath21 and @xmath3 @xcite .
this paper grows out of our early study @xcite and extends it in several ways : we have accumulated more statistics and added more parameters , and we present here an extended set of simulations and details .
we develop a new scale setting procedure , so that we can more confidently measure the critical temperature on a common reference scale among theories with different flavour content .
furthermore , we present new estimates of the critical number of flavours @xmath5 .
partly motivated by the recent work @xcite , we introduce a typical interaction strength @xmath8 at the critical temperature based on our lattice results , and compare it with a four - loop irfp coupling ( @xmath24 ) @xcite and a critical coupling ( @xmath25 ) estimated by using a two - loop schwinger - dyson equation @xcite .
further , we introduce and discuss the finite temperature version of the miransky - yamawaki phase diagram , and propose a strategy to locate the critical number of flavour motivated by the properties of the lattice step - scaling function in the vicinity of the irfp @xcite . some of the new results presented here have been anticipated in a recent proceeding , and talks@xcite .
this paper is organized as follows : in the next section , we will explain the simulation setups . in section [ sec : result
] , we show our results for the chiral crossover at finite @xmath26 , for each @xmath4 , and then , we collect the critical lattice couplings associated with the chiral crossovers at @xmath27 . in section [ sec : as ] , we investigate the asymptotic scaling of our critical couplings at each @xmath4 . in section [
sec : discuss ] , we investigate the @xmath4 dependences of the chiral crossovers , and estimate the lower edge of the conformal window @xmath5 .
finally in section [ sec : sum ] , we provide concluding remarks . the appendix is devoted to the summary tables of the simulation parameters and the numerical results obtained by analysing the simulation outputs .
we investigate finite temperature qcd with different number of flavour @xmath28 by utilising the publicly available milc code @xcite .
the temperature @xmath26 is related to the inverse of the lattice temporal extension , @xmath29 and we control it by varying @xmath30 at fixed @xmath6 .
the number of lattice points in the spacial directions @xmath31 is chosen such that the aspect ratio @xmath32 in all our runs .
for each @xmath4 , we use a single bare fermion mass @xmath33 .
the simulation parameters used in this study are summarised in [ app : sum_table ] .
the setup for the action explained below is the same as the one used for @xmath34 in ref .
@xcite up to the number of flavour .
we use an improved version of the staggered action , the asqtad action @xcite , with a one - loop symanzik @xcite and tadpole @xcite improved gauge action , @xmath35 + \sum_{i = p , r , pg}\beta_i(g^2_{\mathrm{l } } ) \mathrm{re}\bigl[1-u_{c_i}\bigr]\ , \label{eq : action}\end{aligned}\ ] ] where @xmath36 is the lattice bare coupling , and @xmath37 are defined as @xmath38 the plaquette coupling @xmath39 is a simulation input . the @xmath40 $ ] in eq .
( [ eq : action ] ) denotes the matrix for a single flavour asqtad fermion with bare lattice mass @xmath41 , and @xmath42 represents the trace of the ordered product of link variables along @xmath43 , for the @xmath44 plaquettes ( @xmath45 ) , the @xmath46 and @xmath47 rectangles ( @xmath48 ) , and the @xmath49 parallelograms ( @xmath50 ) , respectively - all divided by the number of colours .
the tadpole factor @xmath51 is determined by performing zero temperature simulations on the @xmath52 lattice ( the second column of tables [ tab : nf0_nt4 ] - [ tab : nf8_nt8 ] ) , and used as an input for finite temperature simulations . to generate configurations with mass degenerate dynamical flavours ,
we have used the rational hybrid monte carlo algorithm ( rhmc ) @xcite , which allows to simulate an arbitrary number of flavours @xmath4 through varying the number of pseudo - fermions .
the quenched ( @xmath53 ) system has been realised by using massive bare fermion mass @xmath54 in the four flavour system
. the six flavour system has been realised by using two pseudo - fermions in the rational approximation with a quarter root technique , @xmath55 .
then , we have assumed the rooting does not affect the results within the accuracy of our simulation .
for the other number of flavour ( @xmath56 ) , we do not use the rooting .
we have adjusted the micro - canonical step length @xmath57 and the step length of a single trajectory @xmath58 to realise @xmath59 percent metropolis acceptances .
details are reported in the fourth column of tables [ tab : nf0_nt4 ] - [ tab : nf8_nt8 ] . for each parameter
set , we have collected a number of trajectories ranging from a one thousand to ten thousands - the latter closer to the chiral crossover regime . the focus of this paper is the analysis of the chiral phase transition the fundamental observable is then the order parameter for chiral symmetry , the chiral condensate : @xmath60}\big\rangle \ , \label{eq : pbp}\ ] ] where @xmath61 represents the number of lattice sites in the spatial ( temporal ) direction .
we have measured @xmath62 by using a stochastic estimator with 20 repetitions .
we have also measured connected and disconnected chiral susceptibilities , @xmath63 \rangle \ , \nonumber\\ a^2\chi_{\mathrm{disc } } & = \frac{n_f^2}{16 n_s^3 n_t } \left [ \langle \mathrm{tr } \left[m^{-1}\right ] ^2\rangle - \langle \mathrm{tr } \left[m^{-1}\right ] \rangle^2\right ] \
.\end{aligned}\ ] ] here we have conveniently written the chiral condensate and its susceptibilities in terms of traces of ( products of ) the staggered fermion matrix @xmath12 .
we note that the milc convention for the chiral condensate gives the twice of eq .
( [ eq : pbp ] ) , as will be indicated several times in the following sections for results .
we have measured the susceptibilities @xmath64 and @xmath65 separately .
the disconnected chiral susceptibility is a non - local quantity which can be estimated from the variance of the bulk behaviour of the chiral condensate .
since we have used the stochastic estimator for the chiral condensate measurements , the variance would automatically include part of the connected contributions through random sources multiplying themselves .
following bernard et al .
@xcite , we take into account this effect in our estimate for the disconnected part @xmath65 by considering the only off - diagonal elements of the covariance matrix for the random sources .
the measurements of @xmath66 and @xmath67 allow us to construct two physically relevant quantities : the scalar and pseudo - scalar susceptibilities , @xmath68 their associated cumulant @xmath69 is a probe of the chiral symmetry @xcite . this is owing to the fact that @xmath70 and @xmath71 are related through ward identities to the spacetime volume integral of the scalar ( @xmath72 ) and pseudo - scalar ( @xmath73 ) propagators . in the chiral limit
, the susceptibility ratio @xmath74 should be one in chirally symmetric regime due to the degeneracy of the chiral partners , while it should be zero in the spontaneously broken phase . even including a finite bare fermion mass
, @xmath74 still has a strong signal for the chiral transition or crossover .
in particular , @xmath75 in the chirally symmetric regime holds true till the chiral condensate is dominated by the linear mass term contribution .
it turns out that @xmath76 allows the identification of a pseudo - critical coupling @xmath1 associated with the chiral crossover , which , in the cases we have studied , coincides in the error with the pseudo - critical coupling determined from the maximum of the chiral susceptibility . in the gauge sector
, we measure the polyakov loop , @xmath77 where @xmath78 denotes the trace in colour space , and @xmath79 is the temporal link variable . from the variance of @xmath80
, we also evaluate the susceptibility for the polyakov loops . in the vicinity of the chiral crossover
, we have a long auto - correlation time , and thermalization checks require extra care . here
we explain our analyses by using a typical example : the left panel of fig .
[ fig : traj ] displays the evolution of the chiral condensate on the lattice volume @xmath81 just before the chiral crossover @xmath82 and @xmath83 at @xmath84 , one of the most time - consuming examples in our simulations .
( in order to shorten the simulation time , we started the evolution from thermalized configurations obtained at @xmath85 . )
we have computed the ensemble averages by using the last @xmath86 ( @xmath87 ) trajectories at @xmath82 ( @xmath83 ) and we have confirmed that they are consistent with those obtained by using last @xmath87 ( @xmath88 ) trajectories .
we have then used the latter trajectories to evaluate the average . in the cases we are considering
, this corresponds to the data found in the right - hand side of the vertical green ( dashed ) lines in the left panel of fig .
[ fig : traj ] .
we divide the obtained data set into several bins and utilise the jackknife method in order to take into account the auto - correlation effect in the error estimate .
as a bin - size @xmath89 becomes larger , the jackknife error increases ( the right panel of fig .
[ fig : traj ] ) , which is due to the decrease of the effective number of ( uncorrelated ) data ( @xmath90 , @xmath91 the number of trajectories to calculate the average ) . for a sufficiently large @xmath89 , the jackknife errors at @xmath82 and @xmath83 level
off , giving a reliable error estimate . here is the result obtained from the above procedures : @xmath92 we have performed the analyses explained here for all the various @xmath30 , @xmath4 , and the lattice volumes .
the results are summarised in [ app : sum_table ] .
in this section , we show our simulation results on the chiral and deconfinement crossover for the different number of flavours @xmath4 . j. braun , c. s. fisher , h. gies , phys . rev .
* d84 * ( 2011 ) 034045 ; j. braun and h. gies , jhep * 1005 * ( 2010 ) 060 ; * 0606 * ( 2006 ) 024 .
t. alho , m. jarvinen , k. kajantie , e. kiritsis and k. tuominen , arxiv:1210.4516 [ hep - ph ] .
t. karavirta , a. mykkanen , j. rantaharju , k. rummukainen and k. tuominen , jhep * 1106 * ( 2011 ) 061 ; a. j. hietanen , k. rummukainen and k. tuominen , phys . rev .
d * 80 * ( 2009 ) 094504 ; a. j. hietanen , j. rantaharju , k. rummukainen and k. tuominen , jhep * 0905 * ( 2009 ) 025 .
y. shamir , b. svetitsky and e. yurkovsky , phys .
d * 83 * ( 2011 ) 097502 ; o. machtey and b. svetitsky , phys .
d * 81 * ( 2010 ) 014501 ; y. shamir , b. svetitsky and t. degrand , phys .
d * 78 * ( 2008 ) 031502 .
r. sommer , nucl .
b * 411 * ( 1994 ) 839 , s. necco and r. sommer , nucl . phys .
b * 622 * , 328 ( 2002 ) ; k. jansen _ et al .
_ [ etm collaboration ] , jhep * 1201 * ( 2012 ) 025 ; a. gray , i. allison , c. t. h. davies , e. dalgic , g. p. lepage , j. shigemitsu and m. wingate , phys . rev .
d * 72 * ( 2005 ) 094507 [ hep - lat/0507013 ] . | we study the thermal phase transition in colour su@xmath0 quantum chromodynamics ( qcd ) with a variable number of fermions in the fundamental representation by using lattice monte carlo simulations .
we collect the ( pseudo ) critical couplings @xmath1 for @xmath2 , and @xmath3 ) , and we investigate the pre - conformal dynamics associated with the infra - red fixed point in terms of the @xmath4 dependence of the transition temperature .
we propose three independent estimates of the number of flavour @xmath5 where the conformal phase would emerge , which give consistent results within the largish errors .
we consider lines of fixed @xmath6 in the space of ( @xmath4 , bare lattice coupling ) , and locate the vanishing of the step - scaling function for @xmath7 .
we define a typical interaction strength @xmath8 at the scale of critical temperature @xmath9 and we find that @xmath8 meets the zero temperature critical couplings estimated by the two - loop schwinger - dyson equation or the irfp coupling in the four - loop beta - function at @xmath10 .
further , we study the @xmath4 dependences of @xmath11 where @xmath12 is a uv @xmath4 independent reference scale determined by utilising the coupling at the scale of the lattice spacing .
then , @xmath11 turns out to be a decreasing function of @xmath4 and the vanishing @xmath11 indicates the emergence of the conformal window at @xmath13 .
lattice gauge theory , conformal symmetry , chiral symmetry , finite temperature |
we present a detailed optical spectroscopic study of 20 high - mass x - ray binaries ( hmxbs ) in the small magellanic cloud ( smc ) using the 2df ( two - degree field ) multi - object spectrograph on the 3.9 m anglo - australian telescope ( aat ) . in this study
we have used the 2df spectrograph to confirm the classification of candidate be / x - ray binaries ( be - xrbs ; the most numerous subclass of smc hmxbs ) as emission - line stars and identify their spectral type . although this is only a pilot study , it is at the same time a spectroscopic survey of a relatively large number of be - xrbs and candidates over a large wavelength range .
this provides a deep census of the x - ray binary ( xrb ) population in this galaxy even in low activity levels , which is critical for understanding the characteristics of the overall population .
the smc is one of the best targets to study a , as complete as possible , population of young xrbs .
similar studies in the galaxy are hampered by extinction and distance uncertainties .
the large magellanic cloud ( lmc ) is much more extended than the smc , requiring large area coverage to obtain sufficient numbers of xrbs , while other local group galaxies are too far to reach the quiescent population of hmxbs ( typical @xmath0 ; van paradijs & mcclintock 1995 ) .
in addition , the smc hosts a large number of hmxbs ( e.g. haberl & pietsch 2004 , mcbride 2008 , antoniou 2009 ) when compared to the milky way and the lmc .
this excess is explained by the combination of a star - formation burst at @xmath1 myr at regions where the hmxbs are located ( antoniou 2009 ; consistent with the age of maximum formation of be stars , given by mcswain & gies 2005 ) , and by the low smc metallicity ( antoniou 2009 , mcbride 2008 ) .
simulations show that neither the metallicity nor the star - formation burst alone can produce such a high number of hmxbs in the smc ( dray 2006 ) .
thus , the smc is an excellent laboratory to study the be - xrb populations .
the most representative studies of spectroscopically identified be - xrbs are those of mcbride ( 2008 ) , who investigated the spectral distribution of 37 in total smc be - xrb pulsars , and of negueruela & coe ( 2002 ) , who presented high - resolution blue - band spectroscopy for 13 lmc be - xrbs . however , the main limitation of detailed studies of the hmxb population in the smc is the lack of spectral classification of the optical counterparts of a large fraction of the detected sources , while the vast majority of the studies available in the literature either do not present the spectra and/or the criteria used for the spectral classification .
this severely hampers the construction of a large coherent sample of be - xrbs .
motivated by the limitations of the current studies , and in order to classify the different type of xrbs and separate them from the various interlopers ( background / foreground sources ) , we initiated an optical spectroscopic campaign of a large number of candidate smc be - xrbs , detected either in our observations or other published x - ray surveys of the smc .
the goal of this campaign is to confirm the be classification of each system ( often based on photometric data ) , identify its spectral type , and even identify any possible interlopers .
the structure of this paper is the following . in section [ sample ] , we describe the sample of sources we observed with the 2df spectrograph ; and in section [ obsreduction ] , the observations and data reduction , along with the optical and near - infrared properties of the sources . in section [ criteriasptypes ] , we discuss the criteria used for the classification of the optical spectra , and the comparison of the classification of sources from the present work with those of previous studies . in section [ 2dfdiscussion ]
, we discuss these results in the context of the be stars and be - xrb populations in the galaxy and the lmc , and we investigate correlations between the emission and the orbital period and infrared colors of the systems
. finally , we present the main results of this study in section [ conclusions ] .
our sample is primarily selected from the shallow survey of the smc ( a. zezas 2009 , in preparation ; antoniou 2009 ) , and it also includes candidate and known be - xrbs from the census of haberl & pietsch ( 2004 ; based mainly on observations with ) . the overall strategy of this pilot study was to observe the optical counterparts of confirmed and candidate be - xrbs . in total , we have observed 113 and 27 and sources , respectively .
we note that even though the 2df spectrograph allows the allocation of @xmath2400 fibers , the stellar density of the selected fields , which lie along the smc `` bar '' , did not allow us to observe more objects at a given time due to collision of the fibers .
twenty - nine out of the 113 sources do not have an optical match in the ogle - ii and/or mcps catalogs ( based on the analysis of antoniou 2009 ) , we thus obtained blind spectra , i.e. we placed the fiber at the location of an x - ray source , in order to obtain a spectrum of any object at that position . for the remaining sources , the typical v - band magnitude of the optical counterparts was @xmath3 mag .
in addition to the sources we also observed 9 confirmed and 18 candidate be - xrbs from the compilation of haberl & pietsch ( 2004 ) .
the optical spectra for this work were obtained during service time in 2004 november , with the 2df multiple - fiber spectrograph at the prime focus of the 3.9-m anglo - australian telescope ( lewis 2002 ) .
the observations were taken at airmass between 1.35 and 1.48 , and the seeing was @xmath4 .
we used the @xmath5 grating ( 300b ) , which gave a 4400 wavelength coverage ( from 3640 to 8040 region ) at an instrumental dispersion of 4.3 / pixel .
four 2250s exposures were obtained , providing a total exposure time of 9000s .
the on - target exposures were preceded and followed by arc calibration exposures ( cuar and cuhe ) and fiber - flat - field exposures .
approximately 30 sky " fibers located on star - free regions ( 15 for each spectrograph ) were assigned , to ensure good definition of the sky background .
successful use of the 2df spectrograph depends on the source positions being accurate to better than 0.5 arcsec , in order to avoid significant light loss from the 2-arcsec - diameter fibers .
positions were used from ogle - ii ( which has a small astrometric error of @xmath6 ; udalski 1998 ) whenever possible to fulfill this requirement . in order to achieve accurate pointing and guide the telescope successfully during the observations we chose 4 sources as fiducial stars evenly spaced through the 2df field .
these stars were chosen from the 2mass catalog , however they have consistent coordinates in the ogle - ii catalog and they abide by the suggestions listed in the 2df manual ( e.g. brighter than 15.0 magnitude stars which should not cover more than a one magnitude range ) .
nevertheless , some loss in counts is expected to be caused by the small number of fiducial stars used to guide the field plate and due to the small positional mismatches between fiber and object positions .
the data preparation and spectral extraction were performed using the _ 2dfdr _ software ( lewis 2002 ) .
the main steps include bias subtraction , extraction of the spectra from the ccd image , division by a normalized flat - field , wavelength calibration , calibration of the fiber throughputs , and subtraction of the scaled median sky spectrum ( derived from the sky - fiber spectra ) .
extraction of the spectra , spectral line fitting and ew line measurements were performed with the _ dipso _ v3.6 - 3 and figaro v4.11 packages of starlink . in particular , we measured the center of each line and its full - width at half - maximum ( fwhm ) using a gaussian fit ( _ elf _ subroutine ) , while for the ew we used the _ ew _ subroutine of the _ dipso _ package .
due to the different throughput of each fiber ( which is also wavelength - dependent ) , it is difficult to flux calibrate spectra from a multi - fiber instrument such as 2df , as a flux standard should be -in principle- observed through each fiber .
an average response curve can be applied to give an approximate relative flux calibration , however , the results for individual spectra can vary considerably ( lewis 2002 ) .
therefore , flux calibration was not attempted .
this does not affect our analysis , as our classification criteria are based on the presence or absence of spectral lines rather than their absolute intensity .
whenever comparisons between strengths of lines are used , care is taken to use lines close in wavelength , in which case the fiber response is identical , for all practical purposes .
fibers that were assigned to objects fainter than @xmath7 mag in v - band , yielded low signal - to - noise ( s / n ) ratio spectra .
we defined a nominal @xmath8 ratio of each spectrum using the counts in the 5600-6200 range ( measured using the _ istat _ package of figaro v4.11 ) which can be considered as representative of the local continuum .
spectra with @xmath9400 average counts in this wavelength range were excluded from the present study , as it was difficult to perform accurate spectral classification for them .
this leaves 58 spectra for further consideration .
all but one of these 58 sources exhibit emission
. the presence of emission is one of the main classification criteria for be - xrbs .
we thus focus on the study of emission - line objects . however , the inner regions of the smc show strong and variable diffuse emission from hii regions and supernova remnants ( snrs ) . due to hardware limitations
it is not possible to obtain sky spectra within a few arcseconds from each source , which in turn does not allow us to accurately subtract the variable local background .
therefore , some of our object spectra may be contaminated by the typical interstellar emission lines : @xmath10\ , \lambda5007}$ ] , @xmath11\ , \lambda\lambda6548 , 6584}$ ] , and @xmath12\ , \lambda\lambda6716 , 6731}$ ] .
for this reason we adopt a conservative approach and we focus only on objects with minimal interstellar contamination .
the selection of such objects is based on the width of the line and their ratio .
be stars exhibit broad emission ( e.g. coe 2005 ) , whereas the interstellar emission lines are typically narrow
. therefore , selection of objects with broad emission allows us to identify objects with small interstellar contamination .
we choose a low limit for the fwhm of the line @xmath13 , based on the average fwhm of this line from high s / n spectra of known be - xrbs and be - xrb pulsars , which are observed as part of the present work .
in total , we find 22 sources with fwhm greater than 9.5 .
two of these sources have broad but extremely weak emission when compared to the other sources , thus they are not considered for further investigation .
based on the above criteria our sample is limited to 20 sources .
we note that there could be additional bona - fide be - xrbs in the remaining sample , however the absence of a broad component makes their classification uncertain .
the selection of these sources as bona - fide be - xrbs is confirmed by their high ratio .
be stars typically do not show forbidden high excitation emission lines ( with the only exception of the subclass of b[e ] stars ; e.g. lamers 1998 ) .
therefore , the presence of @xmath146716 emission is a signature of significant contamination from the diffuse interstellar emission .
we determine the average to strength ratio of the diffuse emission from the spectra of the sky fibers which uniformly probe the observed area .
this gives an average local sky " value of to ratio of @xmath15 .
all objects which fulfill the width criterion also have to ratio at least @xmath16 below the average sky ratio .
this gives us further confidence that the emission - line objects we selected for further study are indeed be stars .
in addition , although we can not rule out the possibility that some of the remaining observed sources could be be - xrbs , none of the discarded sources with good s / n has low enough to ratio to fulfill the above criterion .
this indicates that we do not miss a significant population of be - xrbs among the remaining sources .
however , because of these criteria , our sample is biased towards objects with stronger emission .
based on the above analysis , we found 20 objects with good quality optical spectra obtained in this run , and known x - ray properties typical of hmxbs ( e.g. hard x - ray spectrum and/or detection of pulsations ) .
out of the 20 objects , the spectra of which are presented here , only 2 have good quality published optical spectra in the full wavelength range ( @xmath17 ) , while one source has been observed only in the red portion of the optical spectrum around h@xmath18 ( @xmath19 ) .
these 20 sources are presented in table [ otherxrayname ] . in column 1
we list the x - ray source i d as given in antoniou ( 2009 ; ( field id)-(src i d within the field ) ) for sources , while the xmm i d refers to sources from haberl & pietsch ( 2004 ) .
other names for the _ xmm - newton _ sources and their references ( column 2 ) are taken from the latter work . in column 3
we present the x - ray pulsar i d ( defined as the pulse period in seconds ) .
finally , in column 4 we list the associated emission - line object from the catalog of meyssonnier & azzopardi ( 1993 ; hereafter [ ma93 ] ) . with the present observations we will investigate the be - xrb nature of these 20 objects and we will study their spectral characteristics , in conjunction with their optical and near - infrared properties .
the optical and near - infrared counterparts of the studied x - ray sources are present in table [ 2dfoptids ] .
in particular , the optical counterparts of the sources are taken from antoniou ( 2009 ) .
following the same approach for the sources , we cross - correlated their positions with the ogle - ii ( udalski 1998 ) and mcps catalogs ( zaritsky 2002 ) using a search radius of 5 around each x - ray source ( brusa 2007 ) . in column 1 we give the x - ray source i d ( as in table [ otherxrayname ] ) , and in columns 2 and 3 the x - ray source coordinates ( right ascension and declination , respectively ) . in column 4
we present the optical counterpart of the x - ray source from the ogle - ii and/or mcps catalogs .
ogle - ii sources are named as o - f - nnnnnn , where f and nnnnnn are the field and optical source number , respectively ( from udalski 1998 ) , and mcps sources are named as z - nnnnnn where nnnnnn is the line number of the source in table 1 of zaritsky ( 2002 ) . the angular separation ( in arcseconds ) of the counterpart to the x - ray source is given in column 5 .
the optical ( _ v , b - v _ ) data of the sources are presented in columns [ 6 ] and [ 7 ] along with their errors ( these data are taken directly from the original catalogs without applying any reddening or zero - point correction ) : apparent magnitude in the _ v _ band ( column 6 ) , and _ b - v _ color ( column 7 ) . for one source ( xmm-47 )
we found an optical counterpart in slightly larger distance from the x - ray source position ( within 5.3 and 4.9 in the ogle - ii and mcps catalogs , respectively ) .
the near - infrared counterparts of the be - xrbs studied here are taken from the 2mass catalog ( skrutskie 2006 ) .
column 8 lists the closest 2mass counterpart i d within 2 from the x - ray source position , and columns 9 and 10 the j , and k magnitudes along with their errors .
four sources ( 4 - 3 , xmm-15 , xmm-47 and xmm-5 ) do not have a 2mass counterpart within 2 from the x - ray source position .
the full ( 3650-8000 ) wavelength calibrated spectra of the 20 smc be - xrbs are given in figure [ 2dfspecbexrbs ] .
the early - type nature of the stars is apparent in all spectra , as is their strong emission .
the contribution of telluric emission lines is sometimes not well subtracted ( due to the difficulties in sky subtraction in multifiber observations discussed earlier ) leaving emission and/or absorption features around the [ oi]@xmath20 lines . moreover , telluric absorption bands of molecular oxygen are obvious in the red part of our spectra , around 6875 , and 7610 . finally , some diffuse interstellar emission lines can also be seen in some of the spectra , including [ oiii]@xmath21 , [ oii]@xmath22 , [ nii]@xmath23 and [ sii]@xmath24 . in table
[ ewfwhm ] we present for each selected x - ray source ( column [ 1 ] ) the measurements of the center ( column [ 2 ] ) , fwhm ( column [ 4 ] ) , and ew ( column [ 6 ] ) of the emission line , along with their errors ( columns [ 3 ] , [ 5 ] , and [ 7 ] , respectively ) from the interactive fits .
the errors represent the 68% confidence intervals for one interesting parameter .
we found published ew measurements for 8 of our sources .
we present these values in column [ 8 ] along with their source ( given in parenthesis ) .
they are in good agreement , despite the expected scatter , given the variable nature of be stars ( e.g. mcswain 2008 ) , and the different disk inclination in which they may have been observed ( dachs 1986 ) .
the emission - line profiles in the spectra of table [ ewfwhm ] are very simple with no signs of double or split peaks .
however , because of the resolution of the used grating , we may not have been able to resolve any asymmetries or double peaks .
in addition , in column [ 9 ] we give the orbital period of the hmxbs in this sample from the literature ( liu 2005 and references therein ) , and we use these values in order to study the properties of the decretion disk in relation to the parameters of the binary system ( see discussion in section [ porbital ] below ) . in figure [ ewdistr ]
we show the distribution of the ews for be - xrbs ( solid histograms ) and be stars ( open histograms ) .
we note here that by using the term be stars throughout this paper we refer to field and cluster be stars that do not exhibit x - ray emission characteristic of xrbs .
these distributions are shown for the smc ( top panel ) , the lmc ( middle panel ) and the milky way ( bottom panel ) .
the data for the smc be - xrbs are taken from the present study and are supplemented by the catalog of coe ( 2005 ) , while for the smc be stars we used data from the study of the star - cluster ngc330 by martayan ( 2007a ) .
the data for the lmc be - xrbs are taken from liu ( 2005 ; and references therein ) , while the lmc be sample is from the study of the star - cluster ngc2004 by martayan ( 2006a ) .
the data for the galactic be - xrbs are taken from reig ( 2007 ) and liu ( 2006 ; and references therein ) , while for the galactic be stars we used data from ashok ( 1984 ) , supplemented with data from fabregat & reglero ( 1990 ) , dachs ( 1986 ) , and dachs , hummel & hanuschik ( 1992 ) .
we note that the above galactic be samples contain field stars . in order for the above comparisons to be valid , we selected be - xrbs and be stars within the same spectral - type range in each of the 3 galaxies .
be - xrbs in the milky way and the lmc have spectral types up to b2 , while the majority of such systems in the smc is found to have spectral types b3 or earlier ( see discussion in section [ sptypediscuss ] and figure [ bspecsubtdistr ] below ) .
the spectral types of galactic sources with ew measurements used here are taken from andrillat & fehrenbach ( 1982 ) and skiff ( 2007 ; and references therein ) , while for the smc and lmc be stars we used data from martayan ( 2007b ) and ( 2006b ) , respectively .
in total we have used 41 galactic be stars , 21 galactic be - xrbs , 116 smc be stars , 35 smc be - xrbs , 5 lmc be - xrbs and 42 lmc be stars . from figure [ ewdistr ]
we see that the ew distribution of smc be - xrbs peaks at @xmath25 , while that of smc be field stars peaks at @xmath26 .
this is in agreement with reig ( 1997 ) and zamanov ( 2001 ) who reported that be stars in xrbs have on average a lower ew when compared to be stars , due to the truncation of the disk by the compact object . for the lmc and the milky way ,
the comparison is hampered by the small size of the samples .
on the other hand , dachs ( 1992 ) found that be stars with larger ew have larger radii of their emitting circumstellar disks .
grundstrom & gies ( 2006 ) used numerical models to show that the predicted radius of the be disk , based on ew , is monotonically related to the observed disk radius with long baseline interferometry . from the above figure
, we also see that the be stars in the magellanic clouds ( hereafter mcs ) have larger ew values when compared to the galactic sample , thus indicating larger disks around them .
martayan ( 2007a ) , who find similar results , suggest that the higher ews ( especially in the smc ) indicate the presence of more extended circumstellar envelopes around a fraction of be stars .
we also note that some be stars may be in fact in binary systems .
however , even if some of those prove to be binaries , we do not expect their ew distributions to change significantly .
the identification and classification of the optical counterparts of the x - ray sources allow us to identify the interlopers and determine the type of xrbs ( e.g. be - xrbs , supergiant xrbs , low - mass xrbs ) . in figure [ cmd2df ] we present
the extinction corrected absolute _
magnitude vs. _ b - v _ color ( @xmath27 versus @xmath28 ) color - magnitude diagram ( cmd ) of the counterparts of the x - ray sources with 2df spectra ( black circles ) .
the magnitudes , colors and their errors are taken from the ogle - ii catalog ( udalski 1998 ) , except for sources with matches only in the mcps catalog ( zaritsky 2002 ) , based on data from table [ 2dfoptids ] . with gray squares
we also show the ngc330 be stars ( martayan 2007a ) , while with gray small dots we present the ogle - ii stars that lie in our field 4 and have small errors in the _ v _ and _ b _ band ( @xmath29 mag ) . the main sequence ,
the red giant branch and the red clump loci are clearly seen . in this figure , we overplot the isochrones from the geneva database ( lejeune & schaerer 2001 ) for the metallicity of the young smc stars ( @xmath30}=-0.68\pm0.13 $ ] dex , equivalent to @xmath31 ; luck 1998 ) , for ages of 8.7 myr to 275.4 myr .
the position of all the x - ray sources in the cmd agrees well with the ngc330 be stars .
we also see that the be - xrbs from our sample and the be stars from martayan ( 2007a ) are at the right of the main sequence , as expected given that be stars are generally redder than b - type stars ( e.g. mcswain & gies 2005 ) , while they all lie between the @xmath32 myr and @xmath33 myr isochrones .
most be stars exhibit infrared excess ( gehrz , hackwell & jones 1974 ) due to the disk contribution to the continuum emission .
several authors have observed a similar excess in be - xrbs ( e.g. coe 2005 ) . in figure [ jko ]
we present the distribution of the j - k color for the be - xrbs of the present work ( black filled histogram ) .
we also show the j - k color distribution of be stars of luminosity class iii - v ( gray hatched histogram ) and of all non - emission b - type stars ( i.e. excluding the above be stars sample ; gray filled histogram ) using data from the 2df smc spectroscopic survey of evans ( 2004 )
. for clarity , the number of sources in the be - xrb and be stars sample is normalized by a factor of 2 .
we find that the j - k distribution of be - xrbs peaks at @xmath20.3 mag , while that of be stars at @xmath20.5 - 0.9 mag .
non - emission b - type stars show a skewed distribution , which peaks at @xmath34 - 0.1 mag .
the primary spectral classification criteria for ob stars are usually based on metal line strengths except for the earliest subtypes which are based on hei and heii lines ( e.g. lennon 1997 ) .
however , due to the low metallicity of the smc sources , the metal lines are very weak ( see also evans 2004 ) hampering the classification based on them .
therefore , we adopt the classification scheme of evans ( 2004 ) , although , due to the lower resolution of our spectra , it was not always possible to use all the proposed indicators . according to this scheme , the spectral classification is based on the blue part of the spectrum ( 3800-5000 wavelength range ) .
the presence of heii absorption lines ( heii@xmath35 ) indicates spectral types earlier than or equal to b0 .
strong hei lines are also indicative of early spectral types . their strength increases with decreasing temperature , while it peaks around b2 .
the main characteristic of early b - type stars is the absence of lines of ionized helium ( heii ) , which are still present in late o - type stars .
thus , in b0.5 stars , heii@xmath36 and heii@xmath37 are absent , while heii@xmath38 is weak . in b1 stars ,
heii@xmath38 disappears , while siiv@xmath39 lines appear . for spectral types later than b1.5 the oii+ciii@xmath40 blend decreases rapidly , and disappears for stars later than b3 . for spectral types b2 or later mgii@xmath41 and siiii@xmath42 become stronger .
comparison between the strengths of these two lines is used to distinguish between b2 , b2.5 , and b3 types . for spectral types b5 or later
the classification is based mainly on the absence of siiii@xmath42 . in b8 stars
hei@xmath43 is stronger than siii@xmath44 , which in turn is stronger than hei@xmath45 . finally , for b9 stars mgii@xmath41 appears stronger than hei@xmath46 . in table
[ identifications ] we summarize the above line identifications ( column 1 ) for each b spectral type subclass ( column 2 ) .
the resulting spectral types for the 20 objects of table [ otherxrayname ] are shown in table [ spclass ] , while sample spectra in the 3800-5000 wavelength range are shown in figure [ normspectra ] , with characteristic spectral features marked . in table
[ spclass ] we give the x - ray source i d ( column 1 ) , the spectral classification derived from this study ( column 2 ) , and any previous classification of the same sources ( column 3 ) along with their reference ( given in parenthesis ) .
( i.e. tentative be ) input in column 3 indicates sources for which the classification as a be star is based on the positional coincidence of a hard x - ray source with an emission - line object from the catalog of [ ma93 ] and/or with a photometrically classified early - type ( ob ) star ( e.g. antoniou 2009 ) . in total , 5 out of the 20 sources studied here have been previously identified as candidate be stars based only on photometric data and/or positional coincidence with an emission - line object from the catalog of [ ma93 ] .
in addition , two sources with hard x - ray spectra have been previously spectroscopically classified as b - type stars ( i.e. non - be stars ) in the study of evans ( 2004 ) .
furthermore , one object was classified as b - type star showing balmer emission , but no further classification has been provided ( garmany & humphreys 1985 ) .
the present study examines these classifications , since only spectroscopy can unambiguously identify the nature of the donor with that of a be star . due to the absence of flux calibration and the relatively low resolution of our spectra it was not possible to assign with certainty a spectral luminosity class to our objects . for the same reason , evans ( 2004 ) , who have used the same spectrograph ,
have relied on a photometric / spectroscopic approach to estimate the luminosity class . because later - type main - sequence stars are also fainter , and because the present spectroscopic sample is obviously flux limited , there is an obvious selection effect favoring earlier spectral types .
the faintest of these 20 sources ( which has @xmath47 counts in the 5600-6200 wavelength range ) has been classified here as a b0 - 4e star , thus our sample is complete up to mid - b .
we find good agreement between our and other published spectral - type classifications for 9 out of the 15 sources with published classification .
however , for 6 sources ( xmm-3 , xmm-5 , 5 - 2 , 5 - 7 , xmm-39 , and xmm-47 ) we derive different ( or in some cases slightly different ) spectral types .
in addition , due to the low resolution of the spectra used in this work and the low s / n ratio of sources 6 - 1 , 5 - 12 , 6 - 4 , and 3 - 3 we can only assign to them broad spectral types . the blue part of the spectrum ( 3800-5000 ) of the last three sources is presented in figure [ normspectra ] . in summary , for all but four sources ( 5 - 2 , 5 - 7 , xmm-3 , and xmm-47 for which we find 0.5 , 1 , 0.5 and 1 spectral subtype differences , respectively ) we find spectral classifications in good agreement with previous studies .
we estimate that our spectral types are accurate to better than @xmath48 subclass in most cases , especially for earlier than b2 spectral types .
next we discuss in detail the spectral classification criteria for all sources in table [ spclass ] : + * we classify source xmm-3 as a b0.5e type star based on the presence of weak heii@xmath38 , absent heii@xmath37 , and very weak ( almost absent ) heii@xmath36 .
the spectrum of source xmm-3 exhibits strong balmer lines in absorption ( with the strongest one being h@xmath49 ) , and moderate h@xmath50 in emission .
weak hei ( @xmath51 ) lines and oii+ciii@xmath40 blend are also present .
source xmm-3 has been assigned a slightly later spectral type ( b1 - 1.5ve ) by mcbride ( 2008 ) .
* based only on photometric data , source 4 - 8 has been recently classified as a later than b3 emission - line star ( haberl 2008 ) . in the present work , we classify it as a b1.5e star , based on the presence of strong , weak siiii@xmath42 and oii+ciii@xmath40 blend , and the absence of siiv@xmath52 and heii@xmath38 . similarly to source xmm-5
, it shows strong hei+ii@xmath53 and hei@xmath54 lines , while h@xmath50 shows weak emission .
* source xmm-5 has been classified as b0 - 5(ii ) by evans ( 2004 ) .
based on the presence of strong , weak siiii@xmath42 and oii+ciii@xmath40 blend , and the absence of siiv@xmath52 and heii@xmath38 , we classify this source as b1.5e .
other present helium lines include hei+ii@xmath53 and hei@xmath54 .
the blue part of the spectrum ( 3800-5000 ) of source xmm-5 is shown in figure [ normspectra ] .
* sources 4 - 2 , 4 - 5 , 5 - 4 , and 6 - 2 are classified as b1e stars in this work .
siiv@xmath55 lines and oii+ciii @xmath40 blend are present , while heii@xmath56 is absent .
h@xmath57 is partially filled - in by emission in source 4 - 2 , and completely filled - in in sources 5 - 4 and 6 - 2 .
sources 4 - 5 and 5 - 4 are classified here for the first time , while they are listed as candidate be - xrbs associated with objects [ ma93]300 ( haberl & pietsch 2004 ) and [ ma93]798 ( antoniou 2009 ) , respectively .
in addition , sources 6 - 2 and 4 - 2 have been previously classified as b1 iii - ve and b1 - 3 iv - ve stars by mcbride ( 2008 ) . *
we classify sources 4 - 1 and 5 - 16 as b0.5e stars , based on the presence of the hei@xmath58 and heii@xmath59 lines and the absence of the heii@xmath60 lines .
the oii+ciii @xmath40 blend is weak , while in the case of source 5 - 16 h@xmath61 and h@xmath57 are partially filled - in by emission .
coe ( 2002 ) have classified source 4 - 1 as a b0 - 0.5 ve star , while source 5 - 16 has been classified as a o9.5-b1 iv - ve star by mcbride ( 2008 ) . *
our classification of source xmm-15 as a b0.5e star , based on very weak emission in heii@xmath60 and hei@xmath62 , absent heii@xmath38 and siiv@xmath55 , and strong hei@xmath63 and oii+ciii@xmath40 absorption , is in agreement with previous studies ( mcbride 2008 ) . in figure [ normspectra ]
we present the blue part ( 3800-5000 ) of its spectrum .
this source shows h@xmath64 , h@xmath49 and h@xmath65 in absorption , partly filled - in by emission .
h@xmath57 is completely filled - in by emission ( absent from the spectrum ) , while the strong h@xmath50 emission line is the most prominent feature of the blue part of the spectrum .
* source 4 - 3 has not been previously classified .
it is only listed as candidate be - xrb associated with object [ ma93]414 ( haberl & pietsch 2004 ) . in this work ,
we classify it as a b0.5e star , based on the absence of heii@xmath60 lines , and on the presence of weak heii@xmath56 .
* source xmm-17 is classified here as a b0e star , based on the presence of heii@xmath66 lines , in agreement with mcbride ( 2008 ) , who have classified it as a o9.5-b0 ve star .
* source 6 - 1 is classified here as a b1 - 3e star .
the heii@xmath56 line is absent , while siiv@xmath67 is present , indicating a spectral type of b1 .
however , mgii@xmath68 is stronger than siiii@xmath69 , which is common in b3 type stars .
finally , the oii+ciii @xmath40 blend is present but weak , thus indicating a spectral type later than b1.5 . taking into account all the above , we can only assign a broad spectral type of b1 - 3e to source 6 - 1 , which is identical to the classification of mcbride ( 2008 ) . furthermore , h@xmath57 is partially filled - in by emission .
* we classify source 5 - 12 as o9e - b0e , which is the earliest spectral type found in this sample .
we based this classification on the presence of hei@xmath70 and heii@xmath66 lines .
hei+ii@xmath53 and hei@xmath71 are also present . the balmer series h@xmath72 ,
h@xmath49 , h@xmath65 are strongly in absorption .
however , h@xmath57 is party filled - in by emission , while h@xmath50 shows weak emission .
moreover , this spectrum shows strong absorption of the oii+ciii@xmath40 blend , indicative of an early b - type star .
* source 6 - 4 has been broadly classified as a b - type star which shows balmer emission ( garmany & humphreys 1985 ) . in the present work
, we study this object more extensively .
the composite nature of this spectrum is revealed based on the hei@xmath46 and heii@xmath37 lines , that indicate early - type ( ob ) components .
h@xmath50 is almost completely filled - in by emission , while h@xmath72 , h@xmath64 , h@xmath49 , h@xmath65 and h@xmath57 are in absorption .
the strong niv@xmath73 emission is the most prominent feature in the blue portion of its optical spectrum ( 3800-5000 ; figure [ normspectra ] ) .
the nv@xmath74 lines are not present , while heii@xmath38 line is not in broad , low - intensity emission , as it is typical for o3 if * stars ( walborn & fitzpatrick 1990 ) .
we do not detect the weak niii@xmath75 triplet emission , which in combination with strong heii@xmath38 absorption , has been used as a signature of o((f ) ) stars ( e.g. maz - apellniz 2004 ) , but this could be due to the spectral resolution ( most probably blended with nii@xmath76 ) .
in contrast to the above indicators which suggest an o spectral - type , heii@xmath36 is absent , indicating a b - type star .
other helium lines clearly seen in this spectrum , include hei+ii@xmath53 and hei@xmath77 .
taking into account the above spectral features , we conclude that this is a composite spectrum and we assign an o2((f))+obe type to this system .
based on the current spectral resolution , a classification is not straightforward and only higher resolution spectroscopy can unambiguously identify the two components and their correct spectral types .
* we classify source 5 - 2 as a b0.5e star based on the presence of weak heii@xmath38 , and the absence of heii@xmath60 lines .
strong hei+ii@xmath53 and hei@xmath78 absorption lines are in agreement with the above classification .
hei@xmath77 appears weaker , while hei@xmath79 shows strong , broad emission .
this source has been classified as a later type star ( b1 - 2 iv - ve ) by mcbride ( 2008 ) . * in the same work , source 5 - 7 is classified as an o9 ve star ( mcbride 2008 ) .
we find that this is a later ( b0e ) spectral - type star based on the presence of hei and heii lines in the blue part of its spectrum ( 3800-5000 ; figure [ normspectra ] ) .
in particular , heii@xmath60 lines are present , but they are much more weaker than hei@xmath80 .
heii@xmath38 is also weak .
hei@xmath81 lines are clearly seen in this spectrum , while the oii+ciii,@xmath40 blend is also present . the above lines are tell - tale signatures of a b0 star .
* source 3 - 3 , which has been classified as b0-b2 by mcbride ( 2008 ) , shows several conflicting spectral - type indicators . in our spectrum ,
the mgii@xmath41 line , which appears in later than b2 spectral types , is present .
it is also stronger than siiii@xmath42 , indicating a b3 spectral type . for spectral types b5 or later siiii@xmath42 is absent , hence the latest spectral type that can be assigned to source 3 - 3 is b4 .
in addition , heii@xmath38 and the oii+ciii@xmath40 blend are very weak ( almost absent ) , and on that account source 3 - 3 would be of b0.5 - 1.5 spectral type .
on the other hand , siiv@xmath52 is present , which is typical of b1 stars , but not siiv@xmath82 .
furthermore , heii@xmath36 and heii@xmath37 are present , indicating b0 and b0.5 spectral types , respectively . based on the above spectral features , we only assign to source 3 - 3 a wide b0 - 4 spectral type .
we also note that balmer lines h@xmath64 and h@xmath57 are both filled - in by emission .
the spectrum of this source also shows emission in nii@xmath83 and mgi@xmath84 .
* source xmm-39 exhibits the strongest h@xmath50 emission of the sources from the current sample .
moreover , h@xmath65 and h@xmath57 are partly and completely filled - in by emission , respectively .
the optical counterpart of this source has been observed with 2df by evans ( 2004 ) , and it has been classified as a non - emission line b0 - 5(ii ) star .
we classify it as a b1.5e star because we find weak siiii@xmath42 absorption and moderate oii+ciii@xmath40 blend .
this spectrum also shows unidentified emission around 3951 .
* source xmm-47 has been classified as b0.5(iv)e by evans ( 2004 ) , however , following the above criteria , we classify it as b1.5e .
in particular , absence of the siiv@xmath52 @xmath85 line and oii+ciii@xmath40 blend suggests a b1.5 or later spectral type source . on the other hand , no presence of the mgii@xmath41 line is indicative of an earlier than b2 star , while a very weak heii@xmath38 line indicates a b0.5 ( or slightly earlier ) spectral type ( although the presence of this line is ambiguous .
we thus classify source xmm-47 as a b1.5e star .
in the previous sections we presented the spectral classification of optical counterparts of 20 smc hmxbs observed with the 2df spectrograph . we identify all of these sources as be - xrbs .
we also compared these results with previous studies and we found that our classification is in good agreement in all but four cases .
next we discuss the implications of these results in the context of the spectral - type distribution of be - xrbs comparisons in the mcs and the galaxy .
in addition , we investigate the distribution of the ew measurements for be - xrbs and be field stars in these 3 galaxies and their -orbital period ( ) relation .
we also investigate the intrinsic reddening values of the systems presented in this work .
smc and galactic be - xrbs are found to follow the same spectral - type distribution ( mcbride 2008 ) .
the same holds for lmc and galactic be - xrbs ( negueruela & coe 2002 ) .
extending the above comparisons to be stars , coe ( 2005 ) observed somewhat similar distributions for be - xrbs and be stars in the smc , however using photometrically derived spectral types . in the present work we supplement the smc be - xrb sample used in mcbride ( 2008 ) with spectral - type classifications for 10 additional sources , and we investigate differences in the spectral - type distributions of both be - xrbs and be stars in the mcs and the milky way . for sources common in the two studies
, we have used the classification derived from the present work . in figure [ bspecsubtdistr ]
we present the b spectral subtype distributions of be - xrbs ( solid histograms ) and of be stars ( dashed histograms ) in the smc ( top panel ) , the lmc ( middle panel ) , and the milky way ( bottom panel ) .
negative spectral subtypes correspond to o - type stars .
whenever only a broad spectral class was available , we equally divided the contribution in the different subtypes . for the smc ,
the data for be stars are taken from evans ( 2004 ; 168 in total ) , and for be - xrbs from our sample supplemented by the recent study of mcbride ( 2008 ; 46 in total ) .
the galactic be stars ( 148 in total ) are taken from the study of slettebak ( 1982 ) .
for the galactic be - xrbs we used 20 sources from the census of reig ( 2007 ) , 4 additional from the work of mcbride ( 2008 ) , and 2 more from the catalog of liu ( 2006 ) .
for the 14 lmc be - xrbs we used the spectral types from the census of liu ( 2005 ) .
the lmc be stars ( 103 in total ) are taken from martayan ( 2006b ) . in order to check the null hypothesis that different samples of be - xrbs and be stars are drawn from the same populations , we perform the two - sample kolmogorov - smirnov ( ks ) test ( e.g. , conover 1971 ) .
the results of the p - values are presented in table [ kstest ] , a two - way table with the compared distributions for be - xrbs and be stars listed in its rows and columns . because of the small size of the lmc be - xrbs sample we also check these results with the two - sided mann - whitney - wilcoxon test ( bauer 1972 ) .
in addition , because it is very difficult to obtain accurate spectral types for stars later than mid b - type in the mcs , we repeat the above analysis by using the peto & prentice generalized wilcoxon test ( lavalley , isobe & feigelson 1992 ) and assuming lower limits to the spectral subtypes later than b3 in all samples and later than b1 and b2 in the lmc and galactic be - xrb samples , respectively .
the three tests agree only in the case of very small ( @xmath860.1 ) or very large ( @xmath870.9 ) p - values , while intermediate values are not informative .
in particular , small p - values ( @xmath860.1 ) are indicative of the different populations from where these distributions are drawn .
based on these results , we can not definitely say if the smc be - xrbs follow a different spectral - type distribution from galactic and lmc be - xrbs .
this is consistent with mcbride ( 2008 ) , who note that there is indication for similar distributions in the smc and the milky way .
we find similar spectral - type distributions between lmc and galactic be - xrbs ( p - value=1.0 ) , in agreement with negueruela & coe ( 2002 ) , although we note the small size of the samples .
regarding the distributions of be - xrbs and be stars , we find that in the smc the two populations are consistent with each other , in contrast to the milky way populations ( in agreement with negueruela , 1998 ) .
moreover , we find similar be spectral type distributions between the mcs samples , while we do not find evidence for differences between the mcs and the milky way .
previous studies find the highest fraction of be stars at b2 spectral types ( e.g. martayan 2007b ) .
mcbride ( 2008 ) find a spectral distribution cutoff for the smc be - xrbs around b2 and they suggest this might be due to significant angular - momentum losses , even before the be - xrb evolutionary phase ( i.e. during the first stage of mass transfer ) . ekstrm ( 2008 ) believe that the be - star phenomenon may be limited to a given spectral - type range because of the mass loss and the meridional currents . in particular , they argue that for spectral types earlier than the favored one , mass loss prevents the star from approaching the critical velocity limit ( at which the be - star phenomenon appears ) , while for the later spectral types , the meridional currents ( which transfer angular momentum to the surface ) are not efficient enough for accelerating the envelope .
regarding the effects of metallicity in the fraction of be stars in a galaxy , it is now widely accepted that the lower the metallicity the higher this fraction , because of their higher rotational velocities ( e.g. maeder 1999 , wisniewski & bjorkman 2006 , martayan 2006b , 2007a ; after correcting for multi- and single - epoch observations for the milky way and the mcs , respectively ) .
moreover , the be star phase can last longer in low metallicity environments such as the mcs when compared to the galaxy ( martayan 2007b ) .
as pointed out in zorec & briot ( 1997 ) the observed populations of be stars suffer from many incompleteness effects , especially in faint magnitude limits .
since the quantification of these selection effects is still in its infancy and depends on the used samples , we opted to use the available uncorrected samples and simply treat the number of stars at late spectral types as lower limits .
this approach has lower discriminating power but it does not rely on modeling assumption about the selection effects .
furthermore , the above samples include both field and cluster be star populations . in order to check
if there are different spectral type distributions due to their different ages and rotational velocity characteristics , we examined 2 magellanic clouds samples which distinguish between them ( martayan 2007b and 2006b for the smc and the lmc , respectively ) .
we do not find any difference in the spectral type distributions of be field and cluster populations , which could be due to the small size of these samples ( 79 and 52 be stars for the smc , and 27 and 20 for the lmc , respectively ) . in any case
, in the present work we do not distinguish between the different be star populations , especially since all be - xrbs used in our study are located in the field .
although there is a possibility that they were formed in clusters and then ejected due to supernova kicks , there is no way to test it . from the above analysis it is obvious that in comparisons of the spectral - type distribution of different populations
, one should keep in mind that the be phenomenon is a transient phase in the life of some b - type stars , greatly depending on the rotational velocity and perhaps binarity ( e.g. mcswain et al .
because of the transient nature of the be - star disks , we may then get only a fraction of the real underlying be - star population with single - epoch observations , as is the majority in the smc .
the orbital period of an xrb system is related to the size of its orbit , while the ew is considered an indicator of the size of the decretion disk ( e.g. , grundstrom & gies 2006 ; as observed by long baseline interferometry of nearby be stars ) . reig ( 1997 ) found a strong correlation between orbital period and ew , and attributed it to the presence of the neutron star , which appears to act as a barrier which prevents the formation of a extended disk in systems with short orbital period .
figure [ ewporb ] is an updated version of the @xmath88-ew(h@xmath18 ) diagram for be - xrbs including in addition to the 15 galactic be - xrbs ( open circles ) used in the work of reig ( 2007 ) , 14 smc sources ( filled squares ) derived from the present work ( values are taken from table [ ewfwhm ] and references therein ) and 2 recently detected x - ray pulsars ( sxp7.78 , sxp455 ) .
the orbital periods of these 2 sources are taken from liu ( 2005 ; and references therein ) , while for the ew measurements we used the work of coe ( 2005 ) . in figure [ ewporb ]
we only included sources with well determined orbital periods , thus other smc x - ray pulsars ( such as sxp756 , sxp8.80 , sxp46.6 and sxp304 ) with published ew and values are not present .
five x - ray sources ( out of the 20 used in section [ sptypediscuss ] ) do not have published orbital periods , thus they do not appear in this figure .
source 4 - 1 is not included in figure [ ewporb ] because of its complicated nature ( coe 2002 ) .
we also note that the ew measurements used here are the maximum found in the literature ( thus these measurements can be considered upper limits of the disk size ; e.g dachs 1992 ) .
this plot strongly supports the correlation of these two quantities despite the large scatter observed for the shorter period systems and extends it to extragalactic systems . as reig ( 2007 )
points out , the neutron star does not allow the companion star to develop a large decretion disk in cases of small orbital period systems , leading to the tidal truncation of the disk , and in turn smaller ew values .
the values of the linear regression correlation coefficient for the smc , milky way and the combined sample are 0.71 , 0.89 , and 0.81 , respectively .
thus , the smc sample ( solid line ) is less correlated when compared to the galactic sample ( dashed line ) , while the combined data sample ( dotted line ) shows a stronger correlation than the smc data alone because of the increased sample size . from this plot
, we also see that the smc sources tend to have systematically larger ew than galactic sources with similar orbital periods
. this could be due to their faster rotational velocities ( a result of their lower metallicity , as observed for example by martayan 2007b ) and hence larger decretion disks .
however , given the truncation effect discussed earlier , growing a larger disk would require a wider orbit which for the same orbital period would indicate a lower ellipticity .
alternatively , a high mass - loss rate would allow a faster growth of the disk after the truncation .
in addition to the ew , the @xmath89 color is also an indicator of the size of the circumstellar disk . by plotting these quantities , coe ( 2005 )
find a similar range in the values of galactic be stars and smc be - xrbs , but not significant evidence for a correlation between the ew and the @xmath89 color for the be - xrbs . here
we revisit this study using our extended census of be - xrbs and comparing with smc be stars from the study of ngc330 by martayan ( 2007a ) .
the ews for the be - xrbs are taken from table [ ewfwhm ] and the j and k magnitudes from table [ 2dfoptids ] , corrected for reddening assuming e(j - k)=0.56e(b - v ) ( bessell & brett 1988 ) , and e(b - v)=0.09 mag ( following antoniou 2009 ) . the ew measurements for the ngc330 be stars are taken from martayan ( 2007a ) , while the @xmath89 color values are the result of the cross - correlation of the above census with the 2mass catalog .
a comparison of these two quantities does not show any evidence for correlation , which we attribute to the fact that the ir and observations are not simultaneous , and thus the comparison between the two quantities is not direct .
therefore , we do not pursue this comparison further .
instead we focus on the individual distributions of the ew and ( j - k ) color .
as shown in figures [ ewdistr ] and [ jko ] , respectively , the majority of the be - xrbs presented in this work have ews extending up to the mid range of the corresponding values for be stars , while both the be stars and the be - xrbs in the smc have similar values of the @xmath89 color .
when we compare the results for the be - xrbs presented in this work with the study of coe ( 2005 ) , we find compatible values for the ews , however most of the sources in the present work have larger values of the @xmath89 color ( @xmath90 mag ; figure [ jko ] ) than those presented by coe ( 2005 ; @xmath91 mag ) .
this can be the result of the different extinction correction in the two studies .
we have used e(j - k)=0.05 , while coe ( 2005 ) corrected for all reddening effects by using the same color shift used for determining the spectral class of their objects ( shift by ( b - v)=-0.13 from their study ) .
we then investigate the correlation of the ew measurements to the intrinsic reddening caused by the circumstellar disks around be stars .
we define this local " reddening as the difference in the _ b - v _ color of the observed ( extinction corrected ) and theoretical values .
the observed _ b - v _ color is taken from the mcps catalog of zaritsky 2002 ( since all the be - xrbs from the present study have an optical counterpart in this catalog ) after correcting for the extinction to the smc ( e(b - v)=0.09 ; see [ 2dfoptcparts ] ) . in order to derive the theoretical
_ b - v _ color , we first assign an effective temperature to each b subtype using the temperature scale for main - sequence b - type stars at the smc metallicity ( figure 33(d ) of hunter 2007 ) .
we then derived the theoretical _ b - v _ color using the geneva isochrones database ( lejeune & schaerer 2001 ) assuming z=0.004 ( for the smc metallicity ) and the nearest corresponding available effective temperature to the one derived above .
the ew versus intrinsic reddening plot is presented in figure [ localreddening ] , while the values of these quantities are given in table [ deltabv ] . in column 1
we give the source i d , in column 2 the spectral classification from this study and in columns 3 and 4 the extinction corrected @xmath28 color and the corresponding error ( using data listed in table [ 2dfoptids ] ) , respectively . in column 5 we give the nearest b subtype to the classified sources used in order to derive the effective temperature . in column 6
we give the _ b - v _ color from the isochrones , and in column 7 the intrinsic reddening ( extinction corrected photometric minus theoretical _ b - v _ values ) . in table [ deltabv ] and figure [ localreddening ] we did not include source 6 - 4 because its composite spectrum indicates that there might be more than one stars contributing to its emission . from this figure , we do not find any correlation between these two quantities , though they are both related to the decretion disk of the be star .
the large scatter in the plot is mainly due to the fact that the theoretical values for the _ ( b - v ) _ color are in some degree uncertain .
first , we had to assume a mean subtype for sources with uncertain spectral class ( e.g. x - ray sources 3 - 3 , 6 - 1 , 5 - 12 ) .
then , we had to extrapolate the temperature scale of hunter ( 2007 ) to later than b1 spectral subtypes . and finally , we had to choose the nearest effective temperature derived above to the one used in the geneva database .
all the above , in addition to disk inclination effects , can account , at least in part , for the large scatter of this plot , and may explain the fact that we do not find any correlation of the ew and the intrinsic reddening , though they are both related to the size of the circumstellar disk around be stars .
we also note that even if we use the mean _ b - v _ color for each b subtype using data from evans ( 2004 ) , we do not find any correlation either .
the fact that the spectroscopic and photometric observations are not simultaneous may also result in increased scatter .
we have presented the results of a detailed optical spectroscopic analysis of 20 hmxbs in the smc .
these sources have been detected with ( a. zezas 2009 , in preparation ) and _ xmm - newton _
( haberl & pietsch 2004 ) . in the present work ,
we confirm the be nature of all these systems .
in addition , we provide their full wavelength range ( @xmath92 ) spectra .
the classification of these 20 systems is accurate to better than @xmath48 subclass and is proved to be in good agreement with previous studies . in a few cases , only wide range spectral - types could be assigned , and for these systems higher resolution spectroscopy is needed in order to derive more accurate classifications .
we also find similar spectral - type distributions for be - xrbs and be field stars in the smc . with the present available data we can not conclude that the be - xrbs in the smc and the milky way or the lmc follow a different spectral - type distribution .
in addition , we find similar be spectral type distributions between the mcs samples , while we do not find evidence for differences between the mcs and the milky way .
moreover , our results reinforce the relation between the orbital period and the equivalent width of the line that holds for be - xrbs .
this relation is the result of truncated decretion disks in smaller period systems , due to the interaction of the compact object with the circumstellar disk of the be star . for similar orbital periods , smc sources tend to have systematically larger ew than galactic sources .
this could be due to their faster rotational velocities ( a result of their lower metallicity , as observed for example by martayan 2007b ) and hence larger decretion disks .
we also investigate the near - infrared properties of the 20 be - xrbs by cross - correlating their x - ray positions with sources in the 2mass catalog . for sources with a 2mass counterpart we do not find evidence for correlation of the ew values and the j - k color , though they are both related to the size of the decretion disk .
the same holds for the ew values when compared to the intrinsic reddening of the sources ( defined as the observed minus the theoretical _ b - v _ color ) .
we thank the referee , virginia mcswain , for useful comments which have improved this paper .
we would also like to thank rob sharp for performing the 2df observations during service time and nolan walborn for fruitful discussions on the spectral classification .
va acknowledges support from marie curie grant no .
39965 to the foundation for research and technology - hellas , nasa ltsa grant nag5 - 13056 , and nasa grants go2 - 3117x and go8 - 9089c .
az acknowledges support by fp-7 rerpot grant 206469 .
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 .
andrillat , y. , & fehrenbach , c. 1982 , , 48 , 93 antoniou , v. , zezas , a. , hatzidimitriou , d. , & mcdowell , j. c. 2009 , , 697 , 1695 ashok , n. m. , bhatt , h. c. , kulkarni , p. v. , & joshi , s. c. 1984 , , 211 , 471 bauer , d. f. 1972 , journal of the american statistical association , 67 , 687 bessell , m. s. , & brett , j. m. 1988 , , 100 , 1134 brusa , m. , et al .
2007 , , 172 , 353 coe , m. j. , edge , w. r. t. , galache , j. l. , & mcbride , v. a. 2005 , , 356 , 502 coe , m. j. , haigh , n. j. , laycock , s. g. t. , negueruela , i. , & kaiser , c. r. 2002 , , 332 , 473 conover , w. j. 1971 , in practical nonparametric statistics , ed .
w. j. conover ( new york : wiley ) corbet , r. , marshall , f. e. , lochner , j. c. , ozaki , m. , & ueda , y. 1998 , , 6803 , 1 dachs , j. , hanuschik , r. , kaiser , d. , & rohe , d. 1986 , , 159 , 276 dachs , j. , hummel , w. , & hanuschik , r. w. 1992 , , 95 , 437 dray , l. m. 2006 , , 370 , 2079 ekstrm , s. , meynet , g. , maeder , a. , & barblan , f. 2008 , , 478 , 467 evans , c. j. , howarth , i. d. , irwin , m. j. , burnley , a. w. , & harries , t. j. 2004 , , 353 , 601 fabregat , j. , & reglero , v. 1990 , , 247 , 407 filipovi , m. d. , pietsch , w. , & haberl , f. 2000 , , 361 , 823 garmany , c. d. , & humphreys , r. m. 1985 , , 90 , 2009 gehrz , r. d. , hackwell , j. a. , & jones , t. w. 1974 , , 191 , 675 grundstrom , e. d. , & gies , d. r. 2006 , , 651 , l53 haberl , f. , eger , p. , pietsch , w. , corbet , r. h. d. , & sasaki , m. 2008 , , 485 , 177 haberl , f. , & pietsch , w. 2004 , , 414 , 667 haberl , f. , pietsch , w. , schartel , n. , rodriguez , p. , & corbet , r. h. d. 2004 , , 420 , l19 haberl , f. , & sasaki , m. 2000 , , 359 , 573 hunter , i. , et al . 2007 , , 466 , 277 lamers , h. j. g. l. m. , zickgraf , f .- j .
, de winter , d. , houziaux , l. , & zorec , j. 1998 , , 340 , 117 lavalley , m. , isobe , t. , & feigelson , e. 1992 , in asp conf .
25 , astronomical data analysis software and systems i , ed .
d. m. worrall , c. biemesderfer , & j. barnes ( san francisco , ca : asp ) 245 lejeune , t. , & schaerer , d. 2001 , , 366 , 538 lennon , d. j. 1997 , , 317 , 871 lewis , i. j. , et al . 2002 , , 333 , 279 liu , q. z. , van paradijs , j. , & van den heuvel , e. p. j. 2005 , , 442 , 1135 liu , q. z. , van paradijs , j. , & van den heuvel , e. p. j. 2006 , , 455 , 1165 luck , r. e. , moffett , t. j. , barnes , t. g. , iii , & gieren , w. p. 1998
, , 115 , 605 macomb , d. j. , fox , d. w. , lamb , r. c. , & prince , t. a. 2003 , , 584 , l79 maeder , a. , grebel , e. k. , & mermilliod , j .- c . 1999 , , 346 , 459 maz - apellniz , j. , walborn , n. r. , galu , h. .
, & wei , l. h. 2004 , , 151 , 103 marshall , f. e. , et al .
1998 , , 6818 , 1 martayan , c. , floquet , m. , hubert , a. m. , gutirrez - soto , j. , fabregat , j. , neiner , c. , & mekkas , m. 2007a , , 472 , 577 martayan , c. , frmat , y. , hubert , a .- m . , floquet , m. , zorec , j. , & neiner , c. 2006b , , 452 , 273 martayan , c. , frmat , y. , hubert , a .- m . , floquet , m. , zorec , j. , & neiner , c. 2007b , , 462 , 683 martayan , c. , hubert , a. m. , floquet , m. , fabregat , j. , frmat , y. , neiner , c. , stee , p. , & zorec , j. 2006a , , 445 , 931 mcbride , v. a. , coe , m. j. , negueruela , i. , schurch , m. p. e. , & mcgowan , k. e. 2008 , , 388 , 1198 mcswain , m. v. , & gies , d. r. 2005 , , 161 , 118 mcswain , m. v. , huang , w. , gies , d. r. , grundstrom , e. d. , & townsend , r. h. d. 2008 , , 672 , 590 meyssonnier , n. , & azzopardi , m. 1993 , , 102 , 451 negueruela , i. 1998 , , 338 , 505 negueruela , i. , & coe , m. j. 2002 , , 385 , 517 reig , p. 2007
, , 377 , 867 reig , p. , fabregat , j. , & coe , m. j. 1997 , , 322 , 193 russell , s. c. , & dopita , m. a. 1992 , , 384 , 508 santangelo , a. , cusumano , g. , dal fiume , d. , israel , g. l. , stella , l. , orlandini , m. , & parmar , a. n. 1998 , , 338 , l59 sasaki , m. , pietsch , w. , & haberl , f. 2003 , , 403 , 901 schmidtke , p. c. , cowley , a. p. , crane , j. d. , taylor , v. a. , mcgrath , t. k. , hutchings , j. b. , & crampton , d. 1999 , , 117 , 927 skiff , b. a. 2007 , vizier online data catalog , 1 , 2023 skrutskie , m. f. , et al .
2006 , , 131 , 1163 slettebak , a. 1982 , , 50 , 55 stevens , j. b. , coe , m. j. , & buckley , d. a. h. 1999 , , 309 , 421 udalski , a. , szymanski , m. , kubiak , m. , pietrzynski , g. , wozniak , p. , & zebrun , k. 1998 , acta astron . , 48 , 147 van paradijs , j. , & mcclintock , j. e. 1995 , in x - ray binaries , eds . w.h.g .
lewin , j. van paradijs , and e.p.j .
van den heuvel ( cambridge : cambridge univ .
press ) , 58 walborn , n. r. , & fitzpatrick , e. l. 1990 , , 102 , 379 wisniewski , j. p. , & bjorkman , k. s. 2006 , , 652 , 458 yokogawa , j. , imanishi , k. , tsujimoto , m. , koyama , k. , & nishiuchi , m. 2003 , , 55 , 161 zamanov , r. k. , reig , p. , mart , j. , coe , m. j. , fabregat , j. , tomov , n. a. , & valchev , t. 2001 , , 367 , 884 zaritsky , d. , harris , j. , thompson , i. b. , grebel , e. k. , & massey , p. 2002
, , 123 , 855 zorec , j. , & briot , d. 1997 , , 318 , 443 cccc xmm-3 & xmmu j004723.7 - 731226 ( 1 ) = ( rx j0047.3 - 7312 ax j0047.3 - 7312 ) ( 2 ) & sxp264 & 172 + 4 - 8 & cxou j004814.2 - 731004 ( 3 ) = xmmu j004814.1 - 731003 ( 4 ) & sxp25.5 & + xmm-5 & xmmu j004834.5 - 730230 ( 1 ) = ( rx j0048.5 - 7302 ) ( 5 ) & & 238 + 4 - 2 ( xmm-7 ) & cxou j004913.6 - 731138 = xmmu j004913.8 - 731136 ( 1 ) = ( rx j0049.2 - 7311 ) ( 6 ) & sxp9.13 & + 4 - 5 ( xmm-9 ) & cxou j004929.7 - 731059 ( 3 ) = xmmu j004929.9 - 731058 ( 1 ) & sxp893 & 300 + & = ( rx j0049.5 - 7310 ) ( 6 ) & & + 4 - 1 ( xmm-12 ) & cxou j005044.6 - 731605 ( 3 ) = xmmu j005045.2 - 731602 ( 1 ) & sxp323 & 387 + & = ( rx j0050.8 - 7316 = ax j0051 - 733 ) ( 2,7 ) & & + xmm-15 & ax j0051 - 722 ( 8) & sxp91.1 & 413 : : + 4 - 3 ( xmm-14 ) & cxou j005057.2 - 731008 = xmmu j005057.6 - 731007 ( 1 ) & & 414 + & = ( rx j0050.9 - 7310 , ax j0050.8 - 7310 ) ( 2 ) & & + xmm-17 & xmmu j005152.2 - 731033 ( 1 ) = ( rx j0051.9 - 7311 = ax j0051.6 - 7311 ) ( 2,7 ) & sxp172 & 504 + 6 - 1 & cxou j005209.0 - 723804 ( 3 ) & sxp82.4 & + 5 - 12 & cxou j005245.0 - 722844 ( 3 ) & & + 6 - 4 & cxou j005252.2 - 724830 ( 3 ) & & 618 + 5 - 2 & cxou j005323.9 - 722716 ( 3 ) & sxp138 & 667 + 5 - 16 & cxou j005355.3 - 722646 ( 3 ) = 1wga 0053.8 - 7226,xte j0053 - 724 & sxp46.6 & + 5 - 4 & cxou j005446.2 - 722523 ( 3 ) & & 798 + 6 - 2 & cxou j005455.8 - 724511 ( 3 ) & sxp504 & 809 + 5 - 7 ( xmm-30 ) & cxou j005456.3 - 722648 ( 3 ) = & sxp59.0 & 810 + & ( xte
j0055 - 724 = 1sax j0054.9 - 7226 = 1wga j0054.9 - 7226 ) ( 9,10,11 ) & & + 3 - 3 ( xmm-33 ) & cxou j005736.0 - 721934 ( 3 ) = cxou j005736.2 - 721934 ( 11,12 ) & sxp565 & 1020 + xmm-39 & xmmu j010030.2 - 722035 ( 11 ) & & 1208 + xmm-47 & rx j0104.5 - 7221 ( 5 ) & & 1470 + cccccccccc xmm-3 & 00 47 23.7 & -73 12 27 & o-4 - 116979 & 1.52 & @xmath93 & @xmath94 & j00472330 - 7312275 & @xmath95 & @xmath96 + & & & z-1713720 & 1.25 & @xmath97 & @xmath98 & & & + 4 - 8 & 00 48 14.15 & -73 10 4.1 & o-4 - 171264 & 0.63 & @xmath99 & @xmath100 & j00481410 - 7310045 & @xmath101 & @xmath102 + & & & z-1816472 & 0.30 & @xmath103 & @xmath104 & & & + xmm-5 & 00 48 34.5 & -73 02 30 & o-4 - 178950 & 1.85 & @xmath105 & @xmath106 & & & + & & & = o-5 - 43566 & 1.89 & @xmath107 & @xmath108 & & & + & & & z-1857369 & 1.64 & @xmath109 & @xmath110 & & & + 4 - 2 & 00 49 13.8 & -73 11 37 & o-5 - 111490 & 0.84 & @xmath111 & @xmath112 & j00491360 - 7311378 & @xmath113 & @xmath114 + ( xmm-7 ) & & & z-1938257 & 0.75 & @xmath115 & @xmath116 & & & + 4 - 5 & 00 49 29.74 & -73 10 58.5 & o-5 - 111500 & 0.61 & @xmath117 & @xmath118 & j00492984 - 7310583 & @xmath119 & @xmath120 + ( xmm-9 ) & & & z-1971979 & 0.65 & @xmath121 & @xmath122 & & & + 4 - 1 & 00 50 44.61 & -73 16 5.3 & o-5 - 180026 & 0.55 & @xmath123 & @xmath124 & j00504470 - 7316054 & @xmath103 & @xmath125 + ( xmm-12 ) & & & z-2131651 & 0.60 & @xmath126 & @xmath127 & & & + xmm-15 & 00 50 56.9 & -72 13 31 & z-2158744 & 3.22 & @xmath128 & @xmath129 & & & + 4 - 3 & 00 50 57.6 & -73 10 08 & o-5 - 271074 & 2.10 & @xmath130 & @xmath131 & & & + ( xmm-14 ) & & & z-2159045 & 1.93 & @xmath132 & @xmath133 & & & + xmm-17 & 00 51 52.3 & -73 10 33 & o-6 - 22749 & 1.36 & @xmath134 & @xmath135 & j00515203 - 7310340 & @xmath136 & @xmath137 + & & & z-2280409 & 1.32 & @xmath138 & @xmath139 & & & + 6 - 1 & 00 52 8.95 & -72 38 3.5 & o-6 - 77228 & 0.58 & @xmath140 & @xmath141 & j00520896 - 7238032 & @xmath142 & @xmath143 + & & & z-2319498 & 0.64 & @xmath144 & @xmath145 & & & + 5 - 12 & 00 52 45.04 & -72 28 43.6 & z-2406014 & 0.35 & @xmath146 & @xmath147 & j00524508 - 7228437 & @xmath148 & @xmath149 + 6 - 4 & 00 52 52.22 & -72 48 29.8 & o-6 - 147662 & 0.27 & @xmath150 & @xmath151 & j00525230 - 7248301 & @xmath152 & @xmath153 + & & & z-2423181 & 0.56 & @xmath154 & @xmath155 & & & + 5 - 2 & 00 53 23.86 & -72 27 15.5 & z-2498173 & 0.24 & @xmath156 & @xmath157 & j00532381 - 7227152 & @xmath158 & @xmath159 + 5 - 16 & 00 53 55.25 & -72 26 45.8 & z-2573354 & 0.83 & @xmath160 & @xmath161 & j00535518 - 7226448 & @xmath162 & @xmath163 + 5 - 4 & 00 54 46.22 & -72 25 23.0 & o-7 - 70843 & 0.79 & @xmath164 & @xmath165 & j00544633 - 7225228 & @xmath166 & @xmath167 + & & & z-2707354 & 0.83 & @xmath168 & @xmath165 & & & + 6 - 2 & 00 54 55.78 & -72 45 10.7 & o-7 - 47103 & 0.40 & @xmath169 & @xmath170 & j00545586 - 7245108 & @xmath171 & @xmath172 + & & & z-2729974 & 0.61 & @xmath173 & @xmath174 & & & + 5 - 7 & 00 54 56.34 & -72 26 48.4 & o-7 - 70829 & 1.19 & @xmath175 & @xmath176 & j00545618 - 7226478 & @xmath177 & @xmath178 + ( xmm-30 ) & & & z-2730786 & 1.21 & @xmath179 & @xmath155 & & & + 3 - 3 & 00 57 36.00 & -72 19 33.9 & o-8 - 49531 & 0.14 & @xmath180 & @xmath181 & j00573602 - 7219341 & @xmath182 & @xmath183 + ( xmm-33 ) & & & z-3103982 & 0.53 & @xmath184 & @xmath185 & & & + xmm-39 & 01 00 30.2 & -72 20 35 & o-9 - 35989 & 2.04 & @xmath186 & @xmath187 & j01003000 - 7220335 & @xmath188 & @xmath189 + & & & z-3499823 & 2.13 & @xmath190 & @xmath191 & & & + xmm-47 & 01 04 35.7 & -72 21 43 & o-10 - 61612 & 5.28 & @xmath192 & @xmath193 & & & + & & & z-4031467 & 4.91 & @xmath194 & @xmath195 & & & + ccccccccc xmm-3 & 6564.57 & 0.06 & 11.32 & 0.14 & -24.40 & 0.48 & @xmath196 ( 1 ) & 48.8 + 4 - 8 & 6565.61 & 0.12 & 11.08 & 0.26 & -15.68 & 0.33 & & + xmm-5 & 6566.06 & 0.03 & 12.55 & 0.08 & -36.03 & 0.24 & & + 4 - 2 ( xmm-7 ) & 6565.49 & 0.03 & 12.13 & 0.08 & -32.41 & 0.40 & @xmath197 ( 1 ) & 91.5 + 4 - 5 ( xmm-9 ) & 6566.09 & 0.08 & 15.51 & 0.21 & -38.47 & 0.70 & & 91.5 + 4 - 1 ( xmm-12 ) & 6566.27 & 0.08 & 14.11 & 0.21 & -28.30 & 0.82 & @xmath198 ( 1 ) & 0.708 + xmm-15 & 6566.12 & 0.04 & 14.83 & 0.10 & -32.04 & 1.07 & -22 ( 2 ) & 88.4 + 4 - 3 ( xmm-14 ) & 6564.99 & 0.05 & 10.53 & 0.11 & -20.37 & 0.35 & & + xmm-17 & 6566.71 & 0.10 & 11.97 & 0.25 & -16.16 & 0.26 & @xmath199 ( 1 ) & 67.0 + 6 - 1 & 6566.94 & 0.04 & 10.56 & 0.09 & -23.87 & 0.38 & & + 5 - 12 & 6565.67 & 0.26 & 10.44 & 0.62 & -5.38 & 0.11 & & + 6 - 4 & 6565.83 & 0.16 & 13.05 & 0.38 & -6.59 & 0.18 & & + 5 - 2 & 6565.92 & 0.09 & 13.26 & 0.24 & -24.02 & 0.63 & & 125.0 + 5 - 16 & 6566.00 & 0.05 & 11.76 & 0.12 & -19.48 & 0.46 & @xmath200 ( 1 ) & + 5 - 4 & 6566.84 & 0.09 & 12.83 & 0.24 & -36.19 & 1.06 & & + 6 - 2 & 6566.14 & 0.03 & 12.40 & 0.08 & -56.00 & 0.95 & & 261.0 + 5 - 7 ( xmm-30 ) & 6565.67 & 0.04 & 10.90 & 0.10 & -19.97 & 0.66 & @xmath201 ( 2 ) & 123.0 + 3 - 3 ( xmm-33 ) & 6565.78 & 0.09 & 11.98 & 0.22 & -32.08 & 0.45 & @xmath202 ( 1 ) & 95.3 + xmm-39 & 6565.74 & 0.03 & 10.33 & 0.07 & -65.02 & 1.44 & & + xmm-47 & 6565.82 & 0.06 & 11.94 & 0.16 & -38.64 & 0.60 & & + cc heii@xmath36 , heii@xmath37 , heii@xmath38 present & earlier than b0 + heii@xmath37 , heii@xmath38 present & b0 + heii@xmath36 and heii@xmath37 absent , heii@xmath38 weak & b0.5 + heii@xmath38 absent , siiv@xmath39 present & b1 + siiv@xmath52 absent , siiii@xmath42 appear & b1.5 + oii+ciii@xmath40 blend decreases rapidly & later than b1.5 + mgii@xmath41 @xmath9 siiii@xmath42 & b2 + mgii@xmath41 @xmath2 siiii@xmath42 & b2.5 + mgii@xmath41 @xmath203 siiii@xmath42 & b3 + oii+ciii@xmath40
blend disappears & later than b3 + siiii@xmath42 absent & b5 + hei@xmath45 @xmath9 siii@xmath44 @xmath9 hei@xmath80 & b8 + hei@xmath46 @xmath9 mgii@xmath41 & b9 ccc xmm-3 & b0.5e & b1 - 1.5
ve ( 1 ) + 4 - 8 & b1.5e & later than b3e ( 2 ) , be ?
( 3 ) + xmm-5 & b1.5e & b0 - 5 ( ii ) = [ 2df]690 ( 4 ) , be ? ( 5 ) + 4 - 2 ( xmm-7 ) & b1e & b1 - 3 iv - ve ( 1 ) + 4 - 5 ( xmm-9 ) & b1e & be ? ( 6 ) + 4 - 1 ( xmm-12 ) & b0.5e & b0 - 0.5 ve ( 7 ) + xmm-15 & b0.5e & b0.5 iii - ve ( 1 ) + 4 - 3 ( xmm-14 ) & b0.5e & be ? ( 6 ) + xmm-17 & b0e & be ( ( 8) , peculiar ( 9 ) ) + & & o9.5-b0 ve ( 1 ) + 6 - 1 & b1 - 3e & b1 - 5 ( ii)e = [ 2df]5054 ( 4 ) , be ( 8) + & & b1 - 3
iii - ve ( 1 ) + 5 - 12 & o9-b0e & be ? ( 3 ) + 6 - 4 & o2((f))+obe & be ( 5 ) , b type with balmer emission ( 10 ) + 5 - 2 & b0.5e & b1 - 2 iv - ve ( 1 ) + 5 - 16 & b0.5e & o9.5-b1 iv - ve ( 1 ) + 5 - 4 & b1e & be ? ( 3 ) + 6 - 2 & b1e & b1 iii - ve ( 1 ) + 5 - 7 ( xmm-30 ) & b0e & o9 ve ( 1 ) + 3 - 3 ( xmm-33 ) & b0 - 4e & b0 - 2 iv - ve ( 1 ) + xmm-39 & b1.5e & b0 - 5 ( ii ) = [ 2df]1475 ( 4 ) , be ?
( 6 ) + xmm-47 & b1.5e & b0.5 ( iv)e = [ 2df]1905 ( 4 ) , be ? ( 5 ) + ccccccc smc be - xrbs & & 0.92 & 0.92 & 1.00 & 1.00 & 0.31 + lmc be - xrbs & 0.92 & & 1.00 & 0.66 & 0.82 & 0.08 + mw be - xrbs & 0.92 & 1.00 & & 0.66 & 0.82 & 0.08 + smc be stars & 1.00 & 0.66 & 0.66 & & 1.00 & 0.53 + lmc be stars & 1.00 & 0.82 & 0.82 & 1.00 & & 0.66 + mw be stars & 0.31 & 0.08 & 0.08 & 0.53 & 0.66 & + ccccccc xmm-3 & b0.5e & -0.01 & 0.04 & b0.5 & -0.28 & 0.27 + 4 - 8 & b1.5e & 0.17 & 0.06 & b1.5 & -0.26 & 0.43 + xmm-5 & b1.5e & -0.09 & 0.04 & b1.5 & -0.26 & 0.17 + 4 - 2 ( xmm-7 ) & b1 & 0.10 & 0.05 & b1 & -0.27 & 0.37 + 4 - 5 ( xmm-9 ) & b1e & 0.11 & 0.05 & b1 & -0.27 & 0.38 + 4 - 1 ( xmm-12 ) & b0.5e & -0.20 & 0.05 & b0.5 & -0.28 & 0.08 + xmm-15 & b0.5e & -0.17 & 0.06 & b0.5 & -0.28 & 0.11 + 4 - 3 ( xmm-14 ) & b0.5e & -0.01 & 0.06 & b0.5 & -0.28 & 0.27 + xmm-17 & b0e & -0.16 & 0.06 & b0 & -0.28 & 0.12 + 6 - 1 & b1 - 3e & -0.17 & 0.03 & b2 & -0.25 & 0.08 + 5 - 12 & o9e - b0e & -0.09 & 0.09 & o9.5 & -0.28 & 0.19 + 5 - 2 & b0.5e & -0.18 & 0.12 & b0.5 & -0.28 & 0.10 + 5 - 16 & b0.5e & -0.16 & 0.03 & b0.5 & -0.28 & 0.12 + 5 - 4 & b1e & 0.05 & 0.06 & b1 & -0.27 & 0.32 + 6 - 2 & b1e & -0.12 & 0.04 & b1 & -0.27 & 0.15 + 5 - 7 ( xmm-30 ) & b0e & -0.14 & 0.04 & b0 & -0.28 & 0.14 + 3 - 3 ( xmm-33 ) & b0 - 4e & -0.08 & 0.17 & b2 & -0.25 & 0.17 + xmm-39 & b1.5e & -0.15 & 0.14 & b1.5 & -0.26 & 0.11 + xmm-47 & b1.5e & -0.04 & 0.03 & b1.5 & -0.26 & 0.22 + | we present a large sample ( 20 in total ) of optical spectra of small magellanic cloud ( smc ) high - mass x - ray binaries obtained with the 2df spectrograph at the anglo - australian telescope .
all of these sources are found to be be / x - ray binaries ( be - xrbs ) , while for 5 sources we present original classifications .
several statistical tests on this expanded sample support previous findings for similar spectral - type distributions of be - xrbs and be field stars in the smc , and of be - xrbs in the large magellanic cloud and the milky way , although this could be the result of small samples . on the other hand , we find that be - xrbs follow a different distribution than be stars in the galaxy , also in agreement with previous studies .
in addition , we find similar be spectral type distributions between the magellanic clouds samples .
these results reinforce the relation between the orbital period and the equivalent width of the line that holds for be - xrbs .
smc be stars have larger equivalent widths when compared to be - xrbs , supporting the notion of circumstellar disk truncation by the compact object . |
astrometric and radial - velocity ( rv ) planet detections are , from a mathematical standpoint , extremely similar . in each ,
one models 1-dimensional projections of kepler orbits and attempts to fit kepler parameters . in the limit of circular orbits ,
the planet signature is encoded in the simple form @xmath4 where @xmath5 are the astrometric and velocity semi - amplitudes for the two cases , and where @xmath6 designate the period @xmath7 and phase @xmath8 , respectively .
one then goes on to combine information about the star s mass and ( for astrometry ) its distance , to infer the planet mass @xmath9 ( astrometry ) or @xmath10 ( rv ) , where @xmath11 is the inclination . in the astrometric case ,
there are of course two such equations , one for each direction in the plane of the sky , whose ratio gives @xmath12 , and so permit one to break the @xmath10 degeneracy that plagues the intrinsically 1-dimensional rv measurement .
because the form of equation ( [ eqn : basicform ] ) is essentially identical in the two cases , it is generally assumed that the error properties , i.e. , the relation between the measurement errors and the derived - parameter errors , is also the same .
of course , it is well known that the parameter errors have a different dependence on semimajor axis @xmath13 , system distance @xmath14 , etc .
most notably , with other parameters held fixed , astrometric sensitivity increases linearly with @xmath13 whereas rv sensitivity declines as @xmath15 .
but here i am referring to something else .
the differences just mentioned all impact the final result because they change the characteristic signal - to - noise ratio of the experiment , @xmath16 where @xmath17 is the number of measurements and @xmath18 is the error in each measurement ( assumed for simplicity to be all the same ) . here
i will show that the astrometric and rv measurements have substantially different error properties even when snr is identical .
i will work within the framework of the minimum variance bound ( mvb ) , also frequently called the fisher - matrix approximation . as the same approximation will be applied to both techniques
, this will allow me to highlight the difference between them .
of course , if equation ( [ eqn : basicform ] ) really did fully represent both techniques , there could not be any difference in their error properties .
however , the true functional forms of the source motions are actually @xmath19 for simplicity , i will continue to assume circular orbits , and i will assume for the astrometric case that the orbit is seen edge - on .
in fact , the latter is not much of an approximation since in most cases the great majority of the information about the mass and period comes from the major axis of the apparent astrometric ellipse .
finally , i will assume that the data are taken uniformly in time , at intervals that are frequent compared to the orbital period , and that they all have the same error , @xmath18 . for a _ single - planet _
i.e. , a system composed of just a star and a planet , in which the star does not suffer any other accelerations , @xmath20 for the rv case , but @xmath21 for the astrometric case . for rv , @xmath22 is the systemic radial velocity of the system . for astrometry
, @xmath23 is the systemic proper motion of the system , while @xmath22 is the zero point of that motion .
the introduction of the nuisances parameters @xmath22 or @xmath24 is what distinguishes the two cases statistically .
the difference appears only for @xmath25 or @xmath26 , where @xmath27 is the duration of the experiment .
i have implicitly assumed that @xmath28 , and this is the reason that i ignore parallax , which would be an additional nuisance parameter for astrometry , but not for rv . in the fisher - matrix approximation ,
the inverse covariance matrix of the parameters , @xmath29 , is given by @xmath30 where @xmath31 one easily finds that for @xmath32 , @xmath33 and @xmath34 are completely uncorrelated from the other parameters , so that @xmath35 that is , @xmath36 , which is the fractional error in astrometric or rv amplitude depends only on snr , and in a very simple way . similarly , @xmath37 however , as @xmath7 approaches @xmath27 , the parameters of primary interest ( mass and period ) become correlated with the nuisance parameters .
to calculate this effect , i employ equations ( [ eqn : bij ] ) and ( [ eqn : fi ] ) , but using the full 7-parameter kepler formalism , not just the three kepler parameters displayed in equations ( [ eqn : basicform ] ) and ( [ eqn : trueform ] ) .
that is , even though the adopted orbits are circular and edge - on , i allow for free fits to the eccentricity and inclination , and so for correlations between these parameters ( as well as the two remaining kepler parameters ) and the parameters of interest ( amplitude and period ) .
the effect of allowing for these covariances is to increase the errors in semi - amplitude and period by modest amounts relative to what would be obtained using equation ( [ eqn : trueform ] ) .
panels ( a ) and ( b ) of figure [ fig : apm ] show the ratios of the true errors for the amplitude and period , respectively , relative to the naive equations ( [ eqn : massfrac ] ) and ( [ eqn : periodfrac ] ) , for both the rv ( _ green _ ) and astrometric ( _ red _ ) cases .
in fact , as i will discuss in
[ sec : phase ] , for @xmath38 , the errors depends on phase as well as period .
figure [ fig : apm ] therefore shows the root - mean - square of the errors , averaged over all phases .
note that the rv amplitude errors follow the naive form until @xmath39 and then deteriorate relatively gracefully .
by contrast , the astrometric errors begin deviating at @xmath40 and then deteriorate much more quickly . for the period errors ,
deterioration begins at @xmath41 for rv and @xmath42 for astrometry , but the overall pattern is qualitatively similar . the mass estimates for astrometry and rv depend on different combinations of amplitude and period , @xmath43 hence the fractional error in the mass ( or @xmath10 in the case of rv ) is related to the errors in the fit parameters by @xmath44 ^ 2 = { c_{11}\over a_1 ^ 2 } + 2n{c_{12}\over a_1 a_2 } + n^2{c_{22}\over a_2 ^ 2}. \label{eqn : masscovar}\ ] ] in the limit @xmath32 , @xmath45 , but for @xmath38 , the period error and the correlations become important .
figure [ fig : apm]c shows the results of calculations that apply equation ( [ eqn : masscovar ] ) .
of course , the star may have more than one companion ( planetary or otherwise ) , and one may imagine arbitrarily complicated configurations . here
i restrict myself to the next level of complication , a second companion that is sufficiently far away that its effect on the star may be treated as uniform acceleration . even if no such acceleration is identified
, one might decide to fit for it on the grounds that there _ may _ be such a companion that has not been recognized .
i now ask how the inclusion of such an additional nuisance parameter affects the precision of the physical parameters for the planet that was previously treated as isolated .
the number of nuisance parameters is incremented by one in each case . referring to equation ( [ eqn : trueform ] ) , for rv , we now have @xmath21 , with @xmath23 being the systemic acceleration , while for astrometry , @xmath46 , with @xmath47 being the systemic proper - motion acceleration . figures [ fig : apm]a and [ fig : apm]b show the corresponding ratios relative to the naive equations ( [ eqn : massfrac ] ) and ( [ eqn : periodfrac ] ) . in this case
, rv is represented by the _ red curve _ , while astrometry is represented by the _ blue curve_. the fractional mass errors are shown by _
solid curves _ in figure [ fig : apm]c . in this case , both deteriorate rapidly , but the mass - error deterioration begins at @xmath42 for astrometry and at @xmath40 for rv
. one may anticipate that in other , yet more complicated situations , rv will always perform like the naive equations to higher @xmath48 than astrometry , simply because it has one fewer nuisance parameter .
figure [ fig : apm ] shows the root - mean - square errors averaged over phase . in fact , as @xmath48 grows and the rms errors deteriorate , the variations with phase also increase .
figure [ fig : phase ] shows the phase variations of the mass errors for the particular case that the period is exactly equal to the duration of the experiment , @xmath49 .
these large variations imply that figure [ fig : apm ] can not be used to estimate the errors in any particular case ( except if the rms ratio is close to unity ) .
rather , it should be used as a general guide to the reliability of the experiment .
any individual planet detection must be analyzed based on the actual data and the parameters derived .
the results derived here are primarily of interest in regard to future astrometric missions such as gaia and sim .
the first point is that planets with periods that are even slightly longer than the mission can not be reliably detected unless they are many times more massive than the nominal thresholds of detection . and
second , if one is forced to allow for uniform acceleration , then the same statement applies to @xmath50 .
hence , to preserve the discovery space at the longest accessible periods , it is critically important to identify or rule out distant perturbers by supplementary data , such as longer - term rv observations to look for large , distant planetary or stellar companions , or perhaps ao observations to find stellar companions .
this should be relatively straightforward for sim , with its limited number of high - precision targets , but may be more difficult for gaia , which is a survey instrument . | astrometric and radial - velocity planet detections track very similar motions , and one generally expects that the statistical properties of the detections would also be similar after they are scaled to the signal - to - noise ratio of the underlying observations .
i show that this expectation is realized for periods small compared to the duration of the experiment @xmath0 , but not when @xmath1 . at longer periods
, the fact that models of astrometric observations must take account of an extra nuisance parameter causes the mass error to begin deteriorating at @xmath2 , as compared to @xmath3 for rv .
moreover , the deterioration is much less graceful .
this qualitative difference carries over to the more complicated case in which the planet is being monitored in the presence of a distant companion that generates an approximately uniform acceleration .
the period errors begin deteriorating somewhat earlier in all cases , but the situation is qualitatively similar to that of the mass errors .
these results imply that to preserve astrometric discovery space at the longest accessible orbits ( which nominally have the lowest - mass sensitivity ) requires supplementary observations to identify or rule out distant companions that could contribute quasi - uniform acceleration . |
the basic scheme for unification of low - luminosity radio - loud active galactic nuclei ( agn ) assume that radio galaxies of the fanaroff - riley class i ( fr i , fanaroff & riley 1974 ) and bl lac objects differ only because of a different viewing angle to the relativistic jet ( urry & padovani 1995 for a review ) . in bl lacs
, the jet is seen at small angles to the line - of - sight , and the non - thermal emission produced at its very innermost base , on the scale of @xmath2 cm , dominates the observed radiation . in fr
i radio galaxies the jet points at larger angles , and the `` bl lac '' component should be beamed away from the observer .
clearly , the most direct implication of the unified models is that the extended properties ( e.g. radio and optical emission , environment ) of the beamed sources and their putative parent population should appear the same .
therefore , in the absence of any direct detection of the nuclear emission in fr i , the usual way these models are tested involves the comparison of such extended components ( e.g urry et al .
2000 , antonucci & ulvestad 1985 , kollgaard et al .
1996 ) .
however , faint unresolved optical nuclei have been discovered in fr i radio galaxies through hst observations of a complete sample of 3cr sources @xcite .
the presence of a strong correlation between these optical nuclei and the radio cores argues for a common non - thermal synchrotron origin of both components .
therefore , this discovery provides qualitative support for the low - luminosity radio loud agn unification models .
a more quantitative analysis showed that cores of radio galaxies are overluminous with respect to what is expected from mis - oriented bl lac jets , for typical values of the bulk lorentz factor @xcite .
it is therefore plausible that the optical cores of radio galaxies and the radiation observed in bl lacs are not produced in the same region of the relativistic jets : in bl lacs , a `` fast jet spine '' with lorentz factors of @xmath3 dominates , while in fr i we might be observing emission from a slower ( but still relativistic ) shear layer .
modeling of the overall nuclear sed could provide an even more stringent test for the unified scheme , as it represents a fundamental tool for obtaining information on the physical conditions of the emitting region , as it is usually done for blazars ( e.g. ghisellini et al .
1998 ) . unfortunately , the nuclear sed of radio - galaxies are in general not sufficiently well sampled to follow directly this approach and radio galaxies are generally not detected in the gamma - ray band . the first radio galaxy ( and , to the best of our knowledge , the only case to date ) for which a detailed modeling of the overall nuclear sed , from the radio to the gamma ray - band , has been undertaken is centaurus a , the closest radio galaxy @xcite .
the result of such work is that , when considered in the comoving frame of the source , the nuclear sed of cen a can be reproduced adopting the same emission model and physical parameters typical of bl lacs of the same total power , strongly supporting the unification scheme .
in this paper we focus on ngc 6251 , which is a well studied fr i radio galaxy with a well collimated one sided radio jet and a large scale emission .
its total radio power is @xmath4 erg s@xmath5 ( waggett et al .
1977 ) , i.e. close to the fr i / fr ii transition .
however , its radio morphology is well representative of the fr i class .
the projected linear size of the extended radio structure is 1.7 mpc ( for @xmath6 km s@xmath5 mpc@xmath5 ) , which makes ngc 6251 one of the largest known radio galaxies @xcite .
the radio source is associated with a giant elliptical galaxy , which hosts a @xmath7 black hole @xcite . by taking advantage of the recent discovery that the egret gamma - ray source j1621 + 8203
is associated with ngc 6251 @xcite , we perform a detailed analysis of its overall sed as we did for centaurus a. our result is that the nucleus of this low - power radio galaxy can also be interpreted as a misoriented bl lac .
the paper is organized as follows : in sect .
2 we describe the hst , _ beppo_sax and _ chandra _ data and the analysis procedures . in sect .
3 we determine the sed ; in sect .
4 we discuss the nature of the ir - to - uv nuclear source , which is crucial in order to model the overall sed . in sect .
5 we model the sed in the frame of a synchrotron self - compton scenario , and we discuss the resulting model parameters , while in sect .
6 we draw conclusions . a redshift of @xmath8 for ngc 6251 ( ned ) is adopted throughout this paper .
[ gallery ] in order to determine the sed , we first focus on the analysis of the ir - to - uv band and x - ray band observations .
these spectral regions are crucial in order to model the observed radiation , since the data allow us to set the position of the emission peaks , i.e. the frequencies at which most of the energy is released by the source , and thus constrain the model .
we have measured the flux of the central compact source in archival hst images , from the ir to the uv . in table
[ hstobs ] we report the details of the observations , and in fig .
[ gallery ] we show a picture gallery of the nuclear regions of ngc 6251 as they appear in the images .
the f205w and f555w filters might have contribution from emission lines which is however expected to be small ( a few % ) , due to their small equivalent width compared with the width of the filters passband .
the other filters are centered in spectral regions relatively free of strong emission lines .
therefore ( with the possible exception of f205w and f555w ) the flux we measure is continuum emission .
the procedure is as follows .
we perform aperture photometry , evaluating the total counts in a circular region centered on the point source , with different radii for the different filters .
this is important in order to have a reliable determination of the total flux for the point source allowing for slightly different shapes of the point spread function . for the foc uv data both the determination of the background and the choice of the aperture radius is straightforward ,
since the nucleus is by far the dominant component in the image . for the wfpc2 images , due to the presence of the host galaxy stellar component and an extended dusty disk which surrounds the unresolved nucleus ,
the largest source of error is the determination of the background .
thus we have adopted the rather well tested procedure of measuring the flux in a radius of 5 pixels ( corresponding to @xmath9 arcsec ) , measuring the background in the surrounding annulus ( 1 pixel width ) .
this is essentially the same procedure as we used for the analysis of the complete samples of 3cr radio galaxies @xcite . due to the different shape of the psf in the different filters , the lower resolution at ir wavelengths and the larger pixel size of the detector , the nicmos data have to be corrected for psf loss .
we have produced synthetic psfs with the tinytim software @xcite for the different instrument configurations .
we compared the observed shape of the nuclear psf in ngc 6251 with the respective synthetic psf and we selected an appropriate aperture radius for the photometry ( @xmath10 pixels , @xmath11 pixels , @xmath12 pixels for the f205w , f160w and f110w filters , respectively ) .
this procedure ensures that we measure the background as close as possible to the nucleus without suffering contamination from the nucleus itself .
however , we have determined that in this case the aperture photometry radius includes only a fraction of the total counts .
therefore , in order to estimate the total flux of the point source , we have derived the renormalization factors from the analysis of the synthetic psfs ( @xmath13 , @xmath14 , @xmath15 for the f205w , f160w and f110w filters , respectively ) and we have corrected the measured flux accordingly . the effective frequencies of the observations have been calculated using synphot , by taking into consideration the observed slope of the continuum emission .
the count rates have been multiplied by the value of the header parameter photflam , which represents the inverse sensitivity on the filters in the case of a flat continuum ( @xmath16 , @xmath17 , which corresponds to @xmath18 , @xmath19 ) .
we have tested that the observed slope does not affect the values of photflam significantly ( less than @xmath20 ) .
the fluxes measured in the hst observations are reported in table [ hstphot ] .
we estimate that the typical error on the nuclear flux measurements is .5ex@xmath21 .
the ir - to - uv slope is @xmath22 , calculated between 16000 and 2200 .
l l l c c nicmos & nic2 & f205w & go7862 & 1998 jul 06 + nicmos & nic2 & f160w & go7862 & 1998 jul 06 + nicmos & nic2 & f110w & go7862 & 1998 jul 06 + wfpc2 & pc & f814w & go6276 & 1995 jun 28 + wfpc2 & pc & f555w & go6276 & 1995 jun 28 + wfpc2 & pc & f547 m & go6653 & 1996 sep 13 + foc & f/96 & f342w & go6246 & 1996 feb 19 + foc & f/96 & f220w & go6891 & 1997 jul 07 + [ hstobs ] l c c f205w & 21063 & @xmath23 + f160w & 16250 & @xmath24 + f110w & 11777 & @xmath25 + f814w & 8124 & @xmath26 + f555w & 5558 & @xmath27 + f547 m & 5501 & @xmath28 + f342w & 3435 & @xmath29 + f220w & 2414 & @xmath26 + [ hstphot ] we have analyzed unpublished recent _ beppo_sax and _ chandra _ x - ray data .
these observations are available in the respective public archives .
we focus in particular on the _
beppo_sax observations , since ngc 6251 is known to be a bright x - ray source ( 2 - 10 kev flux of @xmath30 ; sambruna , et al .
1999 ) and therefore the _ chandra _ data are significantly affected by pileup .
however , it is extremely useful to compare the two observations . given its arcsec spatial resolution , which is @xmath31 times sharper than that of bepposax , _ chandra _ is indeed the ideal experiment to pinpoint the nuclear emission . in sect .
[ chandra ] we show that the two observations have consistent results , once that the pileup effects in _ chandra _ data are taken into account .
ngc 6251 was observed in july 2001 by the lecs , mecs and pds instruments on board bepposax ( with exposure times of 30.2 , 80.9 and 36.2 ksec , respectively ) , as part of an ao5 go observation ( code 51322001 , p.i .
m. guainazzi ) .
the lecs and mecs proportional counters are operating in the 0.1 - 10 kev and 1 - 10 kev bands , respectively , while the pds instrument is a scintillation detector operating in the high energy band 15 - 300 kev . in our spectral analysis
we consider only the range of energies in which the instruments are best calibrated ( i.e. 0.12 - 4 kev for the lecs , 1.6 - 10 kev for the mecs and 15 - 220 kev for the pds ) .
we perform data analysis by using standard procedures .
we extract the source spectrum from an aperture of 8 and 4 arcmin radius in the lecs and mecs images , respectively , and we evaluate the background from blank sky fields .
the pds data reduction is as performed by the archive pipeline .
we perform the spectral analysis of mecs and lecs data by using xspec v11.0 on rebinned spectra with at least 20 counts per bin to use the @xmath32-statistics in the fitting process .
errors are given at the 90% confidence level for one interesting parameter .
the cross - calibration constant between lecs and mecs , obtained by fitting simultaneously the lecs and mecs data in the common band 1.6 - 4.0 kev , is found to be 0.67 .
the cross - calibration constant between the mecs and the pds , constrained to be in the fiducial range 0.77 - 0.93 , is found to be 0.88 . given the pds large field of view ( 78 arcmin hwzi ) and lack of spatial resolution , other sources of high energy photons might in principle contaminate the pds data .
we search for x - ray sources in the mecs image ( 25 arcmin radius fov ) and , at larger offsets , in the rosat source catalogs . based on the fluxes and spectra of the x - ray sources within the pds fov
, we estimate that more than @xmath33 of the pds photons are produced by ngc 6251 .
the 0.1200 kev spectrum is well described ( @xmath34 ) by a power law with a photon index @xmath35 mildly absorbed by a column density @xmath36 @xmath37 in excess of the galactic value ( @xmath38 @xmath37 ; dickey & lockman 1990 ) .
similar results have also been found by guainazzi et al .
since the nuclear source is embedded in a diffuse gas halo ( e.g. birkinshaw & worrall 1993 ) , we add in the spectral fit a thermal component using the mekal model of xspec . in this case
, we obtain a comparably good fit ( @xmath39 ) , although the estimate of the nuclear absorption rises to @xmath40 @xmath37 , while the power - law spectral index remains unchanged ( @xmath41 ) .
the thermal component is found to have a temperature @xmath42 kev , with a metal abundance @xmath43 .
the latter estimate of the nuclear absorption , the power
law spectral index , as well as the halo temperature and metal abundance , are in good agreement with the 1994 asca observation @xcite @xmath37 , @xmath44 , @xmath45 kev , @xmath46 , while a somewhat steeper powerlaw and lower absorption have been found by turner et al .
( 1997 ) in the same asca dataset ] .
however , we point out that the 2 - 10 kev _ beppo_sax flux ( @xmath47 ) is significantly higher than the flux measured in the asca observation ( @xmath48 ) , suggesting that the source flux is variable in the x - rays .
ngc 6251 was observed by chandra ( 37.4 ksec with acis - i ) in september 2000 , as part of an ao1 go project ( obs .
i d 847 , p.i .
the source is 3 arcmin away from the telescope aimpoint and falls in the gap among the 4 acis - i ccds , where the exposure map is low .
this reduces the number of detected photons , but does not mitigate the pileup effects .
based on the bepposax spectrum and flux , we estimated that in the chandra observation ( with a frametime of 3.24 sec ) the pileup fraction is above 60% ( the pileup fraction decreases to @xmath49 if the asca flux is used ) .
we used ciao v2.2 to perform a standard data reduction and xspec v11.0 for the spectral analysis .
as expected , due to the high pileup fraction , the acis - i spectrum of ngc 6251 is much harder than that observed by bepposax , having a photon index @xmath50 . by using the marx 4.0 simulator we verified that both the spectral slope and the flux are in excellent agreement with the chandra observation , under the assumption that the input spectrum is that observed by bepposax .
we point out that it is appropriate to compare such observations , since the _
beppo_sax spectrum was obtained much closer in time to the chandra data than the asca observation .
in fig . [ 6251sed ] we show the overall nuclear sed of ngc 6251 .
the radio data are taken from @xcite .
the vlba fluxes of the unresolved radio core are a factor of 2.5 lower than the vla both at 6 cm and 18 cm .
the ir - through - uv hst data and the _ beppo_sax x ray observations and data analysis have been discussed in sect.[hstdata ] and section [ xrays ] , respectively .
the egret spectrum is the fit to the spectrum of the source , as taken from @xcite .
the sed in a @xmath51 representation has two broad peaks and , although the peak frequencies are not well constrained , they are clearly located in the far ir - mm region and between the x - rays and gamma - rays . as in the case of centaurus a @xcite
, the sed appears remarkably similar to the typical `` blazar like '' sed . in such objects ,
the low and high energy peaks are believed to be produced by non - thermal synchrotron and inverse - compton radiation , respectively . in the following we discuss in detail the origin of the optical emission and
we show that the ssc emission model can also account for the nuclear radiation of ngc 6251 .
there is growing evidence that the nuclei of fr i radio galaxies are basically unobscured , and the compact nuclei observed in the majority of fr i are non - thermal radiation produced at the base of the relativistic jet .
the main evidence for this is the presence of a strong linear correlation between fr i s optical and the radio nuclear radiation , the latter being definitely of non - thermal origin @xcite .
although the general behavior of the fr i class as a whole seems to be understood , for `` single sources '' a certain ambiguity on the nature of the optical nuclei might still be present .
since the ir through uv radiation is crucial for constraining the frequency of the synchrotron peak , it is very important to assess its nature . on the other hand , for the hard x
ray and gamma ray emission the interpretation as inverse - compton emission appears to be the only viable possibility .
the origin of the bright optical core of ngc 6251 has been discussed by @xcite .
there are two scenarios proposed by these authors : a non thermal process ( most plausibly synchrotron emission ) , or thermal radiation from an optically thick accretion disk .
let us consider the position of ngc 6251 in diagnostic diagrams for the nuclei of radio galaxies . in fig .
[ lum6251 ] we plot ngc 6251 in the plane formed by the luminosity of the optical nucleus of versus the radio core .
the open circles are the nuclei of fr i radio galaxies from the 3cr catalog , as taken from @xcite .
the dashed line is the regression line between the two quantities .
the black point , representative of ngc 6251 , lies just above the correlation .
this is a strong clue that ngc 6251 behaves exactly as other fr i. we stress that the comparison with the 3cr sample is extremely appropriate , since @xcite , have included ngc 6251 in their complete revised 3cr sample ( the 3crr ) . in fig .
[ diagnostics]a we show the position of the nucleus of ngc 6251 in the plane formed by the radio - to - uv vs. the optical - to - uv flux ratios .
as discussed in @xcite , the position of the objects in this plane is sensitive both to absorption and to the different intrinsic spectral properties of the nuclei .
steep spectrum radio quasars ( plotted as stars ) and broad lined radio galaxies ( empty circles ) occupy the same region of the plane , and clearly differentiate from fr i mainly because of the presence of the thermal blue bump in the optical - to - uv spectral range .
ngc 6251 again lies in the region of fr i. however , as already pointed out above , for `` single objects '' we can not exclude on the basis of the radio - optical properties that the optical emission could be scattered light produced by an accretion disk obscured to our line - of - sight . in this case
, we know from @xcite that its representative point would lie on the radio - optical correlation of fig .
[ lum6251 ] _ by chance_. this has been indeed found to be rather common among fr ii in which strong narrow emission lines are observed . in order to remove the degeneracy in the interpretation of the optical emission
, we use the [ oiii]5007 emission line , which is indeed a good indicator of the intrinsic ionizing radiation field . from ground based observations ( @xmath52 aperture , centered on the galaxy nucleus ) the luminosity of the [ oiii]5007 line is @xmath53 erg s@xmath5 @xcite . a useful
diagnostic is the ratio between the flux of the emission line and the observed nuclear continuum emission ( i.e. the `` nuclear '' equivalent width ) .
this has been used to discriminate between objects in which the nuclear source is seen directly and objects in which the ionizing radiation field is obscured to our line - of - sight and observed only through scattered radiation @xcite . in the latter case ,
high values of the equivalent width are expected , since only a few percent of the intrinsic nuclear source is observed . in fig .
[ diagnostics]b we show the position of the source in the plane formed by the [ oiii ] ew plotted against the ratio of the optical nuclear flux to radio core flux @xmath54 .
the shaded area corresponds to the dispersion ( @xmath55 ) of the correlation between these two quantities for fr i galaxies .
the equivalent width of the [ oiii]5007 emission line for ngc 6251 is typical of the other unobscured fr i nuclei .
as already known from fig [ lum6251 ] , the value of @xmath54 is also in the same region as the other fr i. in the light of the above analysis , we conclude that the dominant physical origin for the optical nucleus is non thermal synchrotron emission produced at the base of the jet .
any other thermal contribution to the observed flux should then be considered lower than ( or equal to ) the observed synchrotron radiation . before modeling the nuclear emission
, we must briefly consider the possible presence of absorption on the line - of - sight to the nucleus , since one of the two models for the x - ray emission includes a certain amount of @xmath56 in excess of the galactic value . absorption would indeed naturally steepen the ir - to - uv slope of the continuum emission , and could in principle `` artificially '' produce a two - peak sed .
this is an issue for objects like cen a , because of the presence of the extended dust lane which certainly absorbs the nuclear radiation @xcite .
however , in ngc 6251 there is no apparent reason for absorption in the optical .
firstly , the dusty disk does not appear to hide the nucleus .
although the extinction from the disk is only @xmath57mag ( in the external regions ) , this would significantly affect the uv emission ( @xmath58 , cardelli et al .
if the disk were to cover the nucleus we would expect to see a progressive steepening of the spectral slope as frequency increases , which is not observed .
furthermore , the typical optical - to - uv spectral slope of fr i from the 3cr catalog lies somewhere between @xmath59 @xcite , while only those objects in which absorption features are apparent have steeper slopes .
the value of @xmath60 that we derive for ngc 6251 appears to be consistent with other _ unobscured _ fr is . in fig .
[ uvtrace ] we show the effects of a small amount of optical extinction on the broad band spectral properties of ngc 6251 when compared with other radio galaxies .
we have overplotted the de - reddening line for ngc 6251 on the plane of fig .
[ diagnostics]a , assuming a range of @xmath61 from 0 to 7 .
this confirms that only an @xmath62 mag is allowed . for higher values of extinction , the broad band spectral properties of ngc 6251 would be incompatible with any of the other sources in the same range of redshift and total power .
in particular , it appears that it is not possible to `` drag '' the source into the region occupied by quasars if we only consider the effects of obscuration .
note that since the typical error for each single flux measurement is @xmath63 , the resulting error on the data plotted here is of the order of the size of the symbols .
therefore , we exclude that the nuclear emission is thermal radiation from an accretion disk seen directly . similarly to fig .
[ uvtrace ] we have overplotted the de - reddening curve on the ew([oiii ] ) vs. @xmath54 diagnostic plane ( fig .
[ lo3 ] ) .
an @xmath64 mag would again move the point representative of ngc 6251 in a region of the plane where no sources are found .
the fact that the x - ray data can be modeled with absorption in excess of the galactic value , while in the optical there is no evidence for significant extinction is not surprising .
in fact , the standard conversion between @xmath56 and @xmath61 , adopting local galactic dust to gas ratio ( which in our case would imply @xmath65 mag ) generally does not hold in agns . @xcite
have found that in agns @xmath66 is always lower than the galactic value by a factor of 3 to 100 .
several different explanations for such a discrepancy have been proposed by various authors , e.g. non standard values for the dust - to - gas ratio in the `` absorbing '' structures , a different location for the optical and x - ray sources and/or for the absorbing structures , or the presence of large dust grains which would flatten the extinction curve and significantly lower the @xmath67 ratio @xcite . as noted above , from the analysis of the x - ray data we can not discriminate between the two models that fit the data .
however , for our purpose , the most important parameter we derive from the x - ray data is the slope of the ( non - thermal ) power law component , which remains unchanged in either models . in the following
we discuss the nature of the optical emission through various diagnostics .
then we proceed to model the overall sed in the synchrotron self - compton scenario , with the same method as usually done for blazars .
l l @xmath68 cm & @xmath69 g + @xmath70 erg s@xmath5 & @xmath71 + @xmath72 & @xmath73 + @xmath74 & @xmath75 + [ modpar ] we model the sed of ngc 6251 in the frame of a homogeneous synchrotron self - compton scenario ( ssc ) , which is successfully used to account for the overall emission of blazars ( e.g. ghisellini et al .
details of the model are given in @xcite and a short summary is reported for clarity in appendix [ appendice ] .
the model is particularly appropriate for low - power objects ( bl lacs ) , in which the external radiation field is probably low , and the dominant source of seed photons for inverse - compton scattering is the synchrotron radiation field itself . since the total radio power of ngc 6251 , ( @xmath4 erg s@xmath5 , waggett et al .
1977 ) is in the range spanned by bl lacs , and close to its median value ( see e.g. perlman et al . 1996 , kollgaard et al .
1996 , cassaro et al . 1999 ) , the use of the ssc model appears to be appropriate . in table
[ modpar ] we show the parameters of the model which best reproduce the sed .
all of them are typical of bl lacs , except for the value of @xmath76 , the relativistic beaming factor is defined as @xmath77 , where @xmath78 is the bulk lorentz factor of the jet and @xmath79 is the viewing angle . ] . in order to perform a more detailed analysis ,
we compare our model parameters with the large bl lac sample modeled by @xcite . we use for comparison the compactness parameter @xmath80 , which represents the ratio of the injected power to the size of the source . for ngc 6251
we obtain @xmath81 , which lies at the higher end of the distribution of @xmath82 for bl lacs .
both the magnetic field @xmath83 and the size of the source are well inside the range spanned by bl lacs ( @xmath84 g and @xmath85 cm ) . as already noted above , @xmath76 is lower than in bl lacs , where @xmath86 and the median value is 16 . in the analysis of the sed of centaurus
a , we obtained an even lower value , @xmath87 @xcite . in that case
, we showed that this could be explained by the presence of a low velocity component in the jet , e.g. a shear layer , which could be the origin of the observed nuclear radiation .
this has been interpreted as a further evidence that relativistic jets have at least two velocity components : a fast `` spine '' , with plasma lorentz factors @xmath88 which dominates the bl lac emission , and a slower , mildly relativistic layer , which accounts for the emission of the nuclei of fr i radio galaxies .
considering all possible intrinsic jet velocities , the maximum viewing angle for @xmath89 is @xmath90 obtained for @xmath91 . in ngc 6251 , @xmath92 is still compatible with `` typical '' bl lac lorentz @xmath78 factors but , in this case , the viewing angle to the jet axis would be even smaller ( for @xmath93 and @xmath92 the viewing angle is @xmath94 ) .
the value for the angle of the line - of - sight to the jet axis we derive is rather small , but still compatible with the constraint @xmath95 obtained from vlbi observations of the jet - counterjet ratio @xcite . in absence of a firm determination of the viewing angle to the radio jet axis
, we can not distinguish between the two scenarios proposed above for the location of the emitting region , i.e. the fast `` bl lac '' component or a slower ( but still relativistic ) region of the jet .
our estimate of @xmath79 is substantially lower than both the orientation of the external dusty disk ( @xmath96 ) and the inner gas disk ( @xmath97 ) , as modeled by @xcite .
these authors showed that the structure is warped and tilted , and the axis of the inner disk is significantly twisted with respect to the outer dusty disk .
therefore , our upper limit @xmath90 on the jet direction confirms and extends their picture .
the jet axis is not perpendicular to any of the observed disks : the warping of the accretion structure around the central black holes would continue towards the innermost regions , where the inner accretion disk and the jet are likely to be perpendicular .
the small viewing angle @xmath79 we find would imply a deprojected total size of the source @xmath98 mpc , which would make ngc 6251 by far the largest known radiogalaxy .
however , we point out that the morphology of its jets , on the largest scales , is complex and presents significant bending .
therefore , our determination of @xmath79 , which refers to the inner jet , is probably not appropriate to estimate the overall deprojected size of this radio source .
we have determined the overall spectral energy distribution of the nucleus of ngc 6251 , from the radio to the gamma ray band , by collecting data from the literature and analyzing hst and x - ray data taken with _
beppo_sax and _ chandra_. a significant improvement with respect to previous work is the homogeneous analysis of hst data , which span the range of wavelengths from the ir ( @xmath99 m ) to the uv ( @xmath100 ) and allows us to define a broad peak in the sed at low energies , and the analysis of previously unpublished x - ray data .
the x - ray data , together with the gamma - ray egret detection recently published by @xcite , define a further peak at high energies .
the sed is therefore similar to that of bl lac objects , in agreement with the agn unification model . before proceeding to model
the sed , a key point to be addressed is the nature of the optical source .
we have shown that in diagnostic planes for the nuclear emission of radio galaxies , ngc 6251 follows the general behavior of fr i. therefore , as for the large majority of 3cr fr i radiogalaxies , the unresolved nucleus seen in hst images at all wavelengths can be interpreted as synchrotron emission . at high energies , the _ beppo_sax hard ( @xmath101 ) power - law spectrum and the gamma ray radiation are most plausibly produced by the inverse - compton process . in light of the above results ,
we have modeled the sed in the framework of a synchrotron self - compton scenario , which well accounts for the emission of blazars in general , and especially of bl lac objects .
the parameters of the model are very similar to those found in bl lacs , except for the beaming factor , as qualitatively expected by the unification models .
more quantitatively , @xmath102 implies that the viewing angle to the jet axis can not be wider than @xmath103 , in agreement with the radio information which claims for a small viewing angle . in summary ,
our analysis of the overall sed of ngc 6251 directly supports the fr i bl lac unification scheme and strengthens the results we have obtained for the closest fr i , centaurus a. however , it is still possible that the radiation which dominates bl lacs emission is not produced co - spatially with the nuclear radiation observed in radio - galaxies . because of the larger viewing angle to these latter sources , it is plausible that the `` blazar '' component is beamed away from the observer s line of sight , and the observed radiation is produced in a slower region of the jet , possibly a shear layer , as proposed for other sources @xcite .
this work has been partly supported by an esa post - doctoral fellowship .
the authors thank the referee for insightful comments .
mc whishes to thank g. ghisellini for useful discussion .
the model is described in detail in @xcite and @xcite , in which it has been applied for the first time to reproduce the nuclear sed of a radio galaxy , centaurus a. here we summarize the main assumptions of the model .
the source is a spherical homogeneous region , embedded in a tangled magnetic field .
relativistic electrons are continuously injected at a rate @xmath104 [ @xmath105 s@xmath5 ] @xmath107 between @xmath108 and @xmath109 ( @xmath110 being the lorentz factor ) , and they loose their energy radiatively .
the parameters of the model are : the size of the source @xmath111 , the magnetic field @xmath83 , the injected luminosity @xmath112 , the relativistic beaming factor @xmath113^{-1}$ ] ( where @xmath79 is the angle between the jet axis and the line of sight ) , @xmath108 , @xmath109 and the slope @xmath114 .
the resulting electron distribution at equilibrium is a broken power - law .
electrons are also allowed to escape the emitting region ( thus reducing the particle density , mimicking source expansion ) in a timescale @xmath115 .
we obtain the equilibrium solution of the continuity equation which governs the temporal evolution of the emitting electron distribution @xmath116 .
@xmath117 + q(\gamma , t ) - \frac{n(\gamma , t)}{t_{esc}}=0\ ] ] where @xmath118 is the total ( synchrotron + self compton ) cooling rate .
we solve the equation numerically ( see chiaberge & ghisellini 1999 ) and we calculate self - consistently the produced equilibrium spectrum .
we take into account the effects of the klein - nishina decline , using an approximated step function @xcite . | we determine the nuclear spectral energy distribution ( sed ) from the radio to the gamma
ray band for the fr i radio galaxy ngc 6251 , by using both data from the literature and analyzing hst and x - ray archival data . in the @xmath0 representation ,
the sed has two broad peaks , and it is remarkably similar to those of blazars . we show that the low - energy peak can be explained in terms of synchrotron radiation , while the high energy peak is most plausibly produced by inverse compton scattering .
this brings direct support to the fr i bl lacs unification model .
we model the overall emission in the frame of a synchrotron self - compton scenario , which well describes the sed of bl lacs .
the model parameters we obtain confirm also quantitatively the fr i bl lac unification model and imply a rather small viewing angle to this source ( @xmath1 ) .
ngc 6251 is the second radio galaxy , in addition to centaurus a , for which a similar analysis has been performed .
a significant improvement with respect to the case of cen a is the absence of obscuration in ngc 6251 , which strengthens the overall result . |
galactic formation and evolution is one of the most outstanding problems in modern astrophysics .
one way to address this question is through galactic archaeology , which is an approach to explore the formation and evolution history of the milky way through the @xmath5archeological record provided by its individual stars .
we are coming into a new era of galactic investigation , in which galactic archaeology can be studied with massive spectroscopic surveys .
the spectra generated by these surveys are range from @xmath6 to @xmath7 , for examples , rave @xcite , segue @xcite , apogee @xcite , galah @xcite , lamost @xcite , _ gaia _
@xcite , argos @xcite , 4most @xcite , and weave @xcite .
valuable information , that can be extracted from these spectroscopic surveys , include fundamental stellar parameters , chemical compositions , and radial velocities of individual stars . with the compilations of these massive surveys in the next decade
, our knowledge of galactic formation and evolution will be significantly improved . as a unique campaign among these surveys
, apogee operates in the near - infrared band and is able to target stars at very low galactic latitude @xcite , this penetrating the veil of interstellar dust . besides observing disk stars
, apogee also targets two other components : the halo and the bulge .
it has pencil - beam observations of red giants in the halo that target fields in the bulge .
the primary goal of apogee is to provide constraints on the dynamical and chemical evolution model of our milky way .
numerous galactic structure and stellar population issues can be addressed by the apogee data . in order to take full advantage of this
massive spectroscopic data set , and make a full use of the chemical information , we need to analyze the 6d phase - space distribution of stars in the milky way , which is usually hampered by a lack of reliable distance data .
the most accurate and successful direct distances measurements are from the _ hipparcos _ satellite , which provide trigonometric parallaxes for @xmath8 stars .
however , the _ hipparcos _ measurements are only accurate to a distance around @xmath9 pc @xcite . in the near future , the recently launched _
mission ( perryman 2005 ) will return parallax and proper motion measurements for around @xmath10 stars , which will contribute to this field substantially .
however , even after the _ gaia _ mission is completed , stars with direct distance measurements are still less than 1% of all stars in the milky way .
indirect distance determination methods for more distant stars , especially for apogee stars with heavy extinction in the low latitude disk region , and in the bulge , are still highly desired .
several indirect methods for determining stellar distance have been developed .
red clump ( rc ) stars has been used as standard candles to calculate stellar distances @xcite .
since the rc stage is very short in the whole stellar evolutionary life of a star , in general , this method is only suitable for a small fraction of stars .
asteroseismology can determine stellar parameters , such as stellar mass , radius , and age @xcite , with high precision .
these parameters are then employed , together with other stellar parameters determined with classical methods , e.g. , metallicity and effective temperature , to accurately estimate intrinsic stellar properties , such as distance and extinction .
@xcite have clearly demonstrated that distance can be obtained with high accuracy when coupled with asteroseismic information .
however , the asteroseismology method requires systematic variability studies of its targets .
it not only needs the target stars to have detectable seismic activities , but also requires observing a very time consuming data set .
till now , only small patch of sky coverage has become available with asteroseismology data @xcite , and many stars within the coverage area do have asteroseismic data for accurate distance determinations .
therefore , more general methods of deriving stellar distance are needed . to take advantage of large sets of available spectroscopic and photometric data from recent surveys ,
more general methods to infer stellar parameters , such as extinction , mass , age , distance , etc , have been explored by many authors @xcite .
because of its importance , many among these works focus on distance determination @xcite .
these authors use spectroscopic and photometric quantities to compute the probability distribution of stellar parameters , to infer distance of individual stars .
the logistics of these works are similar , but different authors may implement the method in slightly different ways . for instance
, some use estimated uncertainties of measured stellar parameters , some do not .
authors may treat extinction differently , use different theoretical evolutionary models , use different prior knowledge of the galaxy , etc . in this work ,
we follow the same bayesian method as @xcite to derive distances for around @xmath11 stars in the apogee survey .
the extinction is considered consistently along with the distance determination .
the plan for the paper is as follows . in section 2
, we briefly introduce the bayesian method adopted in this work , while the apogee data from sdss dr12 are described in section 3 . in sections 4 and 5
, we compare our bayesian distances and extinctions with those from other independent methods .
finally , the conclusion is presented in section 6 .
the bayesian method adopted here is the same as used by @xcite , in which the distances and extinctions were calculated for around one million stars in the first data release of the lamost survey @xcite .
the accuracy of the derived distances have been extensively tested with both simulation and several independent measurements utilizing lamost data .
the check shows that a good accuracy can be achieved with the implemented bayesian method in @xcite .
the method is very similar to those in @xcite and @xcite . for clarity
we give a brief description on this method .
for each star we have the relevant observables : effective temperature @xmath12 , metallicity @xmath13 , surface gravity @xmath4 , and near infrared photometries @xmath14 .
these quantities form an observed vector : @xmath15 each star can be related and characterized by a set of @xmath5intrinsic parameters : metallicity @xmath13 , age @xmath16 , initial mass @xmath17 , position on the sky @xmath18 , and distance from the sun @xmath19 .
these quantities form another vector : @xmath20 with the help of trivial bayesian theory , we can derive the posterior probability of @xmath21 , which is the conditional probability of the parameter set @xmath22 given @xmath23 .
@xmath24 @xmath23 and @xmath22 can be connected by theoretical isochrones , with @xmath23 being a function of @xmath22 .
@xmath25 is the probability that the set of observations was made and does not depend on @xmath22 , which is a normalization factor .
@xmath26 is the probability that the set of observations @xmath23 gives the set of parameters @xmath22 .
we assume the uncertainties of the measured observables for each component of @xmath23 can be described as a gaussian distribution with a mean @xmath27 and standard deviation @xmath28 .
@xmath29 @xmath30 is the prior probability we ascribe to the set of parameters , which is an important ingredient in the bayesian method .
@xcite used a three - component prior model of the galaxy for the distribution functions of metallicity , density , and age : @xmath31 where the @xmath32 correspond to thin- , thick - disk , and stellar halo , respectively .
a slightly modified kroupa - type imf is adopted as shown in @xcite and @xcite . the same as @xcite , an extinction prior , which employs the three - dimensional milky way extinction model ,
is used to calculate the extinction by integrating along each line of sight toward individual stars . having the posterior probability distribution function , @xmath33 , the mean and standard deviation for each parameters in @xmath22
can be obtained by taking the first and second moments of this distribution as shown by @xcite . besides providing the distance @xmath34 , this bayesian method can also output another distance estimator , the parallax @xmath35 . in general , there is the relation between these two distance estimators , @xmath36 @xmath37 @xmath38 , which can be attributed to the weights that each estimator attaches to the possibilities of long or short distance@xcite .
@xcite find that @xmath38 is a good distance estimator for rave data . in this work
, we will use the output @xmath36 directly in the following sections .
however , for comparison we also list the results estimated from @xmath39 in table.[table : d ] .
the padova isochrones @xcite are adopted like in @xcite .
the metallicity steps of the isochrones have been carefully chosen by @xcite for avoiding great changes in the observed quantities between adjacent isochrones , and have been extended to low metallicity by @xcite .
since the extinction can be calculated with this bayesian method , an extinction law is a necessary ingredient . at present
, there is no real consensus on the correct extinction law along all the directions of the galaxy , especially toward to the galactic bulge region @xcite .
thus , we decide to use the extinction law of ( * ? ? ?
* ccm ) with @xmath40 as shown in @xcite .
the apache point observatory galactic evolution experiment ( apogee ) as one component of sdss iii @xcite , is a near - infrared ( @xmath41-band ; @xmath42m@xmath43 m ) , high - resolution ( r@xmath4422,500 ) spectroscopic survey .
it targets primarily red giant ( rg ) stars across all galactic environments and stellar populations in the milky way @xcite , using the sloan 2.5 m telescope@xcite and an innovative multi - object ir spectrograph @xcite .
the stellar atmospheric parameters and individual chemical abundances are derived from combined apogee spectra with the apogee stellar parameters and chemical abundances pipeline ( aspcap ; * ? ? ?
* ) , which finds the best matching between pre - calculated synthetic spectra to the observed ones via a @xmath45@xmath46 minimization . to validate aspcap outputs , stellar parameters for members in well studied clusters and surface gravities from asteroseismic analyses
are carefully compared @xcite .
calibration correlations are applied to the original aspcap outputs to obtain calibrated stellar parameters for stars @xcite .
the empirical uncertainty of @xmath12 is about 100 k through comparison of aspcap derived temperatures with photometric temperatures , and 0.11 dex for @xmath4 via comparing calibrated aspcap values with asteroseismic surface gravities for stars in the @xmath47 field .
the internal precision of the apogee abundances is typically @xmath48 dex , but the external accuracy of the abundances is challenging to assess , and may only be good to 0.1 - 0.2 dex @xcite .
it is worthy to note that the applied calibration correlations are derived with members in well - studied clusters and stars in the @xmath47 field ; their reliability outside parameter spaces that these calibrated stars cover , such as surface gravities of metal poor stars , are not validated .
the final sdss - iii apogee public data release , dr12 @xcite , includes catalogs with radial velocity , stellar parameters @xmath49)$ ] , and 15 elemental abundances for over 150,000 stars , as well as more than 500,000 spectra from which these quantities are derived .
we retrieved the calibrated stellar parameter data from the apogee parameter catalogs of sdss dr12 using the casjobs interface .
the two micron all sky survey ( 2mass ) near - infrared photometric data for the same stars in @xmath50 m ) , @xmath51 m ) , @xmath52 m ) are also extracted from the point - source catalogue @xcite .
the following procedure is used to select stars with reliable parameters : first , stars should not be set any flags indicating that the observations or analysis is bad .
second , stars without any calibrated parameters ( such as dwarfs ) are excluded .
we use sdss casjobs to extract the data with the following command scripts : ( aspcapflag & dbo.fapogeeaspcapflag(star_bad ) ) = 0 and teff @xmath53 0 . with these criteria , we select 104816 stars from the apogee catalogue . there are 3084 stars that have been observed more than once . for stars with multiple observations ,
only data with the highest signal to noise ratio are used .
this results in 101726 stars in our final sample .
we note that in the following analysis only stars with @xmath54 k and @xmath55 are used to select stars with reliable parameters @xcite , and there are 99165 stars left after applying this selection .
after further strict criteria are applied , such as @xmath56 k and aspcap_chi2 @xmath57 20 ( 92060 stars left ) , no significant change is noticed in the statistical results .
as discussed above , there are several independent methods which can deliver reliable distances for their suitable targets .
for instance , distances to nearby stars from parallax measurements ; to cluster members from the main - sequence fitting , distances of red clump ( rc ) stars by using them as a distance candle ; distances of giants from asteroseismic information , etc .
our bayesian method , described in section 2 , is a more general one . in principle
, it can be used to derive stellar distances of individual stars which have available spectroscopic and photometric quantities . to validate our derived stellar distances , we compare our derived distances with results from other independent methods for common stars : astrometric parallax distances obtained by esa s _ hipparcos _ mission @xcite ; distances to well studied clusters , distances of rc stars , and distances from asteroseismic analyses .
figures 1 to 6 show these comparisons which are also summarized in table 1 . in summary ,
the systematic offsets , in terms of our derived distances minus others , range from @xmath1 to @xmath58 , with dispersions from 15% to 25% ; considering mutual uncertainties in our derived distances and in others , our distances are consistent with those from other methods .
we discuss these comparisons in more detail below . [ cols="<,<,<,^,^,^,^,^,^,^ " , ] the apogee survey has generated high resolution near - infrared spectra for over @xmath11 stars .
these data provide valuable information on the studies for galactic archaeology . in order to fully exploit these data sets in six - dimension
, we used a bayesian method to derive the distances and extinctions of individual stars by considering photometric and spectroscopic information , as well as prior knowledge on the milky way .
all derived distances and extinctions will be made available as online data .
table [ table : catalogue ] shows the online data format .
the first column is apogee i d , the fourth and fifth columns are derived distance and its associated error respectively , while the eighth and ninth columns are estimated extinction and its error respectively .
the sixth and seventh columns are another distance indicator , parallax and its error output by the bayesian method respectively . to assess the derived distances , we compared our bayesian distances to those from four independent measurements , the _ hipparcos _ parallaxes , stellar cluster distances , red clump star distances , asteroseismic distances from saga catalogue @xcite and apokasc catalogue @xcite .
these independent measurements cover four orders of magnitude in distance , from @xmath59 kpc ( the _ hipparcos _ parallaxes ) to @xmath60 kpc ( star clusters ) .
the results of these validations are all summarized in table [ table : d ] .
we find that the mean relative difference between our bayesian distances and those derived from other methods are @xmath1 to @xmath2 , and that the dispersion ranges from 15% to 25% .
surprisingly , no systematic offset is seen in the comparison with one set of asteroseismic constrained distances @xcite . considering mutual uncertainties in our derived distances and in other measurements ,
we conclude that , statistically , our derived distances are accurate to a few percent although errors for a few individual stars could be large .
therefore , they are suitable for statistical studies on galactic archaeology . we have compared our bayesian extinctions with those derived from the rjce method .
it seems that the extinctions from the rjce method depend on spectral type . for stars with @xmath61 k , @xmath62 , and @xmath63 , the rjce method tends to overestimate extinction , however , in most cases this overestimation is less than 0.11 mag . in @xmath64band , which is still within the error budget suggested by @xcite .
the overestimation of the extinction at large distance can also be attributed to this effect , as giants with even lower @xmath4 are usually the dominant population at this region due to a selection effect .
we also notice that there is a systematic difference at heavy extinction @xmath65 mag . , which could be due to the adopted extinction law . comparing with two dimensional extinction map of @xcite
, we note that the map may overestimate extinction in the nearby regions .
we also compared our results with three - dimensional extinction models of @xcite and maps of @xcite , @xcite , and @xcite . compared with our extinction , both the model of @xcite and map of @xcite
show lower extinction values for distant stars toward the anti - galactic central region .
the model of @xcite matches well with our results of nearby region @xmath66kpc@xmath67 toward the bulge field @xmath68 . for this field
the map of @xcite predicts a steep rise in slope for @xmath69 kpc , while the map of @xcite provides a good fit with our result for distant region ( @xmath70 kpc ) .
the accurate distance and extinction derived in the current work provide valuable information for the apogee stars , and are important for exploiting the data set of the apogee survey when studying the formation and evolution of the milky way .
the anonymous referee is greatly thanked for useful comments and suggestions that helped improve the quality of this paper .
we warmly thank diane feuillet for sharing her _ hipparcos _ target list before publishing it and jon holtzman for the useful discussion .
haibo yuan is thanked for useful discussion .
this research was supported by the national key basic research program of china 2014cb845700 , and by the national natural science foundation of china under grant nos .
11321064 , 11390371 , 11473033 , 11428308 and u1331122 .
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 publication makes use of data products from the wide - field infrared survey explorer , which is a joint project of the university of california , los angeles , and the jet propulsion laboratory / california institute of technology , funded the national aeronautics and space administration .
funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the us department of energy office of science .
the sdss - iii web site is http://www.sdss3.org/. sdss - iii is managed by the astrophysical research consor- tium for the participating institutions of the sdss - iii collabo- ration including the university of arizona , the brazilian partici- pation group , brookhaven national laboratory , carnegie mellon university , university of florida , the french participation group , the german participation group , harvard university , the instituto de astrofisica de canarias , the michigan state / notre dame / jina participation group , johns hopkins university , lawrence berke- ley national laboratory , max planck institute for astrophysics , max planck institute for extraterrestrial physics , new mexico state university , new york university , ohio state university , pennsyl- vania state university , university of portsmouth , princeton uni- versity , the spanish participation group , university of tokyo , university of utah , vanderbilt university , university of virginia , university of washington , and yale university .
99 alam , s. , albareti , f. d. , allende prieto , c. , et al .
2015 , apjs , 219 , 12 ahn c. p. et al . , 2014 , apjs , 211 , 17 allende prieto , c. , majewski , s. r. , schiavon , r. , et al .
2008 , astronomische nachrichten , 329 , 1018 an , d. , terndrup , d. m. , pinsonneault , m. h. , et al .
2015 , apj , 811 , 46 bovy , j. , nidever , d. l. , rix , h .- w . , et al .
2014 , apj , 790 , 127 bertelli , g. , bressan , a. , chiosi , c. , fagotto , f. , & nasi , e. 1994 , a&as , 106 , 275 binney , j. , burnett , b. , kordopatis , g. , et al . 2014a , mnras , 437 , 351 breddels , m. a. , smith , m. c. , & helmi , a. , et al .
2010 , a&a , 511 , a90 burnett , b. , & binney , j. 2010 , mnras , 407 , 339 burnett , b. , binney , j. , sharma , s. , et al .
2011 , a&a , 532 , a113 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , apj , 345 , 245 .
( ccm ) casagrande , l. , silva aguirre , v. , stello , d. , et al .
2014 , apj , 787 , 110 chen , b. q. , schultheis , m. , jiang , b. w. , et al .
2013 , a&a , 550 , a42 chen , b .- q . , liu , x .- w . , yuan , h .- b . , et al
. 2014 , mnras , 443 , 1192 coppola , g. , dallora , m. , ripepe , v. 2011 , mnras , 416 , 1056 .
cui , x .- q . , zhao , y .- h . , chu , y .- q .
, et al .
2012 , research in astronomy and astrophysics , 12 , 1197 cutri , r. m. , skrutskie , m. f. , van dyk , s. , et al .
2003 , the irsa 2mass all - sky point source catalog , nasa / ipac infrared science cutri , r. m. , wright , e. l. , conrow , t. , et al .
2013 , explanatory supplement to the allwise data release products , by r. m. cutri et al .
, 1 dalessandro , e. , salaris , m. , ferraro , f. r. , et al .
2013 , proc.mnras , 430 , 459 dalton , g. , trager , s. c. , abrams , d. c. , et al .
2012 , proc.spie , 8446 , 84460p da silva , l. , girardi , l. , pasquini , l. , et al .
2006 , a&a , 458 , 609 de jong , r. s. , bellido - tirado , o. , chiappini , c. , et al .
2012 , proc.spie , 8446 , 84460 t dkny , i. & kovcs , g. 2009 , a&a , 507 , 803a de silva , g. m. , freeman , k. c. , bland - hawthorn , j. , et al .
2015 , mnras , 449 , 2604 dias , w. s. , alessi , b. s. , moitinho , a. , & lpine , j. r. d. 2002 , a&a , 389 , 871 drimmel , r. , cabrera - lavers , a. , & lpez - corredoira , m. 2003 , a&a , 409 , 205 drimmel , r. , & spergel , d. n. 2001 , apj , 556 , 181 eisenstein , d. j. , weinberg , d. h. , agol , e. , et al .
2011 , aj , 142 , 72 fazio , g. g. , hora , j. l. , allen , l. e. , et al .
2004 , apjs , 154 , 10 freeman , k. c. 2012 , galactic archaeology : near - field cosmology and the formation of the milky way , 458 , 393 feuillet , d. k. , bovy , j. , holtzman , j. , et al .
2016 , apj , 817 , 40 garca prez , et al .
2015 , arxiv:1510.07635 koch , d. g. , borucki , w. j. , basri , g. , et al .
2010 , apjl , 713 , l79 marshall , d. j. , robin , a. c. , reyl , c. , schultheis , m. , & picaud , s. 2006 , a&a , 453 , 635 ness , m. , freeman , k. , & athanassoula , e. 2012a , galactic archaeology : near - field cosmology and the formation of the milky way , 458 , 195 ness , m. , freeman , k. , athanassoula , e. , et al . 2012b , apj , 756 , 22 nishiyama , s. , nagata , t. , kusakabe , n. , et al .
2006 , apj , 638 , 839 nishiyama , s. , nagata , t. , tamura , m. , et al .
2008 , apj , 680 , 1174 nishiyama , s. , tamura , m. , hatano , h. , et al .
2009 , apj , 696 , 1407 holtzman , j. a. , shetrone , m. , johnson , j. a. , et al .
2015 , aj , 150 , 148 harris , w. e. 1996 , aj , 112 , 1487 gonzalez , o. a. , rejkuba , m. , zoccali , m. , et al .
2012 , a&a , 543 , a13 groenewegen , m. a. t. 2008 , a&a , 488 , 935 gilliland , r. l. , brown , t. m. , christensen - dalsgaard , j. , et al .
2010 , pasp , 122 , 131 girardi , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , a&as , 141 , 371 gunn , j. e. , siegmund , w. a. , mannery , e. j. , et al .
2006 , aj , 131 , 2332 lindegren , l. 2010 , iau symposium , 261 , 296 luo , a .-
l . , zhao , y .- h . ,
zhao , g. , et al .
2015 , research in astronomy and astrophysics , 15 , 1095 jacobson , h. r. , pilachowski , c. , & friel , e. d. 2011 , aj , 142 , 59 majewski , s. r. , wilson , j. c. , hearty , f. , schiavon , r. r. , & skrutskie , m. f. 2010 , iau symposium , 265 , 480 majewski , s. r. , zasowski , g. , & nidever , d. l. 2011 , apj , 739 , 25 marigo , p. , girardi , l. , bressan , a. , et al .
2008 , a&a , 482 , 883 marigo , p. & girardi , l. 2007 , a&a , 469 , 239 mszros , s. , holtzman , j. , garca prez , a. e. , et al .
2013 , aj , 146 , 133 nidever , d. l. , holtzman , j. a. , & allende prieto , c. , et al .
2015 , aj , 150 , 173 oconnell , j. e. , johnson , c. i. , pilachowski , c. a. , & butks , g. 2011 , pasp , 123 , 1139 paczynski , b. & stanek , k. z. 1998 , apjl , 494 , l219 perryman , m. a. c. , de boer , k. s. , gilmore , g. , et al .
2001 , a&a , 369 , 339 pinsonneault , m. h. , elsworth , y. , epstein , c. et al .
2014 , apjs , 215 , 19 pont , f. & eyer , l. 2004 , mnras , 351 , 487 robin , a. c. , reyl , c. , derrire , s. , & picaud , s. 2003 , a&a , 409 , 523 robin , a. c. , marshall , d. j. , schultheis , m. , & reyl , c. 2012 , a&a , 538 , a106 rodrigues , t. s. , girardi , l. , miglio , a. , et al . 2014 , mnras , 445 , 2758 santiago , b. x. , brauer , d. e. , anders , f. , et al . 2016 ,
a&a , 585 , a42 sarajedini , a. , dotter , a. , & kirkpatrick , a. 2009 , apj , 698 , 1872 schultheis , m. , zasowski , g. , allende prieto , c. , et al .
2014a , aj , 148 , 24 schultheis , m. , chen , b. q. , jiang , b. w. , et al .
2014b , a&a , 566 , a120 schnrich , r. , binney , j. , & asplund , m. 2012 , mnras , 420 , 1281 skrutskie , m. f. , cutri , r. m. , stiening , r. , et al .
2006 , aj , 131 , 1163 steinmetz , m. , zwitter , t. , siebert , a. , et al .
2006 , aj , 132 , 1645 williams , m. e. k. , steinmetz , m. , binney , j. , et al .
2013 , mnras , 436 , 101 van leeuwen , f. 2007 , astrophysics and space science library , 350 , yanny , b. , newberg , h. j. , johnson , j. a. , et al .
2009 , apj , 700 , 1282 yuan , h. b. , liu , x. w. , & xiang , m. s. 2013 , mnras , 430 , 2188 wang , j. , shi , j. , zhao , y. , et al .
2016 , mnras , 456 , 672 wang , j. x. , ma , j. , wu , z. y. , et al .
2015 , aj , 150 , 61 werner , m. w. , roellig , t. l. , low , f. j. , et al .
2004 , apjs , 154 , 1 wu , z. y , zhou , x , ma , j. , et al , 2009 , aj , 133 , 2061 wu , t. , li , y. , & hekker , s. 2014 , apj , 786 , 10 wilson , j. c. , hearty , f. , skrutskie , m. f. , et al . 2010 , proc , spie , 7735 , 77351c zasowski , g. , johnson , j. a. , frinchaboy , p. m. , et al .
2013 , aj , 146 , 81 zhao , g. , chen , y .- q . , shi , j .-
, et al .
2006 , chjaa , 6 , 265 zwitter , t. , matijevic , g. , breddels , m. a. , et al . 2010 , a&a , 522 , a54 | using a bayesian technology we derived distances and extinctions for over @xmath0 red giant stars observed by the apache point observatory galactic evolution experiment ( apogee ) survey by taking into account spectroscopic constraints from the apogee stellar parameters and photometric constraints from 2mass , as well as a prior knowledge on the milky way .
derived distances are compared with those from four other independent methods , the _ hipparcos _ parallaxes , star clusters , apogee red clump stars , and asteroseismic distances from apokasc @xcite and saga catalogues @xcite .
these comparisons covers four orders of magnitude in the distance scale from 0.02 kpc to 20 kpc .
the results show that our distances agree very well with those from other methods : the mean relative difference between our bayesian distances and those derived from other methods ranges from @xmath1 to @xmath2 , and the dispersion ranges from 15% to 25% .
the extinctions toward all stars are also derived and compared with those from several other independent methods : the rayleigh@xmath3jeans color excess ( rjce ) method , gonzalez s two - dimensional extinction map , as well as three - dimensional extinction maps and models .
the comparisons reveal that , overall , estimated extinctions agree very well , but rjce tends to overestimate extinctions for cool stars and objects with low @xmath4 . [ firstpage ] stars : fundamental parameters stars : distances dust , extinction |
in a transit experiment where the goal is to look for obscuration of light caused by the transiting body , the effect of microlensing by the transiting body is generally thought to be negligible .
the main reason is the following : for orbital separations of a few au , the einstein ring due to a stellar companion ( the lens ) is typically much smaller than the size of the primary ( the source ) , so that the maximum amplification caused by the lens is small .
in addition , the lens has a finite size , which may cause one or more of the rays to be occulted , which makes the effect of microlensing small compared to the obscuration caused by the lens . in principle
however , the effect of microlensing can be important , and this effect becomes more important when the binary companion is a degenerate star ( i.e. massive but small ) or if the orbital separation is large .
gould ( 1995 ) argued that self - lensing in binaries in which at least one member is an ordinary ( non - compact ) star ( assuming @xmath3 0.3 mag amplification ) would be exceedingly rare .
marsh ( 2001 ) argued for better prospects in finding microlensing signatures in main - sequence , white dwarf binaries , but also concluded that finding microlensing in binaries containing neutron stars or black holes is unlikely . at high photometric precision
, however , microlensing effects at _ much _ lower amplitude can be important , and this effect must be taken into account in interpreting the observed light curves . the discovery - class mission , _ kepler : a search for terrestrial planets _ , with an expected launch in 2007 , is a high - precision photometric mission whose main goal is to detect a statistically large number of extrasolar , earth - like planets , if such planets are common around stars in the extended solar - neighborhood .
_ kepler _ will provide such detections by obtaining extensive ( four years ) , near - continuous photometry on the same set of over 100,000 stars at typical transit - depth precisions of 90 parts per million on the 15 minute sampling rate for @xmath4 stars .
an earth - analogue transit will typically last @xmath210 hours and be detected at @xmath26 @xmath5 per event which repeats once per year .
( more details on _ kepler _ , which is a nasa - approved discovery - class mission , can be found at http://www.kepler.arc.nasa.gov/ ) . at such precision
, the effect of microlensing by a transiting compact and massive body can often be significant .
for example , when the lens is a compact object such as a white dwarf or a neutron star , the effect of microlensing can far exceed the transit signal , causing a net positive amplification .
the microlensing contribution due to a brown dwarf at 1 au would be several percent of the transit depth . in this paper , we explore the astrophysical implications of such microlensing in the near - field limit of a lens orbiting the source , with particular emphasis on the upcoming _ kepler _ mission .
gravitational microlensing studies over the past decade have revealed large numbers of events toward the magellanic clouds and the galactic bulge @xcite . with near real - time analysis of the photometry ,
many of the microlensing events are detected early enough to provide alerts allowing for close and intense follow - up monitoring of potentially interesting events .
such intense monitoring has been carried out for several purposes such as , to look for planets associated with the lens ( albrow et al .
2001a ; gaudi et al .
2002 ) ; to determine limb - darkening parameters of the source ( albrow et al .
2001b ) ; to measure lens masses through parallax effects and orbital motions of binary lenses ( an et al .
2002 ) ; to determine the caustic crossing time - scales to determine the lens location ( afonso et al .
2000 ) , etc .
the literature on microlensing both from observational and theoretical perspectives is now quite large and mature ( e.g. schneider et al .
1993 , paczyski , 1996 & gould , 2001 ) .
by contrast , the much simpler effect of geometric obscuration resulting in transits as one object passes in between a source and the observer has received much less attention relative to microlensing ( unless one includes binary star eclipses in this context ) , although there has been a surge of interest on this topic recently . in the case of classical microlensing the events are rare , thus the successful projects monitor very dense stellar fields with large format detectors .
however , the microlensing signal can be quite large , so these projects require only modest precision .
transits of planets generally provide much smaller signals : jupiter transiting the sun , for example , would yield an event of 1% depth .
furthermore , the inclinations are expected to be random , so for planets with orbital radii of @xmath31 au , @xmath699% of the existing systems would not exhibit transits due to the small chance of having the required alignment . for terrestrial planets
the transit depth is even smaller : about one part in 10,000 .
hence , although seeking evidence of earth - like , extrasolar planets ( i.e. , having a size comparable to the earth at a distance of about an au from a solar - analogue host star ) via transits has long been advocated as a means to search for terrestrial planets ( borucki and summers 1984 ) , this is a challenging experiment . with the stable observing conditions provided by observations from space and the ability to mosaic a significant number of ccds it has become feasible to detect true earth - like extrasolar planets with a modest , dedicated mission such as _
kepler_. in a mission with such unprecedented photometric accuracy , it is obviously important to make sure that the signature of an earth - like planet is unique .
since a white dwarf has a radius comparable to that of the earth , a white dwarf in a one year orbit transiting a solar - like star could in principle mimic the signature of a terrestrial planet ( although this can be easily eliminated in most cases by follow up radial velocity measurements ) .
this study is motivated by asking if the photometric signal itself due to microlensing by the white dwarf , even though it is only 1 au from its host star , might be detectable . indeed for the case of a white dwarf in a one year orbit , the geometric obscuration effect is at the level of one part in @xmath7 , but the microlensing effect is much larger , amounting to a few parts in @xmath8 . as we will show later , the light curve caused by the microlensing has a form very similar to the transit light curve with a change of sign .
taking into account the effect of gravitational microlensing , we will show that white dwarfs transiting main sequence stars with orbital periods of @xmath33 days would produce _ positive _ intensity signals , and hence will not be a source of confusion with terrestrial planets for _ kepler_. the purpose of this paper is to establish the framework for calculating microlensing in this near - field regime , to establish domains in source and lens characteristics where the effect is measurable , and to explore whether the near - field microlensing signal can be exploited to return astrophysical information such as the lens mass . in 3 we discuss the amplification in the near - field regime and present a generalized formalism for generating light curves taking the finite source size , the finite lens size , the inclination of the orbit and the limb - darkening parameters into account , and discuss how the simpler cases of point source etc . can be considered as special cases of this generalized case . in 4 , we discuss the amplifications caused by various types of degenerate and non - degenerate secondaries at different orbital separations for different types of sources , and discuss the interesting cases in which the gravitational light amplification and geometric obscuration effects are similar in amplitude .
5 is devoted to a more detailed discussion of special cases such as earth - mass planets , jupiters , brown dwarfs , white dwarfs , neutron stars and black holes as secondaries in orbit around main sequence stars .
we conclude in 6 with a summary of earlier sections and further discussion on the prospects of such observations with the upcoming _ kepler _ mission , and the possible astrophysical returns from such observations .
some early theoretical work involving a finite source ( and a point lens ) was carried out by bontz ( 1979 ) .
a mathematical description leading to an analytical solution for such a case can be found in schneider et al .
( 1992 ) and witt and mao ( 1994 ) . dominik ( 1998 ) and gould & gaucherel ( 1997 ) have independently developed efficient schemes for calculating the finite source effects using green s theorem and line integrals instead of following the conventional approach of 2-dimensional integration over the source . the numerical recipe developed by sahu ( 1994 a
, b ) uses a polar coordinate system centered on the source , which is particularly useful in considering radially - symmetric effects such as the limb darkening , which is followed here .
the effect due to a finite _ lens _ has received much less attention ; nevertheless this has been discussed in some detail by bromley ( 1996 ) and marsh ( 2001 ) , and more recently by agol ( 2002 ) and beskin and tuntsov ( 2002 ) .
the literature dealing with the combined effect of both a finite source and a finite lens is generally lacking .
gould ( 1995 ) considered the effect of microlensing in binaries and concluded that only pairs of pulsars would be interesting .
however , his analysis was done in the frame - work of the standard " microlensing where the source lies within the einstein ring radius of the lens , in which case the amplification may be more than 34% .
marsh ( 2001 ) has generated light curves for compact , degenerate binaries with small orbital separations , where he assumed that the source is uniformly bright and that the fainter of the two images is occulted by the lens . in order to ensure that the results are valid for all cases observed by _
, our calculations take appropriate limb - darkening parameters into account , and do not assume _ a priori _ that the fainter image is occulted . in the following section ,
we discuss in detail the approach adopted here in determining the effect of finite sizes of the source and the lens simultaneously .
our light curve simulations are valid for low - mass as well as high - mass binaries at any orbital separation , and also include an appropriate limb - darkening for the source and different inclinations of the orbital plane .
we note that in the extreme case where the lens is a point lens , the net effect would be the same as that of pure microlensing . in the extreme case where the lens mass is zero
, the effect would be the same as that of a pure transit .
let us first consider the source to be a point source .
let @xmath9 be the distance to the lens ( @xmath10 ) , @xmath11 be the distance from the lens to the source ( @xmath12 ) , and @xmath13 be the distance from the observer to the source as shown in fig .
1 . let @xmath14 represent the distance between the center of the lens and the source projected onto the lens plane .
the radius of the einstein ring at the lens plane , @xmath15 , can be written as @xmath16 where @xmath17 is the the mass of the lens .
the lens produces two images of the source , @xmath18 and @xmath19 , as shown in fig .
1 . at the lens plane ,
these two images are situated on the line joining the source and the center of the lens , one inside and the other outside the einstein ring .
the distance between the lens and the two images can be expressed as ( e.g. see paczyski 1996 for details ) @xmath20\ ] ] .
the amplification of these two images are given by @xmath21 .
thus we see that the image outside the einstein ring is the brighter of the two images , and is at least as bright as the source itself .
the combined amplification @xmath22 for a point source is given by @xmath23 if @xmath24 is replaced by the minimum impact parameter @xmath25 , we get the maximum amplification @xmath26 .
let us now discuss the case of an extended source , which can be treated as an ensemble of point sources .
we will assume the lens to be extended , opaque and spherically symmetric .
for an extended source , it is convenient to work in the lens plane , and fig .
2 schematically shows such a case , where the source is larger than the einstein ring itself .
the two circles a and b represent two parts of the source , whose images are ( a@xmath27 , a@xmath28 ) and ( b@xmath27 , b@xmath28 ) , respectively . the brighter images ( a@xmath27 and b@xmath27 ) are outside the einstein ring , none of which ( a@xmath27 ) is occulted by the lens . the fainter images ( a@xmath28 , b@xmath28 ) are inside the einstein ring , one of which ( a@xmath28 ) is occulted by the lens .
it is worth noting that the images of all the parts of the source are displaced with respect to the original positions such that no two images overlap .
microlensing preserves the surface brightness , and any amplification caused by microlensing is caused by a proportional increase in the size of the image with respect to the original size .
the amplification of an extended source would then be given by ( equation 6.81 of schneider et al .
1993 ) @xmath29 where @xmath30 is the surface brightness profile of the source , @xmath31 is the amplification of a point source at position @xmath32 , and the integration is carried out over the entire surface of the source . for a lens with finite size ,
the above equation must be rewritten as @xmath33 \over \int d^2y \ i(y)}\ ] ] where the integration is carried out over all the points in the image plane which are not occulted by the lens .
note that for a finite - size , uniformly bright source , the maximum amplification ( corresponding to a point lens which is perfectly aligned with the source ) is given by @xmath34^{1\over{2}}\ ] ] where @xmath35 is the radius of the source scaled to the lens plane . in order to carry out the integration given by equation ( 6 ) ,
let us follow fig .
3 where @xmath14 is the distance between the lens and the center of the source projected onto the lens plane , and @xmath36 is the impact parameter ( the minimum distance between the center of the lens and the center of the source at the lens plane as the lens traces the path shown in the figure ) .
let us choose a circular coordinate system and let ( @xmath37 ) be the representative point which is at a distance @xmath32 from the lens . from fig .
3 we see that @xmath38 where @xmath39 i.e. @xmath40 the two images of the point ( @xmath37 ) are formed at distances @xmath41 and @xmath42 from the center of the lens along the line joining the lens and the point ( @xmath37 ) such that @xmath43\ ] ]
. the image is not occulted if @xmath44 where @xmath45 is the radius of the lens .
the amplification corresponding to these two images can be written as @xmath46^{1/2 } } } \pm 0.5\ ] ] in calculating the combined amplification @xmath47 , we consider only the images which satisfy the condition ( 12 ) to ensure that the image is not occulted by the lens . using equation(6 ) , the total amplification can be calculated by integrating over the source using @xmath48 this is an expression suitable for numerical integration which will take the finite size of both the lens and the source into account .
the light curve can be reproduced as a function of time @xmath49 , which is related to @xmath50 through the relation @xmath51 where @xmath52 is the timescale of microlensing ( which is the time taken by the lens to cross its own einstein ring ) .
@xmath32 in turn , is related to @xmath50 through equation 8 and 9 . since we are mostly interested in the effect caused by the secondary in a binary , we need to include the effect of the inclination , which can be incorporated through a suitable choice of the minimum impact parameter .
the light curves can then be calculated with the amplification as a function of the phase ( see fig .
we also provide here the equations that allow the mass of the lens to be derived from photometry in the limit where microlensing amplification dominates the light curve , as for example would be the case for a white dwarf orbiting a main - sequence star at 1 au .
if multiple , periodically repeating events have been observed , kepler s law provides a relation between the orbital period ( @xmath53 ) , semi - major axis ( @xmath54 ) and the sum of the masses : @xmath55 where @xmath56 is the mass of the source which may be constrained from independent observations ( such as spectroscopy and spectral classification and/or with radial velocity observations ) , and @xmath17 is the lens mass for which an estimate is desired . assuming @xmath57 , equation ( 7 ) can be rewritten as @xmath58 where @xmath35 is the radius of the source star scaled to the lens plane , and @xmath59 is the velocity of light .
@xmath60 is the maximum amplification during the microlensing event which we have assumed to be @xmath61 .
note that for source sizes much larger than the einstein ring radius , the maximum amplification during the microlensing event ( @xmath62 ) is nearly equal to @xmath63 given by equation 7 regardless of the impact parameter , provided that the lens actually transits the source .
this is demonstrated by the flat - topped nature of the numerically - simulated lightcurves ( fig .
10 , 14 and 15 ) , and is consistent with discussions in several other papers ( e.g. agol 2002 ) .
thus @xmath64 is generally a good approximation .
if independent observations provide constraints on the source radius , then equation ( 16 ) can be used to solve for orbital separation , @xmath54 , and @xmath17 .
an additional constraint comes from the duration of the transit ( @xmath65 ) which , for a circular orbit and 90@xmath66 inclination , can be written as @xmath67^{1/2}}}}\ ] ] . combining equations ( 16 ) and ( 17 ) to eliminate @xmath35
@xmath68 after a little algebra , this can be rewritten as @xmath69 this equation can be used to derive the mass of the lens if @xmath56 and @xmath54 are known .
we can also combine equations ( 16 ) and ( 17 ) to eliminate @xmath54 , and write @xmath70 this equation can be solved to express the lens mass @xmath71^{1/2 } } - g m_s t_t } \over { 2 g t_t}}\ ] ] this equation can be used to derive the mass of the lens if @xmath56 and @xmath35 are known .
we will return to a discussion of lens mass determination with an example in 5.2 .
since the polar coordinate system we have chosen is centered on the source ( fig .
3 , equation 8) , it is easy to incorporate limb - darkening for the source which has an azimuthal symmetry .
recently , the limb darkening parameters were derived for hd 209458 using the high signal - to - noise observations obtained with the space telescope imaging spectrograph aboard hubble space telescope in the wavelength region 582 to 638 nm ( brown et al .
based on these observations , a limb - darkening of the form @xmath72 was used , where @xmath1 is the cosine of the angle between the line of sight and the normal to the local stellar surface ( claret & gimenez 1990 ) .
we have used @xmath73 and @xmath74 as applicable to hd 209458 in the wavelength region 582 to 638 nm , in generating all the light curves here .
we should note , however , that the formalism discussed in this paper can be used to take any other specific form of the limb - darkening as well . in the limiting case of a point lens
, the results should be identical to the results obtained from other standard treatments of microlensing involving a finite source and a point lens .
in particular , equation ( 7 ) can be directly used to calculate the amplification for the special case where the lens and the source are perfectly aligned .
this provides a direct test ( although only for the special case of a point lens ) which was used to check the algorithm discussed above .
the previous results obtained by sahu ( 1994b ) and confirmed by dominik & hirshfeld ( 1996 ) also served as a test case for a point lens , to check the correctness of the simulated light curves .
we have also checked our calculations with a few special cases calculated by marsh ( 2001 ) and obtained consistent results . in the limiting hypothetical case where the mass of the lens is zero ( but
the size is finite ) , the results obtained here should be identical to the results obtained for pure transit .
the transit light curve simulated by brown et al ( 2001 ) for hd 209458 serves as a test case for this purpose .
our algorithm was used to simulate a light curve for hd 209458 using the same limb - darkening parameters suggested by brown et al ( 2001 ) .
the resultant light curve was found to be identical to the one published by brown et al ( 2001 ) which provided a further check on the validity of the algorithm used here .
the treatment for a zero - mass lens is equivalent to assuming that the images are not deflected , the amplification corresponding to the brighter image is one , and the amplification corresponding to the fainter image is zero .
this , in turn , is equivalent to a restricted version of the algorithm where all deflections and amplifications are ignored .
the light curves were simulated using both these approaches , and the resultant light curves were identical as expected .
the maximum net amplification caused by a secondary companion ( after taking the finite sizes for the source and the lens into account ) generally corresponds to the configuration where the lens and the source are perfectly aligned ( except in a few special cases where the maximum amplification occurs just before or after the egress as explained later ) .
this configuration corresponds to @xmath75 ( equation 8) , which can be used in equations ( 8 - 14 ) to calculate the amplification .
an analytical solution for such a configuration was derived by marsh ( 2001 ) .
the maximum amplification thus calculated for a variety of secondaries ( jupiter , brown dwarf , white dwarf , neutron star and black hole ) at various orbital separations are shown in figures 2 to 7 , assuming @xmath76 .
4 - 8 assume the source to be a solar type star with radius 7 @xmath77 cm .
the radii and the masses assumed for the different secondaries are given in table 1 . [
cols="<,^,^,^,^ " , ] in most cases that we consider , the secondary ( lens ) will provide an amount of light in the system that is small compared to that from the primary ( source ) .
we therefore do not provide explicit corrections to light curves from the luminosity of the lens itself .
we also note the obvious , yet subtle , point that the light from the lens is constantly present before , during and after an occultation / microlensing event .
therefore , during the transit the only effect from the light from the secondary is a dilution of the light curve changes in the same proportion as the secondary contributes to the total light light away from the transit . for a canonical case of a white dwarf orbiting a solar - like star in which the surface brightness of the two components are similar
, the white dwarf will have a flux about @xmath78 that of the primary implying a 0.01% dilution effect .
( @xmath79 might be able to detect the secondary eclipse , and this would be interesting , but correcting for the minor change of inferred microlensing amplitude is not relevant ) .
exceptions to this rule can arise , e.g. when the primary is intrinsically small such as a brown dwarf , the amplifications of which are shown in fig .
9 . in the case of a white dwarf lensing a brown dwarf
, this dilution effect could be a dominant effect which can be applied in a trivial manner using the expected luminosities of the two objects in the observational bandpass .
4 shows the effects of different secondaries for orbital separations of 0 to 1 au and fig .
5 shows the same for orbital separations of 0 to 5 au .
note that the radii of a jupiter and a brown dwarf are identical .
however , while the microlensing contribution due to a jupiter is negligible , the microlensing contribution due to a brown dwarf is appreciable at kepler s sensitivity even at 1 au . as a result ,
the @xmath80mag caused by these two types of secondaries are different at kepler s sensitivity at 1 au , which is shown in fig .
this effect must be taken into account in deriving the radius of a brown dwarf ( more details on such a case are given later ) .
for white dwarfs , neutron stars and black holes , the net effect is dominated by the microlensing at any orbital separation , and the transit contribution is negligible ( figs . 4 - 5 ) . in order to show the _ kepler _ sensitivity
, it had to be multiplied by factors of 10 and 100 in figs . 4 and 5 , respectively .
the net effects caused by the neutron star and the black hole at 1 au , 5 and 30 millimag respectively , are large enough in principle to be detectable even from ground - based observations .
an enlarged view of the effects of the brown dwarf and jupiter are shown in fig . 6 and fig .
7 , for orbital separations of 0 - 5 au and 0 - 100 au , respectively . as seen from fig .
6 , the microlensing contribution due to a brown dwarf is significant even for small orbital separations and must be taken into account in estimating its radius from the observed light curves . even for a jupiter - mass secondary ,
the microlensing contribution starts to become relevant at orbital separations of 5 au or more at _ kepler _ s sensitivity . from fig .
7 we see that the microlensing contribution from a brown dwarf almost exactly cancels out the transit contribution at an orbital separation of @xmath255 au , beyond which the net amplification is positive .
an enlarged view of the effects of the white dwarf , neutron star and black hole are shown in fig . 8 for small orbital separations of 0 to 0.1 au . for neutron star and black hole secondaries ,
the net amplification is positive at any orbital separation .
for a white dwarf , the microlensing contribution almost exactly cancels out the transit contribution at @xmath20.05 au , beyond which the net amplification is positive .
the effect of microlensing is more dramatic if the size of the source is smaller , as is the case for a brown dwarf .
9 shows such a case where we assume the size of the source to be @xmath81 cm , as applicable for a brown dwarf or a jupiter .
the amplification , which is the ratio of the amplified luminosity and the original luminosity , is typically much larger compared to the earlier case of a solar type source .
the expected amplification is a function of inclination , which is shown in fig .
the solid curve is the expected maximum transit depth , and the dotted curve shows the expected amplification at mid point taking both the transit and the microlensing into account . here
the source ( primary ) is assumed to be a solar type star and lens ( secondary ) is a white dwarf at an orbital separation of 1 au . to show the relative importance of microlensing and transit for various kinds of secondaries at different orbital separations , we have plotted in fig .
11 the orbital radius at which the microlensing signal equals the transit signal as a function of the mass of the secondary .
for example , for a white dwarf secondary , the expected transit depth , given by @xmath82 , where @xmath83 and @xmath35 are the radii of the secondary ( lens ) and the primary ( source ) respectively , is 0.0001 and the mass of the lens is 0.6 m@xmath84 .
the corresponding orbital separation at which the microlensing signal equals the transit signal is 0.046 au . for an earth - mass object ,
the corresponding orbital radius is @xmath22300 au .
the radii of brown dwarfs are similar to the radii of jupiters , so the expected light curves due to their transits would be similar .
thus , a brown dwarf is indistinguishable from a jupiter from its transit signal alone .
however , since the mass of a brown dwarf is considerably larger ( 15 - 80 m@xmath85 ) , the microlensing signal from a brown dwarf would be much larger .
for example , fig .
12 shows the light curves as expected from a jupiter and a 50 m@xmath85 brown dwarf at 1 au .
the dashed curve shows the pure transit for both objects and the solid curve shows the combined effect of the transit and microlensing for a brown dwarf . for a jupiter ,
the combined effect of transit and microlensing is indistinguishable from a pure transit curve . for a brown dwarf
, the combined light curve is different from the transit curve .
note that the expected depth in the observed light curve is smaller for a brown dwarf than for a jupiter , but the shapes of the light curves for both objects are very similar .
thus , there is a degeneracy between the light curves due to a jupiter and a brown dwarf .
it is clear that the contribution due to microlensing must be taken into account in the interpretation of the observed light curves , particularly in estimating the mass / radius of the secondary .
13 shows the light curves for a brown dwarf at 55 au which , if observed , would be easy to interpret as due to a brown dwarf since the light curve due to a jupiter is close to the transit curve itself . notice that there is a slight positive amplification at the ingress ( egress ) since the microlensing begins before ( ends after ) the transit contribution begins ( ends ) .
this curve corresponds to an orbital period of 405 yr , and is hence mostly of pedagogical interest since the probability of observing such an event is very small .
interestingly , white dwarfs have radii similar to earth - mass planets . as mentioned earlier , one of the prime objectives of _ kepler _ is to detect earth - mass planets from their transit signals .
there has been some concern that the earth - mass planets would be confused with white dwarfs since their transit signals would be similar .
( the secondary eclipse may provide a clear test to distinguish between a white dwarf and an earth - mass planet , but the secondary eclipse may not be seen in some cases when the orbital eccentricity is high or because the timescale is such that only the primary eclipse is observed ) . however , as we explained in 4 , in the case of a white dwarf the effect of microlensing will dominate the signal .
14 shows the light curves due to a white dwarf secondary at 0.03 au , 0.0463 au and 0.1 au , with corresponding periods of 1.50 , 2.88 and 9.13 days , respectively .
the dashed curves are the light curves expected from pure transit , and the dotted , solid and dot - dashed curves are the corresponding light curves after taking the microlensing contribution into account . only at orbital separations
smaller than 0.0463 au would the expected @xmath80mag be negative , with the shape of the observed light curve similar to the transit curves .
we note that at an orbital separation of less than 0.01 au , the primary is likely to fill its roche lobe ( eggleton , 1983 ) .
this will cause mass - transfer from the primary , which may cause light variations much larger than the transit or the microlensing effect . also , the primary will be tidally distorted producing distinct ellipsoidal variations .
thus , the transit effect may be difficult to observe for separations below 0.05 au , beyond which the @xmath80mag is expected to be positive .
the expected light curve when the white dwarf is at 1 au is shown in fig . 15 .
in such a case , the white dwarf would be detectable through its positive amplification at @xmath6100 @xmath5-level in observations averaged over a single event with _ kepler _ !
it is of interest to consider how accurately _ kepler _ observations of a white dwarf transiting a solar - like star in a 1 au orbit would constrain the lens mass . for such a canonical case ( shown in fig .
15 ) , the maximum amplification is 0.0024 magnitudes .
the precision per 15 minute sample for an assumed @xmath86 = 12 star observed with _
kepler _ is 90 @xmath1mag .
hence , with 40 samples during the 10 hour lensing event , the maximum amplification will be determined to about 15 @xmath1mag , resulting in a relative error on @xmath60 of 0.6% .
given that the mass of the lens scales as the square of the amplification ( equations 15 - 16 ) , and assuming that four events are available for averaging , the relative error contribution for the lens mass from the microlensing observations is also 0.6% .
however , equations ( 15 ) and ( 16 ) also contain the mass and radius of the source star for which we do not have direct values available from the _ kepler _ observations alone .
radial velocity observations of even moderate quality would provide a further constraint on the mass , leaving uncertainties on the radius of the source star , @xmath35 , as the likely largest contributing error source .
equation ( 17 ) provides a direct constraint on @xmath35 simply from the event width , but the unknown orbital inclination must also be allowed for .
spectroscopic measurements might suffice to constrain @xmath35 at about 10% level , thus the lens mass to 20% level . as was done for hd 209458 ( brown _ et al .
_ 2001 ) , constraints on the source stellar radius may also be derived from the light curve itself , since some subtle details will depend in unique ways on the generally degenerate effect of @xmath35 and @xmath17 in fixing the light curve . equation ( 21 )
provides a further constraint on the lens mass if an estimate of the mass of the primary is available .
note that the orbital inclination and limb darkening parameters can also be estimated from the observed light curve ( as was done for hd 209458 ) , and a detailed treatment of the errors would be required in determining the uncertainties in different physical parameters .
fitting the entire light curve would take all the constraints in a self - consistent manner , providing the best determination of the various parameters , including the lens mass to an estimated accuracy of 1 to 5 percent .
the expected transit signals from a neutron star and a black hole are of the order of nano - magnitudes and smaller .
so the transit signal in these cases can be neglected for all practical purposes . in the treatment of microlensing
, we note that the deflection corresponding to the fainter image ( i.e. the smaller of the two deflections ) is larger than the size of the lens even for small orbital separations .
so the finite size of the lens can also be ignored .
thus the expected light curve is the same as what is expected from the pure microlensing of an extended source and a point lens , e.g. as in fig .
15 with amplitude scaled approximately linearly by the lens mass .
note that the expected @xmath80mag from a neutron star and a black hole at an orbital radius of 1 au are 5 and 28 milli - magnitudes respectively , which may be detectable from ground - based observations .
we have shown that near - field microlensing of main sequence stars by orbiting compact companions provides signals that would be easily detected by a space mission such as _ kepler _ dedicated to the detection of extrasolar terrestrial planets via transits . for terrestrial planets with orbital periods up to a few years to which _ kepler _ has sensitivity to transits , the microlensing contribution to the signal will always be entirely negligible .
jovian class planets in orbits of a few au provide a microlensing contribution that just starts to be relevant , albeit still small compared to the inherent transit signal .
white dwarfs , neutron stars , or black holes transiting a main sequence star in an @xmath21 au orbit would provide easily detectable microlensing signatures that swamp any underlying transit signal .
therefore , although the radius of a white dwarf is similar to that of a terrestrial planet , a transiting white dwarf ( with an orbital separation of 0.25 - 2.0 au , which is the region of prime interest for _ kepler _ ) will not provide a source of false alarms for extrasolar planet detection since the resulting signal will be a significant brightening rather than a small diminution in light .
compact objects like white dwarfs , neutron stars , and black holes are now easily found when they reside in close orbits ( @xmath87 0.1 au ) from a main - sequence companion since they induce ellipsoidal variations in the companion .
furthermore , such systems often become x - ray and radio sources through emissions associated with mass transfer ( even when the mass transfer occurs at low levels ) . while the microlensing signal may be detectable even for an orbital separation as small as @xmath20.1 au , other variations ( such as ellipsoidal variations caused by the distortion of the companion , variations from star spots for induced rapid rotation ) clearly call attention to the presence of a degenerate object in such systems . as an example
, we note that kpd 1930 + 2752 is one such interesting case of a sdb binary in which the unseen companion is likely to be a white dwarf with an orbital period of 0.095 days ( maxted et al .
2000 , billeres et al .
2000 ) .
the microlensing signal for compact objects with orbital separations of @xmath31 au will be easily seen by _
kepler_. in such cases , old white dwarfs , neutron stars and black holes would not already have signaled their presence since the wide separation would not induce the canonical signatures discussed above .
however , as with extrasolar planet transits , microlensing from orbiting compact objects will only be observed in a very small fraction of cases in which the inclinations happen to be ideal so that transits can be observed .
the probability of transits is @xmath88 , where @xmath35 is the radius of the primary , and @xmath54 is the orbital separation . for a solar like star and an object at 1 au
this probability is about 0.005 .
thus if all 100,000 stars in the _ kepler _ sample have compact objects orbiting within 0.25 to 2.0 au , some 500 would be detected through their microlensing signatures . if a more physically reasonable fraction of 1% of stars surveyed contain such compact , massive objects then the number detected by _ kepler _ would be 5 .
what , then , are the prospects for observing compact objects via microlensing signatures by _ kepler _ ?
duquennoy & mayor ( 1991 ) and heacox & gathright ( 1994 ) show that about two - thirds of g - type stars are binaries , and that about 20% of them have periods between 0.25 and 2.0 years .
thus , for the _ kepler _ mission to detect a statistically modest total of 5 systems with microlensing in such circumstances , @xmath27.5% of binary systems must contain a compact object ( most likely a white dwarf ) . by comparison ,
the observed white dwarf fraction in the solar - age open cluster m67 is about 9% by mass @xcite , and since incompleteness may still be important in this case , this is a lower limit to the detected cluster white dwarfs .
this suggests that the overall fraction of white dwarfs is sufficient to yield several detections during the _ kepler _ mission .
estimates of the fraction of baryonic mass in compact objects in the local disk typically span values of 10 30% ( e.g. weidemann , 1990 ) .
this , in turn , implies that the detected number of such microlensing systems will be able to place interesting constraints on the fraction of stars having compact companions . for microlensing detections from systems where the orbital period is known from repeated events
, it will be possible to determine the mass of the compact object directly from the amplitude of the microlensing signal ( assuming independent information on the primary main sequence star is available ) .
for the case of white dwarfs , information on the compact object may also be available if the secondary eclipse is observed , and radial velocity followup observations will be able to determine independent mass estimates . | in this paper , we explore the astrophysical implications of near - field microlensing and its effects on stellar transit observations , with a special emphasis on the _ kepler _ mission .
_ kepler _ is a nasa - approved mission whose goal is to detect a large number of extrasolar , earth - like planets by obtaining near - continuous photometry of @xmath0 f , g , and k dwarfs for four years .
the expected photometric precision of _ kepler _ is 90 @xmath1mag ( achieved in 15 minute samples ) , at which the effect of microlensing by a transiting companion can be significant .
for example , for a solar - type primary transited by a white - dwarf secondary , the maximum depth of the transit is 0.01% , which is almost entirely compensated by the microlensing amplification when the white dwarf is at @xmath20.05 au .
the combined effect of microlensing and transit increases to a net amplification of 150 @xmath1mag at an orbital separation of 0.1 au , and 2.4 millimag at an orbital separation of 1 au .
thus , the effect of microlensing can be used to break the degeneracy between a planetary - mass object for which the microlensing effect is negligible , and a more massive object of the same size . for brown dwarfs at orbital separations of a few au
, the effect of microlensing is several percent of the transit depth , and hence the microlensing effect must be taken into account in deriving the physical parameters of the brown dwarf .
the microlensing signal caused by a neutron star or a black hole in a binary can be several millimag , far exceeding the transit depth , and potentially detectable even from ground - based observations .
_ kepler _ will be sensitive to white dwarfs , neutron stars , and black holes in binaries through their microlensing signatures .
these observations can be used to derive the frequency of such compact objects in binaries , and to determine their masses . |
the minimal standard model(msm ) @xcite @xcite has been proved by all precise experimental data that the msm is a very successful model of particle physics . but until now the symmetric breaking structure of the electroweak interactions has not yet been directly explored experimentally .
so the exploration of the sm higgs boson is a major goal of the present and future colliders .
as we know , any enlargement of the higgs sector beyond the single @xmath4 higgs doublet of the msm necessarily introduces other neutral higgs bosons and charged higgs bosons . like
the general two - higgs - doublet model(2hdm ) , the minimal supersymmetric standard model ( mssm ) @xcite @xcite requires the introduction of two higgs doublets in order to preserve supersymmetry .
these two higgs doublets predict some more elementary higgs bosons : one cp - even neutral higgs boson(@xmath5 ) , one cp - odd neutral higgs boson(@xmath6 ) and two charged higgs bosons(@xmath7 ) , which are absent in the msm .
any experimental discovery of these non - sm - like higgs bosons will be the direct verification of these extended versions of the higgs sector .
therefore , the study of various production mechanisms of the non - sm - like higgs bosons at the present and future colliders is well motivated .
searching for the non - sm - like higgs bosons and studying their properties at the future multi - tev hadron colliders , such as the cern large hadron collider ( lhc ) , are possible as expected by supersymmetric ( susy ) theory @xcite @xcite . the gluon fusion mechanism @xmath8 provides the dominant production mechanism of neutral higgs bosons at the lhc in the entire relevant mass range up to about 1 tev for the small and moderate values of @xmath3 in the mssm @xcite .
the heavy neutral higgs boson can be also produced in pair ( @xmath9 ) at the lhc , if it is kinematically allowed @xcite .
studying the process of a heavy charged higgs boson associated with @xmath10 boson is another attractive way in searching for the @xmath7 bosons , because the @xmath11-boson s leptonic decay may be used as a spectacular trigger .
the calculations of the heavy @xmath7 production associated with @xmath12 boson at a future electron - positron collider can be found in refs .
@xcite @xcite @xcite @xcite @xcite .
the complete calculations of the @xmath13 associated production at hadron colliders both in the 2hdm and the mssm are given in refs .
@xcite @xcite @xcite @xcite @xcite .
analogously , @xmath14 associated production would also be an efficient way in searching for the heavy neutral cp - odd higgs boson @xmath6 .
although the @xmath6 boson can be produced in pair at future colliders @xcite @xcite , the @xmath14 associated production will be the kinematically favored mechanism to produce @xmath6 higgs boson for the heavy @xmath6 higgs boson . and
again the leptonic decay of @xmath15 maybe benefit for triggering the @xmath16 associated production events .
the calculations of the @xmath16 associated production at a electron - positron collider were presented in refs .
@xcite @xcite , at a muon collider in ref .
@xcite and at a photon collider in ref .
@xcite , respectively . and
chung kao gave the calculation of @xmath17 associated production via @xmath18 fusion including only quark loop diagrams at the ssc @xcite .
in this paper we concentrate on studying the @xmath17 associated production at the lhc in the mssm , considering both subprocesses @xmath19 and @xmath20 . in the calculation of the loop mediated process @xmath21 , we compare and discuss the cross sections in the general two - higgs - doublet model ( 2hdm ) and the mssm . in section
ii , we present the calculation of the processes @xmath22 and @xmath23 . numerical results and discussion
are given in section iii .
there we use the mssm parameters constrained within the minimal supergravity ( msugra ) scenario @xcite . finally , a short summary is given .
in our calculation we use the thooft - feynman gauge and adopt the dimension regularization scheme in the general 2hdm and the dimensional reduction ( dr ) scheme @xcite in the mssm . in the loop diagram calculation we adopted the definitions of one - loop integral functions in reference @xcite .
the numerical calculation of the vector and tensor loop integral functions can be traced back to scalar loop integrals as shown in the reference@xcite .
the feynman diagrams and the relevant amplitudes are created by feynarts package automatically @xcite .
the numerical calculation of the loop integrals are implemented by using mathematica programs .
we denote the @xmath17 associated production via drell - yan subprocess as @xmath25 due to the feature of the yukawa coupling that the coupling strength between quarks and higgs boson is in proportion to the correspondent quark mass , the cross sections of subprocesses @xmath26 should be much smaller than those of the subprocesses @xmath27 . considering the fact that the luminosity of top ( anti - top ) quark is much lower than that of bottom ( anti - bottom ) quark from a proton
, we conclude that the cross section of the process @xmath28 is approximately equal to the cross section of @xmath29 .
therefore , in this paper we consider only the contributions from the @xmath30 process .
the feynman diagrams of the subprocess @xmath31 at the lowest order are depicted in fig.1 .
the differential cross section of the subprocess @xmath32 can be expressed as @xmath33 where the summation is taken over the spins of the initial and final states , and @xmath34 denotes the two - particle phase space element .
the factor @xmath35 in above equation comes from the averaging over the spins and the colors of the incoming partons .
the matrix element @xmath36 represents the amplitude of the @xmath37(@xmath5 ) exchanging s - channel diagrams(shown in fig.1(a ) ) , @xmath38 corresponds to the amplitude of u- and t - channel diagrams ( shown in fig.1(b ) ) .
the mandelstam kinematical variables are defined as @xmath39 by using the relevant feynman rules , we obtain the explicit expressions of these amplitudes : @xmath40 \frac{(k_1+p_1+p_2)^{\mu}}{\hat{s}-m_{h}^2+m_h\gamma_h i } { \nonumber}\\ & & -\frac{(4\pi\alpha_s)m_b}{4s_{w}^2c_{w}^2}\sin(\beta-\alpha ) \frac{\cos\alpha}{\cos\beta}\epsilon^\mu(k_1 ) \left[\bar{v } ( p_2 ) u(p_1)\right ] \frac{(k_1+p_1+p_2)^{\mu}}{\hat{s}-m_{h}^2+m_h\gamma_h i } \\
a_{(b)}(\hat{s},\hat{t } , \hat{u } ) & = & -\frac{(4\pi\alpha)m_b}{2m_w s_{w}^2c_{w}}\tan\beta\epsilon^\mu(k_1)\frac{1 } { \hat{t}-m_b^2}\left[\bar{v } ( p_2)\gamma^5(m_b-\rlap/{k_1}+\rlap/{p_1})\gamma^\mu(\frac{s_{w}^2}{3}-\frac{p_l}{2 } ) u(p_1)\right ] { \nonumber}\\ & & -\frac{(4\pi\alpha)m_b}{2m_w s_{w}^2c_w}\tan\beta\epsilon^\mu(k_1)\frac{1 } { \hat{u}-m_b^2}\left[\bar{v } ( p_2)\gamma^5(m_b+\rlap/{k_1}-\rlap/{p_2})\gamma^\mu(\frac{s_{w}^2}{3}-\frac{p_l}{2 } ) u(p_1)\right ] { \nonumber}\end{aligned}\ ] ] where @xmath41 and @xmath42 represent the masses of bottom quark and w boson , respectively .
we denote the @xmath17 associated production process via gluon fusions as @xmath44 where @xmath45 , @xmath46 are the color indices of initial gluons . as the subprocess
@xmath47 is loop - induced , the one - loop order calculation can be simply carried out by summing all unrenormalized reducible and irreducible one - loop diagrams and the results will be finite and gauge invariant .
we denote @xmath48 and @xmath49 as the cross sections in the framework of the general 2hdm and the mssm , respectively .
the former is contributed by the feynman diagrams involving only the quark loop diagrams(shown in fig.2 ) and the latter involves the contributions of both the quark and squark loop diagrams(shown in fig.2 - 3 ) .
the possible corresponding feynman diagrams created by exchanging the initial gluons or the two final states , should be also included in fig.2 and fig.3 and involved in our calculation .
we can see that each feynman diagram in fig.2 and fig.3 contains one interacting vertex between ( s)quarks and a higgs boson .
due to the feature of the yukawa coupling as we mentioned above , we can consider only the diagrams which involve the third generation ( s)quark in the calculation of the subprocess @xmath50 .
the cross sections of the subprocess @xmath20 in the general 2hdm and the mssm can be expressed respectively as @xmath51 where the summation is taken over the spins and colors of the initial and final states , and @xmath34 denotes the two - particle phase space element .
@xmath52 represents the amplitude of the diagram of fig.i(j ) .
the factor @xmath53 results from the averaging over the spins and the colors of the incoming partons . with the cross sections of the related subprocesses , the cross section of parent process @xmath54 at the proton - proton collider lhc can be obtained by doing the following integration , @xmath55 where @xmath56 + \left [ f_{j / p}(j , x_1,q^2)f_{i / p}(i,\frac{\tau}{x_1},q^2 ) \right]\right\}\end{aligned}\ ] ] in eq.([integration ] ) @xmath57 and @xmath58 are the colliding proton - proton and parton - parton c.m.s .
energies respectively .
the notation @xmath59 represents the cross section of the parent process @xmath60 .
@xmath61 is the luminosity of incoming partons where @xmath62 can be @xmath63 and @xmath64 , @xmath65 .
@xmath66 and @xmath2 represent the masses of @xmath15 boson and @xmath6 higgs boson .
the definitions of @xmath67 and @xmath68 can be found in ref.@xcite . in our calculation
, we adopt the cteq5 parton distribution function @xcite and take the factorization scale q to be @xmath58 .
the eq.([integration ] ) can be rewritten as @xmath69 where @xmath70 + ( i \leftrightarrow j ) \right\}\end{aligned}\ ] ] when @xmath71 , @xmath48 represents the cross section of the parent process @xmath72 contributed only by quark loop diagrams shown in fig.2 , while @xmath49 represents the cross section contributed by both quark and squark loop diagrams shown in fig.2 and fig.3 . in the next section
we shall take different input data sets to demonstrate the production rates of the parent process @xmath73 .
the numerical results of @xmath48 and @xmath49 would show the importance of squark loop diagrams .
the total cross section of @xmath74 at proton - proton collider should be the summation of @xmath75 and @xmath76 .
quantitatively comparing the @xmath75 with @xmath76 will help us to know in which part of the parameter space the contribution of gluon - gluon fusion process is dominant .
in the numerical calculation , we take the sm parameters as : @xmath77gev , @xmath78gev , @xmath79gev , @xmath80gev @xcite , and take the supersymmetric parameters being constrained within the minimal supergravity ( msugra ) scenario @xcite . in this scenario ,
only five sypersymmetric parameters should be inputed , namely @xmath81 , @xmath82 , @xmath83 , sign of @xmath84 and @xmath3 , where @xmath81 , @xmath82 and @xmath83 are the universal gaugino mass , scalar mass at gut scale and the trilinear soft breaking parameter in the superpotential terms , respectively . in this work , we take @xmath81=120 gev , @xmath83=300 gev and @xmath85 . @xmath82 is obtained quantitatively from the input @xmath86 value .
all other mssm parameters are determined in the msugra scenario by using program package isajet 7.44 . in this program ,
the renormalization group equations ( rge s ) @xcite are run from the weak scale @xmath66 up to the gut scale , taking all thresholds into account in order to get the low energy scenario from the msugra .
it uses two loop rge s only for the gauge couplings and the one - loop rge s for the other supersymmetric parameters . the gut scale boundary conditions are imposed and the rge s are run back to @xmath66 , again taking threshold into account .
here we give some comments about the choice of the decay width values of cp - even neutral higgs bosons @xmath37 and @xmath5 .
we know that some of the feynman diagrams ( shown in fig.1 - 3 ) have s - channel @xmath37 and @xmath5 propagators , which have analytical expressions respectively as @xmath87 @xmath88 it is clear that the cross sections of the subprocess should related to the decay widths of @xmath37 and @xmath5 . in this work
the input parameter @xmath2 is taken in the range of 200 gev to 650 gev .
then we have the following constraints in this parameter space , @xmath89 and by using the package hdecay@xcite in the mssm , we find @xmath90 because @xmath91 , we get @xmath92 .
the propagator of @xmath5 boson can be expressed approximately as @xmath93 it is obvious that the bigger the @xmath94 is , the less sensitive the cross section to the decay widths of neutral higgs bosons @xmath5 and @xmath37 is .
therefore , we choose @xmath95 gev in our numerical calculations .
actually , our final numerical result of the cross section of the process @xmath74 at the lhc , shows also that it is not sensitive to the choice of these two decay widths .
the figures in fig.4 , fig.5 and fig.6 , show the cross sections ( or differential cross sections ) of the process @xmath96 at the lhc as the functions of the cp - odd higgs boson @xmath6 mass , the ratio of the vacuum expectation values @xmath97 and the transverse momentum @xmath98 , respectively .
the curves of the cross sections ( or differential cross sections ) involving the contributions from quark loop diagrams ( in the general 2hdm ) and quark+squark loop diagrams ( in the mssm ) are depicted separately on these figures for comparison of the cross sections in these two models .
and in these three figures the full - lines are for @xmath48 ( or @xmath99 ) , the dotted - lines are for @xmath49 ( or @xmath100 ) .
figure 4 shows the relationship between the cross section of the parent process @xmath101 and @xmath2 with the colliding energy @xmath102tev .
the input msugra parameters are set to be the typical values mentioned in the last subsection ( i.e. @xmath81=120 gev , @xmath83=300 gev and @xmath85 .
@xmath82 is obtained quantitatively from the @xmath86 value ) , and @xmath97=2 , 7 and 32 , respectively . from this figure
, we find that in some parameter space the scalar quark contributions can enhance the cross section obviously , that is to say @xmath103 , while in other regions , we have @xmath104 .
the figure shows that when we have small and moderate @xmath3 values , the scalar quark loop contribution to the @xmath17 associated production at the lhc is most obvious .
we shall also see later from fig.7 that when @xmath105 has small or moderate value , the contributions from gluon fusion subprocess is dominant .
therefore , it is possible to use the experimental measurement of the @xmath17 associated production at the lhc to disentangle the mssm from the general 2hdm in these parameter space regions . in fig.5 ,
the cross section of the parent process @xmath106 at the lhc versus @xmath97 is plotted .
the values of the neutral cp - odd higgs boson @xmath6 mass are set to be 200 gev , 400 gev and 600 gev , respectively .
from the figure , we also find that scalar quark contributions can either enhance or suppress the cross section of the parent process as shown in fig.4 .
fig.5 together with fig.4 , show that when the value of @xmath2 is greater than 400 gev and @xmath107 , the contribution from the scalar quark loop diagrams increases with the decrement of @xmath97 . in fig.5
the two curves for @xmath108 and @xmath109 demonstrate that when @xmath107 , the scalar quark contribution suppresses the cross section , which means @xmath110 , and while the scalar quark contribution enhances the cross section when @xmath111 .
these features can be also seen from fig.4 .
the curve for @xmath112 in fig.4 , demonstrates that @xmath49 is about two third of @xmath48 quantitatively , while the curve for @xmath113 shows @xmath103 .
fig.6 displays the differential cross section @xmath114 of the process @xmath115 at the lhc versus transverse momentum @xmath98 with @xmath116 and the pseudo - rapidity being in the range of @xmath117 .
the @xmath6 mass is set to be 350 gev , and @xmath97 is taken as 2 , 7 and 32 , respectively .
we find that for @xmath112 , the scalar quark contribution suppresses the differential cross section @xmath114 . for @xmath113 and @xmath118gev ,
the scalar quark contribution enhances the differential cross section .
but for @xmath119 , the scalar quark contribution can either enhance or suppress the differential cross section in different @xmath98 regions . in fig.7 , fig.8 and fig.9 , we plot the the cross sections of the process @xmath120 ( @xmath121 ) , and the process @xmath74 contributions from both drell - yan and gluon - gluon fusion subprocesses in the constrained mssm ( @xmath122 ) , as the functions of the cp - odd higgs boson @xmath6 mass , the ratio of the vacuum expectation values @xmath97 and the transverse momentum @xmath98 , respectively . in these three figures ,
the full - lines are for the cross sections or differential cross sections of the process @xmath120 via drell - yang subprocess , the dotted - lines are for the process @xmath123 via both drell - yang and gluon fusion subprocesses . with the comparison between the @xmath124 ( or @xmath125 ) and @xmath121 ( or @xmath126 )
, we can know in which parameter space in the constrained mssm , the contribution from the loop mediated subprocess @xmath127 is important .
fig.7 displays the cross sections of @xmath128 and @xmath129 at proton - proton colliders versus the mass of @xmath6 with @xmath102tev .
we choose @xmath97=2 , 7 and 32 , respectively . from fig.7
, we find that in the region of @xmath130 , the contribution of gluon - gluon fusion subprocess in the mssm enhances the cross section , especially when @xmath131 the contribution of the gluon - gluon fusion subprocess is about @xmath132 of the total cross section @xmath124 .
in fact , the gluon - gluon fusion subprocess is the most important @xmath17 associated production mechanism in this parameter space . from the figure we see also that when @xmath97=32 the difference between @xmath121 and @xmath124 is very small , it means that the contribution of the gluon - gluon fusion subprocess is negligible in this parameter space .
the cross sections of @xmath133 and @xmath134 at the lhc as the functions of @xmath97 with @xmath116 are shown in fig.8 .
the mass of higgs boson @xmath6 is taken as 200 gev , 400 gev and 600 gev , respectively . from this figure
we can find also that gluon - gluon fusion subprocess enhances the cross section of the @xmath17 associated production at the lhc , and will become a very important production mechanism when @xmath135 . in the region of @xmath136 , the cross section of the @xmath17 associated production at the lhc are in the range of @xmath137 fb .
even the @xmath124 can reach 300 fb when @xmath138 and @xmath139gev . so the @xmath17 associated production process may be easily observed experimentally if @xmath97 is large enough .
fig.9 displays the differential cross sections ( @xmath114 ) of @xmath129 and @xmath133 at the lhc as the functions of the transverse momentum @xmath98 with the pseudo - rapidity being in the range of @xmath117 .
we choose @xmath140 gev , and take @xmath97=2 , 7 and 32 , respectively . from this figure
we can see that at high @xmath98 region , when @xmath130 , the difference between @xmath126 and @xmath125 is obvious , even when @xmath97=2 , the @xmath126 can be less than @xmath141 of @xmath125 , which means that the contribution from the @xmath142 process is dominant in this parameter space .
but when @xmath143 , the contribution to the total differential cross sections ( @xmath125 ) is mainly from the drell - yan @xmath17 associated production subprocess , and the contribution from gluon fusion subprocess is negligible .
in this paper , we studied the neutral cp - odd higgs boson @xmath6 production with the association of @xmath15 gauge boson via both drell - yan and gluon - gluon fusion subprocesses in the constrained mssm at the cern lhc .
numerical analysis of their production rates is carried out with some typical parameter sets in the msugra scenario .
our results show that the cross section in the mssm is clearly enhanced by the gluon - gluon fusion subprocess in the parameter space with small or moderate @xmath97 value , and we should consider the gluon - gluon fusion subprocess in this parameter space in the calculation of the @xmath17 associated production at the lhc .
we compared above results of the process @xmath144 in the mssm with those in the general two - higgs - doublet model ( 2hdm ) , where the cross section of subprocess @xmath145 is contributed only by quark loop diagrams .
we find that the contributions from the scalar quark loops in the mssm can either enhance or suppress the cross section obviously and can not be neglected in some parameter space .
the results show also that the @xmath17 associated production at the lhc is strongly related to the parameters @xmath97 and the mass of @xmath6 .
the total cross section increases with increment of @xmath97 , and decreases with increment of @xmath2 .
s25 s. l. glashow , nucl .
* 22 * ( 1961 ) 579 ; s. weinberg , phys .
* 1 * ( 1967 ) 1264 ; a. salam , proc .
8th nobel symposium stockholm 1968 , ed .
n. svartholm(almquist and wiksells , stockholm 1968 ) p.367 ; h. d. politzer , phys . rep .
* 14 * ( 1974 ) 129 .
p. w. higgs , phys .
lett * 12 * ( 1964 ) 132 , phys .
* 13 * ( 1964 ) 508 ; phys . rev .
* 145 * ( 1966 ) 1156 ; f. englert and r.brout , phys .
* 13 * ( 1964 ) 321 ; g. s. guralnik , c. r. hagen and t. w. b. kibble , phys .
* 13 * ( 1964 ) 585 ; t. w. b. kibble , phys . rev .
* 155 * ( 1967 ) 1554 .
h. e. haber and g. l. kane , phys .
* 117 * ( 1985 ) 75 .
j. f. gunion , h. e. haber , g. l. kane and s. dawson , the higgs hunter s guide ( addison - wesley , reading 1990 ) .
h. nilles , phys . rep .
* 110 * ( 1984 ) 1 ; j. rosiek , phys . rev .
* d41 * ( 1990 ) 3464 , hep - ph/9511250(e ) ; m. kuroda , hep - ph/9902340 .
j. f. gunion , h. e. haber and c. kao , phys . rev . *
d46 * , 2907 ; m. spira , fortschr
* 46 * ( 1998)203 ; m. spira , a. djouadi , d. graudenz and p. m. zerwas , nucl .
* b453 * ( 1995 ) 17 .
barrientos - bendezu and b.a .
kniehl , phys .
* d64 * ( 2001 ) 035006 .
a. arhrib , m.c .
peyranere , w. hollik and g. moultaka , nucl .
* b581 * ( 2000 ) 34 . s. kanemura , eur .
j. * c17 * ( 2000 ) 473 .
zhu , hep - ph/9901221 .
heather e. logan , shufang su , hep - ph/0203270 .
f. zhou , w.g .
ma , y. jiang , x.q .
li and l.h .
wan , phys .
* d64 * ( 2001 ) 055005 .
d. a. dicus , j. l. hewett , c. kao and t.g .
rizzo , phys .
* d40 * ( 1989 ) 789 .
a. a. barrientos bendezu and b.a .
kniehl , phys . rev .
* d59 * ( 1998 ) 015009 .
barrientos bendezu and b.a .
kniehl , phys . rev .
* d61 * ( 2000 ) 097701 .
f. zhou , w.g .
ma , y.jiang , l. han and l.h .
wan , phys . rev .
* d62 * ( 2000 ) 015002 .
o. brein , w. hollik and s. kanemura , phys . rev .
* d63 * ( 2001 ) 095001 .
gounaris , p.i .
porfyriadis , eur .
j. * c18 * ( 2001 ) 181 , hep - ph/0007110 .
akeroyd , a. arhrib and m. capdequi peyranere , mod .
* a14 * ( 1999 ) 2093 ; mod .
* a17 * ( 2002 ) 373 , hep - ph/9907542 .
akeroyd , a. arhrib and m. capdequi peyranere , phys . rev .
* d64 * ( 2001 ) 075007 ; phys . rev . *
d65 * ( 2002 ) 099903 , hep - ph/0104243 .
akeroyd , a. arhrib and c. dove , phys . rev .
* d61 * ( 2000 ) 071702 .
gounaris , p.i .
porfyriadis and f.m .
renard , eur .
j. * c20 * ( 2001 ) 659 .
chung kao , phys . rev .
* d46 * ( 1992 ) 4907 , fsu - hep-911205 .
m. drees and s.p .
martin , mad - ph-879 , um - th-95 - 02 , hep - ph/9504324 .
d.m . copper , d.r.t .
jones , p. van nieuwennuizen , nucl .
* b167 * ( 1980 ) 479 ; bernd a. kniehl , phys . rep .
240(1994)211 . g. passarino and m. veltman , nucl . phys .
* b160 * , 151(1979 ) .
j. kblbeck , m. bhm and a. denner , comput .
* 60 * ( 1990 ) 165 ; t. hahn , hep - ph/9905354 .
y. jiang , w.g .
ma , l. han , z.h .
yu and h. pietschmann , phys
* d62 * ( 2000 ) 035006 .
lai , j. huston , and s. kuhlmann , eur .
j. * c12 * ( 2000 ) 375 , hep - ph/9903282 .
d.e . groom , et al . ,
j. c15 , ( 2000 ) 1 ; d. schaile , cern - ppe/94 - 162(11 october 1994 ) .
v. barger , m.s .
berger and p. ohmann , phys . rev .
* d47 * ( 1993 ) 1093 , * d47 * ( 1993 ) 2038 ; v. barger , m.s .
berger , p. ohmann and r.j .
n. phillips , phys . lett .
* b314 * ( 1993 ) 351 ; v. barger , m.s .
berger and p. ohmann , phys .
* d49 * ( 1994 ) 4908 .
m. spira , nucl .
instr . and meth .
a389 * ( 1997 ) 357 ; a. djouadi , j. kalinowski and m. spira , comput .
. commun . * 108 * ( 1998 ) 56 . * fig.1 * the relevant feynman diagrams for the subprocess @xmath146 in the mssm at the tree - level : ( a ) s - channel diagrams .
( b ) u- and t - channel diagrams .
note that fig.1(b ) includes the diagram created by exchanging two final states . * fig.6 * the differential cross sections @xmath149 and @xmath150 of the process @xmath142 , as the functions of the transverse momentum @xmath98 in the msugra scenario at the lhc with @xmath102tev , @xmath151gev and the pseudo - rapidity being in the range of @xmath117 .
the ratio of the vacuum expectation values @xmath3 is taken as 2 , 7 and 32 , respectively . *
fig.7 * the cross sections @xmath121 and @xmath124 of the process @xmath129 as the functions of the mass of higgs boson @xmath6 , the ratio of the vacuum expectation values @xmath3 is taken as 2 , 7 and 32 , respectively . * fig.9 * the differential cross sections @xmath126 and @xmath125 of the process @xmath129 , as the functions of the transverse momentum @xmath98 in the msugra scenario with @xmath102tev , @xmath152 gev and the pseudo - rapidity being in the range of @xmath117 .
the ratio of the vacuum expectation values @xmath3 is taken as 2 , 7 and 32 , respectively . | we investigate in detail the @xmath0 associated production process @xmath1 within the framework of the minimal supersymmetric standard model ( mssm ) at the cern large hadron collider ( lhc ) , considering both contributions from the drell - yan and gluon fusion subprocesses .
we focus on the deviations from the general two - higgs - doublet model ( 2hdm ) arising in the mssm .
we also discuss the contributions of the two @xmath0 associated production subprocesses in the mssm at the lhc , and analyse the dependences of the total cross section on neutral cp - odd higgs boson mass @xmath2 and @xmath3 in the msugra scenario .
we find that the contribution from loop mediated gluon fusion subprocess can be competitive with that from the drell - yan subprocess in some parameter space .
= 0.36 in addtoresetequationsection |
the proximity effect is an important quantum phenomenon which occurs when a superconductor is attached to non - superconducting materials , where the pair amplitude of cooper pairs in the superconductor penetrates into the non - superconducting materials @xcite .
a typical example is the josephson effect , which has been known as one of the macroscopic quantum phenomena , characterized as dc current flowing without a voltage - drop between two superconductors separated by a thin insulator ( _ i _ ) or normal metal ( _ n _ ) @xcite .
the josephson critical current in an superconductor / insulator / superconductor or superconductor / normal metal / superconductor junction forming a josephson junction monotonically decreases with increasing the thickness of _ i _ or _
n _ @xcite .
the proximity effect in @xmath0-wave superconductor / ferromagnetic metal ( _ s_/_f _ ) hybrid junctions has been extensively studied in the last decade because of its fascinating phenomena and potential applications to superconducting spintronics @xcite . due to the proximity effect between _ s _ and _ f _ in a _ s_/_f _
junction , the spin - singlet cooper pairs ( sscs ) penetrate into the _ f _ and acquire a finite center - of - mass momentum proportional to the exchange splitting between up- and down - spin bands in the _
f_. the pair amplitude of ssc shows damped oscillation with increasing the thickness of _
f_. one interesting phenomena induced by the damped oscillatory behavior of the pair amplitude of ssc is a @xmath21-state in a _ s_/_f_/_s _
junction , where the current - phase relation in the josephson junction is shifted by @xmath21 from that of the ordinary _ s_/_i_/_s _ or _
s_/_n_/_s _ junction ( called 0-state ) @xcite .
it is expected that the @xmath21-state can be utilized for an element of quantum computing and circuit @xcite . another intriguing proximity effect in _
s_/_f _ hybrid junctions is the emergence of odd - frequency spin - triplet cooper pairs ( stcs ) , although the @xmath7 is an @xmath0-wave superconductor @xcite . here , the anomalous green s functions of spin - triplet components are odd functions with respect to the fermion matsubara frequency @xmath22 .
it should be noted that the anomalous green s functions in bulk superconductors are generally even functions with respect to @xmath22 . when the magnetization in the _ f _ is uniform in a
_ s_/_f _ junction , not only the ssc , as described above , but also the stc composed of opposite spin electrons ( i.e. , total spin projection on @xmath23 axis being @xmath24 ) penetrates into the _ f _ due to the proximity effect @xcite .
the penetration length of stc with @xmath24 ( and also ssc ) into the _
f _ is very short and the amplitude of stc exhibits a damped oscillatory behavior inside the _ f _ with increasing the thickness of _ f_. the penetration length is determined by @xmath25 , which is typically a order of few nanometers @xcite . here ,
@xmath26 and @xmath27 are the diffusion coefficient and the exchange field in the _ f _ , respectively . on the contrary , when the magnetization in the _ f _ is non - uniform in a _
junction , the stc formed by electrons of equal spin ( @xmath28 ) can also be induced in the _ f_.
this includes cases , for instance , where the @xmath8 contains a magnetic domain wall @xcite , the junction consists of _ f _ multilayers @xcite , the interface of @xmath7/@xmath8 junction is spin active @xcite , and the ferromagnetic resonance occurs @xcite .
although the pair amplitude of stc with @xmath28 monotonically decreases with increasing the thickness of _ f _ , the stc with @xmath28 can propagate into the _ f _ over a distance of the order of @xmath29 ( @xmath30 : temperature ) , which is typically about several dozen nanometers @xcite .
this is approximately 2 orders of magnitude longer than the penetration length of the ssc and the stc with @xmath24 .
therefore , the proximity effect of stcs with @xmath28 is called the long - ranged proximity effect ( lrpe ) . following the theoretical predictions ,
the stc in _ s_/_f _ hybrid junctions has been confirmed experimentally @xcite . the obvious way to observe the lrpe induced by the stc with @xmath28 is to directly measure the josephson current in josephson junctions composed of @xmath8s @xcite .
indeed , the lrpe has been observed in @xmath7/@xmath8 junctions with spin - active interfaces @xcite and in @xmath7/@xmath8 multilayer systems with non - collinear magnetization alignment between @xmath8 layers @xcite .
recently , the variation of superconducting transition temperature ( @xmath31 ) has been observed in _
s_/_f_1/_f_2 type spin valve structures as the direction of magnetizations in the two ferromagnetic metals @xmath81 and @xmath82 is changed @xcite .
this is also due to the lrpe induced by the stc as predicted in the previous theoretical calculation @xcite .
an alternative way to prove the stc is to measure the spin angular momentum carried by cooper pairs because the spin is finite for the stc but is zero for the ssc .
several theoretical studies have already addressed this issue and examined the magnetization induced by the stc in the various geometry of _ s_/_f _ hybrid structures @xcite .
f_/_s_/_f _ junction with a spin valve structure is a typical geometry of such _ s_/_f _ hybrid structures .
when the magnetizations in the two @xmath8s separated by the @xmath7 are non - collinearlly aligned , not only the stc with @xmath32 but also the stc with @xmath28 becomes finite and induces a finite magnetization inside the _ s _ as well as the two _
f_s @xcite .
recently , the magnetization induced by the stc has also been studied in josephson junction type multilayer systems , e.g. , _ s_/_f_/_f_/_s _ , _ s_/_f_/_f_/_s_/_f _ , and rather complex symmetric three terminal
_ s_/_f_/_f_/_s_/_f_/_f_/_s _ junctions
it has been pointed out that such josephson junctions with metallic ferromagnetic multilayers , especially , the symmetric three terminal _
s_/_f_/_f_/_s_/_f_/_f_/_s _ junction may have promising potential for superconducting spintronics applications with low dissipation @xcite .
this is because the magnetization in this junction can be well controlled by changing the superconducting phase difference between the two outmost _
s_s without jule heating . here
, it should be noted that the thickness @xmath33 of s in the middle layer sandwiched by the two ferromagnetic double layers has to be @xmath34 ( @xmath35 : superconducting coherence length ) in order to observe clearly the magnetization in the middle @xmath7 layer induced by the stc @xcite .
however , in this case , the superconductivity in the middle _ s _ layer is violently suppressed . to prevent this from happening , for example , a three terminal josephson junction composed of large superconducting electrodes in the middle @xmath7 layer
is proposed @xcite . in this paper , we focus on a much simpler josephson junction with metallic trilayers , i.e. , a _ s_/_f_1/_n_/_f_2/_s _ josephson junction ( see fig . [ sfnfs - gm ] ) , which nowadays has been able to be fabricated experimentally @xcite , and theoretically examine , by employing the quasiclassical green s function method , the magnetization inside the @xmath1 induced by the odd - frequency stcs composed of electrons of equal and opposite spins . fixing the magnetization in @xmath36 along the @xmath23 direction perpendicular to the junction direction ( @xmath37 direction ) ,
we show that i ) the @xmath37 component of the magnetization in the @xmath1 is always zero , ii ) the @xmath38 component becomes exactly zero when the magnetizations in @xmath81 and @xmath82 are collinear , and iii ) the @xmath23 component is generally finite for any magnetization alignment between @xmath81 and @xmath82 .
we also show that the magnetization in the @xmath1 can be decomposed into two parts , @xmath4 dependent and independent parts , where @xmath4 is the superconducting phase difference between the two @xmath7s in the @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 junction .
the @xmath4 dependent magnetization is induced as a result of finite coupling between the two @xmath7s , while the @xmath4 independent magnetization always exists due to the proximity effect in @xmath7/@xmath8 hybrid junctions .
we find that the @xmath4 independent magnetization decreases slowly with increasing the thickness of @xmath1 , whereas the @xmath4 dependent magnetization decays rather rapidly .
we also investigate the dynamics of the magnetization in the @xmath1 when ac voltage is applied .
because of the ac voltage , the superconducting phase difference @xmath4 is now time dependent and accordingly the @xmath4 dependent part of the magnetization oscillates .
however , we find that the @xmath4 dependent part of the magnetization is converted , when it is time averaged , from the oscillating ac character to the time - independent dc character at specific dc voltages , depending on the frequency of ac voltage , if dc and ac voltages are both applied to the @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 junction .
finally , we argue that the magnetization induced inside the @xmath1 can be large enough to be observed experimentally in realistic settings .
the rest of this paper is organized as follows . in sec .
[ sec : formulation ] , we introduce a simple @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 junction consisting of metallic trilayers and derive the analytical formulation of the magnetization induced inside the @xmath1 on the basis of usadel equation in the diffusive transport limit .
it is clear from this analytical formulation that the magnetization in the @xmath1 is indeed induced by the odd - frequency stcs . in sec .
[ sec : result ] , we show the results of the magnetization as a function of the thickness of @xmath1 for different magnetization alignments of the two @xmath8s .
the @xmath4 dependence of magnetization , including the dynamics when ac voltage is applied , is also discussed .
finally , the magnetization induced by the stcs is estimated for a typical set of realistic parameters in sec .
[ sec : discussion ] .
the summary of this paper is given in sec .
[ sec : summary ] .
the spatial dependence of anomalous green s functions in the @xmath1 is discussed in appendix [ app : agf ] and the local magnetization density induced inside the @xmath1 is examined in appendix [ app : lmd ] .
after introducing the josephson junction studied , we first formulate for this junction the anomalous green s functions in the diffusive transport limit on the basis of the quasiclassical green s function method and then derive the analytical formulae of the magnetization induced inside the @xmath1 . as depicted in fig .
[ sfnfs - gm ] , we consider the @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 junction made of normal metal ( @xmath1 ) sandwiched by two layers of ferromagnetic metal ( @xmath81 and @xmath82 ) attached to @xmath0-wave superconductors ( @xmath7s ) .
we assume that the magnetization in @xmath82 is fixed along the @xmath23 direction perpendicular to the junction direction ( @xmath37 direction ) , while the @xmath81 is a free layer in which the magnetization can be controlled by an external magnetic field , pointing any direction in the @xmath39 plane , parallel to the interfaces , with @xmath40 being the polar angle of the magnetization .
we also assume that the magnetizations in @xmath81 and @xmath82 are both uniform .
the thicknesses of @xmath7 , @xmath81 , @xmath82 , and @xmath1 are @xmath33 , @xmath41 , @xmath42 , and @xmath5 , respectively , with @xmath43 , @xmath44 , and @xmath45 .
furthermore , we assume that @xmath33 is much larger than the superconducting coherent length @xmath35 .
/@xmath81/@xmath1/@xmath82/@xmath7 junction studied , where the normal metal ( @xmath1 ) is sandwiched by two ferromagnetic metals ( @xmath81 and @xmath82 ) attached to @xmath0-wave superconductors ( @xmath7s ) .
arrows in @xmath81 and @xmath82 indicate the direction of ferromagnetic magnetizations .
while the magnetization in @xmath82 is fixed along the @xmath23 direction , the @xmath81 is assumed to be a free layer in which the magnetization can be controlled by an external magnetic field within the @xmath39 plane with @xmath40 being the polar angle of the magnetization .
@xmath33 , @xmath41 , @xmath42 , and @xmath5 are the thicknesses of @xmath7 , @xmath81 , @xmath82 , and @xmath1 , respectively , with @xmath43 , @xmath44 , and @xmath46 .
we assume that the magnetizations are uniform in both @xmath81 and @xmath82 layers , and that @xmath47 .
, width=226 ] in the diffusive transport region , the magnetization inside the @xmath1 is evaluated by solving the linearized usadel equation in each region @xmath48 ( @xmath49 @xmath81 , @xmath1 , and @xmath82 ) @xcite , @xmath50 = \hat 0 \label{usadel},\ ] ] where @xmath51 is the diffusion coefficient in region @xmath48 , @xmath52 with @xmath53 is the fermion matsubara frequency , @xmath54 , and @xmath55 is the @xmath38 ( @xmath23 ) component of pauli matrix .
we assume that diffusion coefficients in @xmath81 and @xmath82 are the same , i.e. @xmath56 . note also that @xmath57 , @xmath58=\hat{q}\hat{r}-\hat{r}\hat{q}$ ] , and @xmath59 is null matrix .
the anomalous part @xmath60 of the ( @xmath61 ) quasiclassical green s function @xcite is given by @xmath62 where the @xmath63 dependence is implicitly assumed .
notice that @xmath64 is the anomalous green s function for the ssc , whereas @xmath65 and @xmath66 represent the anomalous green s functions for the stc with @xmath28 and @xmath32 , respectively .
the @xmath0-wave superconducting gap @xmath67 is finite only in the @xmath7 and assume to be constant , i.e. , @xmath68 the exchange field @xmath69 due to the ferromagnetic magnetization in the _ f_s is described by @xmath70 where @xmath71 , @xmath72 ( see fig .
[ sfnfs - gm ] ) , and @xmath73 is a unit vector in the @xmath38 ( @xmath23 ) direction .
we assume that @xmath74 and @xmath75 are both positive . to obtain the solutions of eq .
( [ usadel ] ) , we impose appropriate boundary conditions @xcite , i.e. , @xmath76 and @xmath77 where @xmath78 and @xmath79 is the conductivity of @xmath81 and @xmath82 ( @xmath1 ) . moreover , in the present calculation , we adopt the rigid boundary condition @xmath80 where @xmath81 is the conductivity of @xmath7 in the normal state and @xmath82 @xcite . assuming that @xmath47 , the anomalous green s function in the @xmath7s attached to @xmath81 and @xmath82 can be approximately given as @xmath83 where @xmath84 ( @xmath85 : real ) and @xmath86 is the superconducting phase in the left ( right ) side of @xmath7s ( see fig .
[ sfnfs - gm ] ) . assuming that @xmath87 , we can preform the taylar expansion for @xmath88 as follows @xcite : @xmath89 using the boundary conditions given in eqs .
( [ bc1 ] ) and ( [ bc5 ] ) for eq .
( [ ff1 ] ) and substituting eq .
( [ ff1 ] ) into eq .
( [ usadel ] ) , @xmath88 can be approximately expressed as @xmath90 \nonumber \\ & - & i { \rm sgn}(\omega_{n } ) \frac{(x - d_{\rm f1})^{2}}{2\hbar d_{\rm f } } \left ( h_{\rm ex}^{y } \left\ { { \hat \tau_{y } , { \hat f}^{\rm s1}(0 ) } \right\ } + \left [ \hat \tau_{z } , { \hat f}^{\rm s1}(0 ) \right ] \right ) \label{ff1x}.\end{aligned}\ ] ] here we also assume that the exchange field @xmath91 in the @xmath81 is much larger than @xmath92 and thus the term @xmath93 is neglected in eq .
( [ ff1x ] ) . similarly , assuming that @xmath94 , we can perform the taylor expansion for @xmath95 and , using the boundary conditions given in eqs .
( [ bc4 ] ) and ( [ bc6 ] ) , @xmath95 can be approximately expressed as @xmath96 \nonumber \\ & - & i { \rm sgn}(\omega_{n } ) \frac{(x - l)^{2 } h_{\rm ex2}}{2\hbar d_{\rm f } } \left [ \hat \tau_{z } , { \hat f}^{\rm s2}(l_{\rm f } ) \right ] \label{ff2x } , \end{aligned}\ ] ] where @xmath97 is also assumed .
the general solutions of @xmath98 are given as @xmath99 and @xmath100 where @xmath101 and @xmath102 .
applying the boundary conditions given in eqs .
( [ bc2 ] ) and ( [ bc3 ] ) to eqs .
( [ fnpm1 ] ) and ( [ fny ] ) , and also using the results in eqs .
( [ ff1x ] ) and ( [ ff2x ] ) , we can obtain the anomalous green s functions in the @xmath1 as @xmath103 -\frac{k_{\rm n}d_{\rm f2}}{\gamma_{\rm f } } { \rm cosh}[k_{\rm n}(x - l ) ] \right ] k_{\omega_{n}}(d ) \nonumber \\ & + & i \frac{\delta_{\rm r}}{e_{\omega_{n } } } \left [ { \rm sinh}[k_{\rm n}(x - d_{\rm f1 } ) ] + \frac{k_{\rm n}d_{\rm f1}}{\gamma_{\rm f } } { \rm cosh}[k_{\rm n}(x - d_{\rm f1 } ) ] \right ] k_{\omega_{n}}(d ) \label{fns } , \\
% f_{ty}^{\rm n}(x ) & = & { \rm sgn}(\omega_{n } ) \frac{\delta_{\rm l}}{e_{\omega_{n } } } \frac{h_{\rm ex}^{y } d_{\rm f1}^{2}}{\hbar d_{\rm f } } \left [ { \rm sinh}[k_{\rm n}(x - l ) ] -\frac{k_{\rm n}d_{\rm f2}}{\gamma_{\rm f } } { \rm cosh}[k_{\rm n}(x - l ) ] \right ] k_{\omega_{n}}(d ) \label{fnty } , % \end{aligned}\ ] ] and @xmath104 -\frac{k_{\rm n}d_{\rm f2}}{\gamma_{\rm f } } { \rm cosh}[k_{\rm n}(x - l ) ] \right ] k_{\omega_{n}}(d ) \nonumber \\ & - & { \rm sgn}(\omega_{n } ) \frac{\delta_{\rm r}}{e_{\omega_{n } } } \frac{h_{\rm ex2 } d_{\rm f2}^{2}}{\hbar d_{\rm f } } \left [ { \rm sinh}[k_{\rm n}(x - d_{\rm f1 } ) ] + \frac{k_{\rm n}d_{\rm f1}}{\gamma_{\rm f } } { \rm cosh}[k_{\rm n}(x - d_{\rm f1 } ) ] \right ] k_{\omega_{n}}(d ) \label{fntz } , % \end{aligned}\ ] ] where @xmath105 and @xmath106 from eqs .
( [ fns])([fntz ] ) , it is immediately found that that @xmath107 describing the ssc is an even function with respect to @xmath22 , whereas @xmath108 describing the stc is an odd function with respect to @xmath22 since @xmath108 is proportional to @xmath109 .
hence , @xmath108 represents the odd - frequency stc .
it should be emphasized here that @xmath110 and @xmath111 whereas @xmath112 is generally finite independently of @xmath91 and @xmath113 .
this is due to the fact that the presence of @xmath8 layers are essential to induce the stc @xcite . on the contrary
, the ssc is always induced inside the @xmath1 in @xmath7/@xmath1 junctions as well as more complex @xmath7/@xmath1/@xmath8 junctions @xcite .
notice also that i ) @xmath114 because the exchange field in the @xmath81 does not have the @xmath37 component . and
ii ) @xmath115 is exactly zero when @xmath116 or @xmath21 as @xmath117 .
the spacial dependence of anomalous green s functions in the @xmath1 is discussed in appendix [ app : agf ] . within the quasiclassical green s function method ,
the magnetization @xmath118 induced inside the @xmath1 is given @xcite as @xmath119 where @xmath120 is the superconducting phase difference between the outmost @xmath7s in the junction and @xmath121 \nonumber \\ \label{m}\end{aligned}\ ] ] with @xmath122 here , @xmath123 is the local magnetization density in the @xmath1 , @xmath124 is the @xmath124 factor of electron , @xmath125 is the bohr magneton , and @xmath126 and @xmath127 are the cross - section area of junction and the volume of @xmath1 , respectively . in the quasiclassical green s function method ,
the density of states @xmath128 per unit volume and per electron spin at the fermi energy is assumed to be approximately the same for up and down electrons in the @xmath1 @xcite .
it is apparent in eq .
( [ m ] ) that @xmath107 and @xmath129 are both required to be nonzero to induce finite @xmath130 .
however , as described in sec .
[ sec : agf ] , nonzero @xmath129 occurs only when @xmath8 layers are involved in the junction and @xmath131 whenever @xmath132 for @xmath133 .
therefore , the origin of the magnetization in the @xmath1 is considered to be due to the stcs induced by the proximity effect @xcite .
note also that because of @xmath114 ( see sec .
[ sec : agf ] ) , @xmath134 and thus @xmath135 are always zero .
therefore , in the following , we only consider the @xmath38 and @xmath23 components of @xmath118 .
more details of @xmath123 are examined in appendix [ app : lmd ] .
substituting eqs ( [ fns])([fntz ] ) into eq .
( [ m ] ) and performing the integration with respect to @xmath37 in eq .
( [ md ] ) , we can obtain the @xmath38 and @xmath23 components of the magnetization induced inside the @xmath1 .
the @xmath38 component @xmath136 of the magnetization is decomposed into two parts , @xmath137 where @xmath138 and @xmath139 here , we have introduced @xmath140 + \left [
1-\left(\frac{k_{\rm n } d_{\rm f2 } } { \gamma_{\rm f } } \right)^{2 } \right ] k_{\rm n } d -\frac{1}{2 } \left [ 1+\left(\frac{k_{\rm n } d_{\rm f2 } } { \gamma_{\rm f } } \right)^{2 } \right ] { \rm sinh}(2k_{\rm n } d ) \end{aligned}\ ] ] and @xmath141 { \rm sinh}(k_{\rm n } d).\end{aligned}\ ] ] similarly , the @xmath23 component @xmath142 of the magnetization is decomposed into two parts , @xmath143 where @xmath144 \label{mz1 } \end{aligned}\ ] ] and @xmath145 here , we have also introduced @xmath146 k_{\rm n } d - \frac{k_{\rm n } d_{\rm f1(2)}}{\gamma_{\rm f } } { \rm cosh}(2k_{\rm n } d ) - \frac{1}{2 } \left [ 1+\left ( \frac{k_{\rm
n } d_{\rm f1(2)}}{\gamma_{\rm f } } \right)^{2 } \right ] { \rm sinh}(2k_{\rm n } d ) \label{eq : ri}.\end{aligned}\ ] ] the @xmath4 independent part of the magnetization , i.e. , @xmath147 and @xmath148 , is due to the proximity effect common in @xmath7/@xmath8 junctions , similar to the one inducing the stcs in @xmath8/@xmath7/@xmath8 and @xmath7/@xmath8/@xmath8 junctions @xcite . on the other hand ,
the @xmath4 dependent part @xmath149 and @xmath150 of the magnetization is induced by the coupling between the two @xmath7s in the junction .
therefore , @xmath151 appears only when ferromagnetic metallic multilayers constitute the josephson junction @xcite .
it should also be noticed that @xmath136 becomes zero when @xmath116 or @xmath21 since @xmath147 and @xmath152 are both proportional to @xmath153 .
in contrast , @xmath142 is generally nonzero for any @xmath154 .
let us first numerically evaluate @xmath136 and @xmath142 in the @xmath1 obtained in eqs .
( [ my])([eq : ri ] ) . for this purpose ,
the temperature dependence of @xmath85 is assumed as @xmath155 where @xmath156 is the superconducting gap at zero temperature and @xmath31 is the superconducting transition temperature @xcite .
figures [ m1][m3 ] show the typical results of the magnetization in the @xmath1 as a function of thickness @xmath5 of the @xmath1 normalized by @xmath157 .
figure [ m1 ] represents the results for @xmath158 where the magnetizations between f1 and f2 are parallel . as shown in fig .
[ m1](a ) , the @xmath38 component @xmath136 of the magnetization is exactly zero since @xmath115 contributing to @xmath136 is zero in the parallel magnetization configuration . on the other hand ,
the @xmath23 component @xmath142 has a finite value , as shown in fig .
[ m1](b ) , because @xmath159 contributing to @xmath142 is non - zero in the parallel magnetization configuration .
furthermore , the induced magnetization @xmath142 is found to be negative , i.e. , opposite to the magnetizations in f1 and f2 .
it is also found in fig .
[ m1](b ) that @xmath160 monotonically decreases with increasing @xmath5 for @xmath161 , but @xmath148 decays rather slowly as compared with @xmath150 .
the difference of the decay rates for @xmath162 as well as the small @xmath5 behavior in @xmath148 and @xmath150 will be further discussed below .
component @xmath136 and ( b ) the @xmath23 component @xmath142 of the magnetization in the @xmath1 for @xmath163 , corresponding to the parallel magnetization configuration between @xmath81 and @xmath82 .
for other parameters , we set @xmath164 , @xmath165 , @xmath166 , @xmath167 , @xmath168 , @xmath169 , and @xmath170 . for comparison , @xmath148 and @xmath150 are also plotted separately in ( b ) .
, width=245 ] figure [ m2 ] shows the results for @xmath171 where the magnetizations between @xmath81 and @xmath82 are antiparallel . as shown in fig .
[ m2](a ) , the @xmath38 component @xmath136 of the magnetization is still exactly zero since @xmath115 contributing to @xmath136 is zero also in the antiparallel magnetization configuration .
however , the @xmath23 component @xmath142 is finite and decreases monotonically with increasing @xmath5 for @xmath172 [ see fig . [ m2](b ) ] .
it is also noticed in fig .
[ m2](b ) that the sign of @xmath142 is positive and is opposite to the one for @xmath158 [ fig .
[ m1](b ) ] . the sign reversal of @xmath142 will be further discussed below .
it is also observed in fig .
[ m2](b ) that @xmath148 decays slowly with respect to @xmath5 as compared with @xmath150 , similarly to the case when the magnetizations in @xmath81 and @xmath82 are parallel .
it should be noted however that @xmath142 becomes exactly zero in the antiparallel magnetization configuration when @xmath173 and @xmath174 @xmath175 because in this case @xmath176 and @xmath177 are both zero , as seen in eqs .
( [ mz1])([eq : ri ] ) [ see also eqs .
( [ ap - mz1])([hex ] ) ] .
component @xmath136 and ( b ) the @xmath23 component @xmath142 of the magnetization in the @xmath1 for @xmath178 , corresponding to the antiparallel magnetization configuration between @xmath81 and @xmath82 .
other parameters are the same as in fig .
[ m1 ] . for comparison , @xmath148 and @xmath150
are also plotted separately in ( b ) .
, width=245 ] figure [ m3 ] shows the results for @xmath179 where the magnetization in @xmath81 is perpendicular to that in @xmath82 . as shown in fig .
[ m3](a ) , the @xmath38 component @xmath136 of the magnetization is now finite since @xmath115 contributing to @xmath136 is nonzero in this case .
it is also found in fig .
[ m3 ] that both @xmath180 and @xmath160 decrease monotonically with increasing @xmath5 for @xmath172 .
moreover , it is clearly observed that the decay rate of @xmath181 with respect to @xmath5 is slower than that of @xmath182 .
this is similar to the other cases discussed above in figs .
[ m1 ] and [ m2 ] .
component @xmath136 and ( b ) the @xmath23 component @xmath142 of the magnetization in the @xmath1 for @xmath183 , corresponding to the case where the magnetization in @xmath81 is perpendicular to that in @xmath82 .
other parameters are the same as in fig .
[ m1 ] . for comparison , @xmath184 and @xmath151
are also plotted separately .
, width=245 ] it is now instructive to consider limiting cases for the magnetization @xmath184 and @xmath151 induced inside the @xmath1 and analyze the qualitative behavior with respect to the thickness @xmath5 of the @xmath1 . for @xmath185 and @xmath186 ,
the @xmath38 components @xmath147 and @xmath149 of the magnetization are approximately given as @xmath187 and @xmath188 whereas the @xmath23 components @xmath148 and @xmath150 of the magnetization are approximately @xmath189 and @xmath190 where @xmath191 and @xmath192 ( @xmath193 at @xmath185 ) .
it is immediately found in eqs .
( [ ap - my1])([ap - mz2 ] ) that @xmath184 decreases rather slowly , i.e. , algebraically , as @xmath194 , whereas @xmath151 decays exponentially .
this is indeed comparable with the numerical results shown in figs .
[ m1][m3 ] .
next , we discuss in the same limiting case the sign change of @xmath142 by flipping the magnetization direction from the parallel to the antiparallel configuration in @xmath81 and @xmath82 . for this purpose , we focus on @xmath195 appearing in eqs .
( [ ap - mz1 ] ) and ( [ ap - mz2 ] ) , the definition being given in eq .
( [ hex ] ) . in the case of parallel magnetization configuration , @xmath142 is always negative , as shown in fig .
[ m1 ] , simply because @xmath195 is positive ( assuming that @xmath196 ) . on the other hand , in the case of antiparallel magnetization configuration
, @xmath197 is negative .
therefore , @xmath142 becomes positive when @xmath198 is larger than @xmath199 , as shown in fig .
[ m2 ] .
let us now consider the opposite limit , i.e. , @xmath200 , at @xmath185 . in this limit ,
the @xmath38 components @xmath147 and @xmath149 of the magnetization are approximately given as @xmath201 and @xmath202\cos \theta , \nonumber \\ \label{my - ap2-ii}\end{aligned}\ ] ] whereas the @xmath23 components @xmath148 and @xmath150 of the magnetization are approximately @xmath203 \label{mz - ap2 } \nonumber \\\end{aligned}\ ] ] and @xmath204\cos \theta .
\nonumber \\ \label{mz - ap2-ii}\end{aligned}\ ] ] it is therefore readily noticed in eqs .
( [ my - ap2])([mz - ap2-ii ] ) that @xmath184 and @xmath151 are linearly dependent on @xmath5 and their slopes are determined by the signs of @xmath205 and @xmath206 .
this is in good qualitative agreement with the numerical results shown in figs .
[ m1][m3 ] . in the previous section ,
we have focused on the @xmath5 dependence of the magnetization induced inside the @xmath1 . here
, we shall demonstrate that the magnetization can also be controlled by the superconducting phase difference @xmath4 in the two @xmath7s .
the most simplest way to tune @xmath4 experimentally is to apply dc bias current to the junction , in which the dc josephson effect can be detected @xcite .
figure [ mq ] shows the @xmath38 and @xmath23 components @xmath207 and @xmath208 of the magnetization induced inside the @xmath1 as a function of @xmath4 for different values of @xmath209 .
figure [ mq ] clearly demonstrates that the magnetization can indeed be controlled by tuning @xmath4 .
it should also be noticed that the magnitude of the magnetization increases with decreasing @xmath209 .
therefore , @xmath209 is an important parameter to increase the magnetization induced inside the @xmath1 .
next , let us discuss an alternative way to control the magnetization induced inside the @xmath1 .
here , we consider the @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 junction subject to both dc and ac external fields , described by the voltage bias model @xcite , as schematically shown in fig .
[ sfnfs - gm2 ] . in this case
, the superconducting phase difference @xmath4 evolves in time @xmath17 according to the following well known formula : @xmath210 where @xmath211 is a time independent constant , @xmath13 is the dc voltage , and @xmath212 and @xmath15 are the amplitude and frequency of the ac voltage , respectively . /@xmath81/@xmath1/@xmath82/@xmath7 junction to observe dynamics of the magnetization induced inside the @xmath1 .
this set up is based on the voltage bias model @xcite , where @xmath13 is the dc voltage , and @xmath212 and @xmath15 are the amplitude and frequency of the ac voltage , respectively .
, width=207 ] substituting eq .
( [ theta - t ] ) into eqs .
( [ my2 ] ) and ( [ mz2 ] ) , and using the generating function of bessel functions , we can easily find that the @xmath4 dependent parts of the magnetization in the @xmath1 are given as @xmath213 and @xmath214 for the @xmath38 and @xmath23 components , respectively , where the @xmath13 and @xmath17 dependence is explicitly shown in the left hand sides . in the above ,
we have also introduced @xmath215 , \label{eq : gamma}\ ] ] where @xmath216 is the josephson frequency , @xmath217 is the flux quantum , and @xmath218 is the bessel function of the first kind ( @xmath48 : integer ) .
let us now consider the time averaged quantity @xcite @xmath219 as shown in eqs .
( [ dmy ] ) and ( [ dmz ] ) , the @xmath4 dependent part of the magnetization clearly oscillates in @xmath17 .
therefore , the time averaged magnetization @xmath220 is zero except for @xmath221 .
this is simply because the coefficient proportional to @xmath17 in the cosine function of eq .
( [ eq : gamma ] ) becomes zero only when this condition is satisfied .
indeed , the characteristic feature of @xmath220 is clearly found in fig .
[ mv ] , i.e. , @xmath220 showing non - zero values only at @xmath222 .
finally , we shall approximately estimate the amplitude of the magnetization induced inside the @xmath1 . as shown in figs .
[ m1][m3 ] , the magnetization in the @xmath1 has a finite value in the length scale of @xmath223 . in dirty normal metals ,
@xmath223 is in a range of several dozen to several hundred nanometers since @xmath224 is about 0.010.1 @xmath225/s and @xmath31 is assumed to be a few kelvins @xcite . as indicated in figs .
[ m1][mq ] , and [ mv ] , the amplitude of the magnetization is estimated to be one to two orders smaller than @xmath226 .
when we use a typical set of parameters , i.e. , the density of states at the fermi energy @xmath227 with the fermi energy @xmath228 ev @xcite ( @xmath48 : the electron mass ) , @xmath229 k for nb @xcite , @xmath230 nm , and @xmath231 nm @xcite , we can estimate that @xmath232 is approximately 31000 @xmath233 .
it is therefore expected that the magnetization induced inside the @xmath1 can be detected with the magnetization measurement by squid @xcite .
we have calculated the magnetization inside the @xmath1 in the @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 josephson junction based on the the quasiclassical green s function method in the diffusive transport limit . by solving the usadel equation , we have found that finite magnetization is induced inside the @xmath1 .
we have shown that the magnetization is due to the odd - frequency stcs formed by electrons of equal and opposite spins , which are induced by the proximity effect in the _ s_/_f_1/_n_/_f_2/_s _ junction . fixing the magnetization in @xmath82 along the @xmath23 direction perpendicular to the junction direction ( @xmath37 direction ) , we have shown that i ) the @xmath37 component of the magnetization in the @xmath1 is always zero , ii ) the @xmath38 component is exactly zero when the magnetization direction between @xmath81 and @xmath82 is collinear , and iii ) the @xmath23 component is generally finite for any magnetization direction between @xmath81 and @xmath82 .
decomposing the induced magnetization into @xmath4 independent and dependent parts , we have found that the @xmath4 independent part of the magnetization decays slowly with increasing the thickness of the @xmath1 , whereas the @xmath4 dependent part of the magnetization decays rather rapidly .
while the @xmath4 independent part of the magnetization is generally induced even in the @xmath7/@xmath8 junctions due to the proximity effect , the @xmath4 dependent part of the magnetization results from the finite coupling between the two @xmath7s in the @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 josephson junction .
we have also found that the time averaged magnetization in the @xmath1 exhibits discontinuous peaks at particular values of dc voltage when dc and ac voltages are both applied to the @xmath7/@xmath81/@xmath1/@xmath82/@xmath7 junction , implying that the ac magnetization oscillation can be converted into the dc component .
we have discussed that the magnetization induced inside the @xmath1 can be large enough to be observed in typical experimental settings .
it is therefore expected that a josephson junction composed of ferromagnetic metallic multilayers such as the one studied here can have a promising potential for low joule heating spintronics devices , where the magnetization can be controlled by varying the superconducting phase difference @xmath4 .
this work is supported by grant - in - aid for research activity start - up ( no .
25887053 ) from the japan society for the promotion of science and also in part by riken ithes project .
t. kontos , m. aprili , j. lesueur , and x. grison , phys .
@xmath238 , 304 ( 2001 ) ; t. kontos , m. aprili , j. lesueur , f. gen@xmath239 t , b. stephanidis , and r. boursier , phys .
@xmath240 , 137007 ( 2002 ) . s. m. frolov and d. j. van harlingen , v. a. oboznov , v. v. bolginov , and v. v. ryazanov , phys . rev .
b @xmath243 , 144505 ( 2004 ) ; s. m. frolov and d. j. van harlingen , v. v. bolginov , v. a. oboznov , and v. v. ryazanov , phys .
b @xmath244 , 020503(r ) ( 2006 ) . .
j. w. a. robinson , s. piano , g. burnell , c. bell , and m. g. blamire , phys .
@xmath245 , 177003 ( 2006 ) ; j. w. a. robinson , s. piano , g. burnell , c. bell , and m. g. blamire , phys .
b @xmath246 , 094522 ( 2007 ) .
m. weides , m. kemmler , h. kohlstedt , r. waser , d. koelle , r. kleiner , and e. goldobin , phys .
@xmath245 , 247001 ( 2006 ) ; m. weides , h kohlstedt , r waser , m. kemmler , j. pfeiffer , d. koelle , r. kleiner , and e. goldobin , appl .
phys a @xmath240 , 613 ( 2007 ) .
m. eschrig , j. kopu , j. c. cuevas , and g. sch@xmath273n , phys .
@xmath254 137003 ( 2003 ) ; m. eschrig , t. l@xmath273fwander , t. champel , j. c. cuevas , j. kopu , and g. sch@xmath248n , j. low temp .
@xmath274 , 457 ( 2007 ) ; m. eschrig and t. l@xmath273fwander , nat .
@xmath275 , 138 ( 2008 ) .
t. s. khaire , mazin a. khasawneh , w. p. pratt , jr . , and norman o. birge , phys . rev .
@xmath284 137002 ( 2010 ) ; c. klose , t. s. khaire , y. wang , w. p. pratt , jr . , n. o. birge , b. j. mcmorran , t. p. ginley , j. a. borchers , b. j. kirby , b. b. maranville , and j. unguris , phys . rev
@xmath285 , 127002 ( 2012 ) . | we theoretically study the magnetization inside a normal metal induced in an @xmath0-wave superconductor / ferromagnetic metal / normal metal / ferromagnetic metal/@xmath0-wave superconductor ( _ s_/_f_1/_n_/_f_2/_s _ ) josephson junction . using the quasiclassical green s function method
, we show that the magnetization becomes finite inside the @xmath1 .
the origin of this magnetization is due to odd - frequency spin - triplet cooper pairs formed by electrons of equal and opposite spins , which are induced by the proximity effect in the _ s_/_f_1/_n_/_f_2/_s _ junction .
we find that the magnetization @xmath2 in the _
n _ can be decomposed into two parts , @xmath3 , where @xmath4 is the superconducting phase difference between the two _ s_s and @xmath5 is the thickness of @xmath1 .
the @xmath4 independent magnetization @xmath6 exists generally in @xmath7/@xmath8 junctions , while @xmath9 carries all @xmath4 dependence and represents the fingerprint of the phase coherence between the two _ s_s in josephson junctions .
the @xmath4 dependence thus allows us to control the magnetization in the @xmath1 by tuning @xmath4 for a fixed @xmath5 .
we show that the @xmath4 independent magnetization @xmath6 weakly decreases with increasing @xmath5 , while the @xmath4 dependent magnetization @xmath9 rapidly decays with @xmath5 .
moreover , we find that the time - averaged magnetization @xmath10 exhibits a discontinuous peak at each resonance dc voltage @xmath11 ( @xmath12 : integer ) when dc voltage @xmath13 as well as ac voltage @xmath14 with frequency @xmath15 are both applied to the _ s_/_f_1/_n_/_f_2/_s _ junction .
this is because @xmath16 oscillates generally in time @xmath17 ( ac magnetization ) with @xmath18/\hbar$ ] and thus @xmath19 , but can be converted into the time - independent dc magnetization for the dc voltage at @xmath20 .
we also discuss that the magnetization induced in the _
n _ can be measurably large in realistic systems .
therefore , the measurement of the induced magnetization serves as an alternative way to detect the phase coherence between the two _ s_s in josephson junctions .
our results also provide a basic concept for tunable magnetization in superconducting spintronics devices . |
carbon stars ( objects with prominent c@xmath3 in their spectra ) have been studied for more than a century @xcite , although faint ( r @xmath7 ) high - latitude carbon stars ( fhlcs ) , which presently number in the hundreds , were not easily found until recently .
most fhlcs are thought to be distant giants , as there is no obvious way for c@xmath3 to reach the photosphere prior to the red giant phase . while relatively rare , these objects are interesting because , for example , of their utility as a halo velocity tracers @xcite .
however , a small number of fhlcs display parallaxes and/or large proper motions , implying main - sequence luminosities ( m@xmath8 ) , and have been designated dwarf carbon ( dc ) stars .
it now appears that a significant fraction of fhlcs are in fact not distant giants , but are nearby dwarfs ( * ? ? ?
* hereafter paper i ) . based on data obtained during the sloan digital sky survey ( sdss ; * ? ? ?
* ) commissioning period , we claimed in paper i that at least @xmath9 of all fhlcs are dwarfs .
here we report the results of an expanded sample of sdss fhlcs . in section 2
, we discuss the observations and selection criteria for our sample , while section 3 describes the classification of the objects as dwarfs or giants .
a key problem remains the derivation of simple luminosity discriminants other than proper motion , as the low resolution ( @xmath10 ) spectra of the giants and dwarfs are so similar .
section 4 discusses photometric and spectroscopic luminosity indicators , while our conclusions are given in section 5 .
the fhlc candidates were chosen by analyzing the imaging data from the sdss camera @xcite , which obtains images in five bands ( _ u , g , r , i , z _ ; * ? ? ?
* ) almost simultaneously .
the instrumental fluxes are calibrated via a network of primary and secondary standard stars @xcite .
observations of previously known fhlcs in the sdss system @xcite were used to determine approximately where in the sdss color - color diagrams fhlcs were expected , and an automated analysis of the photometric database was used to identify objects that appear in those regions .
more details on object selection can be found in paper i. spectroscopic observations were then obtained for as many of the fhlc candidates as possible .
the spectra , obtained with a sdss fiber - fed ccd spectrograph , cover the wavelength range 3800 -9200 with @xmath11/@xmath12 ; see @xcite for details .
although more than 600 spectra are obtained in each observation , our survey is not 100@xmath13 complete due to two factors .
first , there is a minimum separation of the fibers , so a fhlc candidate close to a higher priority science target can not be observed .
second , there are regions where all fibers are utilized by unrelated higher priority programs . nonetheless , the sdss fhlc survey does provide a large homogeneous sample of objects .
as was the case in paper i , we have not relied on an automated algorithm to select true fhlc stars from the sdss spectra , but rather visually examined all spectra .
the sdss first data release dr1 @xcite contains spectra of about 500 unresolved sources that the sdss photometric selection flagged as a possible fhlc .
however , manual examination of these spectra shows that only @xmath14% actually display c@xmath15 bands .
the bulk of the contaminants are late type stars , and a few are high redshift quasars , where prominent l@xmath16 emission simulates a band head . in this respect
the contaminants are identical to those in c star surveys of decades past , made on objective prism plates .
clearly , the efficiency of photometric selection of fhlcs is modest , due to color degeneracy with both interesting and uninteresting objects , galactic and extragalactic .
however , the sdss will at completion still derive a very large , homogeneous sample of fhlcs , in part because multiple other unrelated , high - priority scientific projects in the survey , in particular the search for distant qsos , target many of these candidates for spectroscopy due to the color degeneracy .
even though such programs may have relatively small contamination rates by fhlcs , they target very large numbers of objects . on the basis of the spectroscopic observations ,
251 of the sdss candidates ( including the 39 listed in paper i , which are repeated here for convenience and completeness ) were found to be fhlcs .
ninety seven have data publically available in the dr1 , and the remainder are newly presented here .
the spectra of all fhlcs can be found via anonymous ftp ; sample spectra are displayed in paper i. the 251 objects are listed in table [ tbl-1 ] , which provides astrometric ( coordinates and , when available , proper motions ) and photometric ( both sloan and , when available , 2mass colors ) data .
we also note a luminosity class ( giant or dwarf ) based on the sdss - derived proper motions ( see paper i for details ) : all objects with @xmath17 detections of motion are considered dwarfs .
the spectra are qualitatively similar to those noted in paper i , and we continue to see the broad range of nad strength ( presumptively temperature ) previously noted . however , the fraction of objects with noticeable balmer emission ( 6@xmath13 ) is only about half that seen in our initial sample
. a few special cases of extreme color are worthy of mention . in paper
i , no n - type stars were detected by the selection software ( the 2 objects shown there were found in an sdss vs. 2mass comparison ) , although we noted that their extreme red colors should make them easier to find than the r - type .
our expanded survey has resulted in the detection of one ( but only one ! ) n - type star - sdss j125149.87 + 013001.8 .
a small fraction of the sample are exceptionally blue , and show balmer and ca ii absorption in addition to the c@xmath3 bands .
these `` f / g carbon stars '' , discussed elsewhere @xcite , were initially found serendipitously .
searches of the sdss standard star sample , selected by colors optimized for metal - poor f - subdwarfs , then added @xmath18 more cases .
this small fraction of very red or very blue objects , not found by our nominal fhlc target selection software , stands as yet another warning of incompleteness in this survey .
figures 1 - 3 display color - color diagrams of the sdss fhlcs , and define the location in the diagrams where fhlcs are found . the large scatter in _ ( u - g )
_ is due to the lower sensitivity of the survey in the _ u _ band combined with the faintness of the objects ; however , the one n - type star is clearly separated from the other objects .
there is a tight correlation in the _
( g - r),(r - i ) _ diagram , with the one n - type star again distinctly visible . also , the `` f / g carbon stars '' are clearly distinct , but appear to be an extension of the r - type star correlation .
similarly , in the _ ( r - i),(i - z ) _ diagram , the stars are fairly closely clumped , with the `` f / g carbon stars '' again following the r - type stars ; the n - type star is not as clearly separated from the r - types in this diagram . in paper
i , we noted that our sdss survey is complementary to the 2mass survey , as the former will detect the warmer r - type stars , while the latter will be useful for detecting the cooler n - type stars .
it was also noted that the sdss survey detects objects too faint for 2mass .
we have searched the 2mass all - sky data release point source catalog for each of the sdss fhlcs , and find about @xmath1 of the objects are detected ; the 2mass colors are listed in table [ tbl-1 ] . figure [ fig4 ] shows the distribution of 2mass detections in sdss _
r _ magnitude bins .
not surprisingly , most of the bright objects ( _ r _ @xmath19 ) are detected in 2mass , while most of the faint objects ( _ r _ @xmath20 ) are not . to support our search for luminosity discriminators , we measured equivalent widths for both the discrete lines ( the balmer lines , nad ) , and the strong cah bands ( @xmath21 and @xmath22 ) in each spectrum in which they were present . for the c@xmath3 (
@xmath23 , @xmath24 , @xmath25 , and @xmath26 ) and cn ( @xmath27 ) band heads , we measured a band head intensity corresponding to the average flux at the peak of the band divided by the average flux at the bottom of the band .
the sdss astrometric pipeline @xcite computes j2000.0 positions for all detected objects , and these positions have been compared with usno - b data to obtain proper motions for almost all the fhlcs ( see table [ tbl-1 ] ) ; those without motions are either too faint to obtain reliable positions , or are in crowded fields . all objects with non - zero motions significant at a @xmath28 level are considered dwarfs ; see paper i for a discussion of the validity of using only proper motion to determine the luminosity class .
those objects that do not show motion could either be giants or distant dwarfs whose motions are too small to detect .
a quick examination at the magnitudes of the objects without detected motion shows that only a handful are fainter than the dwarfs , so these objects are likely a combination of true giants and distant dwarfs . to determine which objects are close enough ( if dwarfs ) for a motion to be detectable , we examined the distribution of observed motions , and find that most ( over @xmath5 ) have proper motions @xmath29 mas yr@xmath30 , which we take as our minimum detectable value . after converting the sdss magnitudes to v @xcite , and assuming m@xmath31
, we converted all proper motions to a linear scale , and find that @xmath32 of the objects have tangential velocities @xmath33 km s@xmath30 . finally , we determined the v magnitude an object would have if it were at these limits ( i.e. , would be barely detectable ) , and find that v@xmath34 = 18.6 .
this means that any object brighter than this limit is likely a true giant , which is @xmath35 ( 35 ) of the objects without observed motions ; the nature of the remaining objects must be considered uncertain .
the occasional renegade dwarf may of course by chance have a small tangential component to its motion .
we exclude from this analysis the `` f / g carbon stars '' , which are likely more luminous than the fiducial m@xmath31 , and label them all of uncertain luminosity in table [ tbl-1 ] , unless a positive detection of proper motion is present . for the remaining 74 objects ( excluding the 29 objects that had no proper motion data available ) , it is possible to make a statistical argument concerning the number that should be giants .
the faintest star in our sample has v=21.3 , and it would need to have a tangential velocity of 175 km s@xmath30 to be detectable .
about one - third of the dwarfs have a motion above this limit , so @xmath36 of the uncertain objects should be dwarfs with motions large enough to detect , and hence the lack of detected motion implies they are true giants ( although we do nt know which ones ) .
thus , about half of the objects without detected motions are probably giants .
a different conclusion emerges , however , if a significant fraction of the sample is made of disk dwarfs with space velocities smaller than velocities of halo stars . they will sometimes have tangential velocities less than 50 km s@xmath30 , so some will contaminate the brighter stars that were called giants above .
they will all have tangential velocites less than 175 km s@xmath30 , so they may dominate the fainter stars without significant motions , leaving fewer of the fainter uncertain objects as implied giants than the 24 estimated above . in this case , significantly fewer than half of the objects without detected motions would be giants .
the population models made for paper i indicated that disk dwarfs were required to explain the stars with small radial velocities .
a further analysis of the population mix will be deferred to a future paper .
the total number of dwarfs in our survey ( excluding the `` f / g carbon stars '' ) is 107 , while the total number of non - dwarfs ( excluding the `` f / g carbon stars '' and the 3 objects in the draco dwarf galaxy ) is 117 .
thus , at least @xmath37 of the fhlcs are dwarfs , consistent with the estimate of @xmath38 of paper i. if we assume half the uncertain objects are dwarfs , then at least @xmath39 of the fhlcs are dwarfs .
the recent summary of known dc stars by @xcite lists 14 objects in total not due to sdss .
thus , the homogeneous sample of sdss stars in table [ tbl-1 ] that are unambiguously dwarfs now constitutes @xmath5 of all known dc stars . in paper
i , we derived a surface density of @xmath40 deg@xmath41 , which was consistent with the value of @xmath42 deg@xmath41 of @xcite , derived from a bright , photographic sample . with the same caveats about the completeness of our sample ( see paper i ) , we derived a surface density from our expanded sample ( 224 objects in @xmath433600 deg@xmath44 ) of 0.06 deg@xmath41 ( again excluding the `` f / g carbon stars '' and the 3 draco stars ) .
it is clear that careful incompleteness corrections , not currently available , will be needed to understand the true surface density of the faintest fhlcs .
as previously noted , at least @xmath1 the fhlcs are dwarfs .
therefore the development of a simple observational luminosity discriminant is imperative if we are to , for example , use the giants as halo velocity tracers .
the fact that the dwarfs and giants differ in intrinsic luminosity by @xmath14 mag gives hope that there would be such discriminants .
in fact , several luminosity discriminants have been proposed , although the small number of objects known to date has made it difficult to assess the validity and range of applicability of these discriminants . with our large sample of objects , an objective analysis of the proposed discriminants , as well as
the possibility of uncovering others , is now possible .
numerous authors ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have proposed near - ir colors as a luminosity discriminant , while @xcite showed a tight sequence for n - type stars for his survey objects in the _ ( h - k ) _ , _ ( j - h ) _ diagram .
while we will discuss the luminosity aspect below , we present in figure [ fig5 ] the 2mass color - color diagram for all stars ( @xmath45 stars ) in our survey , in the carbon star catalog ( @xmath46 ; @xmath47 stars ) of @xcite , and all previously reported fhlc dwarfs ( 11 stars ) , summarized by @xcite . while the n - type stars continue to show the trend seen by @xcite ,
the r - type stars are far more scattered , and the `` f / g carbon stars '' appear to form a nice continuation of the c stars sequence .
the one n - type star in our survey fits nicely in with those from @xcite , while the one r - type star from @xcite at _ ( h - k)_=1.0 is classified as r - n , so is likely an n - type ; the r - type star at _
( h - k)_=1.6 is the bright mira variable ru vir .
although near - ir colors have been proposed as a luminosity discriminant , @xcite and @xcite argue on purely physical grounds that _ jhk _ colors alone should not be an unambiguous indicator .
based on 4 objects from paper i , we found that 2 objects were consistent with the _ jhk _ discriminant , while 2 objects were not consistent . in figure
[ fig6 ] , we show an enlargement of the near - ir color - color diagram , with just the sdss stars , as well as the previously known dwarfs , displayed . for _ ( h - k ) _ colors bluer than about 0.3
, there is no separation between giants and dwarfs .
however , for redder colors , the bulk of the objects are dwarfs .
for the ( somewhat arbitrary ) line shown in the figure , 20 ( @xmath48 ) of the objects are dwarfs , 1 ( @xmath49 ) is a giant , and 6 ( @xmath50 ) are uncertain ; if only the confirmed giants and dwarfs are counted , then @xmath51 of the objects are dwarfs . a slight shift of the line to the blue ( or red ) does not significantly change the giant to dwarf ratio .
an examination of figure [ fig5 ] shows that most of the previously known dwarfs do indeed appear to be offset from the bulk of the r - type stars from the @xcite catalog , but with the addition of the sdss fhlcs , it now appears that this discriminant can be refined and confirmed .
@xcite proposed that dwarfs and giants could be separated in the _ ( j - k ) _ , _ ( r - j ) _ diagram . while we do not have r magnitudes for our sdss sample , the sdss _ r _ band should be a reasonable proxy to the johnson r , so we have examined this discriminant with our sample .
figure [ fig7 ] shows that , as with the near - ir colors , this discriminant can identify dwarfs , but is somewhat less effective ( there are 5 giants in the dwarf region ) than using only 2mass colors . in paper
i , we commented that photometric variability may prove to be a simple luminosity criterion , since only giants should show consistent , chaotic variations associated with mass loss ; absence of variability , however , does not imply the object is a dwarf .
many of the sdss objects have multiple photometric measurements in the imaging database , so we have performed a preliminary check for variability . given the faintness of many of the objects ( particularly in _ u _ ) , we define an object as variable if at least 2 of the _ g ,
r _ , and _
i _ magnitudes changed by at least 0.10 mag . with that criteria ,
only 1 of the 49 fhlcs with multiple measurements is variable , and this object is listed in table [ tbl-2 ] .
that object has an uncertain luminosity class , and thus is a strong candidate for a giant .
note , however , that the object could be similar to the dc pg 0824 + 289 @xcite , which shows variability due to heating effects from a hot companion , and so multiple measurements to look for periodic variability would be interesting . in figures 1 - 3 , we plot the 3 sdss color - color diagrams .
an examination of those figures shows that , while the `` f / g carbon stars '' are clearly separated from the r - type stars , there is no indication of a luminosity discriminant .
we also investigated other color - color plots , none of which presented any valid discriminant .
the model for a dwarf c star is that the c@xmath3 was deposited in a previous mass - transfer episode from a companion that is now a faint white dwarf . in some cases ( e.g. ,
pg 0824 + 289 @xcite ; sbs 1517 + 5017 @xcite ) , the white dwarf is hot enough to be detectable in the visible .
however , in most cases , the white dwarf is too cool to be visible in the optical .
we examined those dwarfs in our sample that have the bluest _ u - g _ colors , to look for any spectroscopic evidence of a white dwarf , but none was found . on the other hand , it may be possible to detect the white dwarf in the ultraviolet , where the c star is faint and the contrast may be optimized . to determine how faint ( i.e. , cool ) the white dwarf would typically have to be to be undetected either spectroscopically or photometrically , we selected one blue and one red giant fhlc from our sample , and added a blackbody and/or real white dwarf spectrum of various temperatures .
we find that a 10,000 k blackbody is barely detectable in the red giant , while an 11,000 k blackbody is barely detectable in the blue giant .
however , in the satellite ultraviolet , these cool white dwarfs could be detectable ( e.g. , with the acs sbc camera on _ hst _ ) .
therefore , uv colors may in fact be a useful photometric luminosity discriminant .
numerous spectroscopic luminosity discriminants have been previously proposed .
@xcite proposed that the appearance of a strong c@xmath3 band head at @xmath26 indicated the object was a dwarf , while @xcite suggested the feature was both temperature and luminosity dependent .
however , both studies noted that while the absence of @xmath26 could not be used to firmly classify the object as a giant , the presence of @xmath26 was a good indicator that the object is a dwarf . in figure [ fig8 ]
we show a histogram of the strength of @xmath26 , which demonstrates that this feature is in fact probably not a good luminosity discriminant .
when @xmath26 is weak ( 1.0 - 1.4 ) , about @xmath1 of the objects are confirmed dwarfs , while at its strongest ( @xmath52 ) about a third of the objects are confirmed dwarfs . only in the 1.4 - 1.6 bin
do the dwarfs dominate ( @xmath53 ) . in total ,
@xmath1 of the objects with @xmath26 are confirmed dwarfs , so this feature can no longer be considered a luminosity discriminant .
in paper i we suggested that the presence of h@xmath16 emission could be an indicator that the object is a giant , since dwarfs are unlikely to process active chromospheres or undergo active mass loss .
however , a contrary case ( pg 0824 + 289 ) was noted , where heating of the dwarf by a hot da companion caused the emission . in that case , the white dwarf was visible in the optical spectrum , so h@xmath16 emission in dcs where there is no indication of a hot white dwarf could still be a luminosity discriminant . of the 14 objects in our sample with h@xmath16 emission , 3 are giants ,
3 are dwarfs , and the remaining 8 have are uncertain luminosity .
thus , the use of h@xmath16 emission as a luminosity discriminant is speculative .
the dcs with h@xmath16 emission are rare , and thus should be further observed .
the bands of cah at @xmath54 are normally strong only in k and m dwarfs ( e.g. , * ? ? ?
* ) , which led us to suggest in paper i that the presence of these features was a luminosity indicator in fhlcs . in figure [ fig9 ]
we show plot of the equivalent width of @xmath22 versus that of nad . while it is true that there are no giants whose spectra show cah , there are 10 objects ( out of 16 total ) that have an uncertain luminosity .
so , while promising , this discriminant must still be consider unproven .
it is interesting to note that sharp turn - on of the cah feature at w@xmath55(nad ) = 10 , although there are two objects ( one with weaker nad ( hotter ) which shows cah , one with stronger nad ( cooler ) that does not ) that are exceptions . to estimate the temperature of this turn - on point ,
the temperature index @xcite of every star that showed nad was measured , and a fit of nad as a function of temperature index was made .
the temperature index was converted to a temperature @xcite , with the result that the cah turn - on occurs at t=2900 k. @xcite suggested that dc s have strong @xmath26 band heads given their relatively weak cn band strength . in figure
[ fig10 ] , we plot the strength of @xmath26 versus the strength of the cn band at @xmath27 . unfortunately , there are so few confirmed giants which show cn that a definitive statement about this criterion is not possible .
however , given the range of @xmath26 strengths seen for the dwarfs alone , this discriminant can be considered doubtful . on the other hand
, the strength of the cn band alone , particularly at the weakest and stronger levels , may be a discriminant ( see below ) . with our large sample of fhlcs ,
we have searched for possible new spectroscopic discriminants .
@xcite found a correlation between the strengths of the c@xmath3 bands at @xmath24 and @xmath23 , which they used for detection of carbon stars .
our data also shows this correlation , but there is no separation between dwarfs , giants , and the stars of uncertain luminosity . in a similar vein ,
we show , in figure [ fig11 ] , the equivalent relation for the strongest c@xmath3 bands . again
, there is a correlation , but c@xmath3 band strength does not appear to be a luminosity discriminant .
as previously noted , the strength of the cn band at @xmath27 may be a luminosity indicator .
in figure [ fig12 ] , we plot a histogram of the strength of this feature .
as can be seen , all objects with band strengths greater than 1.6 are either giants ( 6 ) or of uncertain luminosity ( 1 ; the object with the strongest cn ) .
similarly , most of the objects with weak cn are dwarfs ( 19 ) ; there are 2 giants and 1 object of uncertain luminosity .
thus , extremely strong cn seems to imply a giant luminosity , while extremely weak ( but present ) cn implies a dwarf luminosity .
we have increased our sdss survey of faint high - latitude carbon stars to include 251 objects , at least @xmath1 of which are dwarfs . although giant and dwarf carbon stars differ in luminosity by 10 mag , the spectra ( at the sdss spectral resolution of @xmath56 ) are extremely similar .
we have used our expanded sample to search for photometric and spectroscopic discriminants , including an examination of those previously proposed .
we find that sdss colors and the strength of the c@xmath57 band head at @xmath26 are not good luminosity discriminants , while the presence of h@xmath16 emission as a discriminant is speculative .
the suggestion from paper i that the cah bands at @xmath58 indicate that the object is a dwarf must also still be considered unproven . the strength of the cn band at @xmath27 , when it is extremely strong or extremely weak , appears to be a good luminosity discriminant .
however , it is of limited use , as almost half the objects do not show this band . for the redder objects
, the use of _ jhk _ colors appears to allow the identification of dwarfs .
another possible discriminant is ultraviolet colors , where the presumed white dwarf ( which transferred its carbon to the now dc star ) may be detectable .
we thank j. macconnell for helpful comments .
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 nasa and the nsf .
this research has also made use of the simbad database , operated at cds , strasbourg , france , and the nasa / ipac infrared science archive , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the us department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , princeton university , the us naval observatory , and the university of washington .
abazajian , k. , adelman , j. , agueros , m. et al .
2003 , , 126 , 2081 alksnis , a. , balklavs , a. , dzervitis , u. , eglitis , i. , paupers , o. , & pundure , i. 2001 , baltic astron .
, 10 , 1 bothun , g. , elias , j. h. , macalpine , g. , matthews , k. , mould , j. r. , neugebauer , g. , & reid , i. n. 1991 , , 101 , 2220 christlieb , n. , green , p. j. , wisotzki , l. , & reimers , d. 2001 , , 375 , 366 cohen , m. 1979 , , 186 , 837 dunr , n .- c .
1884 , sur les toiles spectres de la troisime classe ( stockholm : norstedt ) fukugita , m. , ichikawa , t. , gunn , j. e. , doi , m. , shimasaku , k. , & schneider , d. p. 1996
, , 111 , 1748 grebel , e. 2004 , private communication green , p. j. , margon , b. , anderson , s. f. , & macconnell , d. j. 1992 , , 400 , 659 gunn , j. e. , carr , m. a. , rockosi , c. m. , sekiguchi , m. et al .
1998 , , 116 , 3040 heber , u. , bade , n. , jordan , s. , & voges , w. 1993 , , 267 , l31 hendon , a. a. & stone , r. c. 1998 , , 115 , 296 hogg , d. w. , schlegel , d. j. , finkbeiner , d. p. , & gunn , j. e. 2001 , , 122 , 2129 jrgensen , u. g. , borysow , a. , & hofner , s. 1998 , in asp conf .
138 , 1997 pacific rim conference on stellar atmospheres , ed .
k. l. chan , k. s. cheng , & h. p. singh ( san francisco : asp ) , 157 joyce , r. r. 1998 , , 115 , 2059 kirkpatrick , j. d. , henry , t. j. , & mccarthy , d. w. 1991 , , 77 , 417 knapp , g. r. et al . 2003 , in preparation krisciunas , k. , margon , b. , & szkody , p. 1998
, , 110 , 1342 liebert , j. , schmidt , g. d. , lesser , m. , stepanian , j. a. , lipovetsky , v. a. , chaffee , f. h. , foltz , c. b. , & bergeron , p. 1994 , , 421 , 733 lowrance , p. j. , kirkpatrick , j. d. , reid , i. n. , cruz , k. l. , & liebert , j. 2003 , , 584 , l95 lupton , r. , gunn , j. e. , ivezic , z. , gnapp , j. r. , kent , s. , & yasuda , n. 2001 , in asp conf .
238 , astronomical data analysis software and systems x , ed
. f. r. harnden , jr . , a. primini , & h. e. payne ( san
francisco : asp ) , 269 macconnell , d. j. 2003 , , 115 , 351 margon , b. et al . 2002 , , 124
, 1651 ( paper i ) mould , j. r. , schneider , d. p. , gordon , g. a. , aaronson , m. , & liebert , j. w. 1985 , , 97 , 130 pier , j. r. , munn , j. a. , hindsley , r. b. , hennessy , g. s. , kent , s. m. , lupton , r. h. , & ivezic , z. 2002 , , 125 , 1559 schroeder , j. et al . 2002 , , 34 , 1126 smith , j. a. , tucker , d. l. , kent , s. m. , et al .
2002 , , 123 , 2121 stoughton , c. et al .
2002 , , 123 , 485 ( early data release ) totten , e. j. & irwin , m. j. 1998 , , 294 , 1 totten , e. j. , irwin , m. j. , & whitelock , p. a. 2000 , , 314 , 630 wing , r. f. & jrgensen , u. g. 1996 , , 28 , 1382 yamashita , y. 1967 , pub .
xiii , 67 york , d. g. , adelman , j. , anderson , j. e. , et al .
2000 , , 120 , 1579 [email protected] & 2000.74 & 18.59 & 2.78 & 1.36 & 0.45 & 0.31 & 0.69 & 0.03 & 1.92 & @xmath60 & 110.1 & d + 000643.14 + 155800.8 & 2001.72 & 19.80 & 2.11 & 1.68 & 0.52 & 0.12 & & & & @xmath61 & 196.8 & d + [email protected] & & 18.15 & 3.16 & 1.47 & 0.57 & 0.24 & 0.58 & 0.56 & 2.06 & @xmath62 & 39.1 & u + [email protected] & 1998.73 & 19.51 & 1.33 & 1.57 & 0.51 & 0.10 & & & & @xmath63 & 175.8 & d + 001716.46 + 143840.9 & 2001.72 & 18.36 & 2.49 & 1.58 & 0.51 & 0.25 & 0.57 & 0.53 & 1.95 & @xmath64 & 149.9 & d + [email protected] & & 18.69 & 2.24 & 0.90 & 0.27 & 0.19 & & & & @xmath65 & 10.9 & u + [email protected] & & 19.44 & 3.81 & 1.89 & 0.47 & 0.28 & & & & @xmath66 & 141.8 & u + 003504.78 + 010845.9 & & 17.28 & 4.17 & 1.91 & 0.68 & 0.50 & 0.89 & 0.45 & 2.66 & @xmath67 & 331.5 & g@xmath68 + 003813.23 + 134551.1 & & 19.32 & 2.29 & 1.60 & 0.51 & 0.31 & & & & @xmath69 & 294.5 & u + 003937.35 + 152910.7 & 1999.78 & 18.61 & 1.68 & 1.05 & 0.30 & 0.17 & & & & @xmath70 & 168.4 & d + [email protected] & & 19.60 & 4.33 & 0.97 & 0.29 & 0.17 & & & & @xmath71 & 149.3 & u + [email protected] & 2000.74 & 19.44 & 4.72 & 1.64 & 0.59 & 0.25 & 0.14 & 0.67 & 2.43 & @xmath72 & 195.4 & d + [email protected] & & 17.54 & 3.14 & 2.00 & 0.82 & 0.45 & 0.72 & 0.32 & 3.20 & @xmath73 & 26.3 & g + [email protected] & 2000.74 & 18.83 & 2.63 & 1.76 & 0.60 & 0.30 & 0.85 & 0.16 & 2.31 & @xmath74 & 108.4 & d + [email protected] & 2000.74 & 16.98 & 2.66 & 1.75 & 0.55 & 0.16 & 0.77 & 0.56 & 1.99 & @xmath75 & 107.7 & d + [email protected] & & 19.23 & 4.26 & 1.46 & 0.54 & 0.14 & & & & @xmath76 & 166.4 & u@xmath77 + 012150.28 + 011302.8 & 1998.72 & 17.00 & 2.73 & 1.80 & 0.48 & 0.23 & 0.86 & 0.45 & 1.90 & @xmath78 & 122.9 & d@xmath79 + [email protected] & & 18.07 & 1.08 & 0.49 & 0.20 & 0.08 & 1.31 & @xmath590.59 & 1.18 & @xmath80 & 99.8 & u@xmath81 + [email protected] & & 17.67 & 1.76 & 1.23 & 0.34 & 0.24 & 0.91 & 0.29 & 1.55 & @xmath82 & 320.2 & g + 012526.74 + 000448.5 & 1998.73 & 19.35 & 1.69 & 1.66 & 0.54 & 0.16 & & & & @xmath83 & 96.9 & d + [email protected] & 2000.74 & 18.82 & 3.04 & 1.74 & 0.58 & 0.30 & 0.39 & 0.66 & 2.25 & @xmath84 & 211.9 & d + 013007.13 + 002635.3 & 1998.73 & 17.63 & 0.41 & 1.07 & 0.44 & 0.38 & 0.71 & 0.54 & 2.04 & @xmath85 & 131.8 & d@xmath86 + + + + + + + + + 013148.46 + 004230.7 & 1998.72 & 19.05 & 1.85 & 1.05 & 0.17 & 0.19 & & & & @xmath87 & 78.1 & d
+ 013249.99 + 000300.4 & & 20.52 & 1.36 & 1.21 & 0.34 & 0.47 & & & & & & ... + [email protected] & & 19.60 & 3.99 & 1.76 & 0.55 & 0.33 & & & & @xmath88 & 152.1 & u + [email protected] & & 17.62 & 1.09 & 0.42 & 0.09 & 0.05 & 0.30 & @xmath590.31 & 1.10 & @xmath89 & 143.1 & u@xmath81 + [email protected] & 1998.72 & 18.15 & 2.69 & 1.68 & 0.48 & 0.26 & 0.66 & 0.13 & 2.04 & @xmath90 & 128.9 & d
+ 020716.86 + 133034.7 & & 19.49 & 2.04 & 1.02 & 0.28 & 0.28 & & & & @xmath91 & 176.2 & u + 021119.82 + 003201.5 & & 19.73 & 3.30 & 1.65 & 0.56 & 0.28 & & & & @xmath92 & 249.2 & u + 023208.60 + 003639.3 & 1998.73 & 19.46 & 2.09 & 1.59 & 0.56 & 0.28 & & & & @xmath93 & 19.2 & d + [email protected] & 1999.79 & 19.47 & 1.94 & 1.11 & 0.27 & 0.17 & & & & @xmath94 & 137.4 & d + 030837.07 + 005157.1 & & 20.63 & 1.92 & 1.89 & 0.44 & 0.41 & & & & & & ... + [email protected] & & 17.96 & 1.16 & 0.53 & 0.19 & 0.08 & 0.27 & 0.56 & 1.18 & @xmath95 & 141.7 & u@xmath81 + [email protected] & & 18.67 & 2.48 & 1.92 & 0.71 & 0.30 & 0.79 & 0.33 & 2.41 & @xmath96 & 116.1 & u + [email protected] & 2000.74 & 16.17 & 1.00 & 0.39 & 0.17 & 0.04 & 0.36 & @xmath590.02 & 1.01 & @xmath97 & 171.4 & d@xmath81 + 073621.29 + 390725.2 & 2000.32 & 18.43 & 3.21 & 1.52 & 0.46 & 0.16 & 0.68 & 0.65 & 1.67 & @xmath98 & 230.1 & d + 073805.73 + 274148.8 & 2001.97 & 18.46 & 2.49 & 1.57 & 0.55 & 0.21 & 0.78 & 0.10 & 2.19 & @xmath99 & 159.4 & d + 074109.31 + 243603.0 & & 19.95 & 2.17 & 1.51 & 0.53 & 0.04 & & & & & & ... + 074638.21 + 400403.5 & & 19.52 & 3.08 & 1.68 & 0.51 & 0.21 & & & & & & ... +
074710.84 + 251619.6 & & 18.12 & 1.06 & 0.42 & 0.15 & 0.10 & & & & @xmath100 & 177.5 & u@xmath81
+ 075116.37 + 391201.4 & & 17.61 & 3.47 & 1.60 & 0.58 & 0.43 & 0.81 & 0.30 & 2.43 & @xmath101 & 168.9 & g + 075228.63 + 280547.5 & 2002.02 & 20.09 & 1.55 & 1.63 & 0.49 & 0.10 & & & & @xmath102 & 191.2 & d + 075953.64 + 434021.1 & & 19.48 & 3.64 & 1.91 & 0.63 & 0.15 & & & & @xmath103 & 324.3 & u + 080040.60 + 292437.9 & & 20.46 & 2.13 & 1.78 & 0.58 & 0.06 & & & & & & ... @xmath104 + + + + + + + + + 080046.72 + 364107.1 & & 17.37 & 1.06 & 0.41 & 0.13 & 0.08 & 0.63 & @xmath590.01 & 1.04 & @xmath105 & 220.4 & u@xmath81 + 080156.39 + 395043.8 & & 18.46 & 4.16 & 1.65 & 0.54 & 0.26 & 0.59 & 0.50 & 2.16 & @xmath106 & 125.7 & u + 080806.06 + 293634.1 & & 18.34 & 3.22 & 1.18 & 0.33 & 0.35 & 0.71 & 0.10 & 1.95 & @xmath107 & 6.6 & u + 080908.15 + 360129.0 & & 18.89 & 0.32 & 0.44 & 0.25 & 0.10 & & & & @xmath108 & 192.4 & u + 082127.90 + 333037.2 & & 19.46 & 2.62 & 1.21 & 0.33 & 0.22 & & & & @xmath109 & 256.9 & u + 082251.41 + 461232.3 & 2000.26 & 17.23 & 2.79 & 1.40 & 0.47 & 0.23 & 0.82 & 0.20 & 1.73 & @xmath110 & 221.3 & d + 082626.76 + 470911.7 & 2000.26 & 17.77 & 2.80 & 1.46 & 0.48 & 0.31 & 0.60 & 0.17 & 1.99 & @xmath111 & 148.9 & d + 084130.12 + 435136.7 & 2001.15 & 19.45 & 2.52 & 1.52 & 0.48 & 0.23 & & & & @xmath112 & 354.7 & d + 084720.16 + 003340.4 & & 20.73 & 1.14 & 1.92 & 0.56 & 0.21 & & & & & & ... +
085212.77 + 383721.9 & 2002.02 & 19.93 & 3.10 & 1.69 & 0.51 & 0.18 & & & & @xmath113 & 204.2 & d + 085853.28 + 012243.5 & 2000.34 & 18.31 & 2.46 & 1.99 & 0.67 & 0.36 & 0.71 & 0.72 & 2.49 & @xmath114 & 124.3 & d + [email protected] & 1999.22 & 18.44 & 3.27 & 1.48 & 0.47 & 0.26 & & & & @xmath115 & 168.4 & d + 090208.05 + 435503.8 & & 17.43 & 1.19 & 0.47 & 0.14 & 0.06 & & & & @xmath116 & 135.2 & u@xmath81 + 091007.61 + 521612.4 & & 17.01 & 2.94 & 1.63 & 0.54 & 0.27 & 0.72 & 0.20 & 2.09 & @xmath117 & 177.4 & g + 091019.21 + 041211.6 & & 19.05 & 3.42 & 1.50 & 0.54 & 0.38 & & & & @xmath118 & 279.3 & u + 091335.63 + 492444.7 & & 19.83 & 2.12 & 1.60 & 0.49 & 0.25 & & & & @xmath119 & 212.6 & u + 091336.99 + 034839.1 & 2001.14 & 18.86 & 2.34 & 1.27 & 0.38 & 0.23 & & & & @xmath120 & 150.0 & d + 092508.21 + 010545.4 & & 17.94 & 2.66 & 1.29 & 0.44 & 0.32 & 0.49 & 0.27 & 2.05 & @xmath121 & 190.6 & g + 092545.46 + 424929.0 & 2002.02 & 17.51 & 1.11 & 0.43 & 0.11 & 0.04 & 0.25 & @xmath590.26 & 0.98 & @xmath122 & 156.6 & d@xmath81 + 093057.12 + 033453.6 & & 19.23 & 2.41 & 1.60 & 0.55 & 0.35 & & & & @xmath123 & 349.8 & u + 093203.67 + 002753.8 & & 20.40 & 1.00 & 1.66 & 0.45 & 0.10 & & & & & & ... + 093741.12 + 470836.2 & 2001.96 & 18.6 & & & & & 0.20 & 0.93 & & @xmath124 & 199.9 & d
+ 094209.94 + 535719.2 & & 16.49 & 2.61 & 1.44 & 0.53 & 0.30 & 0.72 & 0.21 & 1.98 & @xmath125 & 352.4 & g + + + + + + + + 094318.73 + 032743.7 & 2001.14 & 18.11 & 2.70 & 1.38 & 0.49 & 0.13 & 0.54 & 0.04 & 1.66 & @xmath126 & 229.4 & d + 094511.57 + 503214.5 & 2002.02 & 18.86 & 5.21 & 1.32 & 0.40 & 0.22 & & & & @xmath127 & 202.7 & d@xmath128 + 094858.68 + 583020.5 & 2000.26 & 18.77 & 2.55 & 1.15 & 0.33 & 0.22 & & & & @xmath129 & 163.2 & d + 095005.09 + 584123.8 & 2000.26 & 17.03 & 2.93 & 1.43 & 0.47 & 0.18 & 0.67 & 0.41 & 1.76 &
@xmath130 & 162.5 & d + 095516.39 + 012129.8 & & 18.36 & 2.85 & 1.60 & 0.53 & 0.26 & 0.86 & @xmath590.05 & 1.97 & @xmath131 & 100.9 & u + 100132.62 + 020402.7 & & 19.41 & 3.84 & 1.68 & 0.61 & 0.20 & & & & @xmath132 & 237.9 & u + 100432.45 + 004337.7 & 1999.22 & 19.99 & 1.78 & 1.71 & 0.56 & 0.19 & & & & @xmath133 & 169.1 & d + 100627.47 + 462117.4 & & 17.52 & 0.96 & 0.44 & 0.15 & 0.10 & 0.26 & 0.19 & 1.31 & @xmath134 & 164.5 & u@xmath81 + 100638.20 + 473039.5 & 2001.97 & 19.96 & 1.88 & 1.65 & 0.54 & 0.27 & & & & @xmath135 & 161.4 & d + 100812.04 + 012657.2 & 2000.34 & 20.17 & 3.60 & 1.73 & 0.51 & 0.15 & & & & @xmath136 & 190.7 & d + 100913.95 + 514924.6 & 2002.02 & 18.55 & 2.66 & 1.44 & 0.45 & 0.16 & 0.64 & 0.15 & 1.80 & @xmath137 & 217.4 & d + 100958.70 + 010313.2 & 1999.22 & 17.07 & 2.26 & 1.03 & 0.27 & 0.14 & 0.41 & 0.03 & 1.44 & @xmath138 & 171.8 & d
+ 101007.11 + 031444.1 & & 18.96 & 1.97 & 1.78 & 0.63 & 0.40 & 1.03 & 0.21 & 2.25 & @xmath139 & 129.9 & u + 101210.81 + 465249.1 & 2002.04 & 18.36 & 2.76 & 1.39 & 0.46 & 0.29 & 1.01 & 0.39 & 1.66 & @xmath140 & 218.5 & d + 101422.75 + 464737.3 & & 19.77 & 2.34 & 1.52 & 0.52 & 0.09 & & & & @xmath141 & 149.1 & u + 102239.58 + 622344.7 & & 19.88 & 2.24 & 1.79 & 0.37 & 0.34 & & & & @xmath142 & 250.8 & u + 102433.76 + 562855.1 & & 19.51 & 2.37 & 1.78 & 0.56 & 0.20 & & & & @xmath143 & 308.6 & u + 102908.38 + 591853.8 & & 18.93 & 1.91 & 1.36 & 0.46 & 0.18 & & & & @xmath144 & 175.0 & u + 102941.07 + 020613.1 & & 19.48 & 2.62 & 1.75 & 0.58 & 0.26 & & & & @xmath145 & 169.4 & u + 103758.57 + 612559.3 & 2000.26 & 18.08 & 1.00 & 0.36 & 0.19 & @xmath590.01 & & & & @xmath146 & 156.8 & d@xmath81 + 104138.55 + 643219.8 & & 18.48 & 2.45 & 1.89 & 0.75 & 0.42 & 0.68 & 0.58 & 2.58 & @xmath147 & 208.4 & u + 104316.17 + 573755.7 & & 19.96 & 1.54 & 1.55 & 0.54 & 0.11 & & & & & & ... + + + + + + + + + 105629.95 + 012208.8 & & 17.52 & 1.09 & 0.54 & 0.18 & 0.05 & 0.09 & @xmath590.77 & 1.32 & @xmath148 & 180.1 & u@xmath81 + 105806.72 + 032514.7 & & 18.40 & 2.77 & 1.28 & 0.37 & 0.23 & 0.59 & 0.04 & 1.93 & @xmath149 & 160.5 & u + 110558.06 + 510942.9 & 2001.97 & 18.91 & 3.76 & 1.88 & 0.54 & 0.15 & 0.99 & 0.73 & 1.93 & @xmath150 & 204.3 & d + 110607.96 + 495055.0 & 2001.97 & 18.82 & 2.89 & 1.56 & 0.55 & 0.14 & & & & @xmath151 & 200.0 & d + 111833.84 + 563958.1 & 2001.89 & 16.65 & 0.88 & 0.35 & 0.15 & 0.02 & 0.27 & 0.18 & 0.98 & @xmath152 & 222.0 & d@xmath81 + 112034.87 + 555937.2 & 2001.97 & 16.15 & 2.75 & 1.31 & 0.43 & 0.15 & 0.64 & 0.14 & 1.74 & @xmath153 & 200.9 & d
+ 112417.34 + 054921.3 & & 20.30 & 4.73 & 1.19 & 0.31 & 0.24 & & & & @xmath154 & 359.6 & u + 112631.31 + 053705.3 & & 20.26 & 3.42 & 1.79 & 0.55 & 0.23 & & & & @xmath155 & 166.7 & u + 112650.77 + 645646.0 & & 19.09 & 3.15 & 1.39 & 0.45 & 0.25 & & & & @xmath156 & 198.7 & u + 112654.85 + 645155.8 & & 20.30 & 2.27 & 1.78 & 0.57 & 0.09 & & & & @xmath92 & 225.9 & u + 112746.45 + 041502.5 & 2001.14 & 18.92 & 2.99 & 1.54 & 0.49 & 0.29 & & & & @xmath157 & 148.7 & d + 112752.79 + 515815.2 & 2001.97 & 17.68 & 2.89 & 1.21 & 0.38 & 0.24 & 0.61 & 0.27 & 1.66 & @xmath158 & 169.9 & d + 112801.68 + 004034.7 & & 18.81 & 2.46 & 1.41 & 0.38 & 0.23 & & & & @xmath159 & 171.0 & u@xmath160 + 112836.57 + 011331.2 & & 17.23 & 1.03 & 0.43 & 0.17 & 0.09 & 0.60 & 0.01 & 1.05 & @xmath161 & 288.5 & u@xmath81 + 112900.75 + 510742.0 & 2001.97 & 20.14 & 1.81 & 1.44 & 0.48 & 0.04 & & & & @xmath162 & 167.2 & d + 112950.38 + 003344.9 & 1999.22 & 18.28 & 2.27 & 1.40 & 0.41 & 0.33 & 0.77 & 0.17 & 1.81 & @xmath163 & 234.2 & d + 113040.49 + 525039.2 & & 17.34 & 1.14 & 0.45 & 0.13 & 0.03 & 0.28 & @xmath590.36 & 1.01 & @xmath164 & 266.8 & u@xmath81 + [email protected] & 2000.17 & 19.14 & 3.34 & 1.28 & 0.37 & 0.23 & & & & @xmath165 & 230.5 & d
+ 113931.62 + 050231.2 & & 19.59 & 2.03 & 1.69 & 0.56 & 0.39 & & & & @xmath166 & 164.8 & u + 114125.85 + 010504.3 & & 17.29 & 2.12 & 0.89 & 0.15 & 0.29 & 0.42 & 0.35 & 1.50 & @xmath167 & 85.7 & g + 114355.88 + 625924.7 & 2001.38 & 19.41 & 1.62 & 1.55 & 0.51 & 0.32 & & & & @xmath168 & 167.6 & d + 114442.36 + 043641.9 & 2001.29 & 19.17 & 2.31 & 1.39 & 0.31 & 0.29 & & & & @xmath169 & 275.5 & d + + + + + + + + + 114731.68 + 003724.5 & 1999.22 & 18.60 & 1.71 & 1.05 & 0.23 & 0.14 & & & & @xmath170 & 171.9 & d + [email protected] & 2000.17 & 20.00 & 1.43 & 1.89 & 0.49 & 0.38 & & & & @xmath171 & 233.6 & d + 115255.58 + 055337.7 & & 17.39 & 1.09 & 0.59 & 0.15 & 0.10 & 0.32 & 1.01 & 0.94 & @xmath172 & 167.1 & u@xmath81 + 115555.74 + 490012.1 & 2002.22 & 18.81 & 2.11 & 1.76 & 0.83 & 0.06 & 0.76 & 0.66 & 2.07 & @xmath173 & 224.8 & d@xmath128 + [email protected] & & 18.76 & 1.79 & 1.08 & 0.17 & 0.10 & & & & @xmath174 & 84.8 & u + 115929.28 + 640501.8 & 2001.07 & 17.66 & 2.84 & 1.54 & 0.51 & 0.19 & 0.49 & 0.14 & 2.07 & @xmath175 & 233.8 & d + 120031.66 + 035002.8 & & 20.14 & 3.38 & 1.72 & 0.48 & 0.24 & & & & & & ... + 120241.19 + 604322.6 & & 19.85 & 2.57 & 1.43 & 0.44 & 0.12 & & & & & & ... + 120336.68 + 492232.6 & 2002.11 & 19.50 & 2.37 & 1.61 & 0.55 & 0.07 & & & & @xmath176 & 253.4 & d + [email protected] & & 18.24 & 2.90 & 1.40 & 0.46 & 0.35 & 0.85 & 0.24 & 1.84 & @xmath177 & 263.4 & u + 121243.75 + 005722.3 & 1999.22 & 18.90 & 2.27 & 1.17 & 0.31 & 0.28 & & & & @xmath178 & 204.2 & d + [email protected] & & 17.18 & 1.22 & 0.55 & 0.24 & 0.06 & 0.46 & 0.08 & 1.16 & @xmath179 & 322.4 & u@xmath81 + [email protected] & & 17.29 & 1.21 & 0.48 & 0.21 & 0.17 & 0.43 & 0.28 & 1.23 & @xmath180 & 47.9 & u@xmath81 + 122843.93 + 532937.7 & 2001.97 & 19.36 & 3.67 & 1.41 & 0.45 & 0.19 & & & & @xmath181 & 171.8 & d + 123626.98 + 025901.8 & & 18.98 & 2.62 & 1.79 & 0.55 & 0.40 & & & & @xmath182 & 311.7 & u + 123834.40 + 023856.0 & & 18.49 & 0.57 & 1.38 & 0.23 & 0.20 & & & & @xmath183 & 128.8 & u@xmath81 + 124114.82 + 605402.9 & & 19.94 & 1.55 & 1.61 & 0.56 & 0.05 & & & & @xmath184 & 198.2 & u + 124116.94 + 051942.2 & 2001.21 & 19.48 & 1.55 & 1.24 & 0.36 & 0.17 & & & & @xmath185 & 220.4 & d + 124120.87 + 032848.2 & & 20.61 & 1.92 & 0.88 & 0.31 & 0.63 & & & & & & ... + 124215.10 + 635616.6 & 2001.21 & 19.22 & 3.28 & 1.57 & 0.51 & 0.19 & & & & @xmath186 & 171.1 & d
+ 124436.20 + 051309.1 & & 19.49 & 1.95 & 1.65 & 0.52 & 0.31 & & & & @xmath187 & 318.1 & u + 124505.36 + 044053.3 & & 18.31 & 2.01 & 0.94 & 0.19 & 0.23 & & & & @xmath188 & 211.1 & u + + + + + + + + + 125149.87 + 013001.8 & & 16.24 & 4.08 & 3.52 & 1.06 & 0.49 & 1.12 & 0.75 & 4.17 & @xmath189 & 62.2 & g@xmath190 + 125431.75 + 593053.6 & & 19.41 & 3.24 & 1.41 & 0.48 & 0.29 & & & & @xmath191 & 278.4 & u + 125450.29 + 614842.3 & & 19.56 & 3.38 & 1.74 & 0.51 & 0.36 & & & & @xmath192 & 179.6 & u + 125520.36 + 631326.7 & 2001.39 & 17.25 & 2.74 & 1.38 & 0.47 & 0.18 & 0.51 & 0.30 & 1.69 & @xmath193 & 175.2 & g + 130219.89 + 620559.0 & 2001.39 & 18.66 & 2.09 & 1.00 & 0.25 & 0.07 & & & & @xmath194 & 178.2 & d + 130227.53 + 055512.3 & 2001.29 & 19.25 & 4.59 & 1.85 & 0.55 & 0.25 & & & & @xmath195 & 182.7 & d + 130510.57 + 054844.0 & & 20.12 & 2.85 & 1.59 & 0.51 & 0.13 & & & & & & ... +
130744.53 + 600903.7 & & 16.92 & 1.18 & 0.55 & 0.18 & 0.08 & 0.63 & @xmath590.26 & 0.99 & & & ... @xmath81 + [email protected] & 1999.22 & 19.44 & 3.44 & 1.79 & 0.59 & 0.33 & 0.86 & 0.77 & 2.33 & @xmath196 & 248.6 & d + 131841.09 + 055427.4 & & 19.89 & 1.79 & 1.88 & 0.67 & 0.32 & & & & @xmath197 & 171.4 & u + 131921.25 + 025218.9 & & 19.49 & 2.17 & 1.62 & 0.47 & 0.36 & & & & @xmath198 & 208.6 & u + 131935.39 + 585454.0 & & 20.35 & 2.63 & 1.56 & 0.51 & 0.05 & & & & & & ... +
132307.04 + 052048.0 & & 19.34 & 2.16 & 1.06 & 0.28 & 0.21 & & & & & & ... + 132309.75 + 050647.7 & 2001.21 & 18.61 & 2.07 & 1.10 & 0.31 & 0.15 & 0.54 & 0.56 & 1.90 & @xmath199 & 152.7 & d
+ 132604.43 + 605014.0 & & 19.16 & 2.55 & 1.47 & 0.45 & 0.27 & & & & & & ... + 132622.21 + 003743.7 & 1999.22 & 17.22 & 1.02 & 0.51 & 0.19 & 0.07 & 0.21 & 0.22 & 1.39 & @xmath200 & 245.1 & d@xmath81 + 132840.75 + 002716.9 & & 19.40 & 1.91 & 1.17 & 0.18 & 0.19 & & & & @xmath201 & 117.7 & u + [email protected] & & 19.47 & 4.00 & 1.27 & 0.42 & 0.19 & & & & @xmath202 & 265.4 & u + 133034.42 + 030307.0 & & 16.33 & 3.04 & 1.07 & 0.33 & 0.28 & 0.67 & 0.23 & 1.77 & @xmath203 & 66.4 & g + 134745.49 + 655730.7 & & 19.01 & 2.13 & 1.75 & 0.56 & 0.27 & 1.11 & 0.28 & 2.00 & @xmath204 & 150.8 & u + 135057.54 + 050841.0 & 2001.29 & 15.96 & 2.75 & 1.42 & 0.41 & 0.29 & 0.64 & 0.21 & 1.96 & @xmath205 & 279.0 & d + 135111.97 + 043153.0 & 2001.21 & 19.81 & 2.10 & 1.60 & 0.50 & 0.32 & & & & @xmath206 & 210.0 & d + + + + + + + + + 135124.90 + 040920.2 & & 18.67 & 2.77 & 1.30 & 0.40 & 0.25 & 0.78 & 0.54 & 1.91 & @xmath207 & 198.4 & u + 135155.17 + 052009.1 & 2001.21 & 17.34 & 2.90 & 1.77 & 0.56 & 0.24 & 0.90 & 0.35 & 2.00 & @xmath208 & 187.7 & d + [email protected] & 1999.22 & 16.60 & 2.68 & 1.59 & 0.54 & 0.27 & 0.81 & 0.19 & 1.99 & @xmath209 & 236.6 & d
+ 135534.06 + 021151.6 & & 20.67 & 1.08 & 1.66 & 0.32 & @xmath590.09 & & & & & & ... + 140327.81 + 024725.1 & 2000.34 & 19.42 & 1.73 & 1.43 & 0.49 & 0.40 & & & & @xmath210 & 149.9 & d + [email protected] & 2001.39 & 20.04 & 3.14 & 1.21 & 0.45 & @xmath590.10 & & & & @xmath211 & 172.5 & d + 140425.25 + 013252.7 & & 15.93 & 4.33 & 1.53 & 0.56 & 0.43 & 0.66 & 0.26 & 2.40 & @xmath212 & 27.6 & g + [email protected] & 2001.39 & 19.28 & 1.66 & 1.46 & 0.46 & 0.12 & & & & @xmath213 & 276.5 & d + 141111.54 + 614933.7 & 2001.14 & 18.71 & 2.39 & 1.09 & 0.30 & 0.20 & & & & @xmath214 & 251.8 & d + 141324.15 + 015547.8 & 2000.34 & 18.73 & 3.02 & 1.93 & 0.52 & 0.28 & 0.58 & 0.54 & 2.33 & @xmath215 & 290.4 & d + 141412.57 + 030647.6 & & 15.92 & 3.08 & 1.18 & 0.35 & 0.32 & 0.67 & 0.13 & 1.92 & @xmath216 & 310.7 & g + 141421.39 + 043129.3 & & 20.22 & 1.33 & 1.78 & 0.49 & 0.24 & & & & @xmath217 & 270.1 & u + [email protected] & & 19.05 & 2.20 & 1.51 & 0.36 & 0.42 & 1.01 & 0.22 & 2.08 & @xmath218 & 97.7 & u + [email protected] & & 17.69 & 2.77 & 1.13 & 0.33 & 0.27 & 0.51 & 0.03 & 1.89 & @xmath219 & 73.5 & g + 142535.12 + 024730.8 & & 19.19 & 2.66 & 0.98 & 0.20 & 0.25 & & & & @xmath220 & 237.5 & u + 143156.34 + 034228.4 & & 20.42 & 3.45 & 1.82 & 0.51 & 0.20 & & & & & & ... +
143328.12 + 595808.9 & 2001.14 & 18.02 & 0.96 & 0.44 & 0.18 & 0.08 & 0.58 & 0.05 & 1.55 & @xmath221 & 171.0 & d@xmath81 + [email protected] & 2001.39 & 19.44 & 1.62 & 1.85 & 0.60 & 0.19 & 0.52 & 0.96 & 2.48 & @xmath222 & 153.2 & d + [email protected] & & 17.86 & 2.59 & 1.72 & 0.66 & 0.26 & 0.88 & 0.51 & 2.21 & @xmath223 & 304.6 & u + 144428.79 + 021430.7 & & 19.36 & 3.45 & 1.60 & 0.54 & 0.36 & & & & @xmath224 & 310.7 & u + 144448.40 + 043944.2 & 2000.36 & 19.59 & 3.65 & 1.85 & 0.49 & 0.20 & & & & @xmath225 & 291.2 & d + 144841.80 + 042844.5 & & 17.19 & 2.30 & 1.11 & 0.30 & 0.24 & 0.49 & 0.13 & 1.64 & @xmath226 & 291.8 & g + + + + + + + + + 144845.39 + 031059.4 & & 19.21 & 2.15 & 0.97 & 0.23 & 0.22 & & & & @xmath227 & 94.5 & u + [email protected] & & 20.43 & 4.38 & 1.23 & 0.32 & 0.36 & & & & & & ... +
144945.37 + 012656.2 & & 15.64 & 3.45 & 1.26 & 0.39 & 0.41 & 0.75 & 0.18 & 2.11 & @xmath228 & 52.3 & g + [email protected] & 2001.45 & 20.09 & 1.02 & 1.47 & 0.41 & 0.09 & & & & @xmath229 & 256.6 & d + 145318.82 + 600421.1 & 2000.26 & 17.59 & 3.12 & 1.61 & 0.60 & 0.29 & 0.56 & 0.29 & 1.99 & @xmath230 & 267.7 & d + 145459.75 + 563807.2 & 2001.38 & 17.85 & 2.64 & 1.61 & 0.47 & 0.24 & 0.69 & 0.53 & 1.96 & @xmath231 & 179.7 & d + 145506.01 + 002239.1 & & 17.71 & 1.15 & 0.48 & 0.21 & 0.06 & & & & @xmath232 & 198.4 & u@xmath81 + 145512.68 + 614101.1 & & 19.43 & 1.72 & 1.89 & 0.63 & 0.29 & 0.80 & 0.45 & 2.57 & @xmath233 & 270.6 & u + 145933.45 + 033310.2 & 2000.36 & 19.02 & 2.30 & 1.33 & 0.43 & 0.29 & & & & @xmath234 & 191.1 & d + 150051.84 + 022656.1 & & 18.82 & 3.21 & 1.12 & 0.30 & 0.27 & & & & @xmath235 & 174.3 & u + [email protected] & 2001.39 & 17.86 & 2.79 & 1.34 & 0.38 & 0.22 & 0.10 & 0.44 & 1.86 & @xmath236 & 230.3 & d + 150430.04 + 561603.8 & & 19.15 & 2.08 & 1.14 & 0.33 & 0.17 & & & & @xmath237 & 251.6 & u + 150713.81 + 595402.6 & 2001.22 & 19.36 & 3.24 & 1.42 & 0.44 & 0.37 & & & & @xmath238 & 264.0 & d + [email protected] & & 19.20 & 3.60 & 1.74 & 0.50 & 0.29 & & & & & & ... +
152218.79 + 582629.3 & & 19.67 & 2.39 & 1.66 & 0.52 & 0.27 & & & & & & ... + 152352.14 + 460216.5 & 2002.35 & 19.26 & 2.36 & 1.31 & 0.37 & 0.26 & & & & @xmath239 & 144.6 & d
+ 152410.66 + 453657.3 & & 20.00 & 1.49 & 1.86 & 0.62 & 0.28 & & & & @xmath240 & 217.0 & u + 152434.12 + 444956.3 & 2002.35 & 19.33 & 2.22 & 1.74 & 0.55 & 0.23 & & & & @xmath241 & 187.6 & d + 152525.47 + 450706.5 & & 19.54 & 1.39 & 0.68 & 0.29 & 0.11 & & & & @xmath242 & 199.3 & u@xmath243 + 152702.75 + 434517.4 & 2002.35 & 16.25 & 2.99 & 1.68 & 0.62 & 0.37 & 0.69 & 0.29 & 2.25 & @xmath244 & 289.5 & d + 153107.09 + 522200.6 & 2001.39 & 18.24 & 2.31 & 1.20 & 0.36 & 0.23 & 1.03 & 0.12 & 1.49 & @xmath245 & 159.5 & d + 153532.93 + 011016.4 & & 19.29 & 2.17 & 1.96 & 0.71 & 0.31 & 0.58 & 0.55 & 3.01 & @xmath246 & 265.0 & u + + + + + + + + + 153732.19 + 004343.1 & 1999.22 & 17.63 & 2.55 & 1.82 & 0.63 & 0.38 & 0.84 & 0.27 & 2.42 & @xmath247 & 144.3 & d + 154156.85 + 514421.3 & 2001.38 & 17.5 & & & & & 0.50 & 0.26 & & @xmath248 & 199.7 & d@xmath41 + 154426.04 + 023623.7 & 2001.21 & 15.67 & 2.73 & 1.40 & 0.44 & 0.30 & 0.71 & 0.15 & 1.92 & @xmath249 & 221.4 & d + 154839.88 + 411621.1 & 2002.35 & 19.54 & 2.78 & 1.59 & 0.46 & 0.28 & & & & @xmath250 & 251.3 & d + 154900.25 + 514156.0 & 2001.29 & 17.68 & 3.47 & 1.68 & 0.57 & 0.31 & 0.76 & 0.22 & 2.12 & @xmath80 & 222.6 & d + 155043.84 + 571342.5 & & 20.48 & 1.48 & 1.75 & 0.55 & 0.15 & & & & & & ... + 155441.23 + 514601.1 & & 20.81 & 0.83 & 1.83 & 0.49 & 0.16 & & & & & & ... + 155745.94 + 480207.5 & 2001.38 & 19.15 & 2.16 & 1.52 & 0.47 & 0.26 & & & & @xmath110 & 260.2 & d + [email protected] & & 19.06 & 2.96 & 1.70 & 0.60 & 0.22 & & & & @xmath251 & 174.2 & u@xmath128 + 162923.70 + 455729.6 & 2000.26 & 19.67 & 2.62 & 1.68 & 0.50 & 0.16 & & & & @xmath252 & 234.2 & d + 163045.02 + 004025.2 & 1999.22 & 19.99 & 3.24 & 1.80 & 0.57 & 0.15 & & & & @xmath253 & 245.0 & d + 164410.81 + 395642.5 & 2000.26 & 18.73 & 2.70 & 1.70 & 0.52 & 0.22 & 0.21 & 0.87 & 2.19 & @xmath254 & 256.9 & d + 164619.36 + 383333.6 & 2001.38 & 19.63 & 2.62 & 1.46 & 0.37 & 0.25 & & & & @xmath255 & 238.6 & d + 164650.57 + 405509.4 & & 19.53 & 2.73 & 1.58 & 0.48 & 0.28 & & & & @xmath256 & 222.1 & u + 165141.94 + 352012.6 & & 16.65 & 0.89 & 0.36 & 0.14 & 0.06 & 0.40 & 0.03 & 0.95 & @xmath148 & 203.8 & u@xmath81 + 165532.15 + 410112.6 & & 20.10 & 2.32 & 1.64 & 0.51 & 0.19 & & & & & & ... +
165948.02 + 391241.7 & & 19.63 & 3.33 & 1.86 & 0.58 & 0.38 & & & & @xmath257 & 165.7 & u + 170219.09 + 351034.5 & 2001.39 & 16.01 & 2.39 & 1.07 & 0.28 & 0.19 & 0.56 & 0.16 & 1.55 & @xmath258 & 142.3 & d + 171303.09 + 285902.2 & 2001.39 & 18.71 & 2.97 & 1.68 & 0.63 & 0.35 & 0.74 & 0.24 & 2.31 & @xmath259 & 341.5 & d + 171502.33 + 313231.5 & & 19.37 & 3.49 & 1.55 & 0.47 & 0.21 & & & & @xmath260 & 174.8 & u + 171606.44 + 275255.3 & 2001.39 & 19.40 & 2.74 & 1.58 & 0.50 & 0.24 & & & & @xmath261 & 178.8 & d + 171638.50 + 274937.7 & 2001.39 & 17.95 & 2.35 & 1.25 & 0.39 & 0.26 & 0.62 & 0.14 & 1.82 & @xmath262 & 255.6 & d + 171942.38 + 575837.7 & & 16.81 & 3.29 & 1.18 & 0.33 & 0.34 & 0.62 & 0.09 & 2.04 & @xmath263 & 222.2 & g@xmath264 + + + + + + + + 171957.65 + 575005.5 & & 16.47 & 1.25 & 1.25 & 0.28 & 0.41 & 0.67 & 0.25 & 2.08 & @xmath265 & 154.7 & g@xmath266 + 172038.83 + 575934.4 & & 17.76 & 2.66 & 1.17 & 0.29 & 0.24 & 0.42 & @xmath590.17 & 1.85 & @xmath267 & 183.9 & g@xmath268 + 172053.90 + 293416.3 & & 20.15 & 1.97 & 1.79 & 0.59 & 0.26 & & & & @xmath269 & 192.0 & u + 172823.74 + 634507.4 & & 18.73 & 3.03 & 1.06 & 0.30 & 0.15 & & & & @xmath270 & 172.2 & u@xmath128 + 172909.13 + 594034.8 & & 20.19 & 0.16 & 1.85 & 0.24 & 0.11 & & & & 0.0 & & u@xmath271 + 173650.55 + 563800.5 & & 19.19 & 2.53 & 1.73 & 0.67 & 0.32 & & & & @xmath272 & 276.4 & u + [email protected] & & 19.83 & 1.40 & 1.19 & 0.13 & 0.63 & & & & & & ... + 213430.27 + 003119.1 & 2001.72 & 18.97 & 1.52 & 1.80 & 0.59 & 0.41 & 0.68 & 0.30 & 2.43 & @xmath273 & 201.9 & d + [email protected] & & 18.46 & 2.78 & 1.59 & 0.55 & 0.22 & 0.66 & 1.04 & 1.94 & & & ... + 214650.96 + 005829.4 & 2001.72 & 18.73 & 2.17 & 1.32 & 0.39 & 0.33 & @xmath590.76 & 1.87 & 1.90 & @xmath274 & 181.1 & d + 215000.91 + 115343.8 & 2000.74 & 18.66 & 2.60 & 1.71 & 0.59 & 0.27 & 0.74 & 0.41 & 2.19 & @xmath133 & 262.1 & d + [email protected] & & 16.30 & 2.82 & 1.03 & 0.30 & 0.27 & 0.57 & 0.13 & 1.82 & @xmath275 & 175.8 & g + [email protected] & & 16.81 & 1.69 & 0.66 & 0.00 & 0.35 & 0.44 & @xmath590.01 & 1.21 & @xmath276 & 75.4 & u@xmath81 + [email protected] & 1999.79 & 19.25 & 2.47 & 1.44 & 0.46 & 0.25 & & & & @xmath277 & 98.8 & d + 221450.93 + 011250.3 & 1999.78 & 19.87 & 2.57 & 1.62 & 0.56 & 0.11 & & & & @xmath278 & 215.7 & d + [email protected] & & 18.86 & 3.04 & 1.67 & 0.55 & 0.30 & 0.64 & 0.56 & 2.29 & @xmath279 & 102.4 & u + 221854.26 + 010026.1 & & 19.22 & 2.21 & 1.21 & 0.37 & 0.18 & & & & @xmath280 & 166.3 & u + [email protected] & 1998.72 & 19.43 & 3.95 & 1.87 & 0.57 & 0.40 & & & & @xmath281 & 179.5 & d + [email protected] & & 19.36 & 5.55 & 1.93 & 0.60 & 0.15 & & & & @xmath282 & 160.2 & u + 223728.04 + 124213.8 & & 19.77 & 1.89 & 1.61 & 0.51 & 0.20 & & & & & & ... + 223937.82 + 130924.9 & 2000.74 & 17.53 & 2.66 & 1.53 & 0.34 & 0.23 & 0.75 & @xmath590.11 & 1.85 & @xmath283 & 82.6 & d + [email protected] & & 20.16 & 2.60 & 1.48 & 0.30 & 0.13 & & & & & & ... + [email protected] & 2001.72 & 19.64 & 4.33 & 1.79 & 0.76 & 0.48 & & & & @xmath284 & 252.9 & d + 230255.00 + 005904.4 & 1999.79 & 17.69 & 2.42 & 1.57 & 0.54 & 0.32 & 0.81 & 0.20 & 2.07 & @xmath285 & 231.5 & d + [email protected] & & 18.89 & 2.45 & 1.23 & 0.37 & 0.24 & 0.67 & @xmath590.57 & 2.13 & @xmath219 & 106.0 & u + + + + + + 231340.54 + 143600.8 & 2001.72 & 19.7 & & & & & & & & @xmath286 & 169.4 & d@xmath41 + [email protected] & 2000.74 & 19.84 & 2.62 & 1.80 & 0.48 & 0.23 & & & & @xmath287 & 183.2 & d + 232310.27 + 010826.4 & & 20.22 & 0.94 & 1.66 & 0.44 & 0.24 & & & & & & ... +
234038.86 + 150841.8 & & 18.92 & 2.66 & 1.51 & 0.50 & 0.29 & & & & @xmath288 & 357.7 & u + | a search of more than 3,000 square degrees of high latitude sky by the sloan digital sky survey has yielded 251 faint high - latitude carbon stars ( fhlcs ) , the large majority previously uncataloged .
we present homogeneous spectroscopy , photometry , and astrometry for the sample .
the objects lie in the @xmath0 range , and exhibit a wide variety of apparent photospheric temperatures , ranging from spectral types near m to as early as f. proper motion measurements for 222 of the objects show that at least @xmath1 , and quite probably more than @xmath2 , of these objects are actually low luminosity dwarf carbon ( dc ) stars , in agreement with a variety of recent , more limited investigations which show that such objects are the numerically dominant type of star with c@xmath3 in the spectrum .
this sdss homogeneous sample of @xmath4 dc stars now constitutes @xmath5 of all known carbon dwarfs , and will grow by another factor of 2 - 3 by the completion of the survey . as the spectra of the dc and
the faint halo giant c stars are very similar ( at least at spectral resolution of @xmath6 ) despite a difference of 10 mag in luminosity , it is imperative that simple luminosity discriminants other than proper motion be developed .
we use our enlarged sample of fhlcs to examine a variety of possible luminosity criteria , including many previously suggested , and find that , with certain important caveats , _ jhk _ photometry may segregate dwarfs and giants . |
the orbital magnetism of bloch electrons has been an outstanding problem in solid state physics , and attracted renewed interest due to the recent recognition @xcite that the berry phase effect plays a very important role in the orbital magnetism .
the issue was carried out in the powerful semiclassical formalism @xcite , in which the bloch electron for @xmath1-th band is treated as a wave packet @xmath2 with its center ( @xmath3 ) in the phase space .
the orbital magnetic moment characterizes the rotation of the wave packet around its centroid and is given by @xmath4=@xmath5 , where @xmath6 is the charge of the electron and @xmath7 is the velocity operator . by writing the wave packet in terms of the bloch state ,
one obtains ( @xmath8 is abbreviated as @xmath9 ) @xmath10|\nabla_{\mathbf{k}}u_{n\mathbf{k}}\rangle , \label{e1}%\ ] ] where @xmath11 is the periodic part of the bloch state with band energy @xmath12 , and @xmath13 is the crystal hamiltonian acting on @xmath11 .
equation ( [ e1 ] ) can be alternatively derived by taking the differential of the electron energy , which within first order in the perturbative magnetic field @xmath14 turns to be @xmath15=@xmath16 , with respect to @xmath14 .
it was further found @xcite that the presence of a weak magnetic field @xmath14 will result in a modification of the density of states in the semiclassical phase space , @xmath17 , where @xmath18 is the berry curvature in @xmath0-space . due to this weak - field modification , a quantum - state summation @xmath19 of some physical quantity
@xmath20 should be converted to an integral according to @xmath21 .
additional thermodynamic average over bloch bands should be included at finite temperature .
therefore , the total free energy for an equilibrium ensemble of electrons in the weak field may be written as @xmath22 .
\label{e2}%\ ] ] where @xmath23 is the electron chemical potential and @xmath24 .
the equilibrium orbital magnetization ( om ) density is given by the field derivative at fixed temperature and chemical potential , @xmath25 , with the result@xmath26 \nonumber\\ & \equiv\mathbf{m}_{c}+\mathbf{m}_{\mathbf{\omega}},\nonumber\end{aligned}\ ] ] where @xmath27 is the local equilibrium fermi function for @xmath1-th band .
in addition to the conventional term @xmath28 in terms of the orbital magnetic moment @xmath29 , the extra term @xmath30 in eq .
( [ e3 ] ) is a berry phase effect and exposes a new topological ingredient to the orbital magnetism .
interestingly , it is this berry phase correction that eventually enters the thermal transport current @xcite . at zero temperature and magnetic field the general expression ( [ e3 ] ) is reduced to@xmath31 \right ) , \label{e4}%\ ] ] where the upper limit means that the integral is over states with energies below the zero - temperature chemical potential ( fermi energy ) @xmath32 .
the berry phase effect on orbital magnetism was until now partially presented by very few studies .
recent observation of the anomalous nernst effect ( ane ) in cucr@xmath33se@xmath34br@xmath35 compound @xcite was attributed @xcite to the manifestation of the berry phase effect in the om .
also the orbital magnetism was recently studied by use of two - dimensional ( 2d ) haldane model and ferromagnetic _ kagom _ lattice with spin chirality @xcite .
these two models are rare examples to show the zero - field quantum hall effect ( qhe ) @xcite . from ref .
@xcite one learns that the berry phase effect causes the om to display different behavior in metallic and insulating regions .
this difference may be explained in parallel with haldane s recent finding @xcite of the berry phase effect in the intrinsic hall conductivity [ including qhe and anomalous hall effect ( ahe ) ] .
the objective of the present paper is dual .
first we remark that the berry phase effect on the orbital magnetism has been included in the well - known kubo - streda formula @xcite .
therefore , a full quantum - mechanical linear response theory of the om can be developed as a useful complement of the semiclassical formalism , although the latter looks more elegant and practical for calculation on clean samples
. then we examine the 3d problem by studying the orbital magnetism in antiferromagnets on the distorted face - centered - cubic ( fcc ) lattice .
the results reveal that a general _ topological orbital magnetism theory _ that takes into account berry phase effect must now be developed .
the paper is organized as follows . in the next section
, we address that the intrinsic om given in the semiclassical formalism is consistent with the well - known quantum - mechanical kubo - streda formula in the clean - sample limit .
section iii describes the physical model that is used in this work .
the topological property and the consequent intrinsic hall effect associated with the model are also given in this section . in sec .
iv , we present a detailed study of the properties of the om and its effects on transport response in antiferromagnets on the distorted fcc lattice . finally , in sec . v we present our conclusions .
due to the above mentioned modification of the density of states , the particle number in the weak magnetic field ( say , along @xmath36-axis ) is given by @xmath37 it is easy to see that a link between eq .
( [ e3 ] ) and eq .
( [ e5 ] ) is @xmath38 , which is nothing but the usual thermodynamic maxwell relation and therefore should be free from the weak - field limit used by the semiclassical approach .
thus the zero - field om is given by@xmath39 on the other side , the integrand in eq .
( [ e6 ] ) can be written in terms of kubo - streda @xcite formula for electrons as follows@xmath40 where @xmath41 is the hall conductivity and @xmath42 . \label{a1}%\ ] ] here @xmath43 is the operator green function and @xmath44 is the velocity operator . in bloch - state representation
the trace in eq .
( [ e6 ] ) is equivalent to @xmath45 with the hamiltonian transformed to @xmath46 and the velocity to @xmath47=@xmath48 .
replacing @xmath49 in eq .
( [ a1 ] ) by @xmath50 and using the completeness relation of the bloch states , @xmath51=@xmath52 , one has @xmath53 by use of the identity @xmath54 and after a transformation of @xmath9-sum to an integral , eq .
( [ a2 ] ) is ready to be simplified as @xmath55 a comparison of eq .
( [ a4 ] ) with eq .
( [ e1 ] ) immediately gives the following relation@xmath56 where the second line is obtained by using the definition of @xmath28 in eq .
( [ e3 ] ) .
thus one finds that the semiclassical expression for @xmath28 is equivalent to the quantum - mechanical expression for @xmath57 .
note that although @xmath58 is a fermi - surface term , the quantity @xmath59 is a fermi - sea term and all the bloch states below @xmath23 should be accounted when calculating the om . on the other side , in the clean limit
, the kubo formula for the hall conductivity @xmath41 can be written in terms of the berry curvatures @xmath60 @xcite , @xmath61 from eq .
( [ a6 ] ) one has@xmath62 , \label{a7}%\ ] ] which is exactly the semiclassical expression for the berry phase term @xmath63 in eq .
( [ e3 ] ) .
thus the semiclassical om in eq .
( [ e3 ] ) can actually be written as a kubo - streda formula:@xmath64 the other two components @xmath65 and @xmath66 are given in a similar manner .
this equivalence between the semiclassical and quantum - mechanical description for the om is only valid in the intrinsic region and will break down when the impurity scattering effect is included .
thus while the semiclassical formula of the om is more suitably employed to study the intrinsic property of the om , the kubo - streda formula must be used when one takes into account the impurity scattering .
another aspect is that the kubo - streda formula is valid in arbitrary strength of the external magnetic field , while the semiclassical formula only works in the weak field limit . in some special cases , for example ,
when one wants to know the edge state effect on the om in a finite - size sample in a strong magnetic field @xmath67 @xcite , a more apparent approach can be employed by directly calculating the total free energy and then a finite - field om is obtained by a @xmath67-derivative of the free energy .
now we focus our attention to the properties of the om in a specific spin - frustrated system . as in ref .
@xcite , the model we used describes the chiral spin state in the ordered antiferromagnet ( af ) on the three - dimensional fcc lattice .
the af on the fcc lattice is a typical frustrated system , and nontrivial triple-@xmath68 spin structure with finite spin chirality has been revealed by band structure calculation @xcite and observed in experiments @xcite .
the anomalous behaviors in the fcc af were also observed .
for example , there occurs mysterious weak ferromagnetism in nis@xmath33 below the second af transition temperature @xcite .
the hall conductivity in this material is also large and strongly temperature dependent @xcite . in co(s@xmath35se@xmath69)@xmath70 ,
the ahe is enhanced in the intermediate @xmath71 region , where the nontrivial magnetism is realized @xcite .
spin structure on fcc lattice .
( b ) relation of the @xmath72-spin moments @xmath73 ( @xmath74 ) . ] the triple-@xmath68 spin structure on facc lattice is shown in fig . 1 . here
the lattice points are divided into four sublattices with different local spins @xmath73 ( @xmath74 ) on them .
the af nature requires @xmath75=@xmath76 .
the minimization of the 2-spin exchange interaction energy can not uniquely determine the sublattice spin orientation .
the inclusion of higher order ( 4-spin exchange ) interaction @xmath77 with positive @xmath78 gives the ground - state spin configuration @xcite as @xmath79=@xmath80 , @xmath81=@xmath82 , @xmath83=@xmath84 , and @xmath85=@xmath86 , where each direction corresponds to the four corners from the center of a tetrahedron [ fig .
the effective hamiltonian for the hopping electrons strongly coupled to the mean - field effective magnetic field caused by these local spins is given by @xmath87 with @xmath88 . here
the spin wave function @xmath89 is explicitly given by @xmath90 ^{\text{t}}$ ] , where the polar coordinates are pinned by the local spins , i.e. , @xmath91 .
@xmath92 is the angle between the two spins @xmath93 and @xmath94 .
the phase factor @xmath95 can be regarded as the gauge vector potential @xmath96 , and the corresponding gauge flux is related to scalar spin chirality @xmath97=@xmath98 @xcite . in periodic crystal lattices ,
the non - vanishing of the gauge flux relies on the multiband structure with each band being characterized by a chern number .
the chern number appears as a result of the spin - orbit interaction and/or spin chirality in ferromagnets . in ferromagnets
the time - reversal broken symmetry is manifest , while in af the time - reversal operation combined with the translation operation often constitutes the unbroken symmetry . in the latter case ,
the nonzero hall conductivity @xmath99 is forbidden . however , when there are more than two sublattices and the spin structure is noncollinear , this combined symmetry would be absent and finite @xmath99 is not forbidden @xcite .
the net spin chirality for the ideal fcc af lattice in fig . 1 is zero , because the spin chiralities are the vector quantities and the sum of these four vectors is zero .
however , when the lattice is distorted along the [ 1,1,1 ] direction , then the non - zero net spin chirality occurs .
following ref .
@xcite , we express the distortion along the [ 1,1,1 ] direction by putting the transfer integral within the ( 1,1,1 ) plane as @xmath100=1 , while that between the planes as @xmath101=1@xmath102 . as the unit cell is cubic shown in fig .
1 , the first brillouin zone ( bz ) is cubic : [ @xmath103@xmath104 .
from now on , we set @xmath105=@xmath52 .
then the hamiltonian matrix @xmath106 for each @xmath9 is given by@xmath107{cccc}% 0 & e^{-i\frac{\pi}{6}}f_{2 } & e^{i\frac{\pi}{6}}f_{1 } & f_{3}\\ e^{i\frac{\pi}{6}}f_{2 } & 0 & e^{-i\frac{\pi}{6}}f_{3 } & e^{i\frac{2\pi}{3}% } f_{1}\\ e^{-i\frac{\pi}{6}}f_{1 } & e^{i\frac{\pi}{6}}f_{3 } & 0 & e^{-i\frac{2\pi}{3}% } f_{2}\\ f_{3 } & e^{-i\frac{2\pi}{3}}f_{1 } & e^{i\frac{2\pi}{3}}f_{2 } & 0 \end{array } \right ) , \label{e10}%\ ] ] where @xmath108=@xmath109+@xmath110+@xmath111+@xmath110 , @xmath112=@xmath113+@xmath114+@xmath115+@xmath114 , @xmath116=@xmath117+@xmath118+@xmath119+@xmath118 .
in this hamiltonian , the two lower bands are fully degenerate , @xmath120=@xmath121=@xmath122 , while the two upper bands are also degenerate with @xmath123=@xmath124 .
the band structure along high - symmetry lines in the first bz is shown in fig .
= @xmath76 , ( b ) @xmath125=@xmath126 , and ( c ) @xmath125=@xmath127 . ] at @xmath125=@xmath76 [ fig .
2(a ) ] , the upper and lower dispersions touch along the edge of the bz and comprises an assembly of the massless dirac fermions ( weyl fermions ) in @xmath128d . for @xmath125@xmath129@xmath76 ( elongation along [ 1,1,1 ] direction ) , all the weyl fermions along the edge open a gap and turn into the massive dirac fermions [ fig .
therefore the gap fully opens in the density of states centered at zero energy . for @xmath125@xmath130@xmath76 ( suppression along [ 1,1,1 ]
direction ) , all the @xmath128d weyl fermions along the edges open the gap as in the case of @xmath125@xmath129@xmath76 .
however , there occurs two additional @xmath131d weyl fermions [ fig .
2(c ) ] at @xmath132 , where @xmath133 and @xmath134 correspond to the right- and left - handed chirality @xcite . thus unlike @xmath125@xmath129@xmath76
, the gap does not fully open in the case of @xmath125@xmath130@xmath76 due to the presence of a new single contact between the upper and lower bands in the bz .
the normalized eigenvectors are given by@xmath135{cccc}% -\frac{f_{1}\varepsilon_{1\mathbf{k}}+if_{2}f_{3}}{\varepsilon_{1\mathbf{k}% } \sqrt{f_{1}^{2}+f_{2}^{2 } } } , & e^{-i\frac{\pi}{3}}\frac{f_{2}\varepsilon _
{ 1\mathbf{k}}-if_{1}f_{3}}{\varepsilon_{1\mathbf{k}}\sqrt{f_{1}^{2}+f_{2}% ^{2 } } } , & 0 , & -\frac{\sqrt{f_{1}^{2}+f_{2}^{2}}}{\varepsilon_{1\mathbf{k}}}% \end{array } \right ) ^{\text{t}},\label{e11}\\ \begin{array } [ c]{cccc}% e^{i\frac{\pi}{6}}\frac{f_{2}}{\varepsilon_{2\mathbf{k } } } , & e^{-i\frac{\pi } { 6}}\frac{f_{1}}{\varepsilon_{2\mathbf{k } } } , & 1 , & ie^{i\frac{\pi}{6}}% \frac{f_{3}}{\varepsilon_{2\mathbf{k}}}% \end{array } \right ) ^{\text{t}},\nonumber\\ \begin{array } [ c]{cccc}% \frac{f_{1}\varepsilon_{3\mathbf{k}}+if_{2}f_{3}}{\varepsilon_{3\mathbf{k}% } \sqrt{f_{1}^{2}+f_{2}^{2 } } } , & -e^{-i\frac{\pi}{3}}\frac{f_{2}\varepsilon _ { 3\mathbf{k}}-if_{3}f_{1}}{\varepsilon_{3\mathbf{k}}\sqrt{f_{1}^{2}+f_{2}% ^{2 } } } , & 0 , & \frac{\sqrt{f_{1}^{2}+f_{2}^{2}}}{\varepsilon_{3\mathbf{k}}}% \end{array } \right ) ^{\text{t}},\nonumber\\ \begin{array } [ c]{cccc}% e^{i\frac{\pi}{6}}\frac{f_{2}}{\varepsilon_{4\mathbf{k } } } , & e^{-i\frac{\pi } { 6}}\frac{f_{1}}{\varepsilon_{4\mathbf{k } } } , & 1 , & ie^{i\frac{\pi}{6}}% \frac{f_{3}}{\varepsilon_{4\mathbf{k}}}% \end{array } \right ) ^{\text{t}}.\nonumber\end{aligned}\ ] ] the berry curvatures for these four bloch bands are derived to have the form @xmath136 with @xmath137 where ( @xmath138 ) represent a cyclic permutation of ( @xmath139 ) .
let us see the hall conductivity of this system @xcite with @xmath140 .
in the integer filling case , the zero - temperature hall conductivity is a sum of chern invariant @xcite over occupied bloch bands , @xmath141}\frac{dk_{z}}% { 2\pi}c_{n}(k_{z } ) , \label{e14}%\ ] ] where the 2d chern number @xcite @xmath142 is given by @xmath143^{2}}dk_{x}dk_{y}\omega _ { n}^{z}(\mathbf{k})\label{e15}\\ & = -\frac{1}{2\pi}\int_{[-\pi:\pi]^{2}}dk_{x}dk_{y}\text { } \hat{z}% \cdot\left ( \nabla_{\mathbf{k}}\times\mathbf{a}_{n}(\mathbf{k})\right ) .\nonumber\end{aligned}\ ] ] here @xmath144=@xmath145 is the berry phase connection ( vector potential ) for @xmath1-th band . to proceed one
may first transform the integral of @xmath146 over the first bz to the line integral of @xmath147 along the bz boundary by use of stokes theorem , and then apply the complex contour integration technique and residue theorem to sinusoidal functions .
after a straightforward derivation , one obtains the non - zero chern number , @xmath148sgn@xmath149 , @xmath150sgn@xmath149 , where@xmath151 and @xmath152 $ ] .
it is easy to verify that for @xmath125@xmath129@xmath76 , the value of @xmath153 is always positive , independent of @xmath154 . for @xmath125@xmath130@xmath76 ,
the sign of @xmath153 depends on @xmath154 in such a way that @xmath153@xmath130@xmath76 for @xmath155 and @xmath153@xmath129@xmath76 for @xmath156 $ ] .
note that the present choice of the other two bloch states @xmath157 and @xmath158 makes them to have no contribution to the chern number .
however , the above purely mathematical calculation of 2d chern number is not favored by theoretical physicists , who would like to resort to the physical connotation that the vector potential @xmath147 and gauge flux @xmath159 are endowed with . correspondingly
, here we present this gauge - field analysis of the lower band @xmath120 as an example .
the value of 2d chern number @xmath160 , which is confined to the ( @xmath161 ) subspace at fixed @xmath154 , is invariant under gauge transformation @xmath162=@xmath163 , @xmath164 , where @xmath165 is an arbitrary smooth function of @xmath9 .
if the gauge choice for @xmath166 is well - defined everywhere in the whole ( @xmath161 ) subspace in the first bz , then its chern number @xmath160 will obviously be zero .
however , at point @xmath167 , one can find that the wave function @xmath166 in eq .
( [ e11 ] ) is ill - defined since both its denominator and numerator are zero at this point .
this means that the used gauge can not apply to the whole bz and one needs to render a gauge transformation to avoid the singularity at @xmath168 . for this one transforms the wave function to @xmath169{cccc}% -\frac{f_{2}\varepsilon_{1\mathbf{k}}+if_{1}f_{3}}{\varepsilon_{1\mathbf{k}% } \sqrt{f_{2}^{2}+f_{3}^{2 } } } , & e^{-i\frac{\pi}{3}}\frac{f_{3}\varepsilon _ { 1\mathbf{k}}-if_{1}f_{2}}{\varepsilon_{1\mathbf{k}}\sqrt{f_{2}^{2}+f_{3}% ^{2 } } } , & 0 , & -\frac{\sqrt{f_{2}^{2}+f_{3}^{2}}}{\varepsilon_{1\mathbf{k}}}% \end{array } \right ) ^{\text{t}}. \label{e17}%\ ] ] the new eigenvector @xmath162 recovers the well - defined behavior at @xmath168 ; the new singularity brought about is at @xmath170 . thus according to the two different gauge choices , the bz cross section at fixed @xmath154 is now divided into two regions v and v@xmath171 as shown in fig .
3 ( @xmath154=@xmath76 ) = @xmath76 ) of the first bz into two regions v ( red area ) and v@xmath171 ( blue area ) . note that the projection of the singularities @xmath168 and @xmath172 onto @xmath173 subspace vary with @xmath154 . ] .
the wave functions @xmath166 are used onto the region v , while @xmath162 apply to v@xmath171 . note that there remains some freedom in the division of the bz . because @xmath174 and @xmath162 are ill - defined only at @xmath168 and @xmath172 , respectively
, we are free to deform this division as long as @xmath168(@xmath175)@xmath176v(v@xmath171 ) .
this corresponds to the gauge degree of freedom @xcite . at *
* k**@xmath177v@xmath178v@xmath171 , the two choices of wave functions are different by a phase factor @xmath179=@xmath180 , i.e. , @xmath181=@xmath182 , where @xmath183 thus one obtains the value of nonzero 2d chern number for lower band @xmath184 as follows @xmath185 { \small \cdot}d\mathbf{k}\label{e19}\\ & = -\frac{1}{2\pi}\oint_{\partial\text{v}}d\varphi_{1}(\mathbf{k}% ) = -\text{sgn}(g(k_{z})),\nonumber\end{aligned}\ ] ] which is consistent with the explicit calculation based on the complex - contour integration technique .
consider the @xmath23=@xmath76 case , i.e. , the two lower degenerate bands are fully filled while the two upper degenerate bands are empty .
then a further @xmath186-integral of @xmath160 gives the hall conductivity @xmath187 for @xmath125@xmath129@xmath76 and @xmath188 for @xmath125@xmath130@xmath76 .
the asymmetry of @xmath99 between @xmath125@xmath129@xmath76 and @xmath125@xmath130@xmath76 will be explained below together with the behavior of the om .
when the local spins @xmath189 are inverted ( which means that the spin chirality is also inverted ) , then the hall conductivity changes its sign .
now we turn to study the om and its various effects . without loss of generality ,
the present attention is only on the @xmath36-component of the om which is connected with hall conductivity @xmath99 .
first , after a straightforward derivation , one obtains the @xmath0-space orbital magnetic moment as follows @xmath190 see eq .
( [ e13 ] ) for @xmath191 .
one can see that the orbital magnetic moment is identical for upper and lower bands , while the berry curvatures [ eq .
( [ e12 ] ) ] for upper and lower bands differ by a sign . a comparison between eq .
( [ e20 ] ) and eq .
( [ e12 ] ) gives an interesting relation for the present model@xmath192 given the expressions for @xmath193 and @xmath194 , the om @xmath195 can now be systematically studied . at zero temperature , in particular , by substituting eq .
( [ e21 ] ) into eq .
( [ e4 ] ) one finds an important relation@xmath196 which indicates that the om is proportional to the hall conductivity with coefficient ( @xmath197 ) .
equation ( [ e22 ] ) also holds at low temperature .
although this remarkable relation between the om and the hall conductivity is specific to the present hamiltonian model , it definitely tells one that the topological ingredient in the om may be faithfully mapped out through the hall conductivity . if the band is partially filled , then after integrating by parts one finds that eq .
( [ e22 ] ) can be written as a pure fermi - surface integral . in the integer filling case , on the other hand , the om and hall conductivity display a fermi - sea feature .
figure 4 shows @xmath198 and @xmath99 as a function of the distortion @xmath125 . and ( b ) hall conductivity @xmath99 as a function of distortion @xmath125 .
the fermi energy in ( a ) is chosen to be @xmath32=@xmath199 , while in ( b ) the fermi energy is chosen to be @xmath32=@xmath200 which corresponds to case that only the lower bands are fully occupied . ]
the @xmath125@xmath129@xmath76 and @xmath125@xmath130@xmath76 cases are asymmetric by the observation that @xmath99 is quantized for @xmath125@xmath129@xmath76 while non - quantized for @xmath125@xmath130@xmath76 . to understand this asymmetry one may treat the distortion @xmath125 as a control parameter of the 3d band structure and start from @xmath125=@xmath76 ,
at which the upper and lower bands are degenerate along the bz edge , i.e. , ( @xmath201=@xmath202=@xmath203 ) , ( @xmath204=@xmath205=@xmath203=@xmath206 ) , ( @xmath204=@xmath207=@xmath206 ) .
when @xmath125 is varied from @xmath125=@xmath76 to @xmath125@xmath129@xmath76 , then these initial degenerate points _
completely _ split into two groups of dirac point singularities , which play the role of positive and negative monopole sources , respectively [ see , as an example , @xmath168 and @xmath172 points in fig . 2 ] .
the upper and lower bands are now tightly coupled by a series of berry flux loops . along each loop berry curvature flux
@xmath208 passes from the lower bands to the upper bands through one dirac point , then returns through the other corresponding one .
the positive ( negative ) monopoles of the lower bands and the negative ( positive ) monopoles of the upper bands may recombine by a relative displacement of a primitive reciprocal lattice vector @xmath209 . in the present cubic lattice the @xmath209 is along one selective @xmath210-axis ( @xmath211 ) with amplitude @xmath208 ( the lattice
constant @xmath105 has been scaled to be unity ) .
this means that during the dirac point splitting process , the individual chern invariant ( the 3d generalization of 2d chern number ) for the lower and upper bands changes by @xmath212 , respectively , while their sum is conserved . as a result ,
the hall conductivity @xmath99 for lower / upper bands is quantized to be a product of @xmath213 and the following chern invariant@xmath214^{3}}d^{3}\mathbf{k}\omega_{n}^{z}% ( \mathbf{k})=-\frac{1}{2\pi}c_{n}g_{z}. \label{e23}%\ ] ] when @xmath125 is varied from @xmath125=@xmath76 to @xmath125@xmath130@xmath76 , equation ( [ e23 ] ) breaks down , since the 2d chern number @xmath215 is now @xmath0-dependent by the presence of the additional two ( 3 + 1)d weyl fermions at @xmath132 . in this case , although the gap is opened at dirac points along the bz edge , the emergence of new weyl fermions contributes non - zero density of states in the gap .
it is then straightforward to repeat the integration in eq .
( [ e23 ] ) by parts to expose the non - quantized part of the 3d intrinsic hall conductivity as a fermi surface property . in the half filled case ,
i.e. , when the chemical potential is in the mott gap , the hall conductivity vanishes due to the cancellation of the chern invariants of the upper and lower bands .
as a consequence , the om in eq .
( [ e23 ] ) also vanishes .
this result is prominently different from that in ref .
@xcite , in which a finite om with amplitude smoothly varied with distortion @xmath125 was given via a tight - binding calculation .
@xcite employed the kubo - streda formula in calculating the om . as we have shown above , however , the kubo - streda formula and the semiclassical formula give the same expression for the intrinsic om .
so no discrepancy is expected to occur between the two approaches .
the non - zero om in ref .
@xcite in the mott gap was ascribed by the authors to be a bulk fermi sea property . however , the present explicit relation eq .
( [ e23 ] ) together with haldane s argument on the metallic ahe @xcite shows that the non - quantized part of the om in the present model is a fermi - surface berry phase effect , while the quantized part is completely determined by the topology of the filling bands . ,
and ( b ) its two components @xmath216 ( red curve ) and @xmath217 ( blue curve ) as a function of the electron chemical potential @xmath23 for distortion @xmath125=@xmath126 . the shaded area is the gap between the lower and upper bands . to suppress the divergence at band / gap contact
, we have used the temperature of @xmath218=@xmath219 . ] figure 5(a ) shows the @xmath198 as a function of the electron chemical potential @xmath23 for the distortion @xmath125=@xmath126 .
one can see that initially the om rapidly decreases as the filling of the lower bands increases , arriving at a minimum at @xmath23=@xmath220 , a value corresponding to the top of the lower band .
then , as the chemical potential continues to vary in the gap [ shaded region in fig .
5(a ) ] between the lower and upper bands , the om goes up and increases as a linear function of @xmath23 .
this linear relationship in the insulating region is explicit from eq .
( [ e22 ] ) . when the chemical potential touches the bottom of the upper band , then the linear increase in @xmath198 suddenly stops and the om rapidly decreases again by going the chemical potential through the upper bands .
the turning behavior at the band / gap contacts becomes numerically divergent at @xmath218=@xmath76 .
this discontinuity is due to the singular behavior of @xmath221 at the bz edge point @xmath9=@xmath168 , which will play its role when the @xmath0-integral is over the entire bz . the distinct behavior of the om in the metallic and insulating regions , as shown in fig .
5(a ) , reflects the different roles that its two components @xmath216 and @xmath217 play in these two regions . to see this ,
we show in fig .
5(b ) @xmath216 ( blue curve ) and @xmath217 ( red curve ) as a function of @xmath23 .
one can see that @xmath216 and @xmath217 oppose each other , which implies that they are carried by opposite - circulating currents . also one can see that in the band insulating regime , the conventional term @xmath216 keeps a constant during variation of @xmath23 .
this behavior is due to the fact that the upper limit of the @xmath0-integral of @xmath29 is invariant as the chemical potential varies in the band gap . in the metallic region , however , since the occupied states varies with the chemical potential @xmath23 , thus @xmath28 also varies with @xmath23 , resulting in a decreasing slope shown in fig .
the berry phase term @xmath222 also displays different features in the insulating and metallic regions . in the insulating region
, @xmath223 linearly increases with @xmath23 , as is expected from eq .
( [ e22 ] ) . in the metallic region , however , this term keeps a constant with the amplitude sensitively depending on the topological property of the band in which the chemical potential is located . on the whole it reveals in fig .
3 that the metallic behavior of the om is dominated by its conventional term @xmath28 , while in the insulating regime the berry phase term @xmath222 comes to play a main role in determining the behavior of the om .
the above separate discussion of @xmath28 and @xmath222 can be transferred to study the anomalous nernst effect ( ane ) .
the relation between the om and ane has been recently found @xcite . to discuss the transport measurement , it is important to discount the contribution from the magnetization current , a point which has attracted much discussion in the past .
cooper et al .
@xcite have argued that the magnetization current can not be measured by conventional transport experiments .
xiao et al .
@xcite have adopted this point and built up a remarkable picture that the conventional orbital magnetic moment @xmath28 does not contribute to the transport current , while the berry phase term in eq .
( [ e3 ] ) directly enters and therefore modifies the intrinsic hall current as follows@xmath224 in the case of uniform temperature and chemical potential , obviously , the second term is zero and the hall effect of the distorted fcc lattice is featured by the first term in eq .
( [ e24 ] ) . in the following ,
however , we turn to study another situation , where the driving force is not provided by the electric field .
instead , it is provided by a statistical force , i.e. , the gradient of temperature @xmath225 . in this case , eqs .
( [ e24 ] ) and ( [ e3 ] ) give the expression of intrinsic thermoelectric hall current as @xmath226 , where the anomalous nernst conductivity @xmath227 is given by@xmath228 .\nonumber\end{aligned}\ ] ] as a function of the electron chemical potential @xmath23 for @xmath125=@xmath126 and @xmath218=@xmath219 .
the shaded area is the gap between the lower and upper bands . ]
figure 6 shows @xmath227 of the distorted fcc lattice as a function of the chemical potential for @xmath229 and @xmath218=@xmath219 .
one can see that the ane disappears in the insulating regions , and when scanning @xmath23 through the contacts between the bands and gaps , there will appear peaks and valleys .
remarkably , a similar peak - valley structure was also found by the recent first - principles calculations in cucr@xmath33se@xmath230 compound @xcite .
the ane of this compound was recently measured by lee et al .
@xcite as a function of br doping which substitutes se in the compound and changes the chemical potential @xmath23 . due to the scarce data available , until now the peak - valley structure of @xmath227 revealed in fig .
6 and in ref .
@xcite has not been found in experiment , and more direct experimental results are needed for quantitative comparison with the theoretical results .
interestingly , the expression for @xmath227 can be simplified at low temperature as the mott relation @xcite , @xmath231 thus one can see that the low - temperature nonzero ane is a fermi - surface berry phase effect .
another unique feature of @xmath227 is its linear dependence of temperature .
in summary , after pointing out the equivalence of the semiclassical approach and the quantum kubo - streda formula in description of the intrinsic orbital magnetization , we have theoretically studied the properties of the om in antiferromagnets on the distorted 3d fcc lattice .
the distortion parameter @xmath125 in the fcc lattice produces nonzero 2d chern number and results in profound effects on the om properties . an explicit relation between the om and
the hall conductivity in this system has been derived . according to this relation
we have found that the om vanishes when the electron chemical potential is lies in the mott gap , which is in contrast with the results in ref .
we have shown that the two parts @xmath28 and @xmath222 in the om oppose each other , and yield the paramagnetic and diamagnetic responses , respectively .
in particular , due to its fermi - sea topological property , the magnetic susceptibility of @xmath222 remains to be a nonzero constant when the fermi energy is located in the energy gap .
it has been further shown that the om displays distinct behavior in the metallic and chern - insulating regions , because of different roles @xmath28 and @xmath222 play in these two regions .
the anomalous nernst conductivity has been studied , which displays a peak - valley structure as a function of the electron chemical potential .
we expect that these results will be experimentally verified in the spin - frustrated systems .
v. kalmeyer and r. b. laughlin , phys .
lett . * 59 * , 2095 ( 1987 ) ; g. baskaran and p. w. anderson , phys .
b * 37 * , 580 ( 1988 ) ; r. b. laughlin , science * 242 * , 525 ( 1988 ) ; x. g. wen , f. wilczek , and a. zee , phys . rev .
b * 39 * , 11413 ( 1989 ) . | we study the intrinsic orbital magnetization ( om ) in antiferromagnets on the distorted face - centered - cubic lattice .
the combined lattice distortion and spin frustration induce nontrivial @xmath0-space chern invariant , which turns to result in profound effects on the om properties .
we derive a specific relation between the om and the hall conductivity , according to which it is found that the intrinsic om vanishes when the electron chemical potential lies in the mott gap .
the distinct behavior of the intrinsic om in the metallic and insulating regions is shown .
the berry phase effects on the thermoelectric transport is also discussed . |
consider a hamiltonian system with two degrees of freedom .
the hamiltonian @xmath0 depends slowly on coordinate @xmath1 and fast on coordinate @xmath2 : @xmath3 where @xmath2 , @xmath1 are coordinates , and @xmath4 , @xmath5 are their conjugate momenta . the variation of the variables @xmath4 , @xmath2 , @xmath5 , @xmath6 is described by the differential equations @xmath7 the variables @xmath4 , @xmath2 are called fast , and @xmath5 , @xmath6 are called slow variables .
this system is called a slow - fast hamiltonian system .
the system with one degree of freedom in which @xmath8 , @xmath9 is called unperturbed or fast system .
assume that when @xmath10 , the phase portrait of the unperturbed system contains a domain filled by closed trajectories @xcite : here the frequency of motion is non - zero .
so we can introduce action - angle variables @xmath11 the action @xmath12 represents the area surrounded by each trajectory .
it is equal to this area divided by @xmath13 .
denote @xmath14 the hamiltonian @xmath0 expressed in terms of @xmath12 , @xmath5 , @xmath6 . the approximation in which @xmath15 and dynamics of @xmath5 , @xmath6 is described by hamiltonian system with hamiltonian @xmath14
is called an adiabatic approximation . in this approximation ,
@xmath16 @xmath17 we will assume that in the phase portrait of this system there is a domain in which along trajectories @xmath18 , @xmath19 as time @xmath20 .
thus trajectories have a form shown in one of figures below : we consider the case when the adiabatic invariant of the action @xmath12 along a trajectory of the original system possesses limiting values @xmath21 as @xmath20 .
their difference @xmath22 is called an accuracy of persistence of the adiabatic invariant in slow - fast systems .
if this system is analytic , the value of @xmath23 is exponentially small @xcite : @xmath24 the goal of this paper is to find an estimate of constant @xmath25 . after constructing some assumptions
, we will then formulate and prove the conclusion .
for the original hamiltonian @xmath26 with @xmath10 , @xmath2 is the coordinate and @xmath4 is conjugate momentum . in the unperturbed system ,
denote @xmath27 the period of the trajectory and @xmath28 the frequency .
let @xmath29 denote the generating function of canonical transformation @xmath30 .
we have the relations @xcite : @xmath31 for a fixed trajectory @xmath32 , where @xmath33 is a constant , we can express @xmath4 as @xmath34 if @xmath35 .
so the generating function has the following form @xcite : @xmath36 where the constant @xmath37 is the initial value of @xmath38 .
( see , e.g. @xcite ) @xmath39 @xmath40 where @xmath41 now consider another canonical transformation @xmath42 with generating function @xmath43 .
the old conjugate variables are @xmath44 and @xmath45 , and new pairs after transformation are @xmath46 and @xmath47 .
( see , e.g. @xcite ) @xmath48 and new hamiltonian is @xmath49 we will consider the case that there exist limiting values of action variables @xmath12 and @xmath50 as time tends to infinity . the conditions for this and the proof of existence of limits will be added later .
then we will see that the limiting values of @xmath12 and @xmath50 are equal .
so far we have got a new hamiltonian @xmath51 for simplicity , omitting the bar symbols and dependence on @xmath52 , we obtain a hamiltonian system in standard form : @xmath53 which has motion @xmath54
it is assumed , that in adiabatic approximation , @xmath18 , @xmath19 as @xmath20 ( see figure 2 ) .
we will assume also , that when @xmath6 , @xmath5 are changing in accordance with adiabatic approximation , and @xmath4 , @xmath2 are fixed : @xmath55 for definiteness we will consider the case when the trajectories in adiabatic approximation have the form shown in figure 2(1 ) .
result will be valid for the other two cases of figure 2 .
the function @xmath56 is the frequency of unperturbed motion .
let us assume that the following conditions are fulfilled .
* * the functions @xmath57 , @xmath58 can be analytically extended into a complex domain @xmath59 , where @xmath60 is a neighbourhood of a given real point @xmath61 , @xmath62 is a strip of a fixed width about the real axis , and @xmath63 is some domain in the complex plane @xmath64 . the function @xmath65 does not vanish in @xmath66 and @xmath67 .
function @xmath14 satisfies @xmath68 now let us consider approximate hamiltonian @xmath14 which has motion @xmath69 for a fixed @xmath70 , we have @xmath71 .
assume that curves @xmath72 are not closed : introduce the slow time @xmath73 the differential equations of motion are @xmath74 take @xmath75 , @xmath76 is a point in some interval @xmath77 with centre @xmath78 , where @xmath78 is a given real point : we can find the solution for describing the motion for approximate hamiltonian @xmath57 with @xmath79 : @xmath80 from now on we omit the dependence on @xmath76 . * * the solutions @xmath81 , @xmath82 can be analytically extended into a strip @xmath83 we suppose that for any initial point , the solution tends to infinity as time tends to infinity , i.e. @xmath84 @xmath14 satisfies @xmath85 @xmath86 satisfies @xmath87 @xmath65 satisfies @xmath88 here @xmath89 , @xmath90 , @xmath91 , @xmath92 are positive constants . for @xmath93 , @xmath94 , and + @xmath95 , @xmath96 where @xmath97 , @xmath98 , @xmath99 are positive constants .
@xmath100 + implies that @xmath101 @xmath102 since we know that @xmath103 , assume @xmath104 for @xmath105 , @xmath106 , and therefore , @xmath107 let @xmath108 .
the integration @xmath109 does not depend on the path . for @xmath110 , where @xmath111 , @xmath112 are positive constants satisfying @xmath113 -15pt we have @xmath114 @xmath115 as @xmath116
. we can find a positive constant @xmath117 , such that @xmath118 for @xmath119 , it is evident that we can find another positive constant @xmath120 such that @xmath121 so @xmath122 from cauchy estimate @xcite , we can easily get that , in @xmath123 and @xmath124 , @xmath125 @xmath126 now let us prove @xmath127
. from @xmath128 for simplicity , assume @xmath129 , and then @xmath130 in the following figure , mark the boundary of @xmath131 with solid lines , and that of @xmath132 with dashed lines : for @xmath133 , the estimate of @xmath134 on the circle of radius @xmath135 around @xmath136 is + @xmath137 .
if @xmath138 , then @xmath139 so , because of the cauchy estimate @xmath140 if @xmath141 , then because of the cauchy estimate , @xmath142 therefore , @xmath143 we can find a constant @xmath144 such that @xmath145 then @xmath146 therefore , for @xmath147 , @xmath148 , @xmath149 , @xmath150 * * the level lines @xmath151 lie in the domain @xmath131 and have a positive distance from the boundary of @xmath131 .
now consider the exact solution @xmath152 , @xmath153 , @xmath154 , @xmath155 of the hamiltonian system with hamiltonian @xmath53 with real initial conditions @xmath156 ,
@xmath157 , @xmath158 , @xmath159 in @xmath59 at + @xmath160 , and @xmath161 .
the adiabatic invariant of the action @xmath152 satisfies @xmath162 and thus @xmath163 there exist limiting values @xmath164 and their difference @xmath22 satisfies the estimate @xmath165 with the constant @xmath25 introduced in assumption @xmath166
. * remark . *
the method of continuous averaging @xcite gives the same estimate for @xmath25 , as the above - stated theorem .
paper @xcite gives us several examples of slow - fast hamiltonian systems , in one of which the hamiltonian is as follows : @xmath167 here @xmath5 , @xmath6 are slow variables and @xmath12 , @xmath168 are fast ones .
we explicitly treat the case of @xmath169 .
concerning @xmath170 , we need it to be real analytic and tend to zero as @xmath171 . for simplicity
we choose @xmath172 .
the unperturbed hamiltonian system is @xmath173 let @xmath174 be the total energy of the unperturbed system .
we can draw the phase portrait of the system to describe the motion : take @xmath175 as the slow time of motion . from the hamiltonian @xmath176 , we know that @xmath177 so @xmath178 thus we can express @xmath6 and @xmath5 via @xmath179 : @xmath180 with initial data @xmath181 and @xmath182 . after solving differential equations ,
we can obtain the solutions as ( see @xcite ) @xmath183 for simplicity we take the initial value of @xmath175 as @xmath184 . so @xmath185 , @xmath186 .
now after the canonical transformation of variables from @xmath187 to @xmath188 , the hamiltonian @xmath57 becomes @xmath189 and the new hamiltonian is @xmath190 with @xmath191 the differential equations for describing the motion are @xmath192 for @xmath193 , we suppose that @xmath193 is analytic in the strip @xmath194 with constant @xmath195 , and bounded : @xmath196 for @xmath197 , we can easily prove that @xmath198 and @xmath199 @xmath200 take @xmath60 as a neighbourhood of a given real point @xmath61 , and @xmath62 is a strip @xmath194 .
now let us determine the domain of @xmath175 by finding singularities of @xmath197 : @xmath201 points of singularities should satisfy @xmath202 @xmath203 thus @xmath204 the domain of @xmath205 denoted as @xmath206 is a complex ball around a real value @xmath207 .
any value @xmath208 can be taken to determine the singularities .
therefore , we obtain the singularities as @xmath209 from the above conclusion , if we take @xmath175 in the strip @xmath210 where @xmath90 is a positive constant , @xmath197 is analytic for any @xmath211 .
the following figure shows the boundary of domain @xmath212 : we can find the level lines @xmath213 lying in the domain @xmath212 , which have a positive distance from the boundary of this domain .
so @xmath25 has been found as any constant such that @xmath214 according to theorem 1 , the accuracy of conservation of adiabatic invariant @xmath23 has the estimate : @xmath215 this estimate coincides with that in @xcite , where it was obtained by a different method .
the original system is described by hamiltonian @xmath53 with motion in time @xmath216 : @xmath54 as the frequency @xmath217 does not vanish in our domain @xmath66 , we can express @xmath12 from @xmath218 via @xmath168 , @xmath5 , @xmath6 and a constant of energy .
this is the procedure of isoenergetic reduction on the energy level @xmath219 : @xmath220 here @xmath221 is the new hamiltonian , @xmath168 is the new time , and @xmath222 is in a small neighbourhood of @xmath223 , which can be written as @xmath224 , where @xmath225 and @xmath226 is some positive constant .
we do not indicate dependence of @xmath227 , @xmath228 on @xmath52 .
the differential equations of motion are @xmath229 since we know functions @xmath57 and @xmath58 are analytic in @xmath66 , from the implicit function theorem , functions @xmath230 and @xmath231 are also analytic in @xmath232 . now consider the approximate system with hamiltonian @xmath233 introduce slow time @xmath234 , and the motion is @xmath235 we can find the solution for the above approximate system : @xmath236 in what follows , we do not indicate the dependence on constant @xmath76 .
the relation between solutions @xmath237 , @xmath238 and @xmath239 , @xmath240 follows from the formula @xmath241 solution @xmath239 , @xmath240 can be analytically extended into some domain @xmath242 with respect to @xmath243 .
the domain @xmath242 will be introduced in later lemma . as we know
, @xmath88 it implies that @xmath244 uniformly tends to a real function depending on re@xmath245 as time tends to infinity and @xmath70 tends to real .
let us denote the limits as @xmath246 : @xmath247 for @xmath248 close to the imaginary axis , the image of the domain @xmath131 under the map @xmath249 given by the formula @xmath250 is a strip around the real axis that contains a domain @xmath251 where @xmath252 , @xmath253 , @xmath254 , @xmath255 are positive constants and @xmath256 for continuity . straight lines @xmath257 lie in @xmath242 and have a positive distance from the boundary of @xmath242 ( i.e. @xmath258 ) . the function @xmath259 satisfies the estimate : @xmath260 first of all let us list two forms of hamiltonian : @xmath261 therefore , we have @xmath262 so @xmath263 applying the intermediate value theorem here , we get that there exist some value @xmath264 , such that @xmath265 we know from assumptions that @xmath266 therefore , @xmath267 for @xmath94 , @xmath268 , @xmath269 @xmath270 where @xmath271 is a positive constant .
now let us draw the phase portrait of @xmath272 : and introduce new coordinates @xmath273 . here
@xmath274 is exactly the value of @xmath230 on level lines @xmath275 , and @xmath243 is the slow time of motion .
the initial section is @xmath276 , @xmath277 parametrized by @xmath12 .
then construct a transformation @xmath278 .
in the domain @xmath279 , where @xmath242 is what mentioned above and @xmath280 is a neighbourhood of the real point @xmath281 , the mapping @xmath278 is analytic and canonical . with map @xmath278 ,
consider @xmath6 and @xmath5 as functions of @xmath243 , @xmath274 : @xmath282 from the motion above , we have obtained the relations : @xmath283 let us differentiate both sides of the hamiltonian @xmath284 with respect to @xmath274 : @xmath285 replacing with corresponding derivatives , we get @xmath286 that is indeed the jacobi determinant : @xmath287 therefore , @xmath288 is canonical transformation . after this transformation ,
the new hamiltonian has the form @xmath289 function @xmath290 is analytic and has the following estimate : @xmath291 also from cauchy estimate @xcite , for @xmath292 , @xmath268 , @xmath148 , where @xmath293 and @xmath271 are positive constants , we obtain @xmath294 @xmath295 @xmath296 because of the above estimate , we can take another similar isoenergetic reduction through expressing @xmath274 , via @xmath12 , @xmath168 , @xmath243 and constant @xmath76 : @xmath297 here @xmath298 is the new hamiltonian , and @xmath243 is the new slow time .
@xmath12 and @xmath168 are conjugate variables . from implicit function theorem ,
we can obtain the similar conclusions that function @xmath299 is analytic in @xmath300 and estimates of @xmath299 and its derivatives have the following forms : @xmath301 in @xmath302 . here @xmath303 is the same as the domain @xmath304 . in this latest hamiltonian , @xmath305 and
the frequency is @xmath306 .
the level lines @xmath307 also lie in the domain @xmath242 and have a positive distance from the boundary of @xmath242 .
the constant @xmath25 is the same as introduced in assumption @xmath308 .
denoting the new time as @xmath309 , we can write differential equations of motion : @xmath310 now consider the exact solution @xmath311 , @xmath312 with real initial conditions @xmath156 , @xmath157 at @xmath313 .
we obtain the relation which the solution satisfies as follows : @xmath314 from the cauchy criterion @xcite , we can easily prove the existence of limiting values of @xmath311 as @xmath315 .
moreover , we also know differential equations of motion : @xmath316 it is evident that @xmath317 as @xmath315 . meanwhile , @xmath20 .
therefore , the limits when @xmath20 are equal to those when @xmath315 .
let us define values @xmath318
now the system under consideration is described by the hamiltonian @xmath319 where @xmath12 and @xmath168 are conjugate canonical variables .
the action - angle variables @xmath12 , @xmath168 of the unperturbed problem are defined in region @xmath320 .
@xmath243 is the slow time variable defined in region @xmath242 and time @xmath309 .
the function @xmath299 is an analytic , @xmath13periodic in @xmath168 function .
the frequency of unperturbed motion is 1 .
we have proved the following properties : @xmath321 .
the functions @xmath322 , @xmath299 can be analytically extended into a complex domain @xmath300 , where @xmath303 is some neighbourhood of a given real point @xmath61 , @xmath62 is a strip of some fixed width about the real axis , and @xmath242 is a strip @xmath251 where @xmath323 .
the function @xmath299 satisfies the estimate @xmath324 here @xmath89 , @xmath90 , @xmath325 , @xmath92 are positive constant .
the level lines @xmath307 lie in the domain @xmath242 and have a positive distance from the boundary of @xmath242 .
consider a solution @xmath311 , @xmath312 of the hamiltonian system with real initial conditions @xmath156 , @xmath157 at @xmath327 .
for @xmath328 referring to the paper @xcite , we have the estimate of @xmath23 that is exponentially small : @xmath215 actually , all variables @xmath12 , @xmath168 , @xmath5 and @xmath6 mentioned before should contain the bar symbol @xmath329 " .
we obtain in fact the conclusion that @xmath330 now return the bar .
recall that from the original hamiltonian @xmath26 , we transform to two forms of hamiltonian . with generating function @xmath29 where @xmath5 , @xmath6 are considered as parameters , we obtain @xmath14 and with @xmath331 we obtain @xmath332 .
since the limiting values of @xmath50 exist , now let us prove the lemma that the limiting value of @xmath12 is equal to that of @xmath50 and thus exists .
@xmath333 with generating function @xmath334 , the hamiltonian @xmath26 is canonically transformed to @xmath14 and @xmath335 .
as we know @xmath39 so @xmath336 on the other hand , with generating function @xmath331 , the hamiltonian @xmath26 is transformed to @xmath332 and @xmath337 . also we know @xmath338 where @xmath339 and @xmath340 are some intermediate values . as @xmath341 , @xmath342 , thus @xmath343 , as @xmath344 . from assumption @xmath345 , and @xmath346 , as @xmath347 , we know that @xmath348 for big @xmath349 . thus , when @xmath350 , @xmath351 , where @xmath352 , @xmath353 are positive constants . from @xmath354 , @xmath355 , the phase portrait of @xmath356 in @xmath357 plane will close to that of the adiabatic approximation @xmath358 for @xmath359 .
thus for some time moments @xmath360 and @xmath361 , we have @xmath362 then for @xmath363 , we have @xmath364 therefore , @xmath365 as @xmath366 . similarly , @xmath367 as @xmath368 . from @xmath369 taking limiting values on both sides as @xmath20 , we obtain @xmath370 .
we also know that @xmath371 as @xmath370 , it is evident that @xmath372 , @xmath373 as @xmath20 . therefore , @xmath333 thus @xmath12 and @xmath50 have the same limit , and @xmath374 . therefore
, we have the following corollary : the estimate @xmath375 is valid .
i am glad to express my gratitude to professor anatoly neishtadt for providing detailed comments and frequent discussion during the research and completion of this article .
arnold , v.i . ,
kozlov , v.v . and
neishtadt , a.i . :
_ mathematical aspects of classical and celestial mechanics_. second augmented and revised edition .
springer - verlag , 2006 .
neishtadt , a.i .
: _ on the accuracy of persistence of adiabatic invariant in single - frequency systems_. regular and chaotic dynamics , vol . 5 , no 2 , 2000 , pp . 213 - 218 .
arnold , v.i . :
_ mathematical methods of classical mechanics_. second edition .
springer , 2000 .
neishtadt , a.i .
: _ on the change in the adiabatic invariant on crossing a separatrix in system with two degrees of freedom_. pmm u.s.s.r , vol .
51 , no . 5 , 1987 , pp .
586 - 592 .
brown , j.w . and churchill , r.v . : _ complex variables and applications_. sixth edition .
mcgraw - hill , 1996 .
arnold , v.i . : _ ordinary differential equations_. the mit press , cambridge , massachusetts and london , 1973 .
benettin , g. , carati , a. and gallavotti , g. : _ a rigorous implementation of the jeans - landau - teller approximation for adiabatic invariants_. nonlinearity 10 , 1997 , pp .
479 - 505 .
treschev , d.v . : _ the method of continuous averaging in the problem of fast and slow motions_. regular and chaotic dynamics , vol .
3/4 , 1997 , pp . | let the adiabatic invariant of action variable in slow - fast hamiltonian system with two degrees of freedom have two limiting values along the trajectories as time tends to infinity .
the difference of two limits is exponentially small in analytic systems .
an iso - energetic reduction and canonical transformations are applied to transform the slow - fast systems to form of systems depending on slowly varying parameters in a complexified phase space . on the basis of this method
an estimate for the accuracy of conservation of adiabatic invariant is given for such systems . |
suppose we have an interface between two chiral , one - component theories with k - matrices @xmath166 and @xmath167 respectively , which are both positive integers .
the k - matrix at the interface will be treated as @xmath168.$ ] the effective edge theory reads @xmath169 + \ldots \,,\ ] ] @xmath170 = \cos{\left ( \lambda^{t } k \phi \right ) } \,,\ ] ] where @xmath18 represent the excitations originating from the edges states of phases 1 and 2 , with equal - time commutation relations @xmath171 = - 2\,\pi\,\mathrm{i}\,k^{-1}_{i j}\,\delta(x - y ) $ ] , and @xmath36 $ ] is a gap opening interaction . the null vector criterion , for a null vector @xmath172 with @xmath173 is @xmath174 let us now consider this equation in its generality .
to , simplify let @xmath175 such that @xmath176 and @xmath177 we will provide a simple physical interpretation for @xmath49 later , but for now it serves to simplify the gapability constraint .
we can reformulate eq .
[ eq : gapping ] as@xmath178 we immediately see that @xmath49 plays no role in the determination of this condition , and by construction @xmath179 and @xmath180 are relatively prime .
we can now multiply through to find@xmath181 and this implies , using the relatively prime condition , that @xmath179 divides @xmath182 and that @xmath180 divides @xmath183 as such we have the restrictions @xmath184 where @xmath185 and @xmath186 are integers . plugging this into eq .
[ eq : gapping2 ] we arrive at @xmath187 to summarize , we now have the relations @xmath188 now , in order for @xmath35 to be a primitive vector , @xmath189 and @xmath190 can not have common factors , hence @xmath191 if we enforce primitivity .
hence , we arrive at the result that the most generic gappable interface for two , one - component theories is @xmath192.$ ] furthermore , a primitive gapping vector that will gap this theory takes the form @xmath193 with this result we can gain a physical understanding .
if we take the theory @xmath166 and gauge a @xmath194 subgroup of the global particle number conservation symmetry and take @xmath167 and gauge a @xmath195 subgroup of its particle number conservation symmetry then we will arrive at two theories with the same topological order , and which can hence be gapped .
in fact , the gapability equation written as @xmath196 essentially encodes that if we take the theory @xmath167 , confine a @xmath197 group and deconfine a @xmath198 group then we will arrive at the same topological order as @xmath199 from our constraints we see this is exactly true since @xmath200 divides @xmath201 the common factor @xmath11 is interpreted as a common parent theory from which each topological phase @xmath202 can be reached by deconfining discrete @xmath154 and @xmath155 subgroups of @xmath203 respectively .
let us now look at the condensate at the interface .
the gapping term can be written as @xmath204 } \equiv -j\ , \int^{x_{r}}_{x_{l } } dx \cos { \left ( m n \theta \right ) } \\ & \ , \propto -j\ , \int\,dx\ , \left ( \psi_{1 } \right)^{n}\,\left ( \psi^{\dagger}_{2 } \right)^{m } + \textrm{h.c . } \ , .
\end{split}\ ] ] it is important not to factor out @xmath49 when considering the condensate . to see this
let us consider an example for the case when @xmath205 and @xmath206 where @xmath207 , @xmath208 and @xmath209 in phase @xmath133 the statistics are @xmath210 where @xmath211 the only fermion or boson in the quasiparticle set is when @xmath212 is an even integer , i.e. , the local boson in this case , which we will call @xmath213 for phase @xmath214 the statistics are @xmath215 and the quasiparticle with @xmath216 is a ( non - local ) boson which we call @xmath217 . for this system we can form @xmath218 to get a local condensate .
the interaction that will generate this condensate is @xmath219 from the form of the condensate we see that the @xmath64 group structure is broken to @xmath220 and hence we expect there to be @xmath221 parafermions . in the generic case
the symmetry is broken to @xmath222 and hence a @xmath156 parafermion results .
to see this explicitly , note that the second line of makes explicit that the interaction pairs n operators of phase 1 with m operators of phase 2 .
the non - local operator @xmath223 , where @xmath224 shall be identified with a @xmath225 clock variable . from the form of this interaction
, we can immediately write down three symmetry operations and their representations in terms of the bosonic fields : @xmath226 which is a u(1 ) symmetry of the interaction hamiltonian when the topological phases have charge vector @xmath227
. moreover , we have @xmath228 which accounts for a discrete @xmath106 associated with phase 1 and , equivalently , @xmath229 which accounts for a discrete @xmath230 associated with phase 2 .
importantly , while @xmath85 does not transform under the action of @xmath231 , it does so under the other two operators as follows : @xmath232 therefore in the clock representation , the ground state degeneracy of the hamiltonian becomes associated with the symmetry broken " states @xmath233 where @xmath234 .
while it would seem that a particular eigenstate could be identified by the local measurement of @xmath85 , @xmath235 , it is fundamental to recognize that @xmath85 is _ not _ a local operator in terms of the original local degrees of freedom @xmath236 and @xmath59 .
we show the existence of an operator that commutes with the interface hamiltonian @xmath237 and yields an @xmath15-fold degeneracy of the ground state of . for that we seek a unitary transformation @xmath238 } \,\ ] ] parametrized by @xmath239 that commutes with the hamiltonian . with the equal time
commutation relations @xmath240
= -\frac{2\,\pi\,\mathrm{i}}{p\,n^{2}}\,\delta(x - y)$ ] , @xmath241 = \frac{2\,\pi\,\mathrm{i}}{p}\,\delta(x - y)$ ] and @xmath242 = [ \partial_{x}\,\phi_{2}(x ) , \phi_{1}(y ) ] = 0 $ ] , we find that @xmath243 and @xmath244 , where @xmath90 . with that the local operators @xmath245 and @xmath246 transform as : @xmath247 .
it the follows @xmath248 = 0 \quad \longleftrightarrow \quad a_{1 }
+ n a_{2 } = t \in \mathbb{z } \,.\ ] ] following , we can parametrize the operator @xmath116 as @xmath249 @xmath231 can be interpreted as a u(1 ) symmetry generator when the edge modes carry electromagnetic charge with the charge vector @xmath250 : [ eq : transformations of psi under salpha and sbeta ] @xmath251 the operator @xmath67 , on the other hand , reflects the fact that the edge hamiltonian gives rise to an n - particle condensate of phase 2 ; associated to this interaction is the invariance @xmath252 moreover we find from that @xmath253 of particular interest are the operators @xmath254 and @xmath255 where @xmath256 and @xmath257 are parafermion operators@xcite defined at the ends of the interface . the topological property of the gapped interface is manifested by the presence of these zero modes whereby the @xmath15-fold degeneracy is encoded in the @xmath15 eigenvalues of the operator @xmath258 , which commutes with @xmath259 .
furthermore , after mapping parafermion operators into clock operators @xmath260 , the ground state @xmath15-fold degeneracy becomes associated with the symmetry breaking of the clock model , where the generator of the global @xmath106 symmetry is @xmath261 . in the following we discuss this mapping between clock and parafermion operators in further detail .
we now discuss a lattice regularization of the interaction hamiltonian responsible for gapping the interface between phases 1 and 2 .
such regularization serves a useful purpose in allowing a microscopic representation of the parafermion end states , in connection with recent microscopic models of topological parafermion chains @xcite .
we address first the lattice regularization in terms of clock variables . as we have seen , the hamiltonian commutes with the operator @xmath262 defined in eq .
[ eq : definition of s - supplement ] .
this operator can be identified with the generator of @xmath106 transformations of the clock operator @xmath85 via @xmath263 . in this language ,
the ground state degeneracy of the hamiltonian becomes associated with the symmetry broken " states @xmath264 where @xmath82 .
while it would seem that a particular eigenstate could be identified by the local measurement of @xmath85 , @xmath235 , it is fundamental to recognize that @xmath85 is _ not _ a local operator in terms of the original local degrees of freedom @xmath236 and @xmath59 . despite its non - local character ,
it is still useful to identify @xmath85 as a @xmath106 clock operator , with @xmath265 , while acting on the ground state manifold eq .
[ eq : eigenstates of sigma - supplement ] .
note that , since @xmath266 is a bosonic operator , i.e. , @xmath267 $ ] = 0 for any coordinates @xmath268 and @xmath269 , so is @xmath80 : @xmath270 , for @xmath271 . with that we are led to re -
expresses the interaction using the relation @xmath272 eq . motivates introducing a lattice regularization by defining clock operators @xmath273 at every site @xmath274 of the open chain and the canonically conjugated operators @xmath275
satisfying @xmath276 . then a lattice regularization of the interaction reads @xmath277 , whose @xmath15 degenerate ground states satisfy @xmath278 , for @xmath279 , in direct correspondence with . in the lattice representation ,
introduce of operators ( @xmath280 ) at every site of the open chain defined by @xcite @xmath281 these new operators satisfy parafermion statistics @xcite @xmath282 the field theory equivalent of the parafermion operators reads @xmath283 satisfying the commutation relations @xmath284 in the parafermion representation the lattice hamiltonian acquires the form @xmath285 notably , @xmath286 and @xmath287 do not appear in the hamiltonian ; for @xmath288 these dangling " parafermions reduce to the majorana end states in the topological kitaev chain .
@xcite manifestly , @xmath289 = 0 $ ] , where @xmath290 is the non - local operator connecting the parafermion end state through the gapped bulk .
[ eq : 2d topological phase coupled to zn gauge theory - supplement ] @xmath291 where @xmath292 @xmath293 and @xmath294 describe , respectively a @xmath295 fermionic and a @xmath296 bosonic laughlin state coupled to a deconfined @xmath106 gauge theory .
the original charge and flux excitations of the @xmath106 gauge theory are described , respectively , by @xmath112 and @xmath297 for the choice @xmath298 , the gl(3 , @xmath299 ) transformation @xmath300 = -1 \,,\]] implements @xmath301 for the choice @xmath302 , the gl(3 , @xmath299 ) transformation @xmath303 = 1 \,,\]] implements @xmath304 the edge states associated with the bulk theory are described by the @xmath305d theory @xmath306 where @xmath307 sets the relation between the edge fields @xmath308 and @xmath309 of the two representations . as discussed in the main text , in the representation given by @xmath310 , the interaction for the case @xmath311 @xmath312 gaps the extra " trivial modes , which then yields the edge state of phase @xmath133 . alternatively , the interaction @xmath313 gives rise to the edge state of phase 2 .
we can now establish a relation between null vectors in the two representations by noticing that , if @xmath314 is a null vector of @xmath310 , such that , @xmath315 , then @xmath316 is the corresponding null vector of @xmath317 , since @xmath318 . by making use of this relationship
, we find @xmath319 and @xmath320 it is clear then from eqs . and that the vacuum associated with @xmath321 confines the excitations of the @xmath155 gauge theory yielding the original phase 2 . moreover , in the basis given by @xmath317 , the deconfined excitations in the vacuum where @xmath322 are described by vertex operators @xmath323 , with @xmath324 , such that @xmath325
. then computing the braiding statistics of two excitations @xmath326 and @xmath327 we find @xmath328 , which accounts for the chiral phase with @xmath329 .
this result is consistent with the topological state having @xmath330 classes of quasiparticles @xmath331 , where @xmath332 represents the deconfined magnetic flux originating from the @xmath155 gauge theory and @xmath333 represents the fermion ( boson ) of the @xmath334 theory [ for @xmath49 even ( odd ) ] , which becomes a non - local quasiparticle upon coupling to the deconfined gauge theory .
setting @xmath335 for simplicity , suppose a gapped interface between phases 1 and 2 is formed , which physically implies the condensation of the bosonic field @xmath336 , so that @xmath337 . ( equivalently , @xmath338 . ) hence , the gapped spectrum originates from the condensation of @xmath162 , or , equivalently , of @xmath339 .
this anyon condensate implies that the gapped interface behaves as a anyon andreev reflector " in the following sense .
let an excitation @xmath59 approach the gapped edge ( for @xmath335 , this is the only quasiparticle of the phase 2 ) .
moreover , suppose that the pair @xmath340 is produced from vacuum fluctuations on the topological system 1 near the interface .
then due to the nature of the condensation on the interface , the pair @xmath339 can be condensed and absorbed by the interface , while leaving behind the deconfined excitation @xmath341 on the other side of the interface . reversing the logic
, we can ask : what happens if a quasiparticle @xmath342 ( @xmath343 ) belonging to phase 1 approaches the interface where it interacts with the condensate ?
since the only allowed condensation process involves the quasiparticle @xmath341 , the following non - trivial process that can occur : close to the interface , ( 1 ) the incoming quasiparticle decays into @xmath344 and , ( 2 ) the pair @xmath345 is produced from vacuum fluctuations on the topological system 2 near the interface such that , when the condensation @xmath346 occurs , it leaves behind the deconfined excitations @xmath347 and @xmath59 on different sides of the interface . in particular , the case @xmath348 represents an incoming @xmath341 quasiparticle , which can be completely absorbed by the interface while the @xmath59 excitation appears on the other side .
define the operators : @xmath349 } \ , , \end{split}\]] where @xmath350 is a real coefficient and @xmath351 is a finite interval on the line .
it follows that @xmath352 \ , . \end{split}\]]
it can be easily seen from the commutation relations of the edge fields that , if two intervals @xmath353 and @xmath354 are non overlapping or one of intervals is entirely contained within the other , than @xmath355 , which implies that @xmath356 = 0 $ ] . if , however , the intervals @xmath353 and @xmath354 are partially overlapping such that @xmath357 , then it follows that @xmath358 . | we explore a scenario where local interactions form one - dimensional gapped interfaces between a pair of distinct chiral two - dimensional topological states referred to as phases 1 and 2 such that each gapped region terminates at a domain wall separating the chiral gapless edge states of these phases .
we show that this type of t - junction supports point - like fractionalized excitations obeying parafermion statistics , thus implying that the one - dimensional gapped interface forms an effective topological parafermionic wire possessing a non - trivial ground state degeneracy .
the physical properties of the anyon condensate that gives rise to the gapped interface are investigated .
remarkably , this condensate causes the gapped interface to behave as a type of anyon andreev reflector " in the bulk , whereby anyons from one phase , upon hitting the interface , can be transformed into a combination of reflected anyons and outgoing anyons from the other phase .
thus , we conclude that while different topological orders can be connected via gapped interfaces , the interfaces are themselves topological .
topological phases of matter in 2 + 1-d are often characterized by a bulk - boundary " correspondence .
bulk properties such as a topological band structure , quasiparticles exhibiting fractional statistics , or topological ground state degeneracy on manifolds with non - zero genus , go hand in hand with an associated set of boundary / interface states where a topological phase meets a different topological phase such as the vacuum .
@xcite topological phases appear in two general classes : symmetry protected @xcite , or those that have intrinsic " topological order @xcite .
there are several important distinctions between these classes , e.g. , differing constraints on the ability to open a gap in the edge state spectrum . for the first class
, gapped boundaries can exist when the symmetry is broken explicitly or spontaneously . in the latter
, interface states with non - vanishing chirality can not be completely gapped , and , surprisingly , even in the absence of any symmetries , some interface states with _ vanishing _ chirality can not be completely gapped either.@xcite this observation may directly impact experiment since such an ungappable edge may exist in the @xmath0 fractional quantum hall effect , or at the interface between two fractional quantum hall states with , e.g. , filling factors @xmath1 and @xmath2 the latter interface can not be gapped by any local interaction , essentially due to the completely incompatible bulk properties of the two topological phases .
in this article we focus on the complementary effect that allows disparate topological phases to support _ gapped _ interfaces , as they provide a domain for a wide - range of interesting physics .
the existence of such an interface requires that a local gapping condition be satisfied [ see discussion around eq . ] , which , physically amounts to the allowed formation of an anyon condensate " at the interface .
it has been established , for two dimensional abelian topological phases , that each anyon condensate is in one - to - one correspondence with a mathematical structure called a lagrangian subgroup . "
@xcite each lagrangian subgroup is a subset @xmath3 of the set of anyons wherein ( i ) all quasiparticles have mutual bosonic statistics , and ( ii ) every quasiparticle not in @xmath3 has non - trivial statistics with at least one quasiparticle of @xmath3 .
hence , the simultaneous condensation of the quasiparticles in @xmath3 is allowed by ( i ) , and will confine _ all _ the anyons of the theory by ( ii ) . of great interest
are configurations where inequivalent anyon condensates , corresponding to inequivalent choices of @xmath4 are formed in adjoining regions of a topological interface .
indeed , domain walls between these gapped regions have been shown to host non - abelian defect bound states with parafermionic statistics.@xcite such bound states could be used as a platform for realizing topological quantum computation .
@xcite in this letter we characterize a new family of 1d gapped systems that can be formed at the interface between different 2 + 1d abelian topological phases . for our examples ,
we choose single - component chiral phases characterized by the topological invariants ( k - matrices ) @xmath5 and @xmath6 respectively .
hereafter we refer to these as phase 1 and phase 2 .
if these phases arise from quantum hall states then we have @xmath7 , where @xmath8 is the filling fraction that measures the hall conductance in fundamental units .
more generally , for systems without u(1 ) charge conservation symmetry , @xmath9 count the number of distinct bulk quasiparticle types in each phase , and give the topological ground state degeneracy @xmath10 of each system defined on a spatial manifold of genus @xmath11 . for our discussion we will adopt the interface geometry in fig .
[ topological_interface ] .
the bulk topological phases share a gapped interface with each other , and they have a boundary with the vacuum that contains propagating chiral edge modes .
the gapped interface terminates at points separating the gapless edge states of phases 1 and 2 .
our main finding is that such an interface forms a topologically non - trivial , 1d gapped system with a degenerate ground state manifold associated with parafermionic end states .
we stress that the parafermions discussed here are not located at the domain walls between distinct gapped interfaces ; rather , the they are located at domain walls between the _ gapless _ edge states of phases 1 and 2 , as shown in fig .
[ topological_interface ] . therefore the physical scenario discussed here departs significantly from those of refs .
@xcite , and more closely matches the setup of ref .
, though here we are focused more on what is happening in the bulk , rather than the edge as in their discussion . ultimately , our results identify that , while one can find gapped interpolations between 2d phases with different topological order , these are not trivial gapped regions ; they are instead _ topological _ themselves . in the following , we will illustrate our result with a bosonization description built from the pair of counter propagating modes stemming from the edge states of phases 1 and 2 .
we will construct the explicit form of the local , gap - opening interaction , provide a description of the interface anyon condensate , and discuss the onset of the topologically degenerate ground state manifold associated with the expectation value of a non - local operator .
finally , we will discuss the connection between bulk confinement - deconfinement transitions , edge - state transitions , and the bound parafermion modes .
_ * 1 .
luttinger liquid description of the interface * _ in fig .
[ topological_interface](a ) , we consider an array of 2d topological states in phase 1 ( blue ) and phase 2 ( brown ) , surrounded by the vacuum .
as shown in the supplementary material ( sm ) , the most generic gappable interface for one - component states is characterized by @xmath12 and @xmath13 where @xmath14 and @xmath15 are positive integers . at each interface along the x - direction ,
the edge states originating from phases 1 ( right - moving ) and 2 ( left - moving ) are propagating in opposite directions .
the low energy description of a given interface is given by the lagrangian density @xmath16 \,,\ ] ] @xmath17 where @xmath18 represent the bosonic modes originating from the edge states of phases 1 and 2 , v is a velocity matrix and @xmath19 $ ] is a local interaction between the edge modes to be discussed below .
the symmetric ( in this case diagonal ) k - matrix gives the structure of the equal - time commutation relations @xmath20 = - 2\,\pi\,\mathrm{i}\,k^{-1}_{i j}\,\delta(x - y ) \,.\ ] ] to simplify our discussion we will choose @xmath21 and provide the details for @xmath22 in the sm .
this case is also the most experimentally relevant as it includes interfaces between a @xmath23 integer quantum hall state with , e.g. , a @xmath24 fractional quantum hall state when @xmath25 @xcite adopting a compactification radius convention with @xmath26 for the fields @xmath27 , the quasi - particle excitations on the edge are created by the vertex operators @xmath28 , where @xmath29 is an integer - valued 2-component vector . in this representation ,
the exchange statistics associated with taking a quasiparticle @xmath30 adiabatically around another quasiparticle @xmath31 is given by the statistical phase @xmath32 and the ( topological ) spin of each quasiparticle is given by the self - statistics phase @xmath33 .
local excitations of the theory are identified with the vertex operator @xmath34 , where @xmath35 is an integer vector . in eq .
$ ] represents an interaction between the counter - propagating modes that will open a gap at the interface .
this interaction , which involves coupling between _ local _ degrees of freedom , can be parametrized as @xmath37 = -j\,\cos{\left ( \lambda^{t}\,k\,\phi \right ) } $ ] , where @xmath38 is a coupling constant , and @xmath39 is an integer - valued 2-component vector satisfying the null vector criterion @xcite @xmath40 the integer vector @xmath41 is a _ primitive _ solution@xcite of that represents a process where a single local operator @xmath42 of phase 1 couples with @xmath15 local operators @xmath43 of phase 2 , giving rise to the interaction @xmath44 where @xmath45 .
this interaction generates an anyon condensate at the interface as we will now describe .
in phase 1 , there are @xmath46 quasiparticle - types labeled @xmath47 , @xmath48 , and in phase 2 , there are @xmath49 quasiparticle - types labeled @xmath50 , @xmath51 .
the set of anyons forms a discrete lattice@xcite , and each anyon is topologically equivalent to anyons that differ only by the attachment of integer numbers of the local quasiparticles @xmath52 , @xmath53 . in the context of a laughlin fermionic ( bosonic ) state
, @xmath54 represents the local fermion ( boson ) of the @xmath55-th phase .
now we note that in phase 1 the anyon set @xmath56 contains mutual bosons or fermions with spin @xmath57 .
moreover we note that the quasiparticle @xmath58 has the same spin as the local excitation @xmath59 of phase 2 , i.e. , they are both bosons or fermions depending on the parity of @xmath49 .
physically this implies that the composite quasiparticle @xmath60 is a boson that can condense , and generate a fully gapped interface between phases 1 and 2 .
this condensation process , mathematically , is a consequence of the relation @xmath61 , which allows for the existence of a @xmath62-dimensional lagrangian subgroup @xmath3 containing @xmath63 importantly , the interaction involves @xmath15 local particle operators of phase 2 and a single local particle operator of phase 1 .
this interaction breaks the @xmath64 particle conservation symmetries of each phase down to @xmath65 , where @xmath66 means no symmetry .
hence the interaction is invariant under @xmath67 : @xmath68 , @xmath69 , @xmath70 if the phases began with a @xmath71 electromagnetic charge conservation symmetry , then this interaction breaks ( preserves ) the symmetry when the charge vector is @xmath72 ( @xmath73 ) . this discrete symmetry , it turns out , plays a fundamental role in the identification of the gapped interface as a topological parafermion wire similar to those studied in refs . .
the topological properties of the gapped interface can be more transparently revealed by a description in the zero correlation length limit @xmath74 , where the interface hamiltonian _
density _ is given solely by eq .
, thus leading to a gapped interface as depicted in fig .
[ topological_interface](c ) . in this limit
the ground state manifold contains @xmath15 degenerate minima [ @xmath75 , @xmath76 associated with the vacuum expectation value of the composite bosonic operator @xmath77 , which represents a bound state of the quasiparticle @xmath78 with @xmath79 .
we can find a representation of the ground state manifold in terms of the eigenstates of @xmath80 as @xmath81 ( @xmath82 ) .
the structure of the eigenstates is in direct correspondence with symmetry broken states of the ferromagnetic , zero correlation length limit of an @xmath15-state clock model , where @xmath80 naturally acquires the interpretation of a clock operator satisfying @xmath83 and @xmath84 .
however , while it would seem possible to distinguish among the degenerate states by a measurement of @xmath85 , @xmath86 [ which is equivalent to adding a perturbation @xmath87 to the hamiltonian ] , the fact that @xmath85 is a _ non - local _ operator does not permit such a local distinction , and is a hallmark of the topological nature of the system . with this in mind , the eigenstates indicate a degenerate symmetry breaking manifold associated with the global symmetry @xmath88 whereby @xmath89 , for @xmath90 . the topological nature of this system can be made explicit by changing from the clock to the parafermionic representation : @xcite @xmath91 whereby the parafermion operator @xmath92 is a product of the order " , @xmath80 , and the disorder " , @xmath93 , operators . the operators satisfy standard parafermionic commutation relations @xmath94 importantly , we find that the boundary parafermion operators @xmath95 commute with the hamiltonian , and the degenerate ground state manifold is given by the eigenstates of the non - local operator @xmath96 : @xmath97 , @xmath98 where the @xmath99 are linear combinations of the @xmath100 _ * 2 .
edge transitions * _ as indicated in figs .
[ topological_interface](b ) and [ topological_interface](c ) , the formation of the gapped interface prevents the edge modes of phases 1 and 2 from propagating in the x - direction .
any point @xmath101 in the interior of the gapped interface establishes a domain wall between distinct gapped bulk topological phases . in turn , the end states of the gapped interface located at @xmath102 correspond to domain walls between topologically distinct _
gapless _ edge states .
in fact , we shall explicitly demonstrate that , associated to each of the gapless edge transitions there is a parafermion operator with quantum dimension @xmath103 obeying parafermion statistics .
thus , the existence of the parafermions at the end points of the gapped interface , each with quantum dimension @xmath104 is a direct manifestation of the n - fold degeneracy of the ground state manifold of the gapped interface .
similar physics was first explored in ref .
@xcite which focuses on transitions between distinct edge terminations of the _ same _ bulk phase .
here we focus on the interface between different bulk phases , which will have an accompanying transition on the edge .
an important feature of the gappable topological interface is that the bulk phases 1 and 2 can be related to each other by the confinement ( or deconfinement ) of a @xmath105-d @xmath106 gauge theory . in order to see this , imagine phase 2
is coupled to a @xmath106 gauge theory in its deconfined phase .
let the gauge field @xmath107 describe the excitations of phase 2 , and ( @xmath108 ) the excitations of the @xmath106 gauge theory .
hence , the coupled system is described by the abelian chern - simons theory : [ eq : 2d state coupled to zn gauge theory ] @xmath109 where @xmath110 einstein summation convention is assumed for repeated indices and greek indices take values @xmath111 .
in this basis , @xmath112 and @xmath113 represent the original charge and flux excitations of the @xmath106 gauge theory .
it is possible to show that , by a change of basis , k - matrix can be transformed into @xcite @xmath114 where @xmath115 and @xmath116 represents a pauli matrix , i.e. , a trivial sector that can always be gapped out .
thus , eq . explicitly illustrates that phase 1 can be obtained from phase 2 by a gauging mechanism ; reversely , phase 2 descends from phase 1 by confining the @xmath106 gauge theory .
this kind of gauging mechanism has proven useful in understanding the classification of symmetry enriched topological states @xcite and hidden anyonic symmetries@xcite .
using this theory we can now explicitly prove the existence of domain - wall parafermions . to do this
we analyze the edge states obtained from the bulk theory and in order to study the transitions between the edge phases 1 and 2 . for that we model
the gapless edge states propagating along one of the edges , say @xmath117 , with the effective theory @xmath118 where @xmath119 are the edge fields .
the interactions @xmath120 and @xmath121 will be chosen to stabilize the edge phases 1 and 2 , respectively , in different spatial regions . to carry this out we use position - dependent coupling constants @xmath122 and @xmath123 such that , in regions where we want phase 1 , @xmath124 and @xmath125 , while for phase 2 , @xmath126 and @xmath127 .
for the sake of concreteness , we shall take @xmath128 and @xmath129 in , although similar results can be obtained for the @xmath130 case with @xmath131 the interaction choice @xmath132 will gap the trivial modes in @xmath116 , which then yields the edge states of phase @xmath133 .
alternatively , the interaction @xmath134 gives rise to the edge state of phase 2 , that is , it effectively leads to the confinement of the @xmath106 gauge theory . in order to see this , notice that the edge excitations that remain deconfined in the presence of the interaction are described by vertex operators @xmath135 , such that @xmath136 [ with @xmath137 commutes with the argument of the interaction . from this condition , which is satisfied when @xmath138 $ ]
, it is easy to see that the deconfined edge excitations are those of the phase 2 described by @xmath139 .
more intuitively , upon rewriting @xmath140 , with @xmath141 and @xmath142 , is seen as the expected electric "- mass interaction that confines the excitations of the @xmath106 gauge theory .
now let @xmath143 and @xmath144 be the segments in the @xmath117 edge corresponding to phases 1 and 2 , respectively , where @xmath145 is an infinitesimal positive number that regulates the domain wall transitions ( see fig . [ topological_interface](a ) ) .
it then follows that the operators defined as [ eq : definition and algebra of o and t ] @xmath146 } \\ & \ , \mathcal{t}_{i } = \exp \big [ { \frac{\mathrm{i}}{n}\ , \int_{y_{2i } - \varepsilon}^{y_{2i+1}+\varepsilon}\,dy\ , \partial_{y } \left ( l^{t}_{1}\,k'\,\phi\right ) \big ] } \end{split}\ ] ] commute with the edge hamiltonian and satisfy the non - trivial commutation relations @xmath147 the ground state manifold forms a representation of the algebra , which implies a ground state degeneracy of @xmath148 in the presence of @xmath149 domain walls on the boundary , i.e. , @xmath150 gapped interfaces . guided by ,
parafermion operators located on the domain walls between phases 1 and 2 along the @xmath117 edge can be constructed as : [ eq : parafermions definition and algebra ] @xmath151 } \\ & \ , \alpha_{x_{l},2 i + 1 } = e^ { \frac{\mathrm{i}}{n}\ , \big [ l^{t}_{1 } k_{g } \phi^{'}(y_{2i+1 } + \varepsilon ) - l^{t}_{2 } k_{g } \phi^{'}(y_{2i+1 } - \varepsilon ) \big ] } \ , , \end{split}\ ] ] which are readily shown to satisfy @xmath152 for a generic gappable interface between one - component states we have the constraint @xmath153 which implies that the phases must be related by the confinement of a @xmath154 gauge theory , and the subsequent gauging and deconfinement of a @xmath155 symmetry . in these cases
one would find @xmath156 parafermions ( see sm for more detail ) .
a realization of the algebra has been studied in ref .
@xcite , for the transitions been chiral bosonic edge states with @xmath157 and @xmath158 .
while their approach focused solely on the edge transitions of the same bulk phase , our formulation shows that the existence of non - trivial parafermionic modes is a direct consequence of the formation of a gapped interface between different chiral topological states .
hence , we have generalized their result to arbitrary one - component edge transitions , and have shown that such transitions can originate from a bulk phenomenon associated with confinement - deconfinement transitions of discrete gauge theories . additionally , since these parafermions appear at a
t - junction " between two chiral gapless states and the termination of their gapped interface , they represent a completely new physical phenomenon when compared with the cases studied in refs . @xcite .
in closing we note that the gapped interface acts like an anyonic andreev reflector in the bulk .
anyons from , say , phase 1 will hit the interface and be transformed into a combination of outgoing anyons in phase 2 as well as reflected anyons that remain in phase 1 .
for instance , take @xmath159 for simplicity . then
, as , for example , quasiparticle @xmath160 approaches the interface , a vacuum fluctuation can create a ( @xmath161 ) pair in the region of phase 2 immediately adjacent to the interface ; subsequently , the condensation of ( @xmath162 ) leaves behind the quasiparticle @xmath59 in phase 2 , as shown in fig .
[ topological_interface](a ) .
the quasiparticles @xmath163 belonging to phase 1 can be absorbed by the gapped interface and fully transmuted into multiples of the local excitation @xmath59 of phase 2 .
other anyons hitting the interface will be partially transmuted and partially reflected by the condensate .
for example , if @xmath164 hit the surface it could generate a @xmath59 in phase 2 as well as a reflected @xmath165 in summary , we have shown that a gapped interface between different topologically ordered phases can not be topologically trivial itself .
the interpolation between the topological orders generates a quasi-1d topological parafermion phase which exhibits characteristic non - abelian defect modes where the interface intersects the boundary of the system .
although we have only shown this for one - component interfaces , we expect the generalizations to more complicated interfaces to provide a rich set of phenomena .
furthermore , our result may aid in the interpretation of the topological entanglement entropy arising at heterointerfaces of topologically ordered phases as recently calculated in ref . .
we leave this to future work .
* acknowledgments : * we would like to thank j. cano , e. fradkin , m. mulligan and m. stone for useful conversations .
lhs is supported by a fellowship from the gordon and betty moore foundation s epiqs initiative through grant no .
gbmf4305 at the university of illinois .
tlh is supported by the us national science foundation under grant dmr 1351895-car .
99 x .- g .
wen , quantum field theory of many - body systems , oxford , uk : univ .
pr .
( 2004 ) m. z. hasan and c. l. kane , rev .
mod .
phys .
* 82 * , 3045 ( 2010 ) .
j. e. moore , nature ( london ) * 464 * , 194 ( 2010 ) .
x .- l .
qi and s .- c .
zhang , rev .
mod .
phys .
* 83 * , 1057 ( 2011 ) .
a. p. schnyder , s. ryu , a. furusaki , and a. w. w. ludwig , phys .
rev .
b * 78 * , 195125 ( 2008 ) x .- l .
qi , t. l. hughes , and s .- c .
zhang , phys .
rev .
b * 78 * , 195424 ( 2008 ) .
a. kitaev , proceedings of the conference `` advances in theoretical phyiscs : landau memorial conference '' , ( chernogolovka , russia , 2008 ) aip conf
. proc . * 1134 * , 22 ( 2009 ) .
f. pollmann , a. m. turner , e. berg , and m. oshikawa , phys .
rev .
b * 81 * , 064439 ( 2010 ) .
n. schuch , d. perez - garcia , and i. cirac , phys .
rev .
b * 84 * , 165139 ( 2011 ) .
x. chen , z. -c .
gu , z. -x .
liu and x. -g .
wen , phys .
rev .
b * 87 * , 155114 ( 2013 ) .
m. levin and z .- c .
gu , phys .
rev .
b * 86 * , 115109 ( 2012 ) . y .- m . lu and a. vishwanath , phys . rev .
b * 86 * , 125119 ( 2012 ) . a. vishwanath and t. senthil , phys . rev .
x * 3 * , 011016 ( 2013 ) . c. wang , a. potter , and t. senthil , science
* 343 * , 629 ( 2014 ) .
zhen bi , alex rasmussen , and cenke xu , phys .
rev .
b * 91 * , 134404 ( 2015 ) .
x .- g .
wen , adv .
phys . * 44 * , 405 ( 1995 ) .
a. kapustin and n. saulina , nucl .
phys .
b * 845 * , 393 ( 2011 ) .
m. levin , phys .
rev .
x * 3 * , 021009 ( 2013 ) .
m. barkeshli , c .- m .
jian , and x .-
l .
wen , phys .
rev .
b * 88 * , 241103(r ) ( 2013 ) . n. h. lindner , e. berg , g. refael , and a. stern , phys . rev .
x * 2 * , 041002 ( 2012 ) .
d. j. clarke , j. alicea , and k. shtengel , nature commun . * 4 * , 1348 ( 2013 ) .
m. cheng , phys .
rev .
b * 86 * , 195126 ( 2012 ) .
a. vaezi , phys .
rev .
b * 87 * , 035132 ( 2013 ) .
m. barkeshli , c .- m .
jian , and x .-
l .
qi , phys .
rev .
b * 87 * , 045130 ( 2013 ) .
r. s. k. mong , d. j. clarke , j. alicea , n. h. lindner , p. fendley , c. nayak , y. oreg , a. stern , e. berg , k. shtengel , m. p. a. fisher , phys .
rev .
x * 4 * , 011036 ( 2014 ) .
m. n. khan , j. c. y. teo , and t. l. hughes , phys .
rev .
b * 90 * , 235149 ( 2014 ) .
m. n. khan , j. c. y. teo , t. l. hughes , and s. vishveshwara , arxiv:1603.04427 ( unpulished ) . c. nayak , s. h. simon , a. stern , m. freedman , s. das sarma , rev .
mod .
phys . * 80 * , 1083 ( 2008 ) .
j. cano , m. cheng , m. barkeshli , d. j. clarke , and c. nayak , phys .
rev .
b * 92 * , 195152 ( 2015 )
. w. pan , h. l. stormer , d. c. tsui , l. n. pfeiffer , k. w. baldwin , and k. w. west , phys .
rev .
lett . * 88 * , 176802 ( 2002 ) .
w. pan , j. s. xia , h. l. stormer , d. c. tsui , c. vicente , e. d. adams , n. s. sullivan , l. n. pfeiffer , k. w. baldwin , and k. w. west , phys .
rev .
b * 77 * , 075307 ( 2008 ) .
f. d. m. haldane , phys .
rev .
lett . * 74 * , 2090 ( 1995 ) .
m. levin and a. stern , phys .
rev .
b * 86 * , 115131 ( 2012 ) .
n. read , phys .
rev .
lett . * 65 * , 1502 ( 1990 ) .
j. frohlich and a. zee , nucl .
phys . b**364 * * , 517 ( 1991 ) .
x-.g .
wen and a. zee , phys .
rev .
b * 46 * , 2290 ( 1992 ) .
j. frohlich and e. thiram , j. stat . phys . *
76 * , 209 ( 1994 ) .
j. cano , m. cheng , m. mulligan , c. nayak , e. plamadeala , and j. yard , phys .
rev .
b * 89 * , 115116 ( 2014 ) .
a yu kitaev , phys .- usp . *
44 * , 131 ( 2001 ) .
p. fendley , j. stat .
mech .
theor .
exp .
( 2012 ) p11020 .
r. bondesan and t. quella , j. stat .
mech .
( 2013 ) p10024 .
j. motruk , e. berg , a. m. turner , and f. pollmann , phys .
rev .
b * 88 * , 085115 ( 2013 ) .
a. s. jermyn , r. s. k. mong , j. alicea , and p. fendley , phys .
rev .
b * 90 * , 165106 ( 2014 ) .
y. zhuang , h. j. chaglani , n. m. tubman , and t. l. hughes , phys .
rev .
b * 92 * , 035154 ( 2015 ) .
a. alexandradinata , n. regnault , c. fang , m. j. gilbert , and b. a. bernevig , phys .
rev .
b * 94 * 125103 ( 2016 )
. e. fradkin and l. p. kadanoff , nucl .
phys . b**170 * * , ( 1980 ) 1 - 15 . y .- m .
lu and a. vishwanath , phys .
rev .
b * 93 * , 155121 ( 2016 ) .
j. cano , t. l. hughes , and m. mulligan , phys .
rev .
b * 92 * , 075104 ( 2015 ) .
see supplementary material . * supplementary material of _ parafermionic wires at the interface of chiral topological states _ * |
boxiness of galactic bulges is in most cases discussed in connection with the boxiness observed in the edge - on views of disc galaxies .
however , boxiness is encountered also in nearly face - on views of barred galaxies .
this can be either an overall boxiness of the bar , usually referring to rectangularity of the outermost isophotes of the bar component in early - type barred galaxies ( see e.g. * ? ? ?
* ) , or boxiness of the isophotes in the inner regions , i.e. in distances from the centre roughly up to the middle of the semi - major axis of the bar @xcite . as regards the outer boxiness it is plausible to assume that it is related with orbits at the end of the bar , i.e. at regions with a much smaller scale height than that of the regions of the bar that participate in the formation of the peanut
in that respect , for the outer boxiness of the bar the presence of the radial 4:1 resonance brings in the system rectangular - like p.o . @xcite
that attract around them quasi - periodic or sticky non - periodic orbits .
the role of chaotic orbits in reinforcing the outer boxiness of the bars has been indicated by @xcite , @xcite , @xcite and @xcite .
this can lead to the boxy shapes of early type bars . in the study of the outer face - on boxiness of the bars , we deal practically with 2d dynamics .
thus , we find in 2d models , that orbits associated with this feature are also related with the orbits building the chaotic spirals beyond corotation @xcite , i.e. both structures share the same orbital content .
essentially these are orbits on the equatorial plane .
nevertheless , the connection of the rectangularity of a bar with the radial 4:1 resonance has been also detected in 3d models @xcite .
however , the relevant orbits remain at small heights away from the equatorial plane . for the _ inner _ boxiness on the other hand we must use 3d orbits .
if we accept the dimensions of a boxy bulge being as discussed in @xcite ( hereafter paper i ) , the rectangularity of the inner isophotes ( as depicted in the figures of * ? ? ?
* ) corresponds to regions occupied by the b / p bulge in the side - on profiles . from this
arises the question if there exist 3d orbits contributing both to a face - on as well as to a side - on boxy profile .
the existence of this combination is not obvious .
thinking about the shapes of periodic orbits , or about quasi - periodic orbits with similar shapes like the periodic ones , this combination does not exist neither for this model nor for any other of the models in @xcite or in the model used by @xcite .
the periodic orbits of the x1-tree ( x1 , x1v1 , x1v2 , x1v3 etc . ) follow in their projections on the equatorial plane the morphological evolution of the x1 family as the jacobi constant ( hereafter called energy ) increases . we have to reach energies at the radial 4:1 resonance in order to obtain a rectangular shape in their face - on , @xmath1 in our model , projection @xcite .
orbits with such energies are found in the outer parts of the bars .
however , if we consider large @xmath2 or @xmath3 perturbations of planar p.o . in the 4:1 resonance region
, calculating in this way orbits with large @xmath4 values , we enter into chaotic regimes of the models , where orbits do not support any particular structure . by excluding orbits with large energies ,
i.e. at or between the radial 4:1 resonance and corotation , we have to seek after orbits in lower energies , where we encounter the orbits supporting the boxiness of the edge - on profiles .
stable 2d orbits of the planar , `` 3:1-type '' , families could contribute to the reinforcement of boxy structures on the equatorial plane @xcite .
vertically perturbed orbits with @xmath5 initial conditions on their islands of stability are candidates for providing 3d orbits with boxy projections on the @xmath1 plane .
this possibility has to be examined for representatives of these families with appropriate energies .
we already have seen that there are orbits supporting the side - on b / p bulge without having a `` x1-like '' morphology in their face - on projection ( cf figure 19 in paper i ) . in the present paper
we examine the possibility of having among them orbits with boxy @xmath1 projections .
we use in our work the same model as in paper i and we keep the same formalism . in section
[ sec : iboxi ] , starting with 2d orbits on the equatorial plane , we examine their orbital behaviour when they are perturbed away from it . focusing on the non - periodic orbits that support the b / p structure
we first study their dynamics in the typical case where all important families coexist and x1 is stable .
then we investigate as well possible changes introduced in the energy interval of the vertical ilr , where x1 is simple unstable . in section [ sec:3/1 ] we investigate the role of symmetric orbits bifurcated at odd , radial n:1 resonances .
finally , we summaries our results and discuss our conclusions in section [ sec : concl ] .
an energy at which all main families of p.o . contributing to the b / p structure coexist is e@xmath6 @xmath7 . at this energy x1 , x1v1 and x1mul2 are stable , while x1v2 is simple unstable @xcite .
it is a typical energy for understanding the dynamics at the peanut region .
we examine the contribution of non - periodic orbits in the neighbourhood of p.o . from these four families to the reinforcement of inner boxy features , first by successively perturbing the x1 p.o . by increasing @xmath9 .
[ xpx41 ] gives the @xmath10 cross section at e@xmath6 @xmath7 , which has been constructed this way .
we investigate the phase space occupied by these orbits , when they are kicked out of the equatorial plane . in the @xmath10
cross section the initial conditions of all important families of p.o .
involved in the enhancement of the boxy side - on profiles , i.e. x1 , x1v1 , x1v2 and x1mul2 , almost overlap at about @xmath11 .
their location is indicated with a heavy black dot in fig .
[ xpx41 ] . in the same figure
we observe that the perturbed x1 orbits form around the p.o .
a large central region , characterized mainly by order . inside this stability island
we discern three smaller stability islands belonging to a p.o . of multiplicity 3 ,
let us call it tr1 , and further out a chain of seven stability islands ( drawn with green ) that mark the border between the ordered region and a chaotic zone that surrounds it .
embedded in this chaotic zone we observe two sets of islands .
they are related to two p.o . of multiplicity 2 , which we name rm21 and rm22 and
are centred at @xmath12 and @xmath13 respectively .
arrows in fig .
[ xpx41 ] indicate the locations of the orbits tr1 , rm21 and rm22 .
other families of p.o .
existing at this energy , like x2 and x3 , are not discussed at this point and are not indicated in fig .
[ xpx41 ] . by inspection of fig .
[ xpx41 ] these three families of higher multiplicity seem to play a minor role in the enhancement of any morphological feature in the model .
however , our investigation has shown the opposite , so we present here in detail how they contribute to the enhancement of boxiness in both face - on and side - on views . the families rm21 and rm22 have morphologies symmetric with respect to the @xmath14 axis .
they are given together in fig .
[ rrorbs ] , with black and magenta colours respectively . the combination of the two p.o .
provide a morphology ideal for supporting boxy structures in the central parts of face - on bars , first of all on the equatorial plane .
even so , the system should provide for that purpose also non - periodic orbits that will follow similar morphologies . in order to trace the origin of the rm21 and rm22 families
we plot the stability diagram of x1 considering it as being of multiplicity two ( fig .
[ mul2stab ] ) . both rm21 and rm22 families
have the same stability indices , as they are essentially two branches of one family , symmetric with respect to the @xmath15 axis .
they are bifurcated from x1 at e@xmath6 @xmath16 , where the b1 stability index of x1 , considered as being of multiplicity 2 , has a tangency with the b@xmath17 axis . the b1 stability index is associated with radial perturbations , so two 2d families of p.o .
are bifurcated .
they are initially simple unstable ( u ) , since the mother family x1 is u at the energy they are introduced in the system .
the other family of higher multiplicity we have encountered at e@xmath6 @xmath7 is the orbit of multiplicity 3 , tr1 , which is indicated with an arrow , inside the territory of the central stability island in fig .
[ xpx41 ] .
it coexists together with its unstable counterpart of multiplicity 3 , which is located between the three islands of stability ( not indicated in fig .
[ xpx41 ] ) .
the origin of tr1 can be found by considering it as a `` deuxime genre '' family @xcite .
since it is a radial bifurcation of x1 its starting point has to be sought at an energy where the corresponding stability index of the simple periodic mother family will be @xmath18 ( see e.g. the appendix in * ? ? ?
this happens for e@xmath6 @xmath19 ( cf figure 1b in paper i ) . at e@xmath6
@xmath19 tr1 is bifurcated from x1 as stable and extends over a significant range of energies , including the ilr region , until it becomes simple unstable at e@xmath6 @xmath20 .
its face - on portrait is boxy , as can be seen in fig .
in addition its stability secures motion on 4d tori and sticky orbits on them .
perturbations up to @xmath21 of orbits starting on the @xmath10 plane in the region between the dark blue and red coloured invariant curves in fig .
[ xpx41 ] give orbits on triple rotational tori .
the face - on morphologies of these orbits can be thought as thick versions of the tr1 p.o . as depicted in fig .
their side - on projections have a b / p profile , however they remain close to the equatorial plane , reaching heights @xmath22 .
quasi - periodic orbits found by adding @xmath23 to the orbits on the invariant curves in the @xmath10 plane give orbital morphologies similar to those we find when we impose @xmath2 perturbations with slightly thicker side - on views .
the closed elliptical curves around rm21 and rm22 indicate stability only on the equatorial plane
. however , these families are at e@xmath6 @xmath24 unstable to vertical perturbations @xmath25 and thus non - periodic orbits in their neighbourhood in the @xmath26 or @xmath27 directions will be chaotic .
even very small @xmath2 or @xmath3 perturbations of an orbit on an invariant curve around rm21 or rm22 , bring the perturbed orbits away from the @xmath28 plane .
nevertheless , we observe that these orbits during their integration stay confined in some regions of the 4d space of section . as an example we consider an orbit starting _ on _ an invariant curve around rm21 with @xmath29 , which is perturbed by @xmath30 .
we integrated it for time equal to 2000 x1 periods having an energy conservation with a relative error less than 10@xmath31 .
for time corresponding to 10 x1 periods this orbit stays outside the region defined by the seven stability islands in fig .
[ xpx41 ] . then its consequents are found inside the seven islands region and they are projected in the @xmath10 plane around the blue invariant curve . they stay there forming a ring for more than 850 x1 orbital periods .
this is the ring of black points that can be observed in fig .
[ sticky41xpx]a close and around the blue invariant curve region .
the formation of a ring structure in the @xmath10 plane corresponds to a spatial confinement of the orbits in both 3d projections of the 4d space of section , which include this plane , namely @xmath32 and @xmath33 . in both of these two 3d projections we have the formation of a cylindrical structure that extends away from the @xmath10 plane remaining surrounded by the chain of the seven stability islands
this can be seen in fig .
[ sticky41xpx]b and fig .
[ sticky41xpx]c for the @xmath32 and @xmath33 projections respectively .
3d projections including the @xmath34 plane are very illustrative for understanding the dynamical character of these orbits and are described below extensively . for the time
being we want to underline that despite their sticky character these orbits are chaotic . in the example of the orbit we describe ,
this means , ( i ) : : that if we extend the integration time , the consequents of the orbit will cross again the border of the seven stability islands on the @xmath10 projection and will diffuse in a larger volume of the phase space .
nevertheless , later they return and stay trapped on the area of the ring of fig .
[ sticky41xpx]a for hundreds of x1 dynamical times . also during the time an orbit spends away from a sticky zone , i.e. in the example of our orbit away from the ring , it may stay confined for a time interval within this period in another subspace of the space of section . in our case the orbit stays for several hundreds of x1 periods confined within the area of the innermost , red , invariant curve in fig .
[ sticky41xpx]a before it returns again to the ring area and spends another time interval of more than 200 x1 periods in this zone .
( ii ) : : that starting at nearby initial conditions we will have a quantitatively different orbital behaviour .
nevertheless _ qualitatively _ the overall dynamical behaviour is similar .
if we add a tiny @xmath35 perturbation to the initial conditions of the orbit with the black consequents in fig .
[ sticky41xpx]a , b , c we do not find the new orbit being trapped in the same zones of the phase space during the same or close - by time intervals .
however , we find it spending time of the same order ( hundreds of x1 periods ) in roughly the same zone , e.g. the ring area . the trapping of this new orbit in the area around the blue invariant of fig .
[ sticky41xpx]a is given in fig .
[ sticky41xpx]d .
this orbital behaviour is typical not only for orbits starting on the invariant curves of rm21 and rm22 in the @xmath10 plane , but for all orbits we started integrating from the `` chaotic sea '' region of fig .
[ xpx41 ] .
this is to be expected since essentially we deal with a single chaotic sea . in infinite time
a chaotic orbit will visit the whole available volume of the phase space .
however here , we integrate parts of this orbit starting from different initial conditions . even in the region
@xmath36e@xmath6 @xmath37 , where rm21 , rm22 are stable ( fig . [ mul2stab ] ) the above orbital dynamics does not practically change , because the zone of influence around these stable p.o .
in this case we have found a double `` tube torus '' ( i.e a self - intersecting torus in a 3d projection , see * ? ? ?
* ) in the 4d phase space .
we present it in fig .
[ doubtor ] at e@xmath6 @xmath38 since it is the first time we find a tube torus in a family of higher multiplicity .
the tori belong to the perturbed by @xmath39 rm22 orbit at that energy .
the formed `` ribbons '' have an intersection in the 3d @xmath33 projection , but not in the 4d space as one concludes from the different colours ( according to the @xmath27 coordinate ) of the intersecting branches . for slightly larger perturbations of the stable p.o . in the @xmath27 or @xmath26 directions
the orbits diffuse in phase space .
it is evident that the orbits originating in the chaotic sea around the central stability region in the @xmath10 cross section ( fig . [
sticky41xpx]a ) represent a major class of orbits , since they occupy a large volume of the phase space and have similar behaviour in the examined range of energies @xmath40e@xmath6 @xmath41 .
so , we want to investigate if there is a particular group among them contributing to boxy structures in their face - on views .
their e@xmath6 values point to the right size of orbits for supporting the inner boxiness we study .
if so , we want also to find out what is the common feature of such an orbital set .
for example we want to investigate if they reinforce boxy face - on structures when they stay at a particular region of the phase space .
the first thing that we have examined is whether the orbit we have studied starting on the invariant curve of rm21 on the @xmath10 plane , supports a particular morphology during the time it spends in the ring inside the chain of the seven stability islands .
the result is striking as can be seen in fig .
[ upart ] . in fig .
[ upart]a we observe that the face - on projection has clearly a boxy character . since the orbit stays in the ring for more than 850 x1 periods we sample it every 100 points during this time interval in order to plot fig .
[ upart]a .
the dimensions of the box are comparable with the dimensions of the tr1 p.o . in fig .
simultaneously , we observe that the side - on , @xmath42 , projection of the orbit ( fig .
[ upart]b ) reflects the morphology of orbits trapped around x1v1 and x1v1@xmath0 , since it appears as a combination of `` @xmath43 '' and `` @xmath44 '' parts .
we already know the imprint of sticky chaotic orbits to the tori of these families in the @xmath34 plane from paper i ( cf figure 14 in paper i ) .
the spreading of the consequents of the orbit during the time of the 850 x1 periods in the @xmath34 reproduces exactly this typical pattern , as we can observe in fig .
[ upart]c .
we have already seen that during the time an orbit forms a ring of consequents projected inside the area defined by the chain of the seven stability islands in the @xmath10 plane , it can support an inner boxy feature in the face - on view of the model .
next , as a second step , we examine if we have a similar behaviour in the case of orbits having projected consequents during their integration time only in the chaotic zone of fig .
[ xpx41 ] .
such orbits exist indeed . as an example we give the orbit with initial conditions @xmath45 , always at the typical energy e@xmath6 @xmath24 . for time corresponding to about 300 x1 periods ,
its face - on morphology has a boxy character with dimensions similar to the combined rm21,2 p.o .
[ rrorbs ] ) .
this is given in fig .
[ ob41]a . with black
we depict the orbit integrated for 10 periods of x1 , while the red background depicts the same orbit integrated for 300 x1 periods . the side - on morphology
can be observed in fig .
[ ob41]b and it is boxy . although not as sharp as the one in fig .
[ upart]b , this orbit has again a morphology similar to the one produced by a combination of quasi - periodic orbits around x1v1 and x1v1@xmath0 .
it has an overall b / p morphology and it reaches heights close to @xmath46 . let us call `` phase a '' the time interval corresponding to 300 x1 periods within which the orbit has the morphology presented in figs .
[ ob41]a and [ ob41]b . for time
larger than 300 x1 periods , the morphology of the orbit changes in a characteristic way and is given in figs .
[ ob41]c and [ ob41]d .
the red background corresponds now to the part of the orbit in the time interval between 300 and 1000 x1 periods after starting the integration .
we call this period `` phase b '' . with black
we plot the orbit during the first 12 x1 periods during `` phase b '' . the face - on view ( figs .
[ ob41]c ) becomes rounder but evidently it increases the surface density in two stripes parallel to the y - axis at @xmath47 .
they can be better seen in the red background and are indicated with arrows . in the side - on view of the orbit ( fig .
[ ob41]d ) we realize that the denser stripes parallel to the y - axis are like two caps , that restrict the extent of the orbit to @xmath48 . in intermediate heights
the orbit does not follow any particular morphology .
the consequents of the orbit in fig .
[ ob41 ] remain projected _ around _ the central stability island of the @xmath10 plane ( fig .
[ xpx41 ] ) . practically they overlap with the chaotic sea as depicted in fig .
. however , as we move from `` phase a '' to `` phase b '' they drift from the area of the seven stability islands to larger distances .
a few consequents during `` phase b '' are found around the stability island of the planar retrograde family x4 .
( the invariant curves around x4 are not drawn in fig . [ xpx41 ] .
they are located to the left of the figure , reaching @xmath49 values outside the drawn frame ) . to a large extent
the consequents of the orbit in fig .
[ ob41 ] surround the stability region in fig .
[ xpx41 ] _ on _ the @xmath10 plane forming two vaguely defined successive rings .
the projections of the consequents of the orbit of our example in the @xmath34 plane , are given in fig .
they are given together with the consequents of a sticky orbit around x1v1 and x1v1@xmath0 ( magenta dots ) , as well as together with the projections of the rotational tori of the x1mul2 family at the top and the bottom of the figure around the points @xmath50 ( cf .
figure 14 in paper i ) .
the red and the dark blue rings in the central region of the figure are the projections of one of the innermost and one of the outermost tori around x1 respectively .
green points in fig .
[ ob41]e are the consequents of the orbit during phase `` a '' .
they start being projected in the @xmath34 plane inside the blue ring , but then they move outside it , and occupy roughly the same region with the magenta orbit , which is sticky to x1v1 and x1v1@xmath0 . the area covered by the consequents and their distribution on the @xmath34 plane
are very close also to the projected consequents of the orbit in fig .
[ upart]c . during `` phase a ''
the orbit is unambiguously sticky to x1v1 and x1v1@xmath0 . for longer integration times , during `` phase b '' , the consequents form a ring roughly surrounding the green consequents and are plotted in black . their relative location during the two different phases `` a '' and `` b '' , are given also in the embedded frame in the upper right corner of fig .
the dimensions of the embedded frame are the same as of the main frame of fig .
the ring of black points is located practically around the green consequents and its extent in the @xmath27 direction seems to be hindered by the x1mul2 tori .
the weak confinement in two ring structures in the @xmath10 and @xmath34 planes seems to be enough to secure a boxy character in the configuration space , of the kind described in fig .
[ ob41]c and d. the time intervals of hundreds of x1 dynamical times are already longer than the time needed to consider a structure as supported by orbits . however , following the orbit of fig .
[ ob41 ] for even longer times gave further interesting results .
for time larger than 1000 periods of x1 , the consequents in the @xmath34 projection return inside the dark blue ring and form another ring , this time coloured with grey in fig .
the consequents remain trapped on this ring for at least 2500 x1 periods . by the same time they remain trapped on ( rotational ) tori around x4 and would be projected to the left of the `` chaotic sea '' in fig .
[ xpx41 ] ( roughly at @xmath51 ) . during this
last phase no boxiness is supported in the face - on view of the model . instead the orbit forms a counter - rotating disk on the @xmath1 plane . in the @xmath42 projection the orbit builds a thick layer close to the equatorial plane , without any particular substructure .
thus , integrated for more than 1000 dynamical times this orbit does not reinforce the double boxy profile .
this case offers a counter - example of a sticky orbit reinforcing a structure . like quasi - periodic orbits being trapped around a `` wrong '' stable periodic orbit , as is x4 for the building of a galactic bar , during this phase
the sticky chaotic orbit destroys the boxy structure in our example . in the case of sticky orbits ,
our example shows how the same orbit can reinforce or destroy a structure during different phases of its evolution .
the information about the trapping around different tori of p.o . can be obtained only directly by means of the method of the surface of section .
until now it became clear that a large number of planar orbits , which are displaced from the equatorial plane by perturbations in the @xmath26 or @xmath27 directions become sticky to the x1v1 and x1v1@xmath0 tori and during this phase they support a double boxy morphology .
this happens in the energy interval @xmath40e@xmath6 @xmath41 .
however , the time interval during their integration , during which they will be trapped at specific regions of the phase space varies as we change the initial conditions .
immediate support of the double boxiness from the beginning of the integration is obtained by vertically perturbed orbits with @xmath52 initial conditions from the region between the dark blue invariant , just beyond the tr1 islands and the orbit of multiplicity 7 in fig .
[ xpx41 ] . by perturbing initial conditions in this region in the @xmath26 or @xmath27 direction
we find quasi - periodic orbits on 4d tori and orbits sticky to these tori , which have a double - boxy morphology . as a typical example we give the orbit with @xmath53 .
its face - on and side - on views are given in fig .
[ bb41stick]a , b respectively . with black
is given the orbit integrated for 10 periods of x1 at e@xmath6 @xmath24 , while with red is the same orbit integrated for 100 x1 periods .
however , the orbit continues to reinforce the orbit for many hundreds of x1 dynamical times .
we observe in fig .
[ bb41stick]b , that the orbit reaches heights @xmath54 .
the dynamical mechanism is similar to the one we described for the orbits we started integrating in the chaotic sea of fig .
[ xpx41 ] , namely stickiness to the tori of x1v1 and x1v1@xmath0 .
this becomes evident if we look at the @xmath10 and @xmath34 projections of the consequents of the orbit of fig .
[ bb41stick]a , b .
they are given in fig .
[ bb41stick]c , d . especially in the @xmath34 projection ( fig .
[ bb41stick]d ) it is clear that the orbit reproduces the pattern of the sticky orbits we have first encountered in figure 14 of paper i. the advantage of these orbits with respect to other orbits reinforcing the double boxy profiles is that one obtains this morphology immediately after starting integrating them without any delay . in paper
i we designated the energy interval in which x1 becomes simple unstable at the vertical ilr as @xmath55e@xmath56 . unlike the typical case with e@xmath6 @xmath24 we used to discuss the dynamical mechanism in the previous section , in this interval x1
is not stable , but simple ( vertically ) unstable .
nevertheless , in the @xmath10 plane we encounter around x1 invariant curves .
we find that if we perturb in the vertical direction orbits on these invariant curves we may encounter again boxy face - on projections in the configuration space .
the main reason for this is again the proximity of the x1v1 and x1v1@xmath0 tori to the initial conditions of x1 .
there is always a @xmath57 or @xmath3 range for which the perturbed planar orbits are affected by the presence of these tori .
the examples used in paper i for discussing vertically unstable x1 orbits , having 2d regular orbits around them in the @xmath10 plane , are typical cases of such orbits .
the perturbed by @xmath58 x1 orbit at e@xmath6 @xmath59 , has a peanut shape side - on morphology ( figure 5d of paper i ) .
its face - on view is a typical x1 ellipse with a negligible thickness ( fig .
[ borbits]a ) .
if we include @xmath8 perturbations in this orbit and we start increasing them , the @xmath1 projections of the resulting orbits in the beginning are thick ellipses . for @xmath60 ( keeping always @xmath61 ) their face - on projections have a clear boxy character ( fig .
[ borbits]b )
. this boxiness increases with increasing @xmath62 . for @xmath63
the face - on view of the orbits are as for the orbit in fig .
[ borbits]c , for which @xmath64 . the face - on morphology is boxy for all orbits starting with @xmath64 , @xmath65 and @xmath66 .
we give the orbit with @xmath67 in fig .
[ borbits]d ( the orbits increase their projections on the x - axis as @xmath68 increases ) .
it is worth noticing that despite being double - boxy ( in face - on and side - on views ) , and despite their similarity in their face - on projections , the orbits in fig .
[ borbits]b , [ borbits]c and [ borbits]d , have different boxy morphologies in their side - on views . the orbit in fig .
[ borbits]b has a shape like the one we have seen in figure 5d of paper i. the side - on morphologies of the orbits in fig .
[ borbits]c and fig .
[ borbits]d are similar to other profiles we encountered in paper i ( see figures 6a and 6b respectively ) .
all these findings clearly show that there is a class of non - periodic orbits , either quasi - periodic or ( especially ) sticky to quasi - periodic orbits , that reinforce simultaneously boxy side - on and face - on profiles in our model .
such orbits with energies from the radial and vertical ilr region of our model give in their @xmath1 projections boxy orbits that reach distances about 1 kpc from the centre along the y - axis . for larger energies , especially for e@xmath6 @xmath69 , when x1v1 becomes again stable and other p.o . of the x1-tree
are introduced in the system , we find anew non - periodic orbits trapped by x1v1 tori . such orbits may have a double boxy character , but remain narrow and do not participate to the building of the peanut . for example x1v3 or x1v5
orbits perturbed by @xmath26 or @xmath27 can give orbits with boxy face - on profiles , which however remain in small heights .
these are orbits associated with a possible outer boxiness of the bar .
as it is known ( e.g. * ? ? ? * ) in rotating galactic potentials , at the radial n:1 resonances with n being an odd integer , we have a type 3 bifurcation @xcite . in that case ,
pairs of symmetric families are introduced in the system .
usually we encounter such families as bifurcations of x1 at the 3:1 radial resonance . by plotting together both branches of the symmetric periodic orbits belonging to the same family
we create boxy features , as indicated by @xcite .
recently , @xcite have presented chaotic orbits in the neighbourhood of 3:1 p.o .
that reinforce boxy morphologies close to the end of the bar if integrated for long times .
however , for the inner boxiness we study here , the appropriate orbits have to support boxy features in the bar . in other words their @xmath1 projections
they have to lie inside the region occupied by the corresponding projections of the 3d orbits that support the peanut in the side - on view . in our model
the 3:1 families are bifurcated close to the end of the bar . due to their dimensions , possible reinforcement of boxy features by these orbits
will be associated boxiness at @xmath70 .
also in the case of the potential from an @xmath71-body simulation studied by @xcite , such orbits contribute to the outer boxiness of the bar .
however , we have in our model another pair of planar symmetric families , of 1:1 resonance type , which are found in the right energy range . these orbits , called o1 and o1@xmath0 , are two branches of a family bifurcated as stable from x1 at e@xmath6 @xmath72 . they are discussed in @xcite and have a triangular type morphology . by considering both stable symmetric branches of o1 ( see figure 17 in
* ) we construct a skeleton that could support a boxy feature . in the interval @xmath73e@xmath6 @xmath74 , o1 , o1@xmath0 are stable ,
x1 is ( radially ) simple unstable , x1v1/x1v1@xmath0 are complex unstable and x1v3 is simple unstable .
[ ot1fig]a shows a @xmath10 poincar section for e@xmath6 @xmath75 constructed by perturbing the initial conditions of o1 by @xmath8 .
we observe a main chaotic sea with two stability islands belonging to o1 @xmath76 and o1@xmath0 @xmath77 .
there is an increased density of consequents around the islands , indicating the existence of sticky zones .
we followed again the procedure of perturbing initial conditions on the @xmath10 plane by @xmath2 or @xmath3 .
we integrated them for time equal to 10 o1 periods .
we found that the quasi - periodic orbits trapped by either of the symmetric branches of o1 have morphologies similar to those of the p.o . when projected on the @xmath10 plane , but are thicker .
the orbits in fig .
[ ot1fig]b are the quasi - periodic orbits we obtain by perturbing o1 , o1@xmath0 by @xmath78 .
starting from the p.o . , or from initial conditions on an invariant curve around o1/o1@xmath0 , or even from a @xmath10 point belonging to a sticky zone , we find sticky orbits when exceeding a critical @xmath2 or @xmath3 perturbation .
typical examples are given in figs .
[ ot1fig]c and [ ot1fig]d . in fig .
[ ot1fig]c we plot 13 orbits with @xmath79 , while in fig .
[ ot1fig]d 10 orbits with @xmath80 .
most of these orbits are sticky to o1 and o1@xmath0 .
we observe that the boundary of these orbits has an oval shape .
combining sticky orbits we do not find symmetric morphologies in general , even if we perturb by equal @xmath2 or @xmath3 the initial conditions of the periodic orbits .
non - periodic orbits associated with the o1/o1@xmath0 family fail to provide clear boxy features in their face - on views like the orbits in figs .
[ ob41 ] , [ bb41stick ] and [ borbits ] . by comparing the phase space in the case of the latter orbits with that of the orbits in fig .
[ ot1fig ] , we observe that the main difference is the lack of the rotational tori around x1v1 and x1v1@xmath0 . when o1 is introduced in the system x1v1 and x1v1@xmath0
are already complex unstable .
this indicates that the main dynamical phenomenon for having boxiness in the middle of the bar is stickiness to these tori .
in other models in which 3:1 type families exist at the energies we have stable x1v1 and x1v1@xmath0 orbits they may play a role similar to that of rm2 and tr1 in figs .
[ xpx41 ] and [ ob41 ] .
summarizing we can say that as long as a non - periodic orbit in the radial and vertical ilr region of our model is trapped by , or becomes sticky to , x1v1 and/or x1v1@xmath0 tori , it may visit a region of phase space , where the orbits have rectangular - like face - on @xmath1 projections .
rectangularity in the morphology of non - periodic orbits is evidently related with the presence of 2d families of p.o . with boxy shapes in the equatorial plane .
the presence of families like rm21 , rm22 and tr1 in @xmath10 cross sections indicate the prevalence of boxy shapes in large regions of the phase space over a large energy interval .
inner boxiness is generated by perturbing initial conditions of planar orbits at energies where we find orbits with rectangular shapes on the equatorial plane and x1v1/x1v1@xmath0 tori in the phase space . as a result
the orbits participating in the inner boxiness have a double boxy character ( face - on and edge - on )
. among them those that reach larger heights above the equatorial plane tend to become more square .
boxy periodic orbits of higher multiplicity facilitate the prevalence of boxy morphologies , because they influence larger regions of the phase space . in a typical case
we have small stability islands with larger sticky zones around them .
sticky chaotic orbits visit phase space regions that are far from each other .
planar families of symmetric p.o
. introduced at odd radial n:1 resonances do not play an important role for the inner boxiness in our model .
however , they can be significant in models where they co - exist in the same energy intervals with the tori of x1v1 and x1v1@xmath0 . in fig .
[ x1etot ] we put together representative non - periodic orbits that support double boxy profiles , both from the 2:1 resonance region as well as from regions of higher resonances .
the outermost , thicker drawn , black orbit , discernible in figs .
[ x1etot]a and b , is a x1v6 orbit that we identify with the longest bar - supporting p.o . and
is used as a measure of the length of the bar in our model . the longest boxy orbit in fig .
[ x1etot]a , drawn with magenta colour is at e@xmath6 @xmath81 . in fig .
[ x1etot]b , which is an almost side - on view of the same set of orbits , we can see the same magenta coloured orbit remaining in much lower heights than the orbits that build the peanut .
its detailed shape , hidden by the other orbits in the central regions of the model , points to an association with x1v3 . by converting fig .
[ x1etot]a to an image and by applying a smoothing filter , we obtain fig . [ x1etot]c .
we can observe how inner and outer boxiness can be formed in our model by means of quasi - periodic and sticky orbits .
it has to be underlined that orbits at higher energies participating in the formation of the `` x - feature '' in the side - on views , especially by means of orbits associated with the x1v1 and x1v1@xmath82 families , contribute in general less to the surface density of the central parts of the face - on views of the bars and more to their periphery , as long as they retain a clear loop character .
the closer the `` x '' supporting non - periodic orbits remain trapped around the x1v1 and x1v1@xmath82 periodic orbits , the larger is the `` transparency '' towards the face - on rectangular - like periodic orbits at lower energies with smaller heights above the equatorial plane .
this becomes evident in fig .
[ trans ] by means of a set of quasi - periodic orbits with e@xmath6 @xmath83 .
two are trapped around x1v1 and x1v1@xmath82 ( coloured red ) , while the other two ( coloured black ) have boxy face - on projections and are from the @xmath55e@xmath56 energy interval . in ( a ) in @xcite
it is presented a list with 30 galaxies ( see their a1 , a2 figures ) with boxy isophotes in the middle of the bar .
these are moderately inclined galaxies , far from edge - on .
since the bar of our model does not contain asymmetric terms , it is aligned with the major axis of the system .
so the extent and the proportions of the inner and outer boxiness in the face - on views of the bar can be compared with rather symmetric galactic bars , in galaxies with inclinations @xmath8450@xmath85like the one of ngc 4037 or ngc 4123 ( see figures with isophotes in * ? ? ?
* ) . 1 .
the _ inner _ boxy isophotes in the middle of the bars in nearly face - on views of barred galaxies can be supported by 3d quasi - periodic and mainly by 3d chaotic orbits sticky to x1v1 and x1v1@xmath82 p.o . in the ilr region .
inner boxiness is also characterized by the presence of planar periodic orbits with boxy morphology at the appropriate energies . by perturbing vertically such p.o .
we generate a class of non - periodic orbits with double boxy character .
the maximum distance along the bar s major axis reached by these orbits is at or inside the limmits of the b / p bulge .
higher multiplicity p.o .
influence large regions of the phase space , mainly due to the presence of sticky zones around their stability islands .
orbits sticky to multi - periodic orbits have consequents in the sticky zones around different islands . this way a boxy structure can be supported by orbits with initial conditions in regions of the phase space that are far from each other .
2 . at large energies , orbits that reinforce `` x - features '' in the side - on views of b / p bulges by staying close to the periodic orbit ,
enhance locally boxy face - on features without obscuring the sight towards boxy structures formed at lower heights .
the latter are located closer to the centre of the galaxies in the @xmath1 projections .
boxiness at the outer parts of the bars is supported by orbits associated with families of higher resonances of the x1-tree , which remain very close to the equatorial plane .
contopoulos g. , 1983 , apj 275 , 511 contopoulos g. , 1988 , a&a 201 , 44 contopoulos g. , 2004 , `` order and chaos in dynamical astronomy '' , springer verlag , berlin - heidelberg contopoulos g. , harsoula m. , 2013 , mnras 436 , 1201 contopoulos g. , grosbl p. , 1989
, a&ar , 1,261 erwin p. , debattista v.p . , 2013 ,
mnras 431 , 3060 katsanikas m. , patsis p.a . , 2011 , int .
21 - 02 , 467 kaufmann d. , contopoulos g. , 1996 , a&a 309 , 381 ohta k. , hamabe m. , wakamatsu k. , 1990 , apj 357,71 patsis p.a . , 2006 , mnras 369l , 56 patsis p.a .
, katsanikas m. , 2014 , mnras ( in press ) patsis p.a . , athanassoula e. , quillen a.c . , 1997 ,
apj 483 , 731 2 patsis p.a . , skokos , ch . ,
athanassoula e. , 2003 , mnras 342 , 69 poincar h , 1899 , ` les methodes nouvelles de la mechanique celeste ' , vol iii , gauthier - villars , paris skokos ch .
, patsis p.a . ,
athanassoula e. , 2002a mnras 333 , 847 skokos ch . ,
patsis p.a .
, athanassoula e. , 2002b mnras 333 , 861 tsigaridi l. , patsis p.a . , 2013 ,
mnras 434 , 2922 vrahatis m. , isliker h. , bountis t.c . , 1997 , int .
7 , 2707 wozniak h. , 1994 , in `` ergodic concepts in stellar dynamics '' , lnp 430 , 264 , v. g. gurzadyan & d. pfenniger ( eds ) , springer - verlag , berlin heidelberg new york wozniak h. , pfenniger d. , 1999 , cel . | we study the dynamical mechanisms that reinforce the formation of boxy structures in the _ inner _ regions , roughly in the middle , of bars observed nearly _
face - on_. outer boxiness , at the ends of the bars , is usually associated with orbits at the inner , radial 4:1 resonance region and can be studied with 2d dynamics . however , in the middle of the bar dominate 3d orbits that give boxy / peanut bulges in the edge - on views of the models . in the present paper
we show that 3d quasi - periodic , as well as 3d chaotic orbits sticky to the x1v1 and x1v1@xmath0 tori , especially from the inner lindblad resonance ( ilr ) region , have boxy projections on the equatorial plane of the bar .
the majority of vertically perturbed 2d orbits , initially on the equatorial plane in the ilr resonance region , enhance boxy features in face - on bars .
orbits that build a bar by supporting sharp `` x '' features in their side - on views at energies _ beyond _ the ilr , may also have a double boxy character . if populated , the extent of the inner boxiness along the major axis is about the same with that of the peanut supporting orbits in the side - on views . at any rate
these orbits do not obscure the observation of the boxy orbits of the ilr region in the face - on views , as they contribute more to the surface density at the sides of the bar than to their central parts .
[ firstpage ] galaxies : kinematics and dynamics chaos diffusion structure |
the extended radio continuum source g106.3 + 2.7 was first classified as a supernova remnant ( snr ) by @xcite .
this snr was later described by @xcite in more detail .
they separated the comet shaped structure into a small head with high radio surface brightness and a large extended tail of low surface brightness .
a gap in the atomic hydrogen apparently correlated with the snr suggested a distance of about 12 kpc .
this would give the snr a length of 200 pc and they concluded that g106.3 + 2.7 is a supernova remnant in the late stage of its isothermal evolution . at the northern edge of the head
there is a small shell - like structure with a diameter of about 3.5. this source coincides with strong x - ray emission and the @xmath1ray source 3eg j2227 + 6122 .
it was interpreted as the bow shock nebula of a fast moving pulsar or a pulsar wind blown bubble @xcite .
these authors estimate the distance as @xmath2 3 kpc from x - ray absorption values . subsequently this pulsar , psr j2229 + 6114 , was detected in radio and x - rays @xcite .
it has a period of 51.6 ms and a rotational energy loss rate of @xmath3 erg / s . in this paper , by utilizing the data provided by the canadian galactic plane survey ( cgps ) , we analyze the possible relationship between the pulsar nebula and the supernova remnant .
since the number of known combined snrs is rather low , such an analysis would improve our conceptual understanding on the interaction of a pulsar nebula with its environment .
the observations discussed here form part of the cgps , which is described in detail by @xcite . in carrying out the research presented here ,
we have used those parts of the cgps database which derive from observations with the drao synthesis telescope @xcite and co observations from the five college radio astronomy observatory ( fcrao ) @xcite .
angular resolution varies as cosec(declination ) and therefore changes slowly across the final mosaics . before assembly into a mosaic , the data for the individual fields are carefully processed to remove artifacts and to obtain the highest dynamic range , using the routines described by @xcite .
accurate representation of all structures to the largest scales is assured by incorporating data from large single antennas with data from the synthesis telescope , after suitable filtering in the fourier domain .
continuum data are derived from the 408 mhz all - sky survey of @xcite , and from the 1.4 ghz effelsberg survey @xcite .
single antenna data are obtained from a survey of the cgps region made with the drao 26-m telescope @xcite . to improve the resolution of the 1420 mhz continuum observations we also included data taken from the nvss @xcite .
the nvss map was convolved to the resolution of our cgps data .
the resulting map was subtracted from the cgps map and the difference added to the original nvss map .
the final image has a resolution of 45 .
in fig . [ radio ] the radio image of g106.3 + 2.7 at 1420 mhz is shown . in the lower image
the overall structure of the source is clearly visible .
the major parts are a comet shaped structure with a diffuse extended tail of low radio surface brightness and a smaller more structured head with higher radio surface brightness .
the upper image is an enlargement of the latter .
the brightest feature is the `` boomerang '' to the north , described by @xcite .
the point - like source just to its south is most likely an extragalactic source @xcite .
the head itself shows two prominent features . there is a peak in the center , south of the extragalactic source and a shell - like structure on its southern boundary .
also visible is a diffuse bridge extending from the head to the pulsar nebula , suggesting an interaction of those structures .
the snr s radio continuum emission has a spectral index of @xmath4 ( s@xmath5 ) @xcite .
neither the tail nor the head show significant spectral fluctuations .
however , the pulsar nebula seems to have a flat radio spectrum @xcite .
we obtained a flux density of @xmath6 mjy by integration of our 1420 mhz data . together with a flux density of 80 mjy at 4850 mhz taken from @xcite we derive a spectral index of @xmath7 for the pulsar nebula .
this confirms the flat radio spectrum proposed by @xcite and supports its interpretation as a pulsar - powered synchrotron nebula .
[ polar ] displays a map of polarized intensity at 1420 mhz in greyscale together with white contours representing the total power radio continuum .
we find linear polarization in all parts of the object .
the tail has only a small number of polarized patches but these show high percentage polarization of up to 70 % .
the head is polarized except for an area between the pulsar nebula and the center of the head where we also find a peak in total power emission .
the head shows peak polarization of up to 70 % .
the most prominent polarization feature , however , is the pulsar nebula with an integrated percentage polarization of 40 % .
parts of the radio continuum bridge to the pulsar nebula appear to be highly polarized .
these high percentage polarizations indicate low depolarization between the snr and us .
the mean rotation measure towards the `` boomerang '' and the snr g106.3 + 2.7 is around @xmath8 rad / m@xmath9 .
this high rotation measure results in faraday rotation of @xmath10 at a frequency of 1420 mhz . since we find only low depolarization on the `` boomerang '' it requires a very small distance which makes this object a local phenomenon , probably within 2 kpc .
if it was further away , beam depolarization would occur because of superimposed regions of differing faraday rotation within the beam .
the large ob association cep ob2 , which is located at a distance of about 950 pc @xcite , lies in the direction of g106.3 + 2.7 and partly overlaps with the supernova remnant .
the members of this ob association ionize the material between them which leads to the depolarization of extended polarized emission from behind them at low radio frequencies .
this gives us another upper limit for the distance to g106.3 + 2.7 .
in fig . [ hichan ] we plot channel maps of our data cube representing an area around the head of g106.3 + 2.7 at low negative radial velocities . there is a shell - like structure around the eastern edge of the head best seen at -6.41 km / s but also quite prominent at -5.59 km / s and -7.23 km / s .
a shell of is wrapped around the `` boomerang '' , peaking at the same velocity as the shell around the head of the snr .
the coincidence in velocity clearly demonstrates that the snr and the pulsar nebula are associated .
this is the first time that an interaction between a pulsar nebula and cold interstellar material has been observed .
it looks like the pulsar is either pushing the neutral hydrogen outwards or ionizing cold atomic hydrogen in its vicinity with its pulsar wind and thus creating the shell - like structure .
the center velocity of -6.4 km / s puts the snr at a distance of 800 pc . at this distance
the snr has a length of 14 pc and a maximum width of 6 pc .
the pulsar nebula is 0.8 pc wide .
the small shell of atomic hydrogen contains about 0.4 m@xmath0 and the bigger shell around the head about 5 m@xmath0 .
this leads to densities of about 200 @xmath11 within both structures .
@xcite calculated the foreground column density to be @xmath12 @xmath13 from the spectral fit to their x - ray data . based on this value they estimated a distance of 3 kpc .
we integrated the foreground column density from our data using @xmath14 = 1.823\cdot 10^{18 } \int t_b(v ) dv$ ] @xcite with @xmath15 representing the excess brightness temperature and @xmath16 its radial velocity .
the result is @xmath17 @xmath13 .
this equation , however , is only valid if the neutral hydrogen in the foreground is optically thin .
comparison of the brightness temperature of two strong point like background sources in the vicinity of g106.3 + 2.7 ( @xmath18 , @xmath19 and @xmath20 , @xmath21 ) with their absorption profiles indicates that most of the foreground material is optically thick with opacities higher than 3 .
this easily explains the factor of 3 between the calculated foreground @xmath22 and the one derived from the x - ray data . in fig .
[ cochan ] we plot channel maps of our co data cube in the same area and at the same radial velocities as the data in fig .
[ hichan ] . at the same velocities where we found the eastern hydrogen envelope
there is a molecular envelope around the western edge of the head structure .
but in this case we do not see the same structure in all channels .
the location of the co envelope is to the west at -5.6 km / s but rotates to the northwest at -7.2 km / s . since around the eastern edge we find neutral hydrogen in all channels we can assume that the hydrogen completely surrounds this area .
the co , on the other hand , seems to be incomplete giving the expanding supernova remnant space to break through and form the smooth extended tail .
the narrow shell of molecular material contains about 40 m@xmath0 of molecular hydrogen assuming @xmath23 = 2.3\cdot 10^{20}\cdot t_{co}\cdot \delta v$ ] . here
@xmath24 is the column density of molecular hydrogen , @xmath25 [ k ] is the co brightness temperature and @xmath26 [ km / s ] is the velocity width of these structures .
[ bigbubble ] shows an image of the entire region surrounding g106.3 + 2.7 .
the head of the snr and its molecular and atomic envelope are visible in the lower left part of the image as is the small shell surrounding the pulsar wind .
apparently these structures represent only a small part of the entire cold environment .
there is a large bubble centered at @xmath27 and @xmath28 , with an inner diameter of about @xmath29 which translates to about 14 pc at a distance of 800 pc .
the outer diameter is about 20 pc . at the northern and eastern edge of the bubble
we find patches of molecular material .
the supernova remnant has apparently created a little bubble at the lower left edge of the large bubble with its shock wave .
we suggest that a first generation of stars was formed in the center of the big bubble which at that time was a cloudy dense medium . with stellar wind effects and/or supernova explosions
the large bubble was formed .
however , some very dense cloudlets remained because they were too dense and probably also too far away from the location of the exciting stars .
the fact that the remaining patches we observe are located at the outer edges of the bubble supports this , as does an structure surrounding the northeastern patch of molecular material , which is most likely the remains of dissociated molecular hydrogen at the surface of this cloud .
stellar winds and supernova explosions within the large bubble apparently also triggered the creation of the progenitor star of our snr g106.3 + 2.7 . at the time the progenitor star of g106.3 + 2.7 was formed , the environment most likely looked very much like the two northern patches of molecular material .
the new born star was located inside this molecular environment .
it was a massive star , probably an early b or an o - type star .
the small bubble in the cold environment in which we now find the head of the snr was either formed by the progenitor star s stellar wind or by the expanding shockwave of the supernova explosion itself .
usually pulsar nebulae are expanding inside a bubble created by the shock wave of a supernova explosion . in this case
there is no material left to interact with , because all the matter has been swept up and carried away already .
the pulsar in g106.3 + 2.7 is apparently interacting with its dense cold environment . in the following
we investigate two possible scenarios leading to the shell - like shape of the `` boomerang '' pulsar nebula .
since the shape of the object bears a striking resemblance to a bow - shock structure , we first investigate the constraints which are imposed by the observations in order for this hypothesis to be viable . for simplicity , we assume that the explosion which caused the snr took place near the center of the head component and resulted in the candidate pulsar getting a kick velocity which took it to its present position near the bow structure . in view of the observed morphology , we also suppose the motion to be nearly in the plane of the sky and neglect projection effects .
the distance travelled , @xmath30 , is comparable to the radius of the head component @xmath31 , so we set @xmath32 .
letting @xmath33 represent the pulsar travel time and snr age in years , @xmath34 the pulsar space velocity in km / s , and @xmath31 in pc , we have @xmath35 the equality between the pulsar relativistic wind pressure and ism ram pressure implies that @xmath36 here @xmath37 is the ambient number density in @xmath11 , @xmath38 is the distance between the pulsar and the apex of the bow shock in pc , and @xmath39 is the pulsar rotational energy loss rate in @xmath40 erg / s .
snr dynamics give us : @xmath41 1.4 \left ( \frac{\epsilon e_{51}}{\mu n_o}\right ) ^{5/21 } t^{2/7 } & \hbox{for isothermal expansion}\\ \end{array}\right.\ ] ] here the parameter @xmath42 is the ratio of thermal energy in the snr to the initial explosion energy @xmath43 [ @xmath44 erg ] .
the pulsar distance from the center of the head is about 1.5 pc , its distance from the apex of the bowshock is about 0.4 pc and its rotational energy loss rate @xmath45 erg / s @xcite . with an estimated value of 0.3 for @xmath42
we get : @xmath46 1.2\cdot 10^{-6 } v_p^{0.2 } & \hbox{isothermal case } \end{array}\right.\ ] ] any reasonable value for the pulsar velocity ( 100 km / s @xmath47 1000 km / s ) will give us by far the lowest ever recorded explosion energy for the supernova explosion . allowing for inclination effects or substantial variations in the observed parameters
still leaves us with highly unlikely values of the energy parameters . as an alternative ,
@xcite have suggested that the bow - shaped structure can be confined by the thermal pressure of the surrounding medium .
this is indeed possible , but it implies an anomalously low velocity for an object belonging to a high - velocity class . the fact that the spectral index is flat makes it almost unavoidable that the emission originates from the relativistic pulsar wind itself .
however we believe that the confinement is simply achieved because the relativistic wind is encountering the dense snr shell in the northeast . in our picture
the sn explosion took place at or very near the present position of the pulsar . in the northeast
the shock has been quickly decelerated by its encounter with a relatively nearby and particularly high density medium , whereas towards the south it has expanded into a moderately dense medium ( snr head component ) while breaking out in the southwest into a region of much lower density and giving rise to the tail component .
the morphology of the head and the tail structure supports this hypothesis .
the head shows only the `` boomerang '' and the southern shell as indications for the encounter of a shockwave with dense material . we do not find a shell structure towards the east of the head but we find an shell all around the east and the south of the head .
we can explain this by moving the location of the explosion to the current position of the pulsar .
to the north and to the east the shockwave is running into dense material .
the first free path for the shockwave is to the southeast where the southern shell begins .
another point for this scenario is that the distance between the location of the supernova and the shell to the south is now larger that its distance to the beginning of the tail , which should be expected in our picture .
if the explosion occured in the center of the head this would not be the case .
what seems peculiar however is the fact that the little bow - shaped shell in the north does not show enhanced non - thermal radio emission .
this would appear to go against the expected view that an snr should be brighter where the blast wave is moving against dense regions of the ism .
a similar situation has been encountered by @xcite in their analysis of the cta1 snr .
these authors argued that , in overdense regions where the shock velocity is considerably reduced , the fermi acceleration mechanism @xcite might either be too slow or simply break down .
the time efficiency can be quantified by evaluating the timescale @xmath48 for acceleration of an electron to energy @xmath49 , given by @xmath50 @xcite .
hence either a low shock velocity or low ambient magnetic field , or a combination of both may imply that the electrons radiating at a given observing frequency @xmath51 , where @xmath52 , are not yet present in the energy distribution .
an alternative approach is to consider the growth and decay rates of the alfvn waves responsible for the repeated scattering of the relativistic particles back and forth across the shock @xcite . on one hand ,
the growth rate is proportional to the shock velocity and thus decreases as the shock slows down and , on the other hand , the decay rate depends on the first power of the number density of neutral atoms .
again this means that the acceleration mechanism may be quenched whenever the shock wave encounters sufficiently dense clouds .
it is also possible that a thin steep ( spectral index of order -0.5 ) non - thermal layer is present but not detectable because of beam smearing .
the interpretation of the tail component as resulting from a breakout of the snr shock wave is also compatible with theoretical predictions .
the numerical studies of breakout by @xcite have shown that the asymptotic ratio of the velocity of the material breaking out into a region of lower density to that of the shock wave still propagating into the original medium could be expressed as @xmath53 where @xmath54 is the ratio of the densities .
the same ratio at any normalized time @xmath55 , where @xmath56 is the time at which breakout occurred , was shown by @xcite to be satisfactorily represented by the approximate analytical relation @xmath57 it follows from this relation that the ratio @xmath58 where @xmath59 is the length of the tail component , as long as @xmath60 .
using @xmath61 pc and @xmath62 pc gives us a value of 4 for @xmath63 .
the length of the tail is difficult to determine since there is no sharp outer boundary .
the observed value implies @xmath64 .
a value of 8 for @xmath63 would correspond to @xmath65 .
it can also be shown that the breaking out material attains a maximum velocity @xmath66\ ] ] at @xmath67 , where @xmath68 is the velocity just before breakout . for @xmath69 ,
the maximum velocity is about @xmath70 . assuming that the age of the snr is comparable to the characteristic age of the pulsar we take @xmath71 yr @xcite .
for an adiabatic expanding snr we derive @xmath72 . with a density of @xmath73 @xmath11
the explosion energy would be @xmath74 erg .
usually the characteristic age of a pulsar is higher than its real age which means the calculated explosion energy is a lower limit . from the length of the tail ,
an average velocity of @xmath75 is inferred .
equating this to the maximum velocity taken as @xmath76 gives @xmath77 yr , corresponding to @xmath78 pc .
although these numbers should not be taken at face value , they nevertheless indicate that the breakout hypothesis is consistent with the observed morphology .
using the data from the cgps we studied the morphology and the environment of the snr g106.3 + 2.7 and the pulsar nebula associated with the pulsar j2229 + 6114 ( the `` boomerang '' ) .
we conclude that these sources are the result of the same sn event .
the snr morphologically consists of two parts giving the object its cometary shape , in which the tail is due to an outbreak into the interior of a big bubble , whereas the head is created by the expanding shock wave interacting with dense ambient material .
the kinematics of the associated neutral hydrogen and molecular material , @xmath79 km / s , suggests that both objects are located at the same distance of 800 pc , thereby the snr is 14 pc long and 6 pc wide and the pulsar nebula has a diameter of 0.8 pc .
this close distance is also supported by the presence of high polarized emission . by studying the energetics of the snr we found that the `` boomerang '' nebula is created by the relativistic wind of the pulsar . for the `` boomerang '' we found a spectral index of @xmath80 and @xmath81 40 % integrated percentage polarization at 1420 mhz .
the analysis of the cold environment suggests that the creation of the progenitor star was triggered by stellar wind and/or sn explosions of a group of stars which also created the bubble . in this way
this is the first detection of a pulsar nebula apparently interacting with its cold environment .
it is also an example of a pulsar `` displaced '' from the center of its associated snr . in this case , however , the displaced position of the pulsar is not due to high pulsar velocity .
landecker t.l . ,
dewdney p.e . ,
burgess t.a . ,
gray a.d . ,
higgs l.a . ,
hoffmann a.p . ,
hovey g.j . ,
karpa d.r .
, lacey j.d . , prowse n. , purton c.r . , roger r.s .
, willis a.g . , wyslouzil w. , routledge d. , vaneldik j.f . , 2000 ,
a&as 145 , 509 taylor a.r . ,
dewdney p.e . ,
landecker t.l . ,
martin p.g . ,
brunt c. , dougherty s.m . , durand d. , gibson s.j . , gray a.d . ,
higgs l.a . ,
kerton c.r . ,
knee l.b.g . , kothes r. , peracaula m. , purton c.r . , uyaniker b. , wallace b.j . ,
willis a.g . , 2001 ,
aj , submitted | we propose that the pulsar nebula associated with the pulsar j2229 + 6114 and the supernova remnant ( snr ) g106.3 + 2.7 are the result of the same supernova explosion .
the whole structure is located at the edge of an bubble with extended regions of molecular gas inside .
the radial velocities of both the atomic hydrogen and the molecular material suggest a distance of 800 pc . at this distance
the snr is 14 pc long and 6 pc wide .
apparently the bubble was created by the stellar wind and supernova explosions of a group of stars in its center which also triggered the formation of the progenitor star of g106.3 + 2.7 .
the progenitor star exploded at or close to the current position of the pulsar , which is at one end of the snr rather than at its center .
the expanding shock wave of the supernova explosion created a comet shaped supernova remnant by running into dense material and then breaking out into the inner part of the bubble .
a synchrotron nebula with a shell - like structure ( the `` boomerang '' ) of length 0.8 pc was created by the pulsar wind interacting with the dense ambient medium .
the expanding shock wave created an shell of mass 0.4 m@xmath0 around this nebula by ionizing the atomic hydrogen in its vicinity . |
neutron transmutation doped ( ntd ) ge sensors are commonly used as low temperature sensors in the rare decay events like neutrinoless double beta decay @xcite and dark matter search @xcite .
thermal neutron capture by ge atom leads to the production of dopants @xcite like ga , as and se .
this doping method is preferred over the melt growth method due to its homogeneity @xcite .
the temperature dependence of the resistivity of a ntd ge sensor is given by @xcite -0.6 cm @xmath3 where the parameters @xmath4 depends on intrinsic properties of ge , t@xmath5 depends on the doping level and the constant @xmath6 is close to 0.5 @xcite .
it is important to achieve a high sensitivity ( dr / dt ) keeping resistance r of the sensor within the measurable range at mk temperature .
the desired neutron fluence will depend on the actual isotopic composition of the ge wafer .
since the reactor neutrons are not monoenergetic , samples are exposed to neutrons having a broad range of energies ( @xmath7 0.625 ev to 10 mev ) during the irradiation . in particular , high energy neutrons can cause damage to the lattice structure of the crystal . even for thermal neutrons ,
the average recoil energy ( @xmath8 182 ev ) of ge(n,@xmath9 ) reaction is sufficiently high to drive the recoiling ge atom into the interstitial position @xcite . the dopant atom in interstitial positions can not contribute to the total number of carrier concentration @xcite .
thus , defects created during the irradiation can significantly affect the performance of the sensor at low temperature .
it should be mentioned that defect formation may depend on irradiation conditions , since both the flux and energy of fast neutrons are responsible factors . the performance of ntd ge sensors over a wide range of carrier concentration 3 x 10@xmath10 to 1.8 x 10@xmath11 was studied in detail by itoh _ et al . _
generally high temperature annealing is employed to cure the crystal defects and annealing conditions need to be optimized depending on the defect type and the density .
et al . _
@xcite have reported that 30 min annealing in the ar gas at 600@xmath1 c for a hole concentration of 3.5 x 10@xmath12 was sufficient to recover damages , while pasca _ et al . _
@xcite have employed annealing at 400@xmath1 c for 6 hours for carrier concentration in the range of 5 6.8 x 10@xmath13 .
palaio _ et al . _
@xcite have also discussed effects of thermal annealing in the ar gas and have reported that annealing at higher temperatures ( 500@xmath1 c ) resulted in contamination of samples due to rapidly diffusing impurities .
they further observed that the ntd ge samples left at ambient temperature for eight months after irradiation showed improvements similar to thermal annealing .
this is desirable for rare decay studies as annealing can introduce trace impurities in the sensors .
development of ntd ge sensors for mk thermometry in cryogenic bolometer detector tin.tin ( the india - based tin detector ) to search for neutrinoless double beta decay in @xmath14sn has been initiated @xcite .
ntd ge sensors are prepared , by neutron irradiation of device grade ge samples at dhruva reactor , bhabha atomic research centre ( barc ) , mumbai , for the first time .
initial results for a ntd ge sensor performance at 100 mk have been reported in ref .
a systematic study of the fast neutron induced damage in the ntd ge samples for neutron fluence in the range of 2 x @xmath15 to 1.4 x @xmath16 which corresponds to the carrier concentration of ( 0.5 - 3.3 ) x 10@xmath11 is presented in this paper .
the neutron induced damage is investigated using positron annihilation lifetime spectroscopy ( pals ) and channeling in irradiated samples before and after annealing . in this paper section
ii describes the experimental details , data analysis and results are presented in section iii .
conclusions are given in section iv .
for sensor development two sets of semiconductor grade ge samples from university wafers with @xmath17 ( 0.4 mm thick , @xmath18 , 5n purity , n - type ) and @xmath19 ( 1 mm thick , @xmath20 , 5n purity , n - type ) cleavage plane were used .
the ge samples were single side mirror polished .
the cleavage plane @xmath19 is preferred over the @xmath17 in terms of clean cutting of smaller sensors from the irradiated sample .
actual isotopic abundances of both @xmath19 and @xmath17 samples were obtained with the secondary ion mass spectrometry ( sims ) @xcite .
it was observed that @xmath21ge and @xmath22ge , which produce @xmath23ga and @xmath24as , respectively , are fractionally more in @xmath19 compared to that in @xmath17 and hence , the net carrier concentration for a given neutron flunece is expected to be similar .
the production rate of dopants are estimated from the measured isotopic abundance and n - capture cross section taken from ref @xcite .
these estimated production rates per unit neutron fluence and the expected net carrier concentration values are tabulated in table [ table5 ] .
the scanning electron microscope ( sem ) and energy dispersive x - ray analysis ( edax ) studies of the ge wafer prior to irradiation showed no surface impurities other than traces of oxygen .
the neutron irradiation was carried out at dhruva reactor , barc , mumbai ( india ) for samples of different sizes ( 10 x 30 mm@xmath25 or 10 x 10 mm@xmath25 ) in different batches . required neutron fluence was estimated based on measured isotopic abundances but actual neutron fluence varied depending on the reactor power .
typically , samples were irradiated for 4 - 6 days and the estimated thermal as well as fast neutron fluence are given in table [ table1 ] .
a minimum cooldown period of 45 days was required before the irradiated samples could be handled for measurements .
although the samples were cleaned with electronic grade alcohol prior to irradiation ( as described in ref .
@xcite ) , the irradiated samples showed lack of lustre . hence
, the irradiated samples were etched with hf ( 40@xmath26 ) to clean the surface .
it should be mentioned that all measurements have been carried out 4 to 8 months after the irradiation .
[ table5 ] . estimated production rate of dopants ( @xmath27 ) per neutron , obtained by multiplying measured isotopic abundance of ge @xcite with n - capture cross section ( taken from ref .
@xcite ) . net carrier concentration ( p - type ) = @xmath23ga - ( @xmath24as + 2 @xmath28 @xmath29se ) is given in the last column .
[ cols="^,^,^,^,^ " , ] the estimated defect densities for samples b and d , where measurable change in @xmath30 was observed for unannealed samples with respect to virgin samples are tabulated in table [ table4 ] . the defect concentration from the channeling experiment
is estimated to be 1.6 @xmath26 and is consistent with pals measurement ( @xmath31 0.1 @xmath26 ) .
it may be added that the estimated @xmath32 values for 50 kev and 100 kev energy window below the surface peak defined depth are similar within error , indicating no significant dechanneling effects in near surface region .
the development of ntd ge sensors has been initiated for mk thermometry in tin.tin detector . ntd ge sensors are prepared , by thermal neutron irradiation of device grade ge samples at dhruva reactor , barc , mumbai , for the first time .
the carrier concentration in irradiated , annealed samples was estimated from the hall effect measurement at 77 k and is found to be in good agreement with that estimated from the neutron fluence .
fast neutron induced defects have been studied using pals and channeling .
the pals measurements indicated monovacant type defects for all irradiated samples , whereas the channeling studies revealed no significant defects for fast neutron fluence upto @xmath33x 10@xmath34 .
these fast neutron induced defects were observed in the irradiated samples even after about 4 - 8 months of irradiation .
thus unlike in ref .
@xcite , for neutron fluences @xmath31 10@xmath35 studied in the present work leaving the samples at ambient temperature is not sufficient .
it is shown that vacuum annealing at 600 @xmath1c for @xmath2 hours is sufficient to recover the damage in the irradiated samples , as verified from both pals and channeling .
the annealed samples with carrier concentration in the range @xmath36 x 10@xmath11 are found to be suitable for the sensor development .
the sensor made from the annealed ntd ge sample was tested in the temperature range 75 - 250 mk and the results are reported in the ref .
we thank ms . s. mishra for help with sample preparation ; mr .
j. mathew , mr .
a. singh , dr . s.s .
prabhu , prof .
m. deshmukh for assistance with electrical measurements and ms .
chalke for sem measurements .
we are grateful to dr .
datar and dr .
g. ravikumar for useful discussions . c. arnaboldi et al .
, nuclear instruments and methods in physics research section a , @xmath37 ( 2004 ) 775 .
e. pasca , e. oliveri , g. ventura , m. barucci , l. risegari , j. beeman , proc . of 8th int .
conf . adv .
technol . part .
phys . , @xmath38 ( 2004 ) 93 .
o. martineau et al .
, nuclear instruments and methods in physics research section a , @xmath39 ( 2004 ) 426 . e.e .
haller , infrared phys . , @xmath40 ( 1985 ) 257 .
ltoh , e.e .
haller , w.l .
hansen , j.w .
beeman , j.w .
farmer , a. rudnev , a. tikhomirov , v.i .
ozhogin , appl .
lett . , @xmath41 ( 1994 ) 2121 .
woodcraft , r.v .
sudiwala , e. wakui , c. paine , journal of low temperature physics , @xmath42 ( 2004 ) 925 .
schweinler , journal of applied physics , @xmath43 ( 1959 ) 1125 .
hasiguti , defects and radiations effects in semiconductors , ( 1980 ) 46 .
ltoh , w.l .
hansen , j.w .
beeman , e.e .
haller , j.w .
farmer , v.i .
ozhogin , journal of low temperature physics , @xmath44 ( 1993 ) 307 .
k. kuriyama , m. yahagi , k. iwamura , y. kim , c. kim , journal of applied physics , @xmath45 ( 1983 ) 673 .
palaio , s. j. pearton , e. e. haller , journal of applied physics , @xmath46 ( 1984 ) 1437 , doi : 10.1063/1.333397 . v. nanal , epj web of conferences , @xmath47 ( 2014 ) 08005 .
s. mathimalar , v. singh , n. dokania , v. nanal , r.g .
pillay , s. pal , s. ramakrishnan , a. shrivastava , priya maheshwari , p.k .
pujari , s. ojha , d. kanjilal , k.c .
jagadeesan , s.v .
thakare , ieee wolte conference , ( 2014 ) 10.1109/wolte.2014.6881014 .
s. mathimalar , n. dokania , v. singh , v. nanal , r. g. pillay , a. shrivastava , k.c .
jagadeesan , s.v .
thakare , nuclear instruments and methods in physics research section a , doi : 10.1016/j.nima.2014.11.056 .
firestone , v.s .
shirley , eds . , table of isotopes , eighth edition ( wiley - lnterscience , new york , 1999 ) .
f. j. morin , physical review , @xmath48 ( 1954 ) 62 .
rehberg , h.s .
leipner , positron annihilation in semiconductors - defect studies , springer publication ( 1999 ) .
m. satoh , k. kuriyama , journal of applied physics , @xmath49 ( 1990 ) 3890 . | development of ntd ge sensors has been initiated for low temperature ( mk ) thermometry in the india - based tin detector ( tin.tin ) . ntd ge sensors are prepared by thermal neutron irradiation of device grade ge samples at dhruva reactor , barc , mumbai .
detailed measurements have been carried out in irradiated samples for estimating the carrier concentration and fast neutron induced defects .
the positron annihilation lifetime spectroscopy ( pals ) measurements indicated monovacancy type defects for all irradiated samples , while channeling studies employing rbs with 2 mev alpha particles , revealed no significant defects in the samples exposed to fast neutron fluence of @xmath0 .
both pals and channeling studies have shown that vacuum annealing at 600 @xmath1c for @xmath2 hours is sufficient to recover the damage in the irradiated samples , thereby making them suitable for the sensor development .
ntd ge , pals , channeling 61.72.uj , 34.80.-i , 61.85.+p , 72.20.my , 72.20.jv |
new - generation spectrographs , at the major ground - based telescopes , have begun to pour in the hard - disks of astronomers computers an increasing mass of high - quality spectroscopic data .
inverse resolutions as high as @xmath3 can now be easily attained , at least for the brightest ( @xmath4 mag ) objects in the sky , and this is pushing the observation of local and extragalactic stellar systems to an ever unimagined resolution level .
the outstanding performance of instruments like uves and virmos at the eso very large telescope , or hires and sarg at the keck observatory and telescopio nazionale galileo , respectively , urges therefore theoretical tools of comparable accuracy level in order to consistently match and analyse such a huge amount of observational data . in this framework , and to help filling the gap , we undertook a long - term project aimed at providing the community with a systematic theoretical library of high - resolution stellar spectra ( virtually the largest sample currently available in the literature ) in the optical range ( @xmath5 ) and at an inverse resolution of @xmath6 .
this first set of spectral energy distributions relied on the synthe code of kurucz ( 1993 ) , using atlas9 model atmospheres ( kurucz 1995 ) as input .
computations have been carried out at the brera observatory in milan ( italy ) .
the input models follow the classical approximations of steady - state , homogeneous , lte , plane - parallel layers , with a microturbulence velocity of 2 km / s and a mixing - length value @xmath7 . they make use of an improved treatment of convective overshooting in the calculation of the transfer equations ( castelli , gratton , & kurucz 1997 ) .
over 46 million absorption lines are accounted for in the code , including all atomic elements at different ionization states and the most important diatomic molecules .
all line data were extracted from the kurucz ( 1992 ) original database , with the major improvement of the tio contribution , for which we adopted the new list of lines computed by schwenke ( 1998 ) .
the wavelength interval spans the whole optical range , from the balmer break to h@xmath8 .
all the passbands of the lick / ids spectrophotometric indices ( worthey et al .
1994 ) are included , as well as the high - resolution index set defined by rose ( 1994 ) .
spectra are sampled at variable wavelength step maintaining a constant resolution ( @xmath9 ) .
the library extends to effective temperatures as hot as @xmath10 k , and provides a suitable match to o @xmath11 k spectral types .
m stars , cooler than @xmath12 k , are however still lacking in our dataset given the missing contribution of tri - atomic molecular opacity in the input model atmospheres ( kurucz 1992 ) .
gravity spans a wide range across the h - r diagram ( @xmath13 ) , namely from dwarf ( mk class v ) to supergiant stars ( mk class i ) , while metallicity accounts for stars at the two extreme edges of the galaxy [ fe / h ] distribution with @xmath14 }
< + 0.3 $ ] as boundary limits .
a total of 832 synthetic spectral energy distributions ( seds ) have been computed , as summarized in fig .
[ fig : bertonee1 ] and table [ tab : grid ] .
an illustrative example of the model output , sampling the spectral emission around the atomic mg@xmath15 triplet , is shown in fig .
[ fig : bertonee2 ] and [ fig : bertonee3 ] , with varying physical parameters . a complementary project , which will extend our analysis to the ultraviolet interval , between 8504750 with an inverse resolution @xmath16 , is in progress at inaoe and will be completed soon ( rodrguez - merino 2002 ) .
the extended dataset consists of over 1000 seds covering the effective temperature range between 3500 and 50000 k at surface gravity @xmath2 = 1.0 @xmath11 5.0 and chemical compositions @xmath17 }
= + 0.5,\ , 0.0 , -0.5 $ ] and @xmath18 ( rodrguez - merino et al .
2001 ) . in fig .
[ fig : bertonee4 ] we display a subset of spectra computed for solar chemical composition , @xmath19 and different effective temperatures .
ll code : & synthe ( kurucz 1993 ) + input models : & atlas9 ( kurucz 1995 ) + wavelength range : & @xmath20 + resolution : & @xmath21 + wavelength step : & @xmath22 + wavelength points : & 346645 + total absorption lines : & @xmath23 millions + molecules : & c@xmath24 , cn , co , ch , nh , oh , mgh , sih , + & h@xmath24 , sio , tio + total number of spectra : & 832 + effective temperature : & @xmath25 k + surface gravity : & @xmath26 + metallicity : & @xmath27 } \ ; \leq \ ; + 0.3 $ ] +
a comparison with the solar spectrum represents a first mandatory test in order to assess the degree of reliability of our theoretical library .
we present in fig .
[ fig : bertonee5 ] the match of the solar flux atlas by kurucz et al .
( 1984 ) and the corresponding synthetic solar spectrum , which we computed at similar resolution ( i.e. @xmath28 ) .
the input model atmosphere is from the atlas9 and assumes @xmath29 k , @xmath30 , a microturbulent velocity of 1.0 km / s , and chemical abundance according to anders & grevesse ( 1989 ) .
the figure is made up by a dense sequence of positive and negative spikes , which mainly track the residual flux differences in the absorption line profiles .
while positive spikes mainly originate from observed lines that are not present in the synthetic model , the negative peaks come from theoretical absorption lines that are too weak or undetected in the observed spectrum .
note , as a stricking feature in the plot , that both positive and negative spikes tend to decrease at longer wavelength , and no systematic drift is present in the data distribution .
[ fig : bertonee5 ] gives also a direct measure of the intrinsic uncertainty in the input physics of the models ; excluding the telluric bands , clearly affecting the residual distribution about 6300 and 6900 , the rms value in the plot is @xmath31 .
anders , e. , & grevesse , n. 1989 , geochim .
acta , 53 , 197 castelli , f. , gratton , r. g. , & kurucz , r. l. 1997 , , 318 , 841 kurucz , r. l. 1992 , in iau symp .
149 , the stellar populations of galaxies , ed .
b. barbuy & a. renzini , ( dordrecht : kluwer ) , 225 kurucz , r. l. 1993 , cd - rom no .
18 , synthe spectrum synthesis programs and line data kurucz , r. l. 1995 , cd - rom no . 13 , atlas9 stellar atmosphere programs and 2 km / s grid , revised kurucz , r. l. , furenlid , i. , brault , l. , & testerman , l. 1984 , national solar observatory atlas , sunspot , new mexico : national solar observatory rodrguez - merino , l.h .
2002 , ph .
d. thesis , in preparation rodrguez - merino , l. h. , chvez , m. , buzzoni , a. , bertone , e. 2001 , in proc . of new quests in stellar astrophysics : a link between stars and cosmology , ed .
m. chavez , a. bressan , a. buzzoni & d. mayya , ( dordrecht : kluwer ) , 39 rose , j. a. 1994 , , 107 , 206 schwenke , d. w. 1998 , faraday discuss .
, 109 , 321 worthey , g. , faber , s. m. , gonzles , j. j. , & burnstein , d. 1994 , , 94 , 687 | we present an extended theoretical library of over 800 synthetic stellar spectra , covering energy distribution in the optical range ( @xmath0 ) , at inverse resolution r=500000 .
the library , based on the atlas9 model atmospheres , has been computed with the synthe code developed by r. l. kurucz .
the grid spans a large volume in the fundamental parameters space ( i.e. @xmath1 , @xmath2 , [ m / h ] ) , and can be profitably applied to different research fields dealing both with the study of single stars and stellar aggregates , through population synthesis models . a complementary project , in progress ,
will extend the wavelength range to the ultraviolet , down to 850 , at an inverse resolution of r=50000 . |
the phonon concept is a useful and simple concept in nuclear structure physics @xcite .
the low - lying collective @xmath10 and @xmath11 excitations in near - spherical nuclei can be considered as quadrupole and octupole vibrations , which represent the most important vibrational degrees of freedom at low energies .
the bosonic phonon concept suggests the occurrence of multiphonon states , which can decay collectively to states with one phonon less by the annihilation of one phonon . in a harmonic picture
@xmath12-phonon states have excitation energies of @xmath12 times the one - phonon energy .
two of the appealing aspects of the study of multiphonon excitations high above the yrast - line is the investigation of their anharmonicities reflecting the influence of the pauli principle and higher order residual interactions as well as the interaction of multiphonon modes with single particle degrees of freedom . certainly these interactions can cause possible fragmentation of the idealized multiphonon modes .
this fragmentation or the eventual observation of almost unfragmented multiphonon states can tell how well the considered phonon is an eigenmode of the nuclear dynamical system . to judge
the deviation from the idealized multiphonon picture in reality obviously needs the formulation of this idealized situation and the investigation in how far the collective picture can account for the data .
multiphonon and two - phonon states are at present a very actively investigated topic in nuclear structure physics . a two - quadrupole - phonon ( @xmath13 ) triplet with states having spin and parity quantum numbers @xmath14 , @xmath15 , @xmath16
is usually known in near spherical nuclei .
the absence of @xmath17 members of this two - phonon multiplet points at the bosonic character of the identical quadrupole phonons involved .
multi - quadrupole - phonon states ( with @xmath18 ) have also been identified , see e.g. @xcite .
the investigation and identification of double giant dipole resonances ( 2gdr ) , ( @xmath19 ) double - octupole states and ( @xmath20 ) double - gamma - vibrations in deformed nuclei are actively debated in the literature , see e.g. refs . @xcite .
the mixed - multipolarity ( @xmath21 ) quadrupole - octupole coupled quintuplet @xcite with spin and parity quantum numbers @xmath22 @xmath23 is a very interesting example of the coupling of two different phonons .
members of the ( @xmath21 ) two - phonon quintuplet should decay by definition by collective @xmath1 and @xmath0 transitions to the @xmath24 octupole phonon state and to the @xmath4 quadrupole phonon state , respectively .
the ( @xmath21 ) two - phonon states are expected to occur close to the sum energy @xmath25 .
there are only a few examples , e.g. @xmath26sm @xcite , @xmath26nd @xcite , @xmath27ce @xcite and @xmath28cd @xcite , where besides the two - phonon @xmath29 state other multiplet members have been identified experimentally
. the reason might be that all the other states from the @xmath21 quintuplet , except the @xmath29 , are strongly mixed with non - collective excitations , losing their pure quadrupole octupole nature .
this is supported by shell - model calculations performed recently for @xmath30 nuclei @xcite .
thus , the most complete data , which can be used as a testing ground for the concept of quadrupole - octupole coupled two - phonon excitations , are the transition strengths from the mostly unfragmented @xmath29 member of the quintuplet .
the @xmath22 two - phonon state has been very well investigated in magic and near spherical nuclei . in many nuclei
, it has been identified close to the sum energy @xmath31 , indicating a quite harmonic phonon coupling @xcite . according to its definition ,
the two - phonon @xmath29 state should decay by a collective one - phonon @xmath1 transition to the @xmath11 octupole phonon state .
this has been confirmed experimentally @xcite .
the collective @xmath32 @xmath0 transition has not yet been measured due to the possibly dominating @xmath2 and @xmath33 contaminations in this transition .
a particularly interesting feature of the two - phonon @xmath29 state is the existence of a relatively strong @xmath2 transition to the ground state , which can be sensitively measured in photon scattering experiments @xcite .
the strength of this two - phonon annihilating @xmath2 transition approximately equals the strength of the @xmath34 two - phonon changing @xmath2 transition @xcite which strongly supports quadrupole - octupole - collective origin of these @xmath2 transitions .
all the low - energy one - phonon and two - phonon states mentioned above involve isoscalar phonons .
recently the isovector quadrupole excitation in the valence shell , the @xmath35 state , has been identified @xcite in some nuclei from the measurement of absolute transition strengths , see , e.g. , @xcite .
isovector excitations in the valence shell form a whole class of low - lying collective states called , more precisely , mixed - symmetry states @xcite .
the fundamental @xmath35 state decays by a weakly collective @xmath1 transition to the ground state and by a strong @xmath3 transition to the isoscalar @xmath4 state , with an @xmath3 matrix element of the order of one nuclear magneton @xcite . in an harmonic phonon coupling scheme one can expect also the existence of mixed - symmetry two - phonon multiplets , that involve at least one excitation of the mixed - symmetry quadrupole phonon .
two symmetric mixed - symmetric ( @xmath36 ) two - quadrupole phonon excitations with positive parity , namely a @xmath37 state ( the scissors mode ) , and a @xmath38 state have been identified already in the near spherical nucleus @xmath39mo @xcite .
similarly one can expect the existence of a ( @xmath40 ) mixed - symmetry quadrupole - octupole coupled multiplet with negative parity . in a naive phonon coupling picture members of this quintuplet should show a very complex but , nevertheless , simple to understand decay pattern , which can be predicted from the decay properties of the one - phonon excitations ( weakly - collective @xmath1 decay and strong @xmath3 decay of the @xmath35 state , and collective @xmath0 and relatively strong @xmath2 decay of the @xmath24 octupole vibration ) .
for instance , the @xmath41 state should decay by relatively strong @xmath2 , strong @xmath3 , weakly - collective @xmath1 , and collective @xmath0 transitions to the ground and to the @xmath42 scissors state , to the ordinary , isoscalar ( @xmath21 ) multiplet , and to the one - phonon @xmath11 and @xmath35 states , respectively ( see fig . 1 ) .
recently , @xmath7 transitions between the @xmath11 octupole phonon state and the @xmath35 state have been observed experimentally @xcite in near spherical nuclei . for an understanding of these new observations it is necessary to investigate the interrelation of the octupole phonon state and the mixed - symmetry @xmath35 state theoretically .
that is one aim of the present article .
another purpose of this article is the proposal of possible two - phonon states generated by the coupling of the octupole degree of freedom with the proton - neutron degree of freedom and the quantitative investigation of the properties of such constructions .
a convenient and sometimes successful approach to new and complicated physical phenomena is the investigation of symmetries and the algebraic solution of symmetric hamiltonians . in low - energy nuclear structure physics
this task can be achieved by application of the algebraic interacting boson model @xcite .
since we deal simultaneously with the isoscalar octupole vibration and the isovector quadrupole vibration in the valence shell of near spherical nuclei , we need a model which incorporates them both .
we apply for the first time the @xmath5-ibm-2 , which considers @xmath43 monopole ( @xmath44 ) bosons , @xmath45 quadrupole ( @xmath46 ) bosons , @xmath47 octupole ( @xmath48 ) bosons , and the proton - neutron ( @xmath49 - @xmath50 ) degree of freedom @xcite . in this paper
we develop the formalism of the @xmath5-ibm-2 in two particular dynamical symmetry limits relevant for the description of vibrational and @xmath51-unstable nuclei . in section [ sec : qn ] we discuss the group reduction chains of the @xmath5-ibm-2 and the quantum numbers , which are necessary to classify the eigenstates of the dynamically symmetric hamiltonians .
we derive analytical formulae for @xmath7 transition rates .
the @xmath2 transitions are calculated with a two - body operator .
quadrupole - octupole coupled two - body operators have turned out to be essential for the description of @xmath2 transitions generated by quadrupole - octupole collectivity @xcite .
the building blocks of the @xmath5-ibm-2 are @xmath52 creation and @xmath52 annihilation operators @xmath53 where @xmath54 , @xmath55 , @xmath56 , which by definition satisfy standard boson commutation relations : @xmath57 = \delta_{\rho \rho ' } \delta_{l l ' } \delta_{m m ' } \ ; .\ ] ] the total number of bosons is conserved for a given nucleus , i.e. @xmath58 the integers @xmath59 in eq .
( [ n ] ) are the eigenvalues of the boson number operators @xmath60 , where the dot denotes the scalar product and @xmath61 .
the @xmath62 bilinear combinations of proton and neutron boson operators of the type @xmath63 generate the lie algebra u@xmath64(13 ) @xmath65 u@xmath66(13 ) .
the u@xmath64(13 ) @xmath65 u@xmath66(13 ) symmetry algebra has a rich substructure . in the present study
we are interested in the @xmath6-spin symmetric dynamical symmetry limits of the model , which can be characterized by the reduction chains @xmath67 { [ n_{\pi } ] } & & { [ n_{\nu } ] } & & { [ n_1,n_2 ] } & & { [ n_1,n_2 ] } & & { [ m_1,m_2 ] } & { \ldots } & l_d & & l_f & & l \\ \end{array}\ ] ] where the dots denote alternative group reductions discussed later . below each group
we indicate the quantum numbers , associated with it . in total
, the complete classification of the basis states ( [ ds ] ) requires 26 quantum numbers corresponding to the @xmath68 occupation numbers @xmath69 in the @xmath70-scheme .
the @xmath6-spin quantum number @xcite relates in terms of young tableaux to the one- or two - rowed u@xmath71(13 ) representation as @xmath72 the maximum value of the @xmath6-spin quantum number @xmath73 corresponds to the totally symmetric representations of the u@xmath71(13 ) group .
states with smaller @xmath6-spin quantum numbers @xmath74 are called mixed - symmetry states and correspond to non - symmetric representations of the u@xmath71(13 ) group .
since the @xmath5-ibm-2 contains an octupole degree of freedom , it enlarges the diversity of mixed - symmetry states compared to the standard @xmath75-ibm-2 .
this fact can be recognized from the branching rules of u@xmath71(13 ) @xmath76 u@xmath71(6 ) @xmath65 u@xmath71(7 ) ( see appendix a ) .
the simplest two - rowed representation of u@xmath71(13 ) reduces to the u@xmath71(6 ) @xmath76 u@xmath71(7 ) representations in the following way : @xmath77 \eqnum{6.1}\label{6.1 } [ n-1,1 ] & = & [ n-1,1]\otimes [ m ] \\ \eqnum{6.2}\label{6.2 } & & [ n]\otimes [ m-1,1 ] \\ \eqnum{6.3}\label{6.3 } & & [ n]\otimes [ m ] \end{aligned}\ ] ] where , @xmath78 is the total number of @xmath44 and @xmath46 bosons , @xmath79 is the number of @xmath48 bosons , giving thus @xmath80 . as follows from eq .
( 6 ) , three types of the mixed - symmetry states arise . 1 .
the @xmath81 \otimes [ m]$ ] decomposition corresponds to the usual mixed - symmetry states in the @xmath75-space , coupled to @xmath70 octupole - bosons with a wave function , which is symmetric in the @xmath48-sector .
examples are the fundamental @xmath35 state and the @xmath82 `` scissors '' state with no @xmath48-boson at all , which belong to the @xmath83 representation @xmath81 \otimes [ 0]$ ] , or the new negative - parity mixed - symmetry two - phonon states which we denote as @xmath84 because they are generated by the coupling of the lowest @xmath35 state in the @xmath75-sector and one symmetric @xmath48-boson , which belong to the @xmath83 representation @xmath81 \otimes [ 1]$ ] .
note that in the presence of one @xmath48-boson due to the boson number conservation only @xmath85 bosons remain in the @xmath75-sector , which can belong to the symmetric @xmath86 representation @xmath87 = [ n-1]$ ] or to the mixed - symmetry representations , i.e. , to the lowest @xmath81 = [ n-2,1]$ ] .
therefore , the representation @xmath81 \otimes [ 1]$ ] can be considered as @xmath84 two - phonon states .
it turns out from the application of the branching rules given in appendix a that this @xmath84 two - phonon coupling will generate mixed - symmetry states with @xmath6-spin quantum numbers @xmath88 and @xmath89 , because the representation @xmath90 \otimes [ 1]$ ] is also present in the u@xmath91(13 ) representation @xmath92 $ ] , which has a total @xmath6-spin quantum number @xmath93 . in this article , we restrict ourselves to mixed - symmetry states with @xmath88 .
2 . the @xmath87\otimes [ m-1,1]$ ] reduction describes the mixed - symmetry states which appear due to mixed - symmetry in the @xmath48-sector .
we will not consider such kind of states in the present paper .
3 . finally
, @xmath87\otimes [ m]$ ] states are symmetric separately in the @xmath75- and in the @xmath48-sectors , however , they are coupled in a non - symmetric way within the full @xmath5-space .
the simplest example for a mixed - symmetry state if this type is the @xmath94 state , which consists of @xmath95 @xmath48-boson and @xmath96 @xmath44-bosons .
this @xmath94 state is the mixed - symmetry analogue state of the symmetric octupole phonon , the @xmath24 state .
this can be clearly seen from the explicit wave functions given in table [ tab : wvfctsu5 ] .
higher excited mixed - symmetry states of this type can be obtained by replacing in a symmetric way @xmath44-bosons with @xmath46-bosons in the @xmath94 wave function along with proper normalization .
we denote the lowest energy example of such mixed - symmetry states schematically as @xmath97 .
the operator which controls the excitation energy of mixed - symmetry states with @xmath6-spin quantum numbers @xmath98 is the majorana operator . by definition , this is the most general operator which annihilates any totally symmetric state . within the u@xmath71(13 )
@xmath76 u@xmath71(6 ) @xmath65 u@xmath71(7 ) , there are three important types of majorana operators . the well known majorana operator @xmath99 @xcite is associated with the u(6 ) group generated by the @xmath44- and @xmath46-bosons only .
it has the form @xmath100^{(2 ) } \cdot[s_\nu \times \tilde d_\pi -s_\pi \times \tilde d_\nu ] ^{(2 ) } - 2 \sum_{k=1,3 } [ d^+_\nu \times d^+_\pi ] ^{(k ) } \cdot [ \tilde d_\nu \times \tilde d_\pi ] ^{(k ) } \ ; , \ ] ] which influences the states whose wave functions are non - symmetric in the @xmath75-sector of the model space . in eq .
( [ eq : majo6 ] ) and below , @xmath101 denotes the standard tensor product of two irreducible tensorial operators .
it relates to the @xmath86 quadratic casimir operator as @xmath102=n(n+5)-2 \hat{m}_6 \ ; .\ ] ] the majorana operator @xmath103 is associated with the u(7 ) group generated by the @xmath48-bosons only .
it has the form @xmath104^{(k ) } \cdot [ \tilde f_\nu \times \tilde f_\pi ] ^{(k ) } \\ ] ] which is responsible for pushing up the mixed - symmetry states of the second type , i.e non - symmetric in the @xmath48-sector .
it relates to the @xmath105 quadratic casimir invariant as @xmath106=m(m+6)-2 \hat{m}_7 \ ; .\ ] ] finally , the majorana operator which is associated with the full group u(13 ) , reads @xmath107^{(3 ) } \cdot[s_\nu \times \tilde f_\pi -s_\pi \times \tilde f_\nu ] ^{(3 ) } \\ & & + \sum_{k=1,2,3,4,5 } ( -1)^{k+1 } [ d^+_\nu \times f^+_\pi -d^+_\pi \times f^+_\nu ] ^{(k ) } \cdot [ \tilde d_\nu \times \tilde f_\pi -\tilde d_\pi \times \tilde f_\nu ] ^{(k ) } \ ; , \\ & = & \left[f_{\rm max}\left(f_{\rm max}+1\right ) -
\hat{f}^2 \right ] \ ; .\end{aligned}\ ] ] this operator acts on all three types of mixed - symmetry states .
@xmath108 relates to the @xmath109 quadratic casimir operator as @xmath110=n(n+12)-2 \hat{m}_{13 } \ ; .\ ] ] the most general hamiltonian constructed from linear and quadratic casimir operators of the group u@xmath71(6 ) @xmath65 u@xmath71(7 ) reads @xmath111 + \alpha c_2[\mbox{u}_{\pi \nu } ( 5 ) ] + \kappa^\prime c_2[\mbox{su}_{\pi \nu } ( 3)]\\ & + \eta c_2[\mbox{so}_{\pi \nu } ( 6 ) ] + \beta c_2[\mbox{so}_{\pi \nu } ( 5 ) ] + \gamma_dc_2[\mbox{so}^d_{\pi \nu } ( 3 ) ] + \epsilon_fc_1[\mbox{u}_{\pi \nu } ( 7 ) ] \\ & + \kappa c_2[\mbox{so}_{\pi \nu } ( 7 ) ] + \xi c_2[\mbox{g}_{2 \pi \nu } ] + \gamma_fc_2[\mbox{so}^f_{\pi \nu } ( 3 ) ] + \gamma c_2[\mbox{so}_{\pi \nu } ( 3)]\;. \end{array}\ ] ] the definitions of the casimir invariants of lie groups used here are given in appendix b. the structure of the majorana operators has interesting consequences .
the @xmath5-ibm-2 reduces to the simpler version @xmath5-ibm-1 , if the strength constant @xmath112 of the majorana operator @xmath108 is set to infinity . using a finite value of @xmath112 , but @xmath113 will remove only those states which have mixed - symmetry in the @xmath48-sector alone . we will use this choice throughout the paper .
for the ordinary mixed - symmetry states in the @xmath75-sector , which are known already from the @xmath75-ibm-2 , the sum @xmath114 plays the role of the ordinary majorana parameter in the @xmath75-ibm-2 .
the excitation energies of mixed - symmetry states , which contain one @xmath48-boson are in addition also functions of the quantity @xmath115 .
this parameter controls the excitation energy of mixed - symmetry states with negative parity . since we are going to study vibrational or @xmath51-unstable nuclei , we are interested in two well - known dynamical symmetry limits of the positive parity @xmath75-subchain @xcite , namely , @xmath116 which we consider in the following subsections .
the negative parity subchain has a unique reduction @xcite : @xmath117 [ m_1,m_2 ] & & ( \omega_1,\omega_2 ) & & ( u_1,u_2 ) & \{\beta_j\ } & l_f \\
\end{array}\ ] ] where the quantum numbers @xmath118 ( @xmath119 ) are the so - called missing labels , which are necessary to classify completely the g@xmath120 so(3 ) reduction . unlike the case of @xmath5-ibm-1 where only the totally symmetric representations are important @xcite
, we also deal here with two - rowed representations of the algebras in ( [ ds ] ) and ( [ ds2 ] ) .
this requires some modifications .
for example , we have to take into account the exceptional group g@xmath121 . for the reduction of not fully symmetric representations of so(7 ) the group g@xmath121 helps to resolve the labelling problem .
the missing label , which is in general necessary to uniquely define the so(7 ) @xmath122 g@xmath121 reduction , is redundant in our case of only two - rowed representations . in this dynamical symmetry
limit the reduction chain of the positive parity u@xmath71(6 ) subalgebra is @xcite @xmath124 [ n_1,n_2 ] & & \{n_{d_1},n_{d_2}\ } & & [ \tau_1,\tau_2 ] & \{\alpha_i\ } & l_d \\ \end{array}\ ] ] where @xmath125 ( @xmath126 ) are missing labels , necessary to completely classify the so(5 ) @xmath76so(3 ) reduction .
we denote here @xmath78 the total number of @xmath44 and @xmath46 bosons , while the number of @xmath46 bosons only is denoted as @xmath127 .
thus , the total wave function can be written as @xmath128 { \otimes } [ n_{\nu } ] ; [ n_1,n_2][n_1,n_2]\{n_{d_1},n_{d_2}\ } ( \tau_1,\tau_2)\{\alpha_i \ } l_d;[m_1,m_2](\omega_1,\omega_2)(u_1,u_2 ) \{\beta_j\ } l_f;lm \rangle } \ ] ] the quantum numbers of the lowest states are given in table i. in the u@xmath71(5)@xmath65u@xmath71(7 ) dynamical symmetry limit it is rather simple to construct explicitly the wave functions .
we shall do this in order to make clear the variety of mixed - symmetry states appearing in the model . in table
i we present the wave functions for some states of interest .
we discussed above already the one-@xmath48-boson octupole states .
of particular interest are here also the lowest @xmath29 states . from the coupling of one @xmath46-boson quadrupole phonon and one @xmath48-boson octupole
phonon four @xmath29 states emerge .
this is evident in the @xmath70-scheme in @xmath6-space because the four basis wave functions @xmath129 , @xmath130 , @xmath131 , and @xmath132 can be formed .
these basis states can be combined to states with u@xmath71(13 ) symmetry and good @xmath6-spin .
one obtains one symmetric @xmath29 state corresponding to the totally symmetric @xmath133 $ ] irrep of u@xmath71(13 ) with @xmath6-spin quantum number @xmath134 , and three mixed - symmetry states : two of them with @xmath6-spin quantum numbers @xmath135 and one with @xmath93 .
one @xmath135 state corresponds to the @xmath136 \supset [ n-2,1]\otimes [ 1]$ ] reduction and the other to the @xmath136 \supset [ n-1]\otimes [ 1]$ ] reduction in the u@xmath71(13)@xmath76 u@xmath71(6)@xmath65 u@xmath71(7 ) , chain .
the last one arises from the @xmath92 \supset [ n-2,1]\otimes [ 1]$ ] reduction .
according to the discussion above we denote them as @xmath137 , @xmath138 , @xmath139 and @xmath140 , respectively .
the u@xmath71(5)@xmath65u@xmath71(7 ) dynamical symmetry hamiltonian expressed in terms of casimir invariants of first and second order of the relevant groups reads @xmath141 \\ & + \alpha c_2[\mbox{u}_{\pi \nu } ( 5 ) ] + \beta c_2[\mbox{so}_{\pi \nu } ( 5)]+\gamma_dc_2[\mbox{so}^d_{\pi \nu } ( 3 ) ] + \epsilon_fc_1[\mbox{u}_{\pi \nu } ( 7 ) ] \\ & + \kappa c_2[\mbox{so}_{\pi \nu } ( 7 ) ] + \xi c_2[\mbox{g}_{2 \pi \nu } ] + \gamma_fc_2[\mbox{so}^f_{\pi \nu } ( 3 ) ] + \gamma c_2[\mbox{so}_{\pi \nu } ( 3)]\;. \end{array}\ ] ] the eigenvalue problem for the dynamical symmetry hamiltonian ( [ hu5 ] ) can be solved analytically .
the solution of the full u(5 ) hamiltonian reads @xmath142 + \lambda ' n_2 ( n_1 + 1 ) + \lambda '' m_2 ( m_1 + 1 ) \\ & + \epsilon_d(n_{d_1}+n_{d_2})+\alpha [ n_{d_1}(n_{d_1}+4)+n_{d_2}(n_{d_2}+2 ) ] + \beta [ \tau_1(\tau_1 + 3)+\tau_2(\tau_2 + 1)]\\ & + \gamma_d l_d(l_d+1)+\epsilon_f(m_1+m_2 ) + \kappa [ \omega_1(\omega_1 + 5)+\omega_2(\omega_2 + 3 ) ] \\ & + \xi [ u_1(u_1 + 5)+u_2(u_2 + 4)+u_1 u_2]+\gamma_fl_f(l_f+1)+ \gamma l(l+1 ) \ ; .
\end{array}\ ] ] to construct the spectrum one needs to know the branching rules for the groups involved ( see appendix ) .
an energy spectrum corresponding to ( [ hu5 ] ) is shown in fig . 2 .
the parameters of the hamiltonian are specified in the figure caption . in this dynamical symmetry
limit the reduction chain of the positive parity u@xmath71(6 ) subalgebra is @xmath143 [ n_1,n_2 ] & & \langle \sigma_1,\sigma_2 \rangle & & [ \tau_1,\tau_2 ] & \{\alpha_i\ } & l_d \\ \end{array}\ ] ] and the total wave function is as follows : @xmath144 { \otimes } [ n_{\nu } ] ; [ n_1,n_2 ] [ n_1,n_2 ] \langle \sigma_1,\sigma_2 \rangle ( \tau_1,\tau_2)\{\alpha_i\ } l_d ; [ m_1,m_2](\omega_1,\omega_2)(u_1,u_2)\{\beta_j\ } l_f;lm \rangle } \ ] ] the quantum numbers for the lowest states in this dynamical symmetry are presented in table ii .
the so@xmath71(6)@xmath65u@xmath71(7 ) dynamical symmetry hamiltonian reads @xmath145\\ & + \beta c_2[\mbox{so}_{\pi \nu } ( 5)]+\gamma_dc_2[\mbox{so}^d_{\pi \nu } ( 3 ) ] + \epsilon_fc_1[\mbox{u}_{\pi \nu } ( 7 ) ] \\ & + \kappa c_2[\mbox{so}_{\pi \nu } ( 7)]+\xi c_2[\mbox{g}_{2 \pi \nu } ] + \gamma_fc_2[\mbox{so}^f_{\pi \nu } ( 3)]+\gamma c_2[\mbox{so}_{\pi \nu } ( 3)]\;. \end{array}\ ] ] and has the eigenvalues @xmath146
+ \lambda ' n_2 ( n_1 + 1 ) + \lambda '' m_2 ( m_1 + 1 ) \\ & + \zeta [ \sigma_1(\sigma_1 + 4)+\sigma_2(\sigma_2 + 2 ) ] + \beta [ \tau_1(\tau_1 + 3)+\tau_2(\tau_2 + 1 ) ] + \gamma_d l_d(l_d+1 ) \\ & + \epsilon_f(m_1+m_2)+\kappa [ \omega_1(\omega_1 + 5)+\omega_2(\omega_2 + 3 ) ] \\ & + \xi [ u_1(u_1 + 5)+u_2(u_2 + 4)+u_1 u_2 ] + \gamma_fl_f(l_f+1)+ \gamma l(l+1 ) \ ; . \end{array}\ ] ]
the branching rules for the corresponding groups can be found in ref .
@xcite and in the appendix below . a typical energy spectrum corresponding to eq .
( [ eso6 ] ) is shown in fig .
the parameters are specified in the caption . in this subsection
we discuss electromagnetic transitions between low - lying states in the @xmath5-ibm-2 with particular emphasis of those transitions which are new in the present model with respect to the standard @xmath75-ibm-2 and @xmath5-ibm-1 species .
for the study of the interrelation between the @xmath147 octupole phonon and the @xmath148 mixed - symmetry quadrupole phonon and possible two - phonon states generated by the coupling of them , we are interested in @xmath149 , and @xmath0 transition strengths .
the reduced transition probabilities @xmath150 are calculated from the matrix elements of the transition operators @xmath151 , where @xmath152 stands for the radiation character , either @xmath153 or @xmath154 , and @xmath112 denotes the multipolarity .
the electric quadrupole operator reads @xmath155 where @xmath156^{(2 ) } + \chi_{\rho } [ d^{+}_{\rho } \times \tilde d_{\rho } ] ^{(2 ) } + \chi'_{\rho } [ f^{+}_{\rho } \times \tilde f_{\rho } ] ^{(2 ) } \ ; .\ ] ] the magnetic dipole operator has the form @xmath157 where @xmath158^{(1 ) } \ ; , \quad l^f_{\rho } = 2\sqrt{7 } [ f^{+}_{\rho } \times \tilde f_{\rho } ] ^{(1 ) } \ ; .\ ] ] according to the eqs .
( [ te2][l ] ) the addition of the @xmath48 boson enlarges the structure of electromagnetic operators of the standard @xmath75-ibm-2 .
the standard @xmath75-ibm-2 , however , works well for the description of @xmath1 and @xmath3 transitions . in order to reduce the number of model parameters in the @xmath5-ibm-2 we , therefore , neglect below those parts of the @xmath1 and @xmath3 transition operators which involve the @xmath48-boson .
thus , we assume @xmath159 and @xmath160 .
in contrast , the @xmath48-boson is essential for @xmath0 and @xmath2 transitions , which are new with respect to the @xmath75-ibm-2 .
since there is at present no experimental evidence for the existence of states with mixed - symmetry in the @xmath48-boson sector , we shall consider here only @xmath6-scalar octupole transitions , i.e. , we use the following symmetric octupole operator : @xmath161 where @xmath162 stands for @xmath163^{(3 ) } + \chi [ d^{+}_{\rho } \times \tilde f_{\rho } + f^{+}_{\rho } \times \tilde d_{\rho } ] ^{(3 ) } \ ; .\ ] ] the general one - body electric dipole operator has the form @xmath164 where @xmath165^{(1 ) } \ ; . \\\ ] ] while the one - body @xmath3 , @xmath1 , and @xmath0 transition operators work satisfactorily , the one - body @xmath2-transition operator fails to provide an adequate description of low - lying dipole transition strengths in vibrational nuclei @xcite .
this applies for near - spherical nuclei also in the case when a @xmath166-boson is considered as it will be discussed below .
the operator ( [ te11body ] ) has very strict selection rules , conserving the total number of @xmath46- and @xmath48-bosons @xmath167 .
this contradicts the experimental findings where the quadrupole - octupole collective @xmath2 transitions , including the ground state decay of the quadrupole - octupole two - phonon @xmath29 state , are even found to be enhanced , see e.g. refs .
@xcite .
two approaches exist to avoid this problem .
following the first , the @xmath5-ibm is enlarged with a @xmath166 boson ( @xmath168 ) , thus leading to the so - called @xmath169-ibm @xcite .
the inclusion of the @xmath166 boson in the ibm corresponds to the consideration of a collective dipole vibration .
there exists no unambiguous interpretation of the @xmath166-boson origin .
some authors point out the origin of the @xmath166-boson from the collective low - lying valence shell excitation after removing the center - of - mass motion @xcite .
others have associated the @xmath166-boson with the gdr @xcite .
the @xmath166 boson induced one - body electric dipole operator works succesfully for the description of dipole transition strengths in deformed nuclei @xcite . for near - spherical nuclei ,
where the present article is focussing on , the inclusion of the @xmath166 boson does not lift all discrepancies @xcite .
the lowest @xmath29 state is the only strong low - lying @xmath2 excitation in many near - spherical nuclei @xcite .
this fact strongly suggests that this state is identified with the collective negative - parity dipole mode in the case of considering a low - energy @xmath166 boson .
this view , however , renders the well - established agreement of the @xmath29 excitation energy with the quadrupole - octupole two - phonon sum energy a sheer coincidence . also the enhanced @xmath2 strengths from other members of the quadrupole - octupole coupled quintuplet with spin quantum numbers
@xmath170 are not at all simple to understand . on the other hand some other experimental facts indicate different origin of strong @xmath2 transitions in near - spherical nuclei than sheer mixing with the gdr . for example ,
studies of @xmath2-strength distribution in @xmath171ce @xcite report that the lowest @xmath29 state exhibits an enhanced dipole transition to the ground state , while transitions from a number of higher lying @xmath29 states are considerably weaker although they lie closer to the gdr and although they are calculated to have predominantly a simple 1p1h - structure , which would even favor a mixing with the gdr .
thus , the observed concentration of @xmath2-strength in the lowest @xmath29 states should not be simply attributed to admixtures of the gdr into the states wave function if not at the same time the smaller @xmath2 strengths of higher lying @xmath29 states are explained , too .
it is even more striking that a model which can treat all these @xmath2 excitations on the same footing , the microscopic quasiparticle phonon model , supports the interpretation of enhanced low - lying @xmath2 transitions due to quadrupole - octupole collectivity @xcite .
therefore , another approach is preferable .
the multiphonon structure of excitations in vibrational nuclei as well as an octupole or quadrupole - octupole coupled origin of negative - parity states , naturally suggests an @xmath2-transition operator which contains a quadrupole - octupole coupled two - body part .
such a picture is also supported by the empirical correlation between @xmath2-transition strengths of the @xmath172 and @xmath34 transitions found recently for heavy near - spherical nuclei @xcite and by the fore - mentioned microscopic calculations @xcite .
furthermore , it avoids the problems in connection with the introduction of a @xmath166 boson discussed above .
we choose , therefore , to work with a quadrupole - octupole coupled two - body part in the @xmath2 operator .
the two - body dipole operator has already been introduced within the simple quadrupole and octupole phonon model @xcite and earlier within the @xmath5-ibm-1 @xcite .
this ansatz is in good agreement with the available data .
the @xmath2 transition operator considered here has the form @xmath173 \displaystyle \frac{\beta } { 2n } \left [ \left(q^{(2)}_{\pi } { + } \eta q^{(2)}_{\nu } \right ) { \times } \left(o^{(3)}_{\pi } { + } o^{(3)}_{\nu } \right){+ } \left(o^{(3)}_{\pi } { + } o^{(3)}_{\nu } \right){\times } \left(q^{(2)}_{\pi } { + } \eta q^{(2)}_{\nu } \right ) \right]^{(1 ) } \ ; , \end{array}\ ] ] where @xmath174 , @xmath175 , and @xmath176 are given by the eqs .
( [ d ] ) , ( [ q ] ) with @xmath177 , and ( [ o ] ) , respectively , and @xmath178 .
we stress that we introduce only one free parameter more ( @xmath179 ) in the @xmath2 operator with respect to the one - body operator from eq .
( [ te11body ] ) .
below we analyze @xmath2-transitions within two particular dynamical symmetry limits . using the wave functions ( [ wfu5 ] )
we can obtain the matrix elements of the transition operators ( [ te2 ] ) , ( [ tm1 ] ) , ( [ te3 ] ) and ( [ te1 ] ) . the technique for the calculation of such matrix elements
is described in detail in ref .
@xcite . in this particular dynamical symmetry limit ,
it is more convenient to use the explicit expressions of the wave functions constructed from boson operators following the method described in @xcite ( see , e.g. , ref .
the @xmath1 and @xmath3 transition strengths between those states , whose wave functions do not involve any @xmath48 bosons , are exactly the same as in the case of the @xmath75-ibm-2 and can be found in @xcite .
the advantage of the @xmath5-ibm-2 is that it provides the description of @xmath1 and @xmath3 transitions between negative parity states and of @xmath2 and @xmath0 transitions between the states of different parity and different @xmath6-spin .
here we are particularly interested in the electric dipole transitions which connect the octupole @xmath180 state with the symmetric @xmath4 state and with the mixed - symmetry @xmath181 state , as well as in those which connect the @xmath182 two - phonon negative parity multiplet with the ground state band .
we also consider here the new mixed - symmetry quadrupole - octupole coupled dipole states @xmath183 and @xmath184 and their electromagnetic properties ( including @xmath3 strengths ) as examples for mixed - symmetry states with negative parity .
analytical expressions of reduced transition probabilities are summarized in table iii .
it is worthwhile to look closely to the analytical expressions in table iii in order recognize the fingerprints of the simple phonon picture the consideration of which has guided us to the present model .
the @xmath2 transition strengths from the @xmath24 state to the symmetric and mixed - symmetry one - phonon quadrupole excitations , the @xmath4 state and the @xmath35 state , are both of the order of @xmath185 , i.e. , they can be strong on an absolute scale and comparable
. the @xmath186 @xmath6-vector @xmath2 strength can be even larger than that of the strong @xmath6-scalar @xmath34 @xmath2 transition if the effective electric dipole charges @xmath187 and @xmath188 have opposite signs . similarly , the @xmath2 ground state excitation strengths to the @xmath189 state and the @xmath190 state are both of the order of @xmath185 due to the two - body part of the @xmath2 operator .
however , here the @xmath6-scalar @xmath191 @xmath2 strength must be expected to be larger than the @xmath6-vector @xmath2 strength of the @xmath192 transition because the parameter @xmath193 usually takes values between 0 and 1 leading to a reduction of the @xmath6-vector @xmath2 strength which is proportional to the factor @xmath194 .
it is also interesting to compare the properties of the new mixed - symmetry @xmath29 states , @xmath190 and @xmath195 .
the direct @xmath2 ground state excitation strengths to the @xmath195 state is only of the order of @xmath196 and therefore much smaller than the @xmath2 strength to the @xmath190 state , which is of the order of @xmath185 .
the @xmath197 @xmath1 strength is of the order of n , which indicates quadrupole collectivity .
since it is proportional to the square of the difference of the effective quadrupole boson charges @xmath198 this transition should be weakly collective like the known @xmath199 one - quadrupole phonon decay .
in contrast , the @xmath200 @xmath1 strength is of the order of @xmath201 and can not be large .
very interesting are also the @xmath3 transitions between symmetric and mixed - symmetry negative parity states .
the @xmath3 strengths from the @xmath190 state to the quadrupole - octupole coupled @xmath29 and @xmath202 two - phonon states are of the order of @xmath185 and should be comparable to the @xmath203 strong @xmath3 transition .
the @xmath204 @xmath3 transitions are of the order of @xmath196 if the @xmath46-boson part alone is used in the @xmath3 transition operator ( [ tm1 ] ) .
if one allows for finite @xmath205-factors @xmath206 for the @xmath48-boson parts in the @xmath3 operator ( [ tm1 ] ) , they will be responsible for the leading term of the order of @xmath185 which was therefore included in table iii .
this resembles the fact that the @xmath195 state originates from the non symmetric coupling of an @xmath48-boson to the @xmath75-sector . in summary the @xmath190
state shows the behavior , which is mentioned in the introduction to be expected for a two - phonon state formed by the coupling of the fundamental phonons which generate the @xmath24 state and the @xmath35 state . using the wave functions ( [ wfso6 ] ) we have calculated the matrix elements of the transition operators ( [ te2 ] ) , ( [ tm1 ] ) , ( [ te3 ] ) and ( [ te1 ] ) .
we use here the quadrupole operator ( [ q ] ) with @xmath207 as is conventional for the so(6 ) limit .
the analytical expressions for the @xmath1 and @xmath3 transitions between the states which are formed only from @xmath44 and @xmath46 bosons are identical to those of the @xmath75-ibm-2 and can be found in ref .
analytical expressions for @xmath1 and @xmath3 transition strengths between negative parity states and for @xmath2 and @xmath0 transition strengths are summarized in table iv .
the relation and coupling of the collective octupole degree of freedom to proton - neutron mixed - symmetry states is discussed in the framework of the interacting boson model . for this purpose @xmath44-bosons , @xmath46-bosons , and @xmath48-bosons as well as the proton - neutron degree of freedom had to be taken into account leading to the hitherto not considered @xmath5-ibm-2 version of the model .
we developed the formalism of the @xmath5-ibm-2 considering two @xmath6-spin symmetric dynamical symmetry limits which are denoted by u@xmath71(5)@xmath65u@xmath71(7 ) and so@xmath71(6)@xmath65u@xmath71(7 ) .
these analytically solvable limits of the model can be relevant for the description of near spherical , vibrational and @xmath51-unstable nuclei , respectively .
the full dynamically symmetric hamiltonians are formulated , their eigenstates are specified and the analytical expressions for their excitation energies are given .
low - lying , collective , negative parity states are discussed including for the first time negative parity mixed - symmetry states .
analytical expressions for @xmath2 , @xmath3 , @xmath1 , and @xmath0 transition strengths were evaluated .
the @xmath2 operator considered here contains a quadrupole - octupole coupled two - body part besides the one - body term .
evidence for the possible dominance of @xmath6-vector @xmath2 transitions over @xmath6-scalar @xmath2 transitions is found .
the decay properties of @xmath190 states in vibrational nuclei are predicted .
in order to provide a complete classification scheme we need to reduce the representations of u(13 ) group to the kronecker product u(6)@xmath65u(7 ) . here we present the branching rules only for the totally symmetric @xmath133 $ ] and the lowest mixed - symmetry @xmath136 $ ] and @xmath92 $ ] irreducible representations of u(13 ) . the u(6)@xmath65u(7 ) representations are denoted as @xmath208\otimes [ \beta ] $ ] .
those u(6)@xmath65u(7 ) decompositions , which contain @xmath29 states discussed in the text , are underlined .
@xmath209=&[n]\otimes[0 ] , \ ; \underline{[n-1]\otimes[1 ] } , \ldots , [ 1]\otimes[n-1 ] , \ ; [ 0]\otimes[n ] \\[2pt ] [ n-1,1]=&[n-1,1]\otimes[0 ] , \ ; \underline{[n-2,1]\otimes[1 ] } , \ldots , [ 2,1]\otimes[n-3 ] , \ ; [ 1,1]\otimes[n-2 ] \\ & \underline{[n-1]\otimes[1 ] } , \ ; [ n-2]\otimes[2 ] , \ldots , [ 2]\otimes[n-2 ] , \ ; [ 1]\otimes[n-1 ] \\ & [ n-2]\otimes[1,1 ] , \ ; [ n-3]\otimes[2,1 ] , \ldots , [ 1]\otimes[n-2,1 ] , \ ; [ 0]\otimes[n-1,1 ] \\[2pt ] [ n-2,2]=&[n-2,2]\otimes[0 ] , \ ; [ n-3,2]\otimes[1 ] , \ldots , [ 3,2]\otimes[n-5 ] , \ ; [ 2,2]\otimes[n-4 ] \\ & \underline{[n-2,1]\otimes[1 ] } , \ ; [ n-3,1]\otimes[2 ] , \ldots , [ 3,1]\otimes[n-4 ] , \ ; [ 2,1]\otimes[n-3 ] \\ & [ n-2]\otimes[2 ] , \ ; [ n-3]\otimes[3 ] , \ldots , [ 3]\otimes[n-3 ] , \ ; [ 2]\otimes[n-2 ] \\ & [ n-3,1]\otimes[1,1 ] , \ ; [ n-4,1]\otimes[2,1 ] , \ldots , \ ; [ 2,1]\otimes[n-4,1 ] , \ ; [ 1,1]\otimes[n-3,1 ] \\ & [ n-3]\otimes[2,1 ] , \ ; [ n-4]\otimes[3,1 ] , \ldots , [ 2]\otimes[n-3,1 ] , \ ; [ 1]\otimes[n-2,1 ] \\ & [ n-4]\otimes[2,2 ] , \ ; [ n-5]\otimes[3,2 ] , \ldots , [ 1]\otimes[n-3,2 ] , \ ; [ 0]\otimes[n-2,2 ] \\ [ 2pt ] \end{array}\ ] ] the branching rules for the chains u(6)@xmath76 u(5)@xmath76 so(5 ) @xmath76so(3 ) and u(6)@xmath76 so(6)@xmath76 so(5 ) @xmath76so(3 ) reduction chains can be found elsewhere @xcite .
the partial classification scheme for the chain u(7)@xmath76 so(7)@xmath76g@xmath121 @xmath76so(3 ) is given in table vii .
the casimir invariants of lie algebras exploited here can be expressed in terms of boson creation and annihilation operators as follows @xcite : @xmath210= & \left(s^{+}\cdot s\right ) % + \left(d^{+}\cdot \tilde d\right)+\left(f^{+}\cdot \tilde f\right ) \ ; , \\
% c_2\left[\mbox{u(13)}\right]= & ( s^{+ } \cdot s ) \cdot ( s^{+}\cdot s ) % + \sum\limits_{k=0}^4 \left[d^{+ } \times \tilde d\right]^{(k)}\cdot % \left[d^{+ } \times \tilde d\right]^{(k ) } % + \sum\limits_{k=0}^6 \left[f^{+ } \times \tilde f\right]^{(k)}\cdot % \left[f^{+ } \times \tilde f\right]^{(k ) } \ ; , \\ c_1\left[\mbox{u(7)}\right]=&\left(f^{+}\cdot \tilde f\right ) \ ; , \\
c_2\left[\mbox{u(7)}\right]=&\sum\limits_{k=0}^6 \left[f^{+ } \times \tilde f\right]^{(k)}\cdot \left[f^{+ } \times \tilde f\right]^{(k ) } \ ; , \\ c_1\left[\mbox{u(6)}\right]=&\left(s^{+}\cdot s\right )
+ \left(d^{+}\cdot \tilde d\right ) \ ; , \\
c_2\left[\mbox{u(6)}\right]= & ( s^{+ } \cdot s ) \cdot ( s^{+}\cdot s ) + \sum\limits_{k=0}^4 \left[d^{+ } \times \tilde d\right]^{(k)}\cdot \left[d^{+ } \times \tilde d\right]^{(k ) } \ ; , \\
c_2\left[\mbox{su(3)}\right]= & 2 \left(\left[d^{+}{\times } s + s^+{\times } \tilde d \right]^{(2 ) } -\frac{\sqrt{7}}{2}\left[d^{+}{\times } \tilde d \right]^{(2)}\right ) \left(\left[d^{+}{\times } s + s^+{\times } \tilde d \right]^{(2 ) } -\frac{\sqrt{7}}{2}\left[d^{+}{\times } \tilde d \right]^{(2)}\right)\\ & + \frac{15}{2}\left[d^{+}{\times } \tilde d \right]^{(1)}\cdot \left[d^{+}{\times } \tilde d \right]^{(1 ) } \ ; , \\
c_2\left[\mbox{so(6)}\right]= & \left[d^{+}{\times } s + s^+{\times } \tilde d \right]^{(2 ) } \cdot \left[d^{+}{\times } s + s^+{\times } \tilde d \right]^{(2 ) } + 2 \sum\limits_{k=1,3 } \left[d^{+}{\times } \tilde d \right]^{(k)}\cdot \left[d^{+}{\times } \tilde d \right]^{(k ) } \ ; , \\
c_2\left[\mbox{g}_2\right]=&\sum\limits_{k=1,5 } \left[f^{+}{\times } \tilde f \right]^{(k)}\cdot \left[f^{+}{\times } \tilde f \right]^{(k ) } \ ; .
\\ \end{array}\ ] ] the exceptional algebra g@xmath121 has been introduced in physics by racah @xcite .
the eigenvalues of the casimir invariants from ( [ cas_inv ] ) in an irreducible representation given by a @xmath211 $ ] young tableau are @xmath212\rangle = \sum\limits_{i=1}^{n } f_i \ ; , \\ \langle c_2\left[\mbox{u}(n)\right]\rangle = \sum\limits_{i=1}^{n } f_i ( f_i+n+1 - 2 i ) \ ; , \\ \langle c_2\left[\mbox{su(3)}\right]\rangle = \lambda^2 + \mu^2 + \lambda \mu + 3 \lambda + 3 \mu \ ; , \qquad ( \lambda = f_1-f_2 , \mu = f_2 )
\\ \langle c_2\left[\mbox{so}(2n+1)\right]\rangle = \sum\limits_{i=1}^{n } 2 f_i ( f_i+2 n+1 - 2 i ) \ ; , \\ \langle c_2\left[\mbox{so}(2n)\right]\rangle = \sum\limits_{i=1}^{n } 2 f_i ( f_i+2 n-2 i ) \ ; , \\ \langle c_2\left[\mbox{g}_2\right]\rangle = f_1(f_1 + 5)+f_2(f_2 + 4)+f_1f_2 \ ; . \end{array}\ ] ] r.f .
casten , j. jolie , h.g .
brner , d.s .
brenner , n.v .
zamfir , w .- t .
chou and a. aprahamian , phys .
b 297 ( 1992 ) 19 ; phys . lett .
b 300 ( 1993 ) 411 .
r. schmidt , th .
blaich , th.w .
elze , h. emling , h. freiesleben , k. grimm , w. henning , r. holzmann , j.g .
keller , h. klingler , r. kulessa , j.v .
kratz , d. lambrecht , j.s .
lange , y. leifels , e. lubkiewicz , e.f .
moore , e. wajda , w. prokopowicz , ch .
schutter , h. spies , k. stelzer , j. stroth , and w. walus , h.j .
wollersheim , m. zinser and e. zude , phys .
70 ( 1993 ) 1767 .
t. aumann , p.f .
bortignon and h. emling , ann .
nucl . part .
48 ( 1998 ) 351 . c. volpe , ph .
chomaz , m.v .
andres , f. catara and e.g. lanza , nucl .
phys . a 647 ( 1999 ) 246 .
p. kleinheinz , j. styczen , m. piiparinen , j. blomqvist and m. kortelahti , phys .
48 ( 1982 ) 1457 .
m. yeh , p.e .
garrett , c.a .
mcgrath , s.w .
yates and t. belgya , phys .
76 ( 1996 ) 1208 .
brner , j. jolie , s.j .
robinson , b. krusche , r. piepenbring , r.f .
casten , a. aprahamian and j.p .
draayer , phys .
66 ( 1991 ) 691 .
r. a. gatenby , j. r. vanhoy , e. m. baum , e.l .
johnson , s.w .
yates , t. belgya , b. fazekas , a. veres and g. molnar , phys .
c 41 ( 1990 ) r414 . s. j. robinson , j. jolie , h. g. brner , p. schillebeeckx , s. ulbig and k.p .
lieb , phys .
73 ( 1994 ) 412 .
s. f. hicks , c. m. davoren , w. m. faulkner and j. r. vanhoy , phys . rev .
c 57 ( 1998 ) 2264 .
j. r. vanhoy , j. m. anthony , b. m. haas , b.h .
benedict , b.t .
meehan , s.f .
hicks , c.m .
davoren and c.l .
lundstedt , phys . rev .
c 52 ( 1995 ) 2387 .
p. e. garrett , h. lehmann , j. jolie , c.a .
mcgrath , m. yeh and s.w .
yates , phys .
c 59 ( 1999 ) 2455 .
l. esser , r.v .
jolos and p. von brentano , nucl .
phys . a 650 ( 1999 ) 157 .
metzger , phys .
c 17 ( 1978 ) 939 .
herzberg , i. bauske , p. von brentano , th .
eckert , r. fischer , w. geiger , u. kneissl , j. margraf , h. maser , n. pietralla , h.h .
pitz and a. zilges , nucl .
a 592 ( 1995 ) 211 .
j. bryssinck , l. govor , d. belic , f. bauwens , o. beck , p. von brentano , d. de frenne , t. eckert , c. fransen , k. govaert , r .- d .
herzberg , e. jacobs , u. kneissl , h. maser , a. nord , n. pietralla , h.h .
pitz , v.yu .
ponomarev and v. werner , phys .
c 59 ( 1999 ) 1930 .
b. fazekas , t. belgya , g. molnar , a. veres , r.a .
gatenby , s.w .
yates and t. otsuka , nucl .
phys . a 548 ( 1992 ) 249 .
garrett , h. lehmann , c.a .
mcgrath , minfang yeh and s.w .
yates , phys .
c 54 ( 1996 ) 2259 .
i. wiedenhver , a. gelberg , t. otsuka , n. pietralla , j. gableske , a. dewald and p. von brentano , phys .
c 56 ( 1997 ) r2354 .
n. pietralla , d. belic , p. von brentano , c. fransen , r .- d .
herzberg , u. kneissl , h. maser , p. matschinsky , a. nord , t. otsuka , h.h .
pitz , v. werner and i. wiedenhver , phys .
c 58 ( 1998 ) 796 .
n. pietralla , c. fransen , d. belic , p. von brentano , c. friessner , u. kneissl , a. linnemann , a. nord , h. h. pitz , t. otsuka , i. schneider , v. werner and i. wiedenhver , phys .
83 ( 1999 ) 1303 .
n. pietralla , c. fransen , p. von brentano , c. friener , a. gade , a. linnemann , p. matschinsky , i. schneider , v. werner , i. wiedenhver , d. belic , u. kneissl , a. nord and h. h. pitz , contribution to the international conference on _ nuclear models 1998 _ , october , 12 14 1998 , camerino , italy , in press .
herzberg , p. von brentano , j. eberth , j. enders , r. fischer , n. huxel , t. klemme .
p. von neumann - cosel , n. nicolay , n. pietralla , v. yu . ponomariev , j. reif , a. richter , c. schlegel , r. schwengner , s. skoda , h. g. thomas , i. wiedenhver and a. zilges , phys .
b 390 ( 1997 ) 49 . ll & @xmath213 { \otimes } [ n_{\nu } ] ; [ n_1,n_2],[n_1,n_2]\{n_{d_1},n_{d_2}\ } ( \tau_1,\tau_2)\{\alpha_i\ } l_d;[m_1,m_2](\omega_1,\omega_2)(u_1,u_2 ) \{\beta_j\ } l_f;l \rangle } } $ ] + @xmath214 & = @xmath215 \otimes [ n_{\nu } ] ; [ n],[n]\{0\}(0)0;[0](0)(0)0;0 \rangle = |s_{\pi } ^{n_{\pi } } s_{\nu } ^{n_{\nu } } ; 0 \rangle $ ] + @xmath216 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n]\{2\}(0)0;[0](0)(0)0;0 \rangle $ ] + & @xmath218 + @xmath219 & @xmath217 \otimes [ n_{\nu } ] ; [ n-1,1],[n-1,1]\{1,1\}(1,1)1;[0](0)(0)0;1 \rangle $ ] + & @xmath220 + @xmath221 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n]\{1\}(1)2;[0](0)(0)0;2 \rangle $ ] + & @xmath222 + @xmath223 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n]\{2\}(2)2;[0](0)(0)0;2 \rangle $ ] + & @xmath224 + @xmath225 & @xmath217 \otimes [ n_{\nu } ] ; [ n-1,1],[n-1,1]\{1\}(1)2;[0](0)(0)0;2 \rangle $ ] + & @xmath226 + @xmath227 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n]\{3\}(3)3;[0](0)(0)0;3 \rangle $ ] + & @xmath228 + & @xmath229 + & @xmath230 + @xmath231 & @xmath217 \otimes [ n_{\nu } ] ; [ n-1,1],[n-1,1]\{1,1\}(1,1)3;[0](0)(0)0;3 \rangle $ ] + & @xmath232 + @xmath233 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n]\{2\}(2)4;[0](0)(0)0;4 \rangle $ ] + & @xmath234 + @xmath235 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n]\{3\}(3)6;[0](0)(0)0;6 \rangle $ ] + & @xmath236 + & @xmath237 + & @xmath238 + @xmath239 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n-1]\{0\}(0)0;[1](1)(1)3;3\rangle $ ] + & @xmath240 + @xmath241 & @xmath217 \otimes [ n_{\nu } ] ; [ n-1,1],[n-1]\{0\}(0)0;[1](1)(1)3;3 \rangle $ ] + &
@xmath242 + @xmath243 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n-1]\{1\}(1)2;[1](1)(1)3;1\rangle $ ] + & @xmath244 + & @xmath245 + @xmath246 & @xmath217 \otimes [ n_{\nu } ] ; [ n-1,1],[n-2,1]\{1\}(1)2;[1](1)(1)3;1 \rangle $ ] + & @xmath247 + & @xmath248 + @xmath249 & @xmath217 \otimes [ n_{\nu } ] ; [ n-1,1],[n-1]\{1\}(1)2;[1](1)(1)3;1 \rangle $ ] + & @xmath250 + & @xmath251 + @xmath252 & @xmath217 \otimes [ n_{\nu } ] ; [ n-2,2],[n-2,1]\{1\}(1)2;[1](1)(1)3;1 \rangle $ ] + & @xmath253 + & @xmath254 + @xmath255 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n-1]\{1\}(1)2;[1](1)(1)3;2 \rangle $ ] + & @xmath256 + & @xmath257 + @xmath258 & = @xmath215 \otimes [ n_{\nu } ] ; [ n],[n-1]\{1\}(1)2;[1](1)(1)3;3 \rangle $ ] + & @xmath259 + & @xmath260 + @xmath261 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n-1]\{1\}(1)2;[1](1)(1)3;4 \rangle $ ] + & @xmath262 + & @xmath263 + @xmath264 & @xmath217 \otimes [ n_{\nu } ] ; [ n],[n-1]\{1\}(1)2;[1](1)(1)3;5\rangle $ ] + & @xmath265 + & @xmath266 + @xmath213 { \otimes } [ n_{\nu } ] ; [ n_1,n_2],[n_1,n_2 ] \langle \sigma_1,\sigma_2 \rangle ( \tau_1,\tau_2)\{\alpha_i \ } l_d ; [ m_1,m_2](\omega_1,\omega_2)(u_1,u_2)\{\beta_j\ } l_f;lm \rangle } } $ ] + @xmath267 + @xmath268 + @xmath269(0)(0)0;1 \rangle $ ] + @xmath270 + @xmath271 + @xmath272(0)(0)0;2 \rangle $ ] + @xmath273(0)(0)0;3 \rangle $ ] + @xmath274 + @xmath275 + @xmath276(1)(1)3;3 \rangle $ ] + @xmath277 + @xmath278(1)(1)3;1 \rangle $ ] + @xmath279(1)(1)3;1 \rangle $ ] + @xmath280(1)(1)3;1 \rangle $ ] + @xmath281&=&@xmath282 + @xmath283&=&@xmath284 ^ 2 $ ] + @xmath285&=&@xmath286 ^ 2 \frac{n_{\nu } \,n_{\pi } } { n^2 } $ ] + @xmath287&=&@xmath288 + @xmath289&=&@xmath290 ^ 2 \frac{n-1}{n^4}$ ] + @xmath291&=&@xmath292 ^ 2 $ ] + @xmath293&=&@xmath294\frac{1}{n^2 } \right]^2 $ ] + @xmath295&=&@xmath296 ^ 2 $ ] + @xmath297&=&@xmath298 ^ 2 $ ] + @xmath299&=&@xmath300 + @xmath301&=&@xmath302 + @xmath303&=&@xmath304 + @xmath305&=&@xmath306 + @xmath307&= & @xmath308 + @xmath309&=&@xmath310 + @xmath311&=&@xmath312 + @xmath313&=&@xmath314 + @xmath315&=&@xmath316 + @xmath317&=&@xmath318 ^ 2 \frac{n_{\pi } \,n_{\nu } } { n^2}$ ] + @xmath319&=&@xmath320 ^ 2 \frac{1}{n^2}$ ] + @xmath321&=&@xmath322 + @xmath323&=&@xmath324 ^ 2 \frac{n_{\pi } \,n_{\nu } } { n^2 } $ ] + @xmath325&=&@xmath326 ^ 2 \frac{n_{\pi } \,n_{\nu } } { n^2 } $ ] + @xmath327&=&@xmath328 + @xmath329&=&@xmath330 ^ 2 \frac{n_{\pi } \,n_{\nu } } { n^4 ( n-1)}$ ] + lcl @xmath331&=&@xmath332 + @xmath333&=&@xmath334 ^ 2 \frac{(n+3)(n+4)}{n^2(n+1 ) } $ ] + @xmath335&=&@xmath336 ^ 2 \frac{(n+2)(n+3)n_{\nu } \,n_{\pi } } { n^2(n+1)^2}$ ] + @xmath337&=&@xmath338 ^ 2 \frac{n-1}{n(n+1 ) } $ ] + @xmath339&=&@xmath340 + | the interrelation between the octupole phonon and the low - lying proton - neutron mixed - symmetry quadrupole phonon in near - spherical nuclei is investigated .
the one - phonon states decay by collective @xmath0 and @xmath1 transitions to the ground state and by relatively strong @xmath2 and @xmath3 transitions to the isoscalar @xmath4 state .
we apply the proton neutron version of the interacting boson model including quadrupole and octupole bosons ( @xmath5ibm-2 ) .
two @xmath6-spin symmetric dynamical symmetry limits of the model , namely the vibrational and the @xmath7-unstable ones , are considered .
we derive analytical formulae for excitation energies as well as @xmath8 , and @xmath9 values for a number of transitions between low - lying states . corresponding author : + nadya a. smirnova + centre de spectromtrie nuclaire et de spectromtrie de masse , + bat .
104 , orsay campus , f-91405 , france + tel : ( + 33 ) ( 0)1 69 15 48 57 + fax : ( + 33 ) ( 0)1 69 15 50 08 + e - mail : [email protected] + |
open clusters have been suspected as sites for the high energy processes due to the presence of large concentration of matter and also early type compact objects such as massive stars , compact massive binaries , pulsars and their nebulae and supernova remnants .
in fact , up to now tev @xmath0-ray emission has been detected from the direction of 3 open clusters ( i.e. cyg ob 2 - aharonian et al .
2002 , westerlund 2 - aharonian et al . 2007 , and westerlund 1 - abramowski et al .
also gev @xmath0-ray emission , extending up to @xmath1100 gev , has been recently reported from the direction of the supermassive binary system eta carinae within the carina nebula complex ( tavani et al .
2009a , abdo et al .
2009a , farnier et al . 2011 ) .
a part of this gev emission shows evidence of modulation with the binary period ( farnier & walter 2011 , reitberger et al .
therefore , it has to be produced not far from the binary system itself .
eta carinae binary system has not been detected up to now in the tev @xmath0-ray range ( abramowski et al .
the modulated gev @xmath0-ray emission from binary system cyg x-3 has been recently detected by the agile and fermi - lat telescopes ( tavani et al .
2009b , abdo et al .
2010 , bulgarelli et al .
this emission is transient and appears close to the quenched radio state just before the major radio flares observed from this system .
the gev @xmath0-ray emission has not been identified up to now in the fermi - lat data from the binary system wr 20a within the open cluster westerlund 2 .
two new sources reported by fermi collaboration in this direction have been identified with the @xmath0-ray pulsars ( abdo et al .
2009b , saz parkinson et al .
2010 ) .
a few scenarios have been considered as possible explanations of the high energy emission from the open clusters .
it has been proposed that strong winds produced by massive stars in open clusters are able to accelerate particles at shock waves formed during collisions of the winds with dense matter of the open cluster ( e.g. vlk & forman 1982 , cesarsky & montmerle 1983 , giovannelli et al .
1996 , torres et al .
2004 , bednarek 2007 ) .
@xmath0-rays and neutrinos can be also produced by particles accelerated at the shocks formed within massive binary systems ( e.g. eichler & usov 1993 , white & chen 1995 , benaglia & romero 2003 , bednarek 2005a , reimer et al .
2006 , pittard & dougherty 2006 , bednarek & pabich 2011 ) .
massive stars within open clusters end their life at a relatively short time producing supernova remnants ( snrs ) and pulsar wind nebulae ( pwne ) .
interaction of snrs shock waves with dense clouds within the open cluster can also accelerate relativistic particles that are able to produce high energy radiation ( aharonian & atoyan 1996 ) .
a part of @xmath0-ray emission from the open clusters can be also produced within pwne or as a result of the interaction of particles accelerated within pwne with dense matter of an open cluster ( bednarek 2003 , bartko & bednarek 2008 , abramowski et al .
2011 , ohm et al .
2013 , aliu et al .
it is not clear at present which of those scenarios provide a bulk of the high energy @xmath0-ray emission observed from a few open clusters mentioned above .
it is likely that in specific open cluster a few processes can be important since different types of objects ( massive star winds , snrs , pwne ) might inject relativistic particles with comparable power , i.e. of the order of @xmath2 erg . in the present paper we investigate in detail the high energy radiation expected in hadronic processes within and around the massive binary systems surrounded by the large concentration of matter .
it is assumed that nuclei are accelerated in the region of colliding winds within the binary stars ( e.g. see bednarek 2005b ) .
these nuclei can severely disintegrate , in the interaction with the radiation field of massive stars and with the matter of the stellar winds , injecting neutrons and protons . charged protons diffuse through the open cluster producing @xmath0-rays and neutrinos in collisions with the matter of the stellar wind cavity and dense environment of the open cluster . on the other hand , neutrons move balistically through the wind cavity and decay into protons at some distance from the binary system which can be still within the wind cavity or already within the open cluster .
protons from their decay can also contribute to the high energy @xmath0-ray and neutrino spectrum .
as an example , we perform calculations of the @xmath0-ray and neutrino fluxes produced in the clusters surrounding the eta carina supermassive binary system and wr 20a binary system in westerlund 2 open cluster .
this emission is confronted with the present measurements of the gev - tev @xmath0-ray emission from these clusters , sensitivities of the future cherenkov telescope array ( cta ) and the icecube neutrino telescope .
we consider a massive binary system in which one or both companion stars belongs to the class of the wolf - rayet ( wr ) stars .
wr type star produces fast and dense wind , due to the huge mass loss rate , which can be of the order of @xmath3 m@xmath4 yr@xmath5 .
the winds propagate with the characteristic velocities of the order of @xmath6 km s@xmath5 .
the density of the wind drops with the distance from the star according to , @xmath7 where @xmath8 cm is the radius of the star , and @xmath9 is the distance from the star in units of the stellar radius .
note that at the distance of the order of a parsec , density of the wind becomes very low , e.g. for 1 pc the wind density becomes @xmath10 @xmath11 .
so then , the outer regions of the wind cavity are filled with very rare but high velocity gas .
we consider consequences of acceleration of nuclei within the binary system located in the open cluster .
nuclei ( from the stellar winds ) can be accelerated within the collision region of the winds ( e.g. eichler & usov 1992 ) .
the recconnection of the magnetic field and the diffusive shock acceleration process can play important role in this place .
the details of the acceleration process , propagation of hadrons and their subsequent interaction within the binary system and its surrounding are discussed in the next section .
we show that nuclei can efficiently disintegrate in the dense radiation and matter of the winds from massive stars . as a result
, neutrons are injected .
they decay at the distance from the binary system which is determined by their lorentz factors .
protons , extracted from nuclei , lose energy on interaction with the dense wind close to the binary system .
they also suffer adiabatic energy losses in the expanding wind within the wind cavity .
we consider high energy processes in which @xmath0-rays and neutrinos are produced in hadronic collisions in the above described scenario .
the schematic representation of the processes around the binary system is shown in fig . 1 .
note that hadrons might be also accelerated in such scenario at the shock structure formed in the collision of the stellar wind with the circumstellar gas .
we expect that this acceleration process is less efficient than the acceleration process occurring within the binary system .
the product of the magnetic field strength and the characteristic dimension of the shock , determining the maximum energies of accelerated particles , is expected to be lower at the shock produced by the wind in surrounding gas due to the mainly radial of the magnetic field within the binary system and its toroidal structure far away from the binary .
massive stars produce fast and dense winds which can extract significant amount of original mass from the star during its lifetime .
in fact , the intensive mass loss rates of the order of a few @xmath21 m@xmath4 yr@xmath5 are characteristic for the wolf - rayet ( wr ) and o type stars . during the main sequence stage ,
the outer parts of stars are completely lost and only inner parts , composed of heavy nuclei , are left .
therefore , the winds of early type stars are expected to be mainly composed from nuclei heavier than hydrogen such as helium to oxygen .
massive stars are frequently found within the massive binary systems in which strong winds collide providing conditions for acceleration of nuclei to large energies .
below , we generally consider the process of acceleration of nuclei at the shocks or the reconnection regions created by these colliding winds .
the propagation and interaction of nuclei with the stellar radiation field results in their photo - disintegration to neutrons , protons and secondary nuclei .
we perform numerical simulations of nuclei in the radiation field of the massive stars in order to determine the rate of injection of different nuclei . note that density of stellar photons in the vicinity of massive star , @xmath22 ph .
@xmath11 ( where @xmath23 k is the surface temperature of the massive star ) , is a few orders of magnitude larger than density of matter in the stellar wind ( see eq .
[ eq1 ] ) . therefore
, it is expected that nuclei with sufficiently large energies interact at first in collisions with stellar photons rather than with the matter of the stellar wind .
after that , nuclei can suffer significant fragmentation in collisions with the matter of the stellar wind .
this second process is independent on energy of nuclei .
therefore , lower energy nuclei can also suffer strong disintegration process in collisions with dense matter of the stellar winds in the case of stars with exceptionally strong winds such as considered in this paper ( i.e. wr and o type stars which mass loss rate is above @xmath24 m@xmath4 yr@xmath5 ) .
neutrons released from nuclei in collisions with the matter have the spectrum similar to the spectrum of primary nuclei .
therefore , neutrons with low energies can decay relatively close to the binary system where the density of matter is still high . on the other hand
, high energy neutrons can even reach dense regions outside the wind cavity of the massive binary system . in this paper
we calculate @xmath0-ray and neutrino spectra produced by protons extracted from nuclei and also from protons from neutrons decaying at some distance from the binary system .
these neutrons decay within the wind cavity and also within the dense open cluster in which binary system is immersed .
we estimate the maximum energies of hadrons accelerated in the shock region of colliding winds following the conditions considered e.g. in bednarek & pabich ( 2011 ) and bednarek ( 2005b ) .
the maximum energies of hadrons are determined by comparing their acceleration time scale and their escape time scale from the acceleration region or collision time scale with the matter of the stellar wind .
the acceleration time scale is given by , @xmath25 where @xmath26 and @xmath27 are the energy and the lorentz factor of particles , @xmath28 is the acceleration rate , @xmath29 is the acceleration coefficient , @xmath30 is the larmor radius of hadron in the magnetic field at the shock @xmath31 , @xmath32 is the velocity of light , @xmath33 is the elemental charge , and @xmath34 and @xmath35 are the mass and charge numbers of nuclei .
we apply @xmath36 .
@xmath31 , is related to the surface magnetic field of the massive star by assuming its dipolar structure close to the surface up to @xmath37 and radial structure at larger distances ( usov & melrose 1992 ) ) . for the distance of the collision region from the star equal to @xmath38
, @xmath31 drops to @xmath39 .
the advection time scale of hadrons from the collision region is estimated from , @xmath40 where @xmath41 cm is the distance of the collision region from the star and @xmath42 .
we estimate the diffusion time scale of hadrons ( in the bohm approximation ) in the acceleration region @xmath43 s , where @xmath44 is the lorentz factor of hadrons and @xmath45 g. this diffusion time is larger than the above estimated advection time scale even for the most energetic hadrons considered in this scenario .
then , the maximum energy of hadrons accelerated at the region of colliding stellar winds is obtained from comparison of eq . 3 and eq .
is then given by , @xmath46 where the acceleration coefficient is estimated to be @xmath47 .
llllllllllll + name & b & r@xmath48 & @xmath49 & r@xmath50 & t@xmath48 & @xmath51 & @xmath52 & @xmath53 & @xmath54 & @xmath55 & @xmath56 + unit & g & cm & km / s & @xmath57 & k & m@xmath4 & m@xmath4/yr & erg s@xmath5 & & + + wr & @xmath58 & @xmath59 & @xmath58 & 2 & @xmath60 & 5 & @xmath61 & @xmath62 & @xmath10.3 & @xmath63 & @xmath64 + + wr 20a & @xmath65 & @xmath66 & @xmath65 & 2 & @xmath67 & 30 & @xmath68 & @xmath62 & @xmath12.4 & @xmath69 & @xmath70 + + eta car & @xmath71 & @xmath72 & @xmath73 & 1.4 & @xmath67 & 80 & @xmath74 & @xmath75 & @xmath14.1 & @xmath76 & @xmath77 + [ tab1 ] on the other hand , hadrons lose also energy on collisions with the matter of the stellar wind .
the hadron - hadron energy loss time scale can be estimated from , @xmath78 where @xmath79 @xmath80 is the cross section for proton - proton collision , and @xmath81 is the in - elasticity coefficient .
the maximum energies of hadrons allowed by collisional energy losses are , @xmath82 we have calculated the maximum energies of hadrons due to these two processes for three example massive stars within the binary systems .
the results are shown in table 1 . in the case of the parameters of the wr type star
as observed in cyg x-3 binary system , the maximum energies of accelerated hadrons are limited by advection process .
in the case of the binary system containing wr 20a stars , the maximum energies of hadrons are comparable in the case of advection process and collisional energy losses .
but , in the case of eta carinae the maximum energies of hadrons are limited by collisional energy losses . in all considered cases hadrons can reach energies of the order of @xmath110@xmath83 gev . in all calculations of the @xmath0-ray and neutrino spectra ( presented below ) , it is assumed that nuclei are injected with the power law spectrum ( spectral index equal to 2 ) , extending up to the lower value between those given by eqs . 5 and 7 .
the power in these hadrons is normalized to a part , @xmath84 , of the power provided by the stellar wind of the massive star , @xmath85 ( see table 1 for specific binaries ) . in table 1
we also report other basic parameters of these three binary systems .
their orbital periods are equal to 4.8 hrs for cyg x-3 binary system ( becklin et al .
1973 ) , 3.675 days for wr 20a ( rauw et al . 2004 ) and 2027.7 days for eta carinae ( damineli et al .
the process of disintegration of nuclei has been extensively studied in the past in the context of the origin and propagation of the highest energy cosmic rays ( e.g. stecker 1969 , tkaczyk et al .
1975 , puget et al . 1976 ) .
this process has been also proposed to be important in compact sources in which dense soft radiation field is expected , i.e. in supernova remnants ( e.g. pollack & shen 1969 ) , binary systems ( e.g. karakua et al .
1994 ) , dense stellar clusters ( e.g. anchordoqui et al .
2007a , b ) , or the galactic center ( e.g. kusenko et al . 2006 ) .
@xmath0-ray emission from de - excitation of fragments of nuclei , which survived disintegration , has been calculated in some of these works following the prescription emphasized by balashov et al .
this @xmath0-ray production process is generally relatively inefficient since only a small part of the initial energy of hadrons is converted into radiation .
however , it may play a role in the case of sources with strong soft radiation field but without large amount of matter . here
we consider production of @xmath0-rays and neutrinos by secondary products of disintegrated nuclei in their interaction with dense matter surrounding the source of relativistic nuclei .
we neglect possible direct production of neutral high energy radiation in the interaction of hadrons with the stellar radiation ( @xmath86 ) since , as we have shown above , nuclei are not accelerated to large enough energies to reach the threshold for this process in the environment of the binary systems . in order to have an impression about the role of the fragmentation process of nuclei in the radiation field of massive stars we consider a simple case of their injection from a point - like source not far from the stellar surface , e.g. at the distance of two stellar radii .
we calculate the optical depths for fragmentation of nuclei injected at different initial directions , as a function of their energy and two mass numbers corresponding to the helium and the oxygen nuclei ( see fig . 2 ) .
the approximation of the cross section for photo - disintegration process of nuclei is applied as described in karakua & tkaczyk ( 1993 ) .
nuclei are considered to propagate in the vicinity of three representative massive stars with different parameters : wr type star as observed in the cyg x-3 binary system , wr star in the wr 20a binary system and supermassive star in the eta carinae binary system ( see fig . 2 ) .
note that only nuclei with lorentz factors above @xmath87 can be efficiently disintegrated .
this critical lorentz factor can be estimated from @xmath88 , where @xmath89 mev is the minimum energy of stellar photon in the nuclei rest frame which is able to effectively photo - disintegrate the nuclei , @xmath23 k is the surface temperature of the star , and @xmath90 is the boltzmann constant . in this figure
we also determine the number of nucleons dissolved from the primary nuclei in the photo - disintegration process .
the investigation of the results presented in fig . 2 , in the context of the maximum energies of nuclei expected in our model ( see table . 1 ) , allows us to conclude that in the case of wr 20a binary system the process of photo - disintegration of nuclei can be neglected . in the case of wr star with parameters observed in cyg x-3 binary system and eta carinae ,
significant number of nuclei with energies in the range @xmath91 will lose nucleons in collisions with stellar radiation . in order to conclude whether nuclei can be efficiently disintegrated in collisions with the matter of the stellar wind
, we estimate the optical depth for relativistic hadrons on collisions with the matter of the wind . this optical depth can be estimated from , @xmath92 where @xmath93 ( in cm ) is the radius of the binary system at which hadrons are injected within the wind .
it is assumed to be equal to the shock radius .
the optical depths for protons on the interaction with the matter of the wind are shown in table 1 for three example parameters of the massive stars .
note that the optical depth for nuclei with specific mass number @xmath94 in collisions with nuclei with the mass number @xmath95 scales as , @xmath96 , where @xmath97 is between @xmath98 to @xmath99 ( see dpmjet calculations given in mori 2009 ) . for the case of the helium nuclei ( @xmath100 )
, the optical depths will be a factor of @xmath101 larger ( @xmath102 ) and for the oxygen nuclei a factor of @xmath103 larger ( @xmath104 , see mori 2009 ) .
therefore , we conclude that in the case of considered binary system relativistic nuclei should be completely disintegrated .
large values of the optical depths for collisions of protons with the matter of the stellar winds indicate that these relativistic hadrons should efficiently lose energy on production of high energy @xmath0-rays and neutrinos in the close vicinity of the massive binary system .
on the other hand , neutrons , released from these nuclei , should move rectilinearly and decay at large distances from the binary system within the wind cavity or , in the case of the most energetic neutrons , directly in the region of the open cluster .
protons , from decay of these neutrons , can also produce high energy radiation at large distances from the binary systems , i.e in the region with low level of the soft radiation field and low density of matter in respect to the density of the wind within the binary system .
accelerated nuclei initiate a sequence of processes within the binary system and its surrounding . as we have shown above nuclei
suffer complete disintegration if injected within massive binary systems containing luminous stars characterized by dense stellar winds ( wr type stars ) .
unstable neutrons , from photo - disintegration of these nuclei , decay at some distance from the binary system . depending on their energy
, they can decay within the stellar wind region ( the wind cavity ) or outside the stellar wind shock , i.e. within the open cluster .
protons , from neutrons decaying within the stellar wind , are expected to suffer adiabatic energy losses during the fast expansion of the wind .
these protons are also partially advected with the wind to the open cluster .
neutrons with large enough energies can also decay outside the wind cavity in the volume of the open cluster .
protons from their decay diffuse gradually through the open cluster suffering some collisions with the distributed matter . on the other hand , primary nuclei , accelerated within the binary system , and also protons from their fragmentation are captured by the magnetic field in the dense stellar wind .
these hadrons can interact with the matter of the wind . as a result of these interactions , an additional population of neutrons
is produced not far from the vicinity of the binary system where the matter in the stellar wind is still relatively dense .
these neutrons decay within or outside the wind cavity as described above .
we conclude that the simple scenario which postulates acceleration of hadrons within the massive binary system immersed in the open cluster provides variety of conditions for production of high energy radiation in different parts of open cluster characterized by different conditions , e.g the wind cavity of the binary system and dense medium of the open cluster .
we are interested in the @xmath0-ray and neutrino emission produced in such scenario in the context of recent observations of the tev @xmath0-ray emission from open clusters ( e.g. westerlund 2 and the carina complex ) .
let us at first estimate whether charged hadrons ( accelerated within the binary system and products of their fragmentation ) can be captured in the stellar wind .
we compare the larmor radius of charged hadrons with a specific energy with the characteristic distance scale which is the distance at which hadrons are located from the binary system , @xmath105 . the larmor radius of hadrons with the lorentz factor @xmath106 is given through the expression @xmath107 cm , where @xmath108 g is the magnetic field strength in the wind at the distance @xmath105 from the binary system . the magnetic field in the stellar wind
is expected to have complicated structure as a function of the distance from the star ( dipolar , radial and at farther distances toroidal ) .
this structure becomes even more complicated in the case of stars within the compact binary system when both stars have large velocities .
we assume that the magnetic field is radial at distances below @xmath109 , i.e. in this region @xmath110 . at larger distances
the magnetic field becomes toroidal .
then , for distances greater than @xmath109 , the magnetic field strength can be approximated by @xmath111 g , where @xmath112 g is the surface magnetic field of the wr type star . applying this simple scaling ,
the condition for capturing of protons , @xmath113 , is fulfilled for @xmath114 .
protons with energies fulfilling this condition are captured in the stellar wind .
they are expected to suffer strong adiabatic energy losses during gradual expansion of the wind from the star . on the other hand , hadrons with larger energies
can leave the stellar wind region ( and the wind cavity ) without significant energy losses .
they also propagate almost along the straight lines without significant collisions with the matter of the wind .
however , hadrons with such large energies are not expected to be accelerated at the collision region within the binary system ( see table 1 ) .
hadrons captured in the stellar wind lose energy on the adiabatic process .
we can determine the lorentz factor of hadrons at a specific distance from the binary system by taking into account adiabatic and collisional energy losses , @xmath115 where @xmath54 is given by eq .
[ eq8 ] , and @xmath116 is the in - elasticity coefficient in proton - proton collisions . depending on the parameters of the considered scenario , either adiabatic losses or collision losses
determine the lorentz factors of hadrons at a specific distance of the wind from the binary system .
the lorentz factors of hadrons at a specific distance from the binary system , in the case of only collisional and also collisional and adiabatic energy losses , for three example stars within the binary systems @xmath84 car , wr 20a and wr star in cyg x-3 , are shown in fig .
3 . there are huge differences in the values of these lorentz factors with and without adiabatic energy losses . therefore , adiabatic energy losses play an important role in the process of production of high energy radiation in the open clusters . 6.2truecm
in this section we calculate the spectra of @xmath0-rays and neutrinos produced within the wind cavity .
as shown above , primary nuclei are completely fragmented on separate nucleons ( protons and neutrons ) in collisions with the matter and radiation already within the binary systems .
these nucleons are accelerated with the power law spectrum .
protons from disintegration of nuclei are captured by the magnetic field of the stellar winds .
they are advected from the binary system with the velocity of the wind .
they produce high energy radiation in collisions with the matter of the stellar wind during their propagation within the wind cavity .
these @xmath0-rays are produced isotropically since the production mechanism ( hadronic collisions ) is independent on the radiation field of the massive star .
on the other hand , neutrons , extracted from nuclei , move rectilinearly through the wind cavity .
they decay at distances from the binary system which depend on their lorentz factors .
secondary protons , from decaying neutrons , are also captured in the stellar wind .
all these secondary protons suffer collisional and adiabatic energy loses due to the expansion of the wind . in order to obtain the @xmath0-ray spectra , escaping from the wind cavity region to the observer
, we have to consider possible absorption of @xmath0-rays produced in hadronic collisions in the radiation field of the massive stars .
therefore , we calculate the optical depths for @xmath0-rays .
these absorption effects has to be taken into account when calculating the @xmath0-ray spectra produced by hadrons not far from the binary system .
6.5truecm significant amount of @xmath0-rays produced in hadronic collisions ( as described above ) originate close enough to the massive star that their absorption in the stellar radiation can play important role .
the details of the absorption process are determined by the geometry of the binary system ( production place of @xmath0-rays in respect to the star ) and the parameters of the massive star .
its importance in the case of binary systems without precise orbital parameters can be approximately evaluated by calculating the average reduction factors for @xmath0-ray photons , i.e. the probability of escape averaged over the whole range of possible propagation angles in respect to the direction towards the star .
we calculate the reduction factors for @xmath0-rays injected isotropically at the distance r from the massive star from , @xmath117 where @xmath118 is the angle dependent optical depth for the @xmath0-ray photon injected at the distance r , @xmath119 , and at the angle @xmath120 in respect to the direction defined by the injection place and the center of the star ( see e.g. bednarek 1997 , bttcher & dermer 2005 , dubus 2006 , zdziarski et al .
these optical depths can be easily re - scaled for stars with different parameters following simple prescription given by eq . 2 in bednarek & pabich ( 2010 ) .
the absorption process concerns mainly @xmath0-rays with energies above @xmath121 gev .
we calculate the optical depths for @xmath0-rays as a function of their energies and the distance from the massive star . as we noted above
, the exact values of the optical depths depend on the propagation angle of @xmath0-rays in respect to the direction towards the star .
therefore , we show the reduction factor values averaged over the injection angles assuming isotropic injection of @xmath0-rays from a point source at a specific distance from the star . the results of these example calculations are shown in fig . 4 for the wr 20a and eta carinae massive stars .
it is clear that @xmath0-rays with energies above several gev are efficiently absorbed even if they are produced at relatively large distances from the binary system .
we consider the production of radiation by protons from nuclei accelerated within the binary system .
these protons are advected outside the binary system with the velocity of the stellar wind .
the number of hadrons which interact during the advection process with the stellar wind , on the range of distances from the binary system between @xmath93 ( considered as the injection place ) and r , is determined by the factor @xmath122 $ ] , where @xmath123 ( see eq . 8) , and @xmath124 cm .
the spectrum of hadrons , which interact at the distance @xmath105 from the binary system , is @xmath125 where @xmath126 is the injection rate of hadrons with the lorentz factors @xmath27 from the binary system , @xmath127 is the jacobian of transformation of hadron energy ( see eq .
[ eq9 ] ) . in order to calculate the @xmath0-ray and neutrino spectra produced by protons
, we have to integrate the above injection rate of protons ( eq . [ eq11 ] ) over the whole dimension of the wind cavity and over the spectrum protons ( given by eq .
11 ) , @xmath128 the spectra of @xmath0-rays and neutrinos are calculated by applying the scaling break model for hadronic collisions developed by wdowczyk & wolfendale ( 1987 ) , which is suitable in the considered energy range of relativistic hadrons .
the distribution of protons within the wind cavity , @xmath129 , is given by eq .
the spectra are calculated for the range of energies of hadrons @xmath130 and @xmath131 as reported in table 1 .
@xmath132 and @xmath133 are the energies of @xmath0-ray photons and neutrinos , respectively .
the example spectra of @xmath0-rays , calculated for two binary systems ( wr 20a and eta carinae ) , are shown in fig .
we show the @xmath0-ray spectra with ( thick curves ) and without ( thin curves ) absorption effects in the radiation of the massive stars .
as expected , the absorption of @xmath0-rays produced by protons , extracted from nuclei within the inner part of the wind cavity , is very strong .
6.5truecm 6.5truecm neutrons , extracted from nuclei within the binary system , propagate along the straight lines and gradually decay into protons in the dense medium surrounding the binary system . depending on the lorentz factors of neutrons , these protons appear within the wind cavity or outside the wind cavity ( i.e. within the open cluster ) . in this subsection , we calculate the spectra of protons from neutrons decaying within the wind cavity .
these spectra are used for the calculation of @xmath0-ray and neutrino spectra produced within the wind cavity .
the optical depth for neutrons , which move along straight lines through the wind cavity , on the interaction with the matter of the wind is equal to @xmath134 , where @xmath135 is defined by eq .
1 . it is a factor of @xmath136 smaller than the optical depth for protons given in table 1 .
therefore , the interaction rate of neutrons moving along the straight lines through the wind cavity can be neglected for the considered range of the parameters characterizing the binary systems .
the neutron decay rate at the distance , @xmath137 , from the binary system ( equal to the rate of creation of protons from their decay ) is given by , @xmath138 where @xmath139 is the neutron injection rate from the binary system ( equal to the acceleration rate of nuclei ) , @xmath140 s is the neutron decay time , and the lorentz factor of secondary protons ( @xmath141 ) is assumed to be equal to the lorentz factor of neutrons ( @xmath106 ) .
protons , from decaying neutrons , are confined by the magnetic field of the stellar wind .
therefore , they suffer adiabatic and collisional energy losses as considered for protons from fragmentation of nuclei in the previous section .
the collision rate of protons at the distance , @xmath142 , from the binary system , is given by the formula similar to that given by eq .
[ eq11 ] , @xmath143 where @xmath144 , and @xmath145 .
the @xmath0-ray and neutrino spectra , produced by these secondary protons in the wind cavity , can be calculated by integrating the above formula over the spectrum of protons , their creation distance in the cavity measured from binary system , @xmath137 , and their interaction distance @xmath105 , @xmath146 where @xmath147 is given by eq .
[ eq14 ] , and the spectra of @xmath0-rays and neutrinos are calculated as described below eq .
we have performed calculations of the @xmath0-ray spectra produced by protons from neutrons decaying within the wind cavity for the parameters of two considered binary systems ( see thick solid curves in fig .
these spectra are clearly below the @xmath0-ray spectra produced by protons directly extracted from nuclei .
this effect is due to the appearance of protons ( from decaying neutrons ) at a relatively large distances from the binary system where the density of the stellar wind is already low . on the other hand ,
these @xmath0-ray spectra show much smaller absorption effects in the stellar radiation due to their production at larger distances from the massive star .
the @xmath0-ray spectra produced in the wind cavity by two considered above populations of protons are compared with the available observations of the two considered binary systems in the gev - tev energy range and with the sensitivity of the planned cta . as mentioned in the introduction , the tev @xmath0-ray source has been reported in the direction of the open cluster westerlund 2 which contains binary system wr 20a ( aharonian et al .
2007 , abramowski et al . 2011 ) .
the @xmath0-ray spectrum of the source in the direction of this open cluster has the spectral index 2.58 in the energy range @xmath11 - 10 tev .
the nature of this source is at present unknown .
it is supposed that this emission can be related to the binary system wr 20a , the pulsar wind nebula ( pwne ) around psr j1022 - 5746 or maybe also to the dense molecular clouds present in this open cluster .
we compare the @xmath0-ray spectrum expected from the wind cavity around the binary system wr 20a with the above mentioned observations .
the results are shown in fig .
6 . in order to be consistent with the tev observations , reasonable value for the energy conversion efficiency from the stellar wind to relativistic nuclei
is required ( @xmath148 ) . according to our calculations
, @xmath0-rays produced in the wind cavity should be mainly responsible for a part of this emission at the highest observed energies , i.e. @xmath1 10 tev . as we show below ,
the lower energy part of the @xmath0-ray emission from westerlund 2 could be either produced by protons which escape from the wind cavity into dense regions of the open cluster , as considered in sect . 7 , or it comes from the other sources present within the westerlund 2 ( e.g. other massive binary systems or pwne ) .
we also compare our calculations with the observations of the binary system eta carinae and the surrounding dense complex carina nebula .
the source in this direction has been detected in the gev energies by the agile ( tavani et al .
2009 ) and fermi telescopes ( abdo et al .
@xmath0-ray emission from this source shows two components spectrum , first one extending to a few gev and the second one extending up to @xmath1 100 gev ( farnier et al .
the highest energy component show evidences of variability with the orbital period of the binary system ( walter & farnier 2011 , reitberger et al .
this component is expected to be produced within the binary system . on the other hand
, the lower component seems to be steady .
we compare the @xmath0-ray spectrum expected in terms of our model from the wind cavity around the binary system in fig .
the spectrum is normalized to the observed lower energy component extending to a few gev .
since the @xmath0-ray emission produced within the wind cavity is strongly absorbed in the stellar radiation , the @xmath0-ray spectrum expected in our model in the tev energy range is steep ( spectral index close to -4 , see fig . 6 ) .
therefore , this emission is clearly below the present upper limit reported by the hess collaboration in the tev energy range ( abramowski et al .
moreover , due to the steepness of this spectrum , future detection of this emission component by the cta is rather problematic ( fig .
in this section we calculate the @xmath0-ray ( and neutrino ) spectra produced by protons which originate as a decay products of neutrons within the wind cavity , but they are advected with the stellar wind into the surrounding open cluster .
we also calculate the @xmath0-ray emission from relativistic protons which originate from the most energetic neutrons decaying directly outside the wind cavity , i.e. within the open cluster .
it is assumed that the open cluster has the basic parameters of the order of , the density of matter 10 @xmath11 , the magnetic field strength @xmath149 g , and the radius 20 pc . as we show below
, the combination of these parameters determines the escape of protons from the open cluster and , as a consequence the @xmath0-ray production rate and their spectra in collisions with the matter of the open cluster .
a part of protons , which is advected with the stellar wind , can arrive up to the wind cavity border .
they can be injected into the dense open cluster surrounding the binary system .
however , only protons from decaying neutrons can still have large energies allowing them to produce the tev @xmath0-rays and neutrinos .
protons , from direct disintegration of nuclei , appear close to the binary system and therefore suffer huge adiabatic energy losses .
the spectrum of protons injected from the wind cavity ( with the radius @xmath150 ) into the open cluster can be calculated from , @xmath151 where @xmath152 is given by eq .
13 and the jacobian is @xmath153 .
we calculate the spectra of protons advected into the open cluster for two considered binary systems ( wr 20a and eta carinae ) .
two different values for the density of matter within the open cluster , which determine the outer radius of the wind cavity , are assumed .
the results are shown in fig .
7 ( see dashed curves ) .
as expected , for denser matter within the open cluster ( corresponding to smaller radius of the wind cavity ) the spectra have larger intensities and the maximum in the spectra are shifted to lower energies .
these effects are due to the lower adiabatic energy losses in the case of wind cavities with smaller radii . in order to calculate the @xmath0-ray and neutrino spectra produced by these protons in the open cluster
, we have to estimate their residence time within the open cluster , due to the diffusion process , and compare it with the collisional time scale .
we assume that protons diffuse in the turbulent medium of the open cluster with the rate well determined by the bohm diffusion coefficient , @xmath154 @xmath80s@xmath5 , where the magnetic field strength within the open cluster is @xmath155 g. then , the average diffusion time scale of protons within the open cluster , with the characteristic dimension @xmath156 pc , is estimated on , @xmath157 s. the collision time of these protons can be estimated from @xmath158 yrs , where the density of surrounding matter is @xmath159 @xmath11 .
the interaction process of these protons within the open cluster depends on the lifetime of the massive binary system and the diffusion time scale of protons from the open cluster .
the time scale for collisions of protons with the matter is , @xmath160 s. by comparing the diffusion time scale with collision time scale , we estimate critical energy of relativistic protons , @xmath161 below which they can interact efficiently within the open cluster .
we assume that protons with such energies reaches steady state equilibrium in which the rate of proton injection equals the rate of proton interaction .
we calculate the @xmath0-ray and neutrino spectra produced by protons with the spectra shown in fig . 7 in the range limited by eq .
note that the winds from the massive stars in the binary system do not change significantly the distribution of the matter within the open cluster since the mass swept - up by the stellar wind is @xmath162 m@xmath4 .
this is only a small amount of the total mass of the open cluster which is typically expected in the range @xmath65 m@xmath4 to a few @xmath163 m@xmath164
. the @xmath0-ray and neutrino spectra produced by protons , advected from the wind cavity , are calculated by integration of the above derived proton spectrum , @xmath165 where @xmath166 is given by eq .
( 16 ) .
the most energetic neutrons , extracted from nuclei within the binary system , can decay directly outside the wind cavity , since their propagation distance @xmath167 becomes comparable to the radius of the wind cavity .
the spectrum of protons from decay of these neutrons is given by , @xmath168 where @xmath169 is the spectrum of neutrons extracted from nuclei , @xmath170 is the radius of the wind cavity ( see eq . 2 ) .
this neutron spectrum has exactly the same shape as the spectrum of accelerated nuclei , i.e. the power law type with the spectral index equal to 2 ( see section . 3 ) .
it is assumed that lorentz factors of neutrons are equal to lorentz factors of parent nuclei .
proton spectra , from neutrons decaying inside the open cluster ( i.e. outside the wind cavity ) , are shown for both considered binary systems in fig .
7 ( dotted curves ) .
relative contribution of protons from these neutrons , in respect to protons advected from the wind cavity , depends on the parameters defining the acceleration scenario , the wind of the companion stars and the parameters of the open cluster .
the largest contribution is expected in the case of binary systems containing massive stars with strong surface magnetic field ( large lorentz factors of nuclei ) and relatively weaker stellar winds ( smaller radius of the wind cavity ) .
therefore , more neutrons should decay directly within the open cluster ( outside the wind cavity ) in the case of the wr type stars than in the case of the most massive stars of the eta carinae type which have very large mass loss rates but weaker surface magnetic fields .
the @xmath0-ray and neutrino spectra , produced by these protons , are calculated by integrating formula given by eq .
6.5truecm we calculate the @xmath0-ray spectra produced by protons in hadronic collisions with the matter of the open cluster , applying their diffusion and interaction model as described above . as before ,
we consider the open clusters which contain the wr 20a and eta carinae binary systems .
the calculations of the @xmath0-ray spectra are performed for the same parameters of the acceleration scenario as discussed for the @xmath0-ray production in the wind cavity regions of these two binaries .
the results are shown for two different values of the parameter @xmath171 which determines the conditions within the open cluster . in the case of the open cluster containing the binary system
wr 20a , the @xmath0-ray spectrum expected from the open cluster westerlund 2 is consistent with the positive detection of this open cluster by the hess collaboration ( abramowski et al .
2011 ) , provided that the parameter @xmath172 ( see fig .
note that @xmath0-rays produced in this case are expected to contribute mainly to the energy range around @xmath173 tev .
this is clearly below the energy range at which the @xmath0-ray emission from the wind cavity is expected .
@xmath0-rays , expected from the open cluster in terms of our model for westerlund 2 , are predicted to be still detected by the cta provided that the parameter @xmath174 .
we have also compared the predictions of our model with the observations of the carina complex containing the eta carinae binary system .
hess collaboration has derived an upper limit on the @xmath0-ray flux from the extended source from this direction ( abramowski et al .
2012b ) , which is about a factor of five higher than the upper limit on the point source towards the eta carinae itself .
we have compared the calculated @xmath0-ray spectra with the hess upper limit in fig .
these spectra are still consistent with the upper limit provided that the value of the parameter @xmath175 .
@xmath0-rays produced in terms of our model should contribute only to the sub - tev energy range .
we conclude that the @xmath0-ray emission , produced by relativistic protons in the carina complex , is expected to be still detectable by the cta provided that the open cluster is characterized by the parameter @xmath176 .
however , this @xmath0-ray component is expected to extend only through the low energy range of the cta sensitivity , i.e. at @xmath1100 gev .
detection of neutrinos , produced in hadronic interactions between relativistic protons and the matter of the stellar wind and/or open cluster , could provide additional constraints on the high energy processes in the considered scenario .
therefore , we calculate the neutrino spectra from the wind cavity and the open cluster applying derived above spectra of relativistic protons in these regions . the neutrino spectra from the wind cavity ( on the left ) and the open cluster ( on the right ) are shown in fig . 9 for the wr 20a binary system ( solid ) and eta carinae binary system ( dashed ) .
they are obtained for the parameters of the model as used for the corresponding @xmath0-ray spectra shown above , i.e. for the energy conversion inefficiencies as described in fig .
6 and the distances to these binary systems equal to @xmath1 2 kpc .
these spectra are compared with the atmospheric neutrino background ( anb ) and the upper limit on the neutrino flux from discrete sources reported by the antares collaboration ( adrian - martinez et al .
2012 ) and the 5 yr sensitivity of the icetop + icecube ( aartsen et al .
the neutrino spectra , expected from these two binary systems immersed within the open clusters , are on the level of the atmospheric neutrino background . however , they are about an order of magnitude below the present upper limit of the flux from discrete sources by antares collaboration .
it looks that nuclei are not energetic enough in order to produce neutrino fluxes which could be potentially detected by the ice cube telescope .
we have also calculated the expected neutrino flux produced by relativistic protons within the open cluster applying the normalization obtained from the comparison of the @xmath0-ray fluxes with observations of these two open clusters ( see fig .
these neutrino fluxes are predicted on much lower level than those expected from the wind cavities of these binary systems .
we conclude that considered binary systems are not expected to produce detectable neutrino fluxes from the directions of the open clusters westerlund 2 and carina complex .
note however , that neutrino fluxes should be larger if the specific open cluster contains many massive binary systems .
therefore , our results do not completely exclude future detection of neutrinos from massive stellar clusters at distances of a few kpc from the vicinity of the sun .
star forming regions contain plenty of extreme objects , such as massive binary systems and remnants of their evolution ( supernova remnants , pwne , x - ray binaries ) .
these objects might be responsible for the acceleration of hadrons to tev - pev energies .
in fact , gev - tev @xmath0-ray emission has been already detected from a few open clusters . in this paper
we concentrate on the processes due to the presence of massive binary systems in dense regions ( open clusters ) .
we have formulated a model for the interaction of the binary system with the matter of the open cluster .
it is postulated that nuclei ( from helium to oxygen ) can be efficiently accelerated within the binary system .
we show that these nuclei are disintegrated in the dense stellar radiation and the matter of the stellar wind injecting relativistic protons and neutrons .
the fate of these secondary particles is followed in detail in the region of the stellar wind cavity and in the surrounding dense open cluster .
we calculate the expected @xmath0-ray and neutrino emission produced in the interaction of these particles with the matter of the stellar wind and surrounding open cluster .
the results of calculations are shown for the case of the two well known massive binary systems ( wr 20a and eta carinae ) , which have been recently reported as a @xmath0-ray sources in the gev - tev energy range .
our calculations show that the largest fluxes of @xmath0-rays are produced by protons close to the binary system where the density of stellar wind is the largest .
however , the absorption of gamma - rays in the stellar radiation field has the important effect on the gamma - ray spectrum , produced by protons extracted directly from the nuclei .
therefore , the @xmath0-ray spectra produced in these regions are clearly steeper than the injection spectra of protons ( equal to the spectrum of accelerated nuclei ) .
on the other hand , @xmath0-ray spectra , from hadronic collisions of protons which are decay products of neutrons extracted form nuclei , are mainly produced at larger distances from the binary systems since neutrons move ballisticaly through the wind cavity and decay at relatively large distances .
they interact with the matter of the wind with lower density but do not suffer strong absorption in the stellar radiation field . in total , these @xmath0-ray spectra have clearly lower level than spectra produced in previous process
. significant amount of protons , which appeared in the wind cavity as a result of decay of neutrons , is advected with the stellar wind to the open cluster .
we calculate the spectra of these protons taking into account their adiabatic and collisional energy losses .
protons , directly extracted from nuclei , can not be advected to the open cluster with large energies due to the huge adiabatic energy losses . therefore
, the radiation produced by them can be safely neglected .
another population of relativistic protons is provided by the most energetic neutrons which can decay directly into the open cluster , i.e. outside the wind cavity .
we take these protons into account when calculating the radiation produced in the open cluster .
the fate of relativistic protons in the open cluster depends on the cluster parameters .
protons diffuse outside the cluster and interact with the cluster matter .
we show , that in the case of bohm diffusion approximation , the escape / interaction conditions of protons depend on the parameter @xmath171 , which is the combination of the parameters characterizing the open cluster , its radius , magnetic field strength and density of matter .
we show that for likely parameters of the open cluster , protons with largest expected energies escape from the cluster practically without interaction with the matter . only lower
energy protons are captured in the open cluster and lose energy on the production of @xmath0-rays and neutrinos .
note that protons escaping from the clusters are expected to have lorentz factors in the range @xmath177 .
therefore , we expect that open clusters , containing massive binary systems , might become interesting sources of relativistic protons in the galaxy .
we confronted the @xmath0-ray emission , expected in terms of this model , with the observations of the open clusters containing binary systems wr 20a ( westerlund 2 ) and eta carinae ( carina complex ) .
it is concluded that protons within the specific open cluster can contribute to the observed tev @xmath0-ray spectrum ( mainly at its lower energy part ) observed from westerlund 2 .
this @xmath0-ray emission is consistent with the upper limits on the tev flux from carina complex .
we also determine the minimum constraints on the parameters of these two clusters for which they will be detectable by the planned cta .
we have also calculated the neutrino spectra expected in this model for these two binary systems .
unfortunately , these neutrino fluxes are about two orders of magnitude below the present upper limit on neutrino emission from discrete sources provided by antares collaboration ( adrian - martinez et al . 2012 ) .
these neutrinos will be also difficult to observe with the icecube telescope since their fluxes , produced in the wind cavity regions , are comparable to the atmospheric neutrino background .
they are clearly below this background in the case of neutrinos produced in the open cluster itself .
in fact many massive binary systems can be present at the same time in a specific open cluster .
therefore , in principle larger @xmath0-ray and neutrino fluxes might be expected from specific open clusters than calculated here fluxes from isolated binary systems .
however the environment of the open cluster with many massive binary systems might be very complicated due to likely interaction between different wind cavities .
therefore , simple re - scaling of the @xmath0-ray and neutrino fluxes by the number of massive binaries in specific open cluster seems to be not adequate .
a more complicated scenario should be considered in such a case . | tev gamma - ray emission has been recently observed from direction of a few open clusters containing massive stars .
we consider the high energy processes occurring within massive binary systems and in their dense environment by assuming that nuclei , from the stellar winds of massive stars , are accelerated at the collision region of the stellar winds .
we calculate the rates of injection of protons and neutrons from fragmentation of these nuclei in collisions with stellar radiation and matter of the winds from the massive companions in binary system .
protons and neutrons can interact with the matter , within the stellar wind cavity and within the open cluster , producing pions which decay into @xmath0-rays and neutrinos .
we discuss the detectability of such @xmath0-ray emission by the present and future cherenkov telescopes for the case of two binary systems eta carinae , within the carina nebula , and wr 20a , within the westerlund 2 open cluster .
we also calculate the neutrino fluxes produced by protons around the binary systems and within the open clusters .
this neutrino emission is confronted with antares upper limits on the neutrino fluxes from discrete sources and with the sensitivity of icecube . |
the nucleon elastic electromagnetic form factors are fundamental quantities needed for an understanding of the nucleon s electromagnetic structure .
the sachs electric , @xmath8 , and magnetic , @xmath9 , form factors @xcite , defined in terms of linear combinations of the dirac and pauli form factors , are of particular physical interest , as their evolution with @xmath3 , the square of the four - momentum transfer , is related to the spatial distribution of charge and current within the nucleon . as such , precise measurements of these form factors over a wide range of @xmath3 are needed for a quantitative understanding of the electromagnetic structure not only of the nucleon , but also of nuclei ( e.g. , @xcite ) .
further , in the low - energy regime of the nucleon ground state , the underlying theory of the strong interaction , quantum chromodynamics ( qcd ) , can not be solved perturbatively , and a proper description of even the static properties of the nucleon , the lowest stable mass excitation of the qcd vacuum , in terms of the qcd quark and gluon degrees of freedom still stands as one of the outstanding challenges of hadronic physics .
indeed , one of the most stringent tests to which non - perturbative qcd ( as formulated on the lattice or in a model of confinement ) can be subjected is the requirement that the theory reproduce experimental data on the nucleon form factors ( e.g. , @xcite ) . because of the lack of a free neutron target
, the neutron form factors are known with less precision than are the proton form factors , and measurements have been restricted to smaller ranges of @xmath3 .
a precise measurement of the neutron electric form factor , @xmath6 , has proven to be especially elusive as the neutron s integral charge is zero .
prior to the realization of experimental techniques utilizing polarization degrees of freedom , values for @xmath6 were extracted from measurements of the unpolarized quasielastic @xmath1h@xmath10h cross section and the deuteron elastic structure function @xmath11 . those results for @xmath6 deduced from measurements of the quasielastic @xmath1h@xmath10h cross section provided little information on @xmath6 , as all results were consistent with zero over all ranges of @xmath3 accessed , @xmath12 ( gev/@xmath4)@xmath1 ( e.g. , @xcite ) .
similarly , results for @xmath6 deduced from measurements of @xmath11 , although establishing @xmath13 for @xmath14 ( gev/@xmath4)@xmath1 , were plagued with large theoretical uncertainties ( @xmath15% ) related to the choice of an appropriate @xmath16-potential for the deuteron wavefunction ( e.g. , @xcite ) . with the advent of high duty - factor polarized electron beam facilities and
state - of - the - art polarized nuclear targets and recoil nucleon polarimeters , experimental efforts over the past 15 years have now yielded the first precise determinations of @xmath6 .
our experiment @xcite was designed to extract the neutron electric to magnetic form factor ratio , @xmath0 , from measurements of the neutron s recoil polarization in quasielastic @xmath1h@xmath2h kinematics at three @xmath3 values of 0.45 , 1.13 , and 1.45 ( gev/@xmath4)@xmath1 .
these results were published rapidly by madey _
et al_. @xcite ; here we provide a more detailed report of the experiment and analysis procedures .
the remainder of this paper is organized as follows .
we begin , in section [ sec : neutron - form - factors ] , with a brief overview of the experimental techniques utilizing polarization degrees of freedom that have been employed for measurements of the neutron form factors .
we continue with an overview of our experiment in section [ sec : experiment ] , and then discuss our neutron polarimeter in section [ sec : neutron - polarimeter ] .
details of the analysis procedure are discussed in section [ sec : data - analysis ] .
our final results are then presented in section [ sec : final - results ] and compared with selected theoretical model calculations of the nucleon form factors .
finally , we conclude with a brief summary in section [ sec : summary - conclusions ] . a more detailed account of the discussion that follows may be found in @xcite .
we will use the following notation for the electron kinematics : @xmath17 will denote the four - momentum of the initial electron , @xmath18 will denote the four - momentum of the scattered electron , @xmath19 will denote the electron scattering angle , @xmath20 will denote the energy transfer , @xmath21 will denote the three - momentum transfer , and @xmath22 will denote the square of the spacelike four - momentum transfer in the high - energy limit of massless electrons .
the electron scattering plane is defined by @xmath23 and @xmath24 .
the polarization of the recoil nucleon , @xmath26 , in elastic polarized - electron , unpolarized - nucleon scattering is well - known to be of the form @xcite @xmath27 where @xmath28 denotes the unpolarized cross section , @xmath29 denotes the helicity - independent recoil polarization , @xmath30 denotes the helicity - dependent recoil polarization , and @xmath31 denotes the electron helicity .
the polarization is customarily projected onto a @xmath32 unit vector basis , with the longitudinal component , @xmath33 , along the recoil nucleon s momentum ; the normal component , @xmath34 , perpendicular to the electron scattering plane ; and the transverse component , @xmath35 , perpendicular to the @xmath33-component in the scattering plane . in the one - photon exchange approximation , @xmath36 , and @xmath30
is confined to the scattering plane ( i.e. , @xmath37 ) . the transverse , @xmath38 , and longitudinal , @xmath39 , components are expressed in terms of kinematics and nucleon form factors as @xcite @xmath40g_{m}^{2 } } , \\ % p^{(h)}_{n } & = & 0 , \\ p^{(h)}_{\ell } & = & p_{e}\frac{2g_{m}^{2}\tau\sqrt{(1+\tau ) + ( 1+\tau)^{2}\tan^{2}\displaystyle{\frac{\theta_{e'}}{2 } } } \tan\displaystyle{\frac{\theta_{e'}}{2 } } } { g_{e}^{2}+\left[\tau+2\tau(1+\tau ) \tan^{2}\displaystyle{\frac{\theta_{e'}}{2}}\right]g_{m}^{2 } } , \nonumber \\\end{aligned}\ ] ] where @xmath41 denotes the electron beam polarization , @xmath42 , and @xmath43 denotes the nucleon mass .
access to both @xmath44 and @xmath45 via a secondary analyzing reaction in a polarimeter is highly advantageous , as the analyzing power of the polarimeter , denoted @xmath46 , and @xmath41 cancel in the @xmath47 ratio , yielding a measurement of @xmath48 that is relatively insensitive to systematic uncertainties associated with these quantities . for the case of the neutron form factor ratio ,
as suggested by arnold , carlson , and gross @xcite and first implemented experimentally by ostrick _
et al_. @xcite , a vertical dipole field located ahead of a polarimeter configured to measure an up - down scattering asymmetry sensitive to the projection of the recoil polarization on the @xmath35-axis permits access to both @xmath38 and @xmath39 . during transport through the magnetic field , the recoil polarization vector will precess through some spin precession angle @xmath49 in the @xmath35-@xmath33 plane , leading to a scattering asymmetry , @xmath50 , which is sensitive to a mixing of @xmath38 and @xmath39 , @xmath51 in the above , @xmath52^{1/2}$ ] , and we define the phase - shift parameter @xmath53 according to @xmath54 the above formalism is directly applicable to an extraction of the proton form factor ratio , @xmath56 , from measurements of the proton s recoil polarization in elastic @xmath57h@xmath58 scattering .
an extraction of the neutron form factor ratio , @xmath0 , from measurements of the neutron s recoil polarization in quasielastic @xmath1h@xmath2h scattering is , however , complicated by nuclear physics effects , such as final - state interactions ( fsi ) , meson exchange currents ( mec ) , isobar configurations ( ic ) , and the structure of the deuteron .
the pioneering study of the sensitivity of the quasielastic @xmath1h@xmath2h reaction to the neutron form factors , reported by arenhvel @xcite , revealed that for perfect quasifree emission of the neutron ( i.e. , neutron emission along the three - momentum transfer @xmath59 ) , @xmath38 is proportional to @xmath6 , but is relatively insensitive to fsi , mec , ic , and the choice of the @xmath16-potential for the deuteron wavefunction .
a more detailed study of the @xmath1h@xmath2h reaction reported by arenhvel , leidemann , and tomusiak @xcite found that these results also apply to @xmath39 .
similar findings were subsequently reported by @xcite . in appendix
[ sec : appendix - a ] , we present a detailed discussion of the formalism for the kinematics and recoil polarization observables for the quasielastic @xmath1h@xmath2h reaction
. in particular , we provide there a definition for @xmath60 , the polar angle between the proton momentum and @xmath59 in the recoiling neutron - proton center of mass frame ( hereafter , @xmath61-@xmath62 c.m . frame ) , a variable to which we will refer frequently throughout this paper .
( perfect quasifree emission of the neutron is defined by @xmath63 . )
we follow this , in appendix [ sec : appendix - b ] , with a discussion of the sensitivity of the recoil polarization components to fsi , mec , ic , and the choice of the @xmath16-potential for the deuteron wavefunction at and away from perfect quasifree emission . the cross section in the one - photon exchange approximation for elastic polarized - electron , polarized - nucleon scattering is well - known to be of the form @xcite @xmath65.\ ] ] here , @xmath66 and @xmath67 denote , respectively
, the polar and azimuthal angle between the target nucleon polarization vector and @xmath59 , and @xmath68 denotes the polarized - electron , polarized - nucleon beam - target asymmetry , which is a function of kinematics and the nucleon form factors .
the sensitivity of @xmath69 to the form factors is enhanced if the target polarization is oriented in the electron scattering plane either parallel or perpendicular to @xmath59 ; in the former ( latter ) case , the expression for @xmath69 is identical to that for @xmath70 ( @xmath38 ) and will be denoted @xmath71 ( @xmath72 ) .
similar to the recoil polarization technique , measurements of both @xmath72 and @xmath71 are desirable as the target polarization ( analog to the analyzing power ) and beam polarization cancel in the @xmath73 ratio , again yielding a measurement of @xmath48 that is relatively free of systematic uncertainties .
[ cols="<,<,^,^,^,^,^",options="header " , ] [ tab : summary - final - results ] uncertainties in the neutron spin precession angle were estimated @xcite via a calculational scheme that employed the reconstructed kinematics from the experimental data as the source of the neutron momentum vectors incident on the charybdis dipole field .
spin vectors were transported through the field using the same magnetic spin transport algorithms developed for our two simulation programs .
this technique provided a measure of the sensitivity of the precession angle to details of the field map .
the uncertainties in the mean values of the precession angles derived from these studies ( at the level of @xmath76 ) were combined in quadrature with two other sources of uncertainty .
first , as discussed in section [ sec : neutron - polarimeter - charybdis ] , we observed small differences between the measured field integrals for opposite magnet polarities and also between the field integrals derived from our measured maps and the calculated ` tosca ` maps .
these uncertainties were estimated to be on the level of @xmath77 .
second , as also discussed in section [ sec : neutron - polarimeter - charybdis ] , the field was mapped only along the central axis ; therefore , we assigned further uncertainties ( at the level of @xmath76 ) for incomplete knowledge of the field beyond the central axis .
our best estimates of the total uncertainties in the precession angle were then propagated through the form factor ratio , eq.([eq : phase - shift - delta ] ) .
radiative corrections were calculated specifically for the kinematics of this experiment by afanasev _
et al_. @xcite .
the primary effect of radiative corrections on the recoil polarization components @xmath38 and @xmath39 was found to be depolarization of the electron such that both components of the recoil polarization should be increased by @xmath78% , @xmath79% , and @xmath80% at @xmath81 , 1.132 , and 1.450 ( gev/@xmath4)@xmath1 , respectively ; however , these corrections nearly cancel in the form factor ratio such that the net effect is small at @xmath81 ( gev/@xmath4)@xmath1 and negligible at the two higher @xmath3 points . the residual non - cancellations of the corrections in the form factor ratio were taken to be the systematic uncertainties we quote in table [ tab : systematic - uncertainties ] .
the timing calibrations we deemed suitable for certain running conditions ( e.g. , periods in between changes to the high - voltages for the pmts ) were obtained using a subset of the data for that particular running period . to assess the dependence of our results for the scattering asymmetries on the choice of the subset of data employed for the timing calibration , various calibrations were generated from different subsets of the available data .
excellent agreement was always found between the results for the scattering asymmetries obtained from analyses using these different calibrations ; however , we did find a @xmath82% sensitivity of our results to the choice of the subset of data employed for the timing calibration .
we deemed two other possible sources of systematic uncertainties to be negligible .
first , we demonstrated quantitatively that our scattering asymmetries were insensitive ( within statistical errors ) to a possible geometric asymmetry in the polarimeter ( i.e. , a spin - averaged `` top - bottom '' asymmetry ) by varying our software energy thresholds on the top ( bottom ) rear array while maintaining a constant threshold on the bottom ( top ) rear array .
second , analysis of our data taken with the `` dummy targets '' ( see section [ sec : scattering - chamber - cryotargets ] ) showed that the level of contamination within our @xmath83 $ ] ns ctof window from scattering in the target cell windows was negligible ( @xmath84% ) . at @xmath87 , 1.132 , and 1.450 ( gev/@xmath4)@xmath1 extracted from the various analyses summarized in table [ tab : summary - final - results ] .
the data points shown for the three analyses at each @xmath88 point have been slightly displaced about the actual @xmath89 value for clarity .
the solid curve is the galster parametrization @xcite . ] our final results for @xmath0 and
@xmath6 extracted from three different analyses are tabulated in table [ tab : summary - final - results ] and compared in fig.[fig : point - pwba - fsi - gen - results ] .
the three analyses are for : ( 1 ) elastic @xmath90 scattering and infinitesimal hms and npol point acceptances ; ( 2 ) quasielastic @xmath1h@xmath2h scattering and acceptance - averaging of the pwba model ; and ( 3 ) quasielastic @xmath1h@xmath2h scattering and acceptance - averaging of the fsi+mec+ic+rc model .
we note that in our first publication @xcite we used values for @xmath91 taken from @xcite ; here we use slightly different values for @xmath91 taken from @xcite .
the total systematic errors we quote for @xmath6 are the quadrature sum of the experimental sytematic errors and the relative uncertainties in @xmath91 .
the availability of precise new data on nucleon form factors has stimulated much more theoretical work in the last few years than we can review here ; our selection of models is not intended to be complete .
although the present experiment is limited to @xmath6 , we believe that comparison with models must consider all four sachs form factors . in this section
, we compare representative models with selected data .
the data from this experiment are shown as filled circles in figs .
[ fig : vmd][fig : qm ] while data from other experiments are shown as open circles .
we selected @xcite for @xmath92 , @xcite for @xmath56 , @xcite for @xmath93 , @xcite for @xmath6 , and @xcite for @xmath91 .
models based upon vector meson dominance ( vmd ) postulate that the virtual photon couples either directly to an intrinsic nucleon core or through the mediation of a neutral vector meson , initially limited to the lowest @xmath94 , @xmath95 , and @xmath96 mesons .
the earliest versions assumed that the core is a structureless dirac particle .
et al_. @xcite assigned the core a form factor and employed a model of the @xmath95 width .
gari and krmpelmann @xcite then permitted different dirac and pauli intrinsic form factors and introduced modifications that ensured consistency with pqcd scaling at large @xmath97 and logarithmic running of the strong coupling constant .
bijker and iachello @xcite adopted the gari and krmpelmann ( gk ) pqcd prescriptions and refit their model to modern data , still using a common intrinsic form factor .
this fit , using a total of 6 free parameters , is compared with the data in fig .
[ fig : vmd ] .
finally , lomon @xcite produced a more flexible set of fits using a model described as `` gk extended '' ; the gkex(02s ) version is also shown in fig .
[ fig : vmd ] .
the lomon model uses two intrinsic form factors , the gk prescription for the pqcd limit , and includes @xmath98 and @xmath99 couplings in addition to the customary @xmath95 , @xmath94 , and @xmath96 couplings .
the @xmath95 width is included but the @xmath100 and @xmath101 structures are not .
the fit achieved by this extended model , with 13 free parameters , is clearly superior , especially at large @xmath97 .
the bijker and iachello model describes the qualitative behavior of @xmath92 , but its transition between @xmath102 at low @xmath97 and the nearly linear decrease for @xmath103 ( gev/@xmath4)@xmath104 is too gradual . nor
does it reproduce the slope in @xmath105 for @xmath106 ( gev/@xmath4)@xmath104 . both of these features are fit well by the lomon model .
unfortunately , the neutron data do not discriminate between these models very strongly .
the bijker and iachello model provides a slightly better fit to the present @xmath6 data , but the lomon fit was performed before these data became available ; it is likely that only a slight parameter adjustment would be needed to achieve a comparable fit without sacrificing the fits to other form factors .
it will be interesting to see whether the rather large values for @xmath107 for @xmath108 ( gev/@xmath4)@xmath104 predicted by the bijker and iachello model are confirmed by upcoming experiments @xcite that will probe @xmath6 to @xmath109 ( gev/@xmath4)@xmath1 .
note , however , that the bijker and iachello fit is systematically above the @xmath91 data for the same kinematics , @xmath108 ( gev/@xmath4)@xmath104 .
the role of the pion in mediation of the long - range nucleon - nucleon interaction clearly demonstrates its importance in understanding form factors for low @xmath97 .
typical pion cloud models describe nucleon form factors using diagrams in which the virtual photon couples to either a bare nucleon core or to the nucleon or the pion loop in a single - pion loop .
some models also permit excitation of the intermediate state and include additional contact terms .
a relatively simple example is the adelaide version @xcite of the cloudy bag model ( cbm ) in which the core is based upon the bag model , intermediate excitation is neglected , and relativistic corrections are made using a simple ansatz for lorentz contraction @xcite .
predictions from lu _
et al_. @xcite using a bag radius of 0.8 fm are compared with the data in fig .
[ fig : pion ] .
although density - dependent extensions of this model , described as the quark - meson coupling ( qmc ) model , have been used to study the sensitivity of recoil polarization in nucleon electromagnetic knockout to medium modifications of the nucleon form factors @xcite , its description of free form factors is rather poor and one must hope that the density dependence of @xmath48 ratios is more accurate . alternatively , the light front cloudy bag model ( lfcbm ) of miller @xcite maintains poincar invariance by formulating wave functions using the light - front approach .
this version should then be applicable to higher @xmath97 .
there are only 4 adjustable parameters and the results for set 1 are compared with data in fig.[fig : pion ] .
a previous version of this model @xcite provided one of the earliest predictions of the sharp slope in @xmath56 for @xmath110 ( gev/@xmath4)@xmath104 , but the agreement with recent recoil - polarization data is only qualitative .
the lfcbm calculation for @xmath111 also decreases too rapidly at large @xmath97 .
calculations using this model agree relatively well with the @xmath6 data for @xmath112 ( gev/@xmath4)@xmath104 but are too small at lower @xmath97 .
interestingly , this model predicts much stronger values for @xmath107 at large @xmath97 than the lomon parametrization .
however , the lfcbm calculations for three of the four form factors show complicated and rather implausible shapes for @xmath113 ( gev/@xmath4)@xmath104 that disagree strongly with data .
chiral effective field theory @xcite provides a more systematic procedure that includes intermediate excitation and can be extended to two pion loops @xcite .
alternatively , two - loop contributions can be evaluated in dispersion theory @xcite .
recently it has become possible also to include both pion loops and vector meson diagrams in a consistent manner @xcite ; however , we do not show curves here because this approach remains limited to @xmath114 ( gev/@xmath4)@xmath104 .
friedrich and walcher @xcite performed a phenomenological analysis of the nucleon electromagnetic form factors using a parametrization motivated by pion cloud models .
the core form factor is represented by two dipole form factors with different ranges while the pion cloud contribution , represented as a `` bump '' at low @xmath97 , is described by two gaussians .
these fits , with 5 free parameters for @xmath6 and 6 for each of the other form factors , are also compared with data in fig .
[ fig : pion ] .
the quality of these fits is generally satisfactory , but it is not clear that the postulated oscillation in @xmath107 is warranted by the available data ; considerably better experimental precision at @xmath115 ( gev/@xmath4)@xmath104 would be needed to justify such a structure .
( data from @xcite ) .
solid curve : a fit based upon the pion cloud model of kaskulov and grabmayr @xcite .
short - dashed curve : parametrization of friedrich and walcher @xcite .
long - dashed curve : re - analysis by glazier _
et al_. @xcite using the friedrich and walcher model .
the dash - dotted curve is the original galster parametrization @xcite . ]
a closer look at the @xmath6 data is given in fig.[fig : kaskulov ] .
the original friedrich and walcher fit ( short - dashed curve ) used a very preliminary version of the data from the present experiment and falls systematically below the final data for this and other more recent experiments for @xmath116 ( gev/@xmath4)@xmath104 . a re - analysis using final data for this experiment plus new data @xcite was made by glazier _
et al_. @xcite and is shown as the long - dashed curve featuring a bump for @xmath115 ( gev/@xmath4)@xmath104 superimposed upon a much flatter core form factor . with 5 parameters it is obviously possible to fit the data very well , perhaps too well the simple two - parameter fit of kelly @xcite based upon the galster parametrization already provides @xmath117 without distinguishing between soft and hard structures .
the data presently available do not require this complication .
data at higher @xmath97 should test whether such a hard core is needed but significantly more precise data for low @xmath97 would be needed to establish the soft pion cloud contribution to @xmath6 . finally , kaskulov and grabmayr @xcite used a chiral quark model ( @xmath49qm ) to derive a relationship @xmath118 between @xmath6 , the pion form factor @xmath119 , and the core form factor @xmath120 for the 3-quark component of the nucleon .
the coefficient @xmath121 is a weighted average over spectroscopic factors for @xmath122 and @xmath123 intermediate states in the one - pion loop contribution to the self - energy but is treated as an adjustable parameter . if one stipulates a monopole for @xmath124 and a dipole for @xmath125 , the neutron electric form factor @xmath126 with @xmath127 reduces to a galster - like form with up to 3 free parameters ( @xmath121 , @xmath128 , @xmath129 ) ; however , @xmath121 is largely determined by the neutron radius @xmath130 if we further assume that @xmath128 within a loop is the same as that for pion electroproduction , only @xmath129 remains to be fit to data for @xmath6 .
thus , using fixed parameters @xmath131 and @xmath132 suggested by kaskulov and grabmayr , we fit @xmath133 ( gev/@xmath4)@xmath104 to the current @xmath6 data .
the value given in @xcite for @xmath129 is slightly smaller because they used the same preliminary data as @xcite that are smaller than the final results .
our fit is shown in fig .
[ fig : kaskulov ] and is practically indistinguishable from the two - parameter galster fit given in @xcite .
the kaskulov and grabmayr model has the same physical basis as that of friedrich and walcher , but is much more constrained ; nevertheless , it fits the @xmath6 data quite well .
this result suggests that the radius of the @xmath134 nucleon core is @xmath135 the predictions of several recent relativistic constituent quark models are compared with the data in fig .
[ fig : qm ] .
all employ a linear confining potential .
the solid curves show calculations of the pavia - graz collaboration @xcite that used the point - form spectator approximation ( pfsa ) for pointlike constituent quarks and a goldstone boson exchange interaction fitted to spectroscopic data .
no additional parameters were adjusted to fit the form factors .
the data for @xmath136 are reproduced very well and the data for magnetic form factors are also described relatively well for @xmath137 ( gev/@xmath4)@xmath104 , but the calculated value of @xmath138 decreases too rapidly for larger @xmath97 .
the prediction for @xmath139 lies well below the data for @xmath110 ( gev/@xmath4)@xmath104 .
the long - dashed curves show calculations of simula @xcite , based upon the model of cardarelli _
et al_.@xcite , that used the light - front approach and the one - gluon exchange ( oge ) interaction . here
, constituent - quark form factors were fitted to data for @xmath113 ( gev/@xmath4)@xmath104 and the calculations were extrapolated to larger @xmath97 .
this approach provides good fits up to about 4 ( gev/@xmath4)@xmath104 .
finally , the dash - dotted curves show the results for a semirelativistic hypercentral constituent quark model ( hcqm ) @xcite where the constituent - quark form factors , chosen as linear combinations of monopole and dipole forms , were also fitted to recent data . of the selected quark model calculations ,
their results clearly achieve the best overall agreement with the data .
finally , the most recent lattice qcd calculations of nucleon form factors were reported by the qcdsf collaboration @xcite using nonperturbatively improved wilson fermions in the quenched approximation .
unfortunately , straightforward chiral extrapolation @xcite does not provide adequate agreement with data for @xmath140 ( gev/@xmath4)@xmath104 .
et al_. @xcite proposed a model - dependent extrapolation procedure based upon the lfcbm .
this extrapolation is quite severe because the lattice calculations remain limited to quark masses that correspond to @xmath141 gev , lattice spacings with @xmath142 fm , and volumes that might not fully contain the pion cloud ; therefore , comparison with data is probably premature .
we reported values for the neutron electric to magnetic form factor ratio , @xmath0 , deduced from measurements of the neutron s recoil polarization in quasielastic @xmath1h@xmath2h kinematics at three acceptance - averaged @xmath3 values of 0.45 , 1.13 , and 1.45 ( gev/@xmath4)@xmath1 . in the one - photon exchange approximation for elastic scattering from a free neutron , the polarization vector of the recoil neutron is confined to the scattering plane and consists of a longitudinal component , @xmath143 , and a transverse component , @xmath144 .
the use of a deuteron target to access the neutron form factor ratio via the quasielastic @xmath1h@xmath2h reaction has the advantage , as established by arenhvel _
et al_.@xcite , that both @xmath38 and @xmath39 are relatively insensitive to final - state interactions ( fsi ) , meson - exchange currents ( mec ) , isobar configurations ( ic ) , and to theoretical models of deuteron structure .
a high - luminosity neutron polarimeter designed specifically for our experiment , jefferson laboratory e93 - 038 , was used to measure neutron polarization - dependent scattering asymmetries proportional to the projection of the polarization vector on the transverse axis .
a dipole magnet located upstream of the polarimeter was used to precess the neutron polarization vector in the transverse - longitudinal plane , thereby permitting access to the ratio @xmath145 .
values for the scattering asymmetries were extracted from neutron time - of - flight measurements in our polarimeter via the cross ratio technique .
the merit of the cross ratio technique is that the scattering asymmetries are independent of the luminosities for the two electron beam helicity states , and independent of the efficiencies and acceptances of the top and bottom halves of the polarimeter .
systematic uncertainties in our results are minimal as the analyzing power of the polarimeter and the polarization of the electron beam cancel in the form factor ratio .
further , other sources of uncertainty , such as radiative corrections and neutron depolarization by lead shielding , are small as they nearly cancel in the ratio . to account for the finite experimental acceptance and nuclear physics effects ( i.e. , fsi , mec , and ic ) , we used two independent simulation programs to average theoretical @xmath1h@xmath2h recoil polarization calculations computed according to the model of arenhvel _
et al_. @xcite over the acceptance .
the results from these two simulation programs agreed to better than 1% at our two lower @xmath3 points and 2% at our highest @xmath3 point .
further , by averaging two different sets of theoretical calculations assuming different parametrizations for @xmath6 , our acceptance - averaged and nuclear physics corrected values for @xmath6 were found to be insensitive to the choice of the @xmath3 dependence of @xmath6 .
our results for @xmath6 and data on the nucleon form factors were compared with selected theoretical model calculations .
all of the model calculations based upon vector meson dominance and those emphasizing the pion cloud presented here provide qualitative agreement with some of the four nucleon form factors , but no model achieves simultaneous agreement with all four form factors .
the predictions of several recent relativistic quark models also achieve qualitative agreement with the data , with the most successful models utilizing form factors for the constituents ; the results from a chosen model assuming pointlike constituents are not as successful .
although a comparison between data and the results of lattice qcd calculations is probably premature , the recent precise data obtained from experiments employing polarization degrees of freedom will no doubt serve as a future challenging test of qcd as formulated on the lattice . in conclusion , our results at @xmath146 and 1.45 ( gev/@xmath4)@xmath1 are the first direct measurements of @xmath6 using polarization degrees of freedom in the @xmath147 ( gev/@xmath4)@xmath1 region and are the most precise determinations of @xmath6 over all ranges of @xmath3 .
the achievement of relative statistical uncertainties in the form factor ratio @xmath0 of 8.4% and 9.5% , respectively , at these two @xmath3 points , together with relative systematic uncertainties on the level of 2% , was a triumph for our high figure - of - merit and high luminosity neutron polarimeter .
this work was supported in part by grants from the u.s
. national science foundation , the u.s .
department of energy , and the deutsche forschungsgemeinschaft .
two of us ( r.m . and t.e . ) acknowledge support during the planning phase of e93 - 038 from the nuclear and high energy physics center at hampton university .
the southeastern universities research association ( sura ) operates the thomas jefferson national accelerator facility under the u.s .
department of energy contract no .
de - ac05 - 84er40150 .
we thank the jefferson laboratory hall c engineering and technical staff for their skillful assistance during our experiment .
our notation for the kinematics and nucleon recoil polarization for the quasielastic @xmath1h@xmath2h reaction follows that of arenhvel _
et al_. ( e.g. , @xcite ) . for ease of notation ,
all kinematic quantities in the center - of - mass ( c.m . ) frame of the recoiling neutron - proton ( @xmath61-@xmath62 ) system will carry a superscript @xmath150 ; however , kinematic quantities referred to the laboratory frame will not be adorned with a superscript @xmath151 . a schematic diagram of the kinematics for the electrodisintegration of the deuteron in the one - photon exchange approximation is shown in fig .
[ fig : deuteron - electrodisintegration - schematic ] .
our notation for the electron kinematics is as usual , and we assume the electron scatters from a deuteron with initial four - momentum @xmath152 . following the break - up of the deuteron , the proton and neutron exit with three - momenta @xmath153 and @xmath154 , respectively .
as is customary , we use @xmath155 ( @xmath156 ) to denote the polar angle between @xmath153 ( @xmath154 ) and @xmath59 in the laboratory frame , and a reaction plane is defined by any two of @xmath59 , @xmath153 , and @xmath154 . as is shown in fig.[fig : deuteron - electrodisintegration - schematic ] , the reaction plane is tilted at a dihedral angle @xmath157 with respect to the scattering plane .
it should be noted that in the @xmath61-@xmath62 c.m .
frame , this dihedral angle , @xmath158 , is , obviously , just equal to @xmath157 .
the @xmath61-@xmath62 c.m .
frame is reached via a boost along @xmath59 . in the laboratory frame ,
the @xmath61-@xmath62 final state has an invariant mass , @xmath159 , of @xmath160 , where the relative @xmath61-@xmath62 energy in the laboratory frame , @xmath161 , is @xmath162 . with these definitions ,
it is clear that the lorentz factor for the boost from the laboratory frame to the @xmath61-@xmath62 c.m .
frame is @xmath163 we denote the polar angle between the relative @xmath61-@xmath62 motion in the c.m .
frame , @xmath164 ( assuming equal nucleon masses ) , and @xmath165 as @xmath60 .
as can be shown easily , this angle can be written solely in terms of the laboratory frame observables @xmath166 , @xmath167 , @xmath168 , and @xmath94 as @xmath169 where @xmath170 clearly , @xmath171 ( @xmath172 ) corresponds to perfect quasifree emission of the proton ( neutron ) ; however , it should be noted that there is vanishing phase space for perfect quasifree emission .
the five - fold differential coincidence cross section for the electrodisintegration of the deuteron in polarized - electron , unpolarized - deuteron scattering is of the simple form @xcite @xmath173 as the electron asymmetry , @xmath174 , is the only polarized contribution to the cross section . as usual
, @xmath28 denotes the unpolarized cross section .
the above expression for the cross section can also be written in terms of structure functions as @xcite @xmath175 where the @xmath176 structure functions are evaluated in the @xmath61-@xmath62 c.m .
frame , the @xmath177 are elements of the virtual photon density matrix and functions of kinematics , and @xmath178 is a function of kinematics .
it should be noted that the above expression for the cross section is differential in @xmath179 , @xmath180 , and @xmath181 .
the jacobian , @xmath182 , which transforms @xmath183 is given by @xcite @xmath184 here , @xmath185 is as given in eq .
( [ eq : gamma - lab - npcm ] ) , @xmath186 is the lorentz factor for the boost which takes the neutron from its rest frame to the @xmath61-@xmath62 cm frame , @xmath187 and @xmath188 is the lorentz factor for the boost which takes the neutron from its rest frame to the laboratory frame , @xmath189 .
\label{eq : gamma - n - lab}\ ] ] @xmath190 , @xmath191 , and @xmath192 are the velocities associated with @xmath185 , @xmath193 , and @xmath188 , respectively .
the nucleon recoil polarization in the @xmath61-@xmath62 c.m .
frame , @xmath194 , is of the form @xcite @xmath195,\ ] ] where @xmath29 and @xmath30 denote , respectively , the helicity - independent and helicity - dependent recoil polarization . written in terms of @xmath196 structure functions ,
the helicity - independent polarization components are @xmath197 and the helicity - dependent polarization components are @xmath198 the boost from the laboratory frame to the @xmath61-@xmath62 c.m . frame is along @xmath59 , which is not , in general , parallel to either nucleon s momentum vector ; therefore , the recoil polarization components in the laboratory frame are related to the recoil polarization components in the @xmath61-@xmath62 c.m .
frame via a relativistic wigner spin rotation . as the nucleons momenta span
the @xmath35-@xmath33 plane , the @xmath34-component is unchanged , while the @xmath35- and @xmath33-components mix according to @xmath199 where @xmath200 , @xmath201 denotes a matrix element of the wigner rotation matrix , @xmath202 and @xmath203 , the wigner rotation angle for the neutron , is expressed in terms of kinematics as @xcite @xmath204.\ ] ] here , @xmath205 and @xmath206 denote , respectively , the polar angle of the neutron s momentum vector relative to @xmath59 in the @xmath61-@xmath62 c.m .
frame and the laboratory frame . for non - relativistic boosts ( i.e. , @xmath185 , @xmath193 , and @xmath188 all @xmath207 )
, it is clear that we recover the non - relativistic result , @xmath208 .
also , it is obvious that for perfect quasifree emission ( i.e. , @xmath209 or @xmath210 ) , the recoil polarization components in the @xmath61-@xmath62 c.m.frame are identical to those in the laboratory frame .
to demonstrate the sensitivity of @xmath38 to the value of @xmath6 and the insensitivity of @xmath38 and @xmath39 to fsi , mec , ic , and the choice of the @xmath16 potential , we present several examples of @xmath1h@xmath2h recoil polarization calculations performed within the pwba and fsi+mec+ic+rc models of arenhvel _
et al_.@xcite in figs.[fig : phx - bonn - fsi - mec - ic - compare - galster ] , [ fig : phxphz - bonn - compare - pwba - fsi - mec - ic ] , and [ fig : phxphz - fsi - mec - ic - compare - potential ] .
we have ( arbitrarily ) chosen to show examples of these calculations for the central kinematics of our @xmath211 ( gev/@xmath4)@xmath1 point ( i.e. , @xmath212 gev , @xmath213 gev , @xmath214 ) .
first , fsi+mec+ic+rc calculations of @xmath38 are shown in fig.[fig : phx - bonn - fsi - mec - ic - compare - galster ] as a function of @xmath60 for three values of @xmath6 scaled by the galster parametrization : 0.5 , 1.0 , and 1.5 . a strong ( nearly linear )
sensitivity of @xmath38 to the value of @xmath6 is seen at and near quasifree emission .
second , the insensitivity of @xmath38 and @xmath39 to fsi , mec , and ic for quasifree emission is shown in fig.[fig : phxphz - bonn - compare - pwba - fsi - mec - ic ] where little difference between the pwba and fsi+mec+ic+rc calculations is observed at and near quasifree emission .
finally , we compare fsi+mec+ic+rc calculations of @xmath38 and @xmath39 for the argonne v18 @xcite , bonn @xcite , nijmegen @xcite , and paris @xcite @xmath16 potentials in fig.[fig : phxphz - fsi - mec - ic - compare - potential ] .
again , at and near quasifree emission , there is little model dependence .
( top panel ) and @xmath39 ( bottom panel ) for the argonne v18 , bonn , nijmegen , and paris potentials .
the results shown are for the central kinematics of our @xmath211 ( gev/@xmath4)@xmath1 point and @xmath215 . ]
d. r. yennie , m. m. lvy , and d. g. ravenhall , rev .
* 29 * , 144 ( 1957 ) ; f. j. ernst , r. g. sachs , and k. c. wali , phys . rev . * 119 * , 1105 ( 1960 ) ; r. g. sachs , phys . rev . * 126 * , 2256 ( 1962 ) ; l. n. hand , d. g. miller , and r. wilson , rev .
phys . * 35 * , 335 ( 1963 ) .
b. plaster , ph.d .
thesis , massachusetts institute of technology ( 2003 ) ; s. tajima , ph.d . thesis ,
duke university ( 2003 ) ; w. tireman , ph.d . thesis , kent state university ( 2003 ) ; g. maclachlan , ph.d .
thesis , ohio university ( 2004 ) ; a. aghalaryan , ph.d . thesis ,
yerevan physics institute ( in preparation ) ; e. crouse , ph.d .
thesis , the college of william and mary ( in preparation ) .
s. taylor , e93 - 038 technical reports ( unpublished ) : + http://www.jlab.org/@xmath216plaster/reports/charybdis1.ps ; http://www.jlab.org/@xmath216plaster/reports/charybdis2.ps ; http://www.jlab.org/@xmath216plaster/reports/charybdis3.ps .
j. w. watson _
et al_. , nucl .
instrum .
methods phys .
res . a * 272 * , 750 ( 1988 ) ; r. madey _ et al_. , ieee trans .
sci * 36 * , 231 ( 1989 ) .
t. eden _ et al_. , nucl .
instrum .
methods phys .
a * 338 * , 432 ( 1994 ) .
r. f. wagenbrunn , s. boffi , w. klink , w. plessas , and m. radici , phys . lett .
b * 511 * , 33 ( 2001 ) ; s. boffi , l. ya .
glozman , w. klink , w. plessas , m. radici , and r. f. wagenbrunn , eur .
phys . j. a * 14 * , 17 ( 2002 ) . | we report values for the neutron electric to magnetic form factor ratio , @xmath0 , deduced from measurements of the neutron s recoil polarization in the quasielastic @xmath1h@xmath2h reaction , at three @xmath3 values of 0.45 , 1.13 , and 1.45 ( gev/@xmath4)@xmath1 . the data at @xmath5 and 1.45 ( gev/@xmath4)@xmath1 are the first direct experimental measurements of @xmath6 employing polarization degrees of freedom in the @xmath7 ( gev/@xmath4)@xmath1 region and stand as the most precise determinations of @xmath6 for all values of @xmath3 . |
physical theories are merely approximations to the natural world and the physical constants involved can not be known without some degree of uncertainty .
properties of a model that are sensitive to small changes in the model , in particular changes in the values of the parameters , are unlikely to be observed .
it can thus be reasoned that one should search for physical theories which do not change in a qualitative matter under a small change of the parameters .
such theories are said to be physically _
stable_. this concept of the physical stability of a theory can be given a mathematical meaning as follows .
a mathematical structure is said to be mathematically stable for a class of deformations if any deformation in this class leads to an isomorphic structure .
more precisely , a lie algebra is said to be stable if small perturbations in its structure constants lead to isomorphic lie algebras .
the idea of mathematical stability provides insight into the validity of a physical theory or the need for a generalization of the theory .
if a theory is not stable , one might choose to deform it until a stable theory is reached .
such a stable theory is likely to be a generalization of wider validity compared to the original unstable theory .
lie algebraic deformation theory has been historically successful .
snyder @xcite in 1947 showed that the assumption that spacetime be a continuum is not required for lorentz invariance .
framework however leads to a lack of translational invariance , which later in the same year , yang @xcite showed can be corrected if one allows for spacetime to be curved .
yang , in the same paper also presented the complete lie algebra associated with the suggested corrections .
it was mendes @xcite who in the last decade concluded that when one considers the poincare and heisenberg algebras together , the resultant poincare - heisenberg algebra is not a stable lie algebra .
mendes showed however that the algebra can be stabilized , requiring two additional length scales .
the stabilized algebra is the same as the algebra obtained by yang in 1947 .
it was faddeev @xcite and mendes who noted that , in hindsight , stability considerations could have predicted the relativistic and quantum revolutions of the last century .
chryssomalakos and okon @xcite showed that by a suitable identification of the generators , triply special relativity proposed by kowalski - glikman and smolin @xcite can be brought to a linear form and that the resulting algebra is again the same as yang s algebra .
more recently , stability considerations have led to the stabilized poincare - heisenberg algebra ( spha ) as the favorite candidate for the lie algebra describing physics at the interface of gr and qm .
chryssomalakos and okon @xcite showed uniqueness of the spha .
incorporating gravitational effects in quantum measurement of spacetime events renders spacetime non - commutative and leads to modifications in the fundamental commutators @xcite . in 2005 , ahluwalia - khalilova showed that the fact that the heisenberg fundamental commutator , @xmath1=i\hbar$ ] , undergoes non - trivial modifications at the interface of gr and qm , suggests quantum mechanics and relativity will undergo numerous corrections and modification including modification of the position - momentum heisenberg uncertainty relations . on top of this ,
spacetime in spha requires two new length scales , one in the extreme short distance range , the other on the cosmological scale .
we denote these new length scales @xmath2 and @xmath3 respectively ( following the notation used in @xcite ) .
in addition , spha points to the existence of another dimensionless constant @xmath4 ( or @xmath5 in @xcite ) which , if nonzero , will radically affect some of the quantum relativistic notions .
the presence of @xmath4 has been noted in @xcite with different emphasis .
chryssomalakos and okon @xcite show that it is always possible to gauge away this dimensionless constant by a suitable redefinition of the generators .
the overall meaning of @xmath4 seems to not be well understood in the literature .
we show that clifford algebra allows one to obtain an explicit expression for @xmath4 in terms of an angle parameter @xmath6 .
this is a step toward understanding how @xmath4 will affect various quantum relativistic notions .
our approach to the spha is quite different in that we adopt a clifford algebra perspective .
a physical theory should have its roots in the geometry underlying the theory and be represented in a way that captures this geometry accordingly .
clifford ( geometric ) algebra is the appropriate tool that allows us to naturally encompass the underlying geometry of the space we are working in .
although a brief overview of clifford algebra will be presented in section 3 of this paper , we omit any in - depth discussion of clifford algebra and how it arises from geometry .
the reader is instead referred to the following texts by lounesto @xcite , hestenes @xcite and doran and lasenby @xcite .
the spha has fifteen generators .
we wish to represent these generators by elements of the clifford algebra @xmath0 which is a sixteen dimensional algebra .
the non - scalar basis elements of @xmath0 can be used to generate the spha by taking commutators with the clifford product .
( @xmath0 is chosen over @xmath7 or other sixteen dimensional clifford algebras for reasons explained in @xcite . )
it is interesting to note that using this clifford algebra , one avoids all stability considerations .
this is because the clifford algebra @xmath0 is semi - simple and therefore stable .
some further discussion on this is reserved for the next section .
ahluwalia - khalilova @xcite noted that incorporating gravitational effects in quantum measurement of spacetime events renders spacetime non - commutative and leads to modifications in the fundamental commutators .
this non - commutativity of space - time and modifications to the fundamental commutators arise naturally from clifford algebra .
there thus seems to be a number of reasons why one should consider using clifford algebra in theories of quantum gravity .
since we are logically led to consider two additional length scales as well as a dimensionless constant in the proposed new algebra for kinematics at the interface of gr and qm , it is important to note that on heuristic grounds amelino - camelia @xcite and smolin @xcite have also considered such a path .
however , in the work presented here such resulting modifications to the quantum and spacetime structure arise not as an ad hoc speculation but from the principle of lie algebraic stability .
the reader is referred to @xcite for additional discussion of these and related issues .
as noted by mendes @xcite and faddeev @xcite , in hindsight , the paradigm of algebraic deformations to obtain stable theories has the power to have predicted the relativistic and the quantum revolutions of the last century .
this section will show how algebraic deformations lead from the poisson algebra of classical mechanics to the heisenberg algebra of quantum mechanics and from the galilean algebra of galilean relativity to the poincare algebra of special relativity .
the theory of lie - algebraic deformations is not discussed in detail here . for a thorough and complete treatment the reader
is referred to gerstenhaber @xcite , nijenhuis and richardson @xcite and chryssomalakos and okon @xcite .
when considering the poisson and galilean algebras , one finds that the algebraic structures are unstable .
it is however possible to stabilize both of the algebras .
doing so requires two deformation parameters .
these turn out to be the physical constants @xmath8 and @xmath9 for the galilean algebra and poisson algebra respectively .
chryssomalakos and okon @xcite explain that both the galilean and poisson algebra cases , the deformed algebras are isomorphic for all non - zero values of @xmath8 and @xmath9 .
the values of these deformation parameters are determined by experiment .
neither the heisenberg nor the poincare algebras preserve their stability at the interface of gr and qm . a first attempt to find an algebra describing physics at the interface of gr and qm may be to take the direct sum of the poincare and heisenberg algebras to give the poincare - heisenberg algebra but this algebra however is not stable .
mendes @xcite and chryssomalakos and okon @xcite have both emphasised that one of the important criteria to consider for a theory to be physically viable , is the stability of the underlying lie algebras , and so the poincare - heisenberg algebra fails to be the algebra we desire .
the poincare - heisenberg algebra can however be stabilized , requiring two additional length scales and a dimensionless constant in the same manner that the stabilization of the algebras of galilean and classical kinematics requires two constants @xmath8 and @xmath9 .
the resulting algebra is the stabilized poincare - heisenberg algebra ( spha ) , the algebra of kinematics at the interface of gr and qm .
it will be shown later that the spha is precisely the algebra obtained by taking commutators of the elements of the clifford space - time algebra @xmath0 .
as noted above , the stabilised poincare - heisenberg algebra comes from combining quantum mechanics and relativity to get a stable theory of kinematics at the interface of gr and qm .
the algebra is given below in a similar form to that of chryssomalakos and okon @xcite ( which in turn is a somewhat more abstract and mathematical version than the one given by ahluwalia - khalilova @xcite which focuses more on the physical aspects of the algebra ) .
@xmath10=-(\eta_{\nu\rho}j_{\mu\sigma}+\eta_{\mu\sigma}j_{\nu\rho}-\eta_{\mu\rho}j_{\nu\sigma}-\eta_{\nu\sigma}j_{\mu\rho})\end{aligned}\ ] ] @xmath11=-(\eta_{\nu\lambda}p_{\mu}-\eta_{\mu\lambda}p_{\nu})\end{aligned}\ ] ] @xmath12=-\hbar(\eta_{\nu\lambda}x_{\mu}-\eta_{\mu\lambda}x_{\nu})\end{aligned}\ ] ] @xmath13=q\alpha_1 j_{\mu\nu}\end{aligned}\ ] ] @xmath14=q\alpha_2 j_{\mu\nu}\end{aligned}\ ] ] @xmath15=q\eta_{\mu\nu}m+q\alpha_3 j_{\mu\nu}\end{aligned}\ ] ] @xmath16=-\alpha_1 x_{\mu}+\alpha_3 p_{\mu}\end{aligned}\ ] ] @xmath17=-\alpha_3 x_{\mu}+\alpha_2 p_{\mu}\end{aligned}\ ] ] @xmath18=0\end{aligned}\ ] ] this paper is primarily concerned with finding a clifford representation of the spha .
for this reason we focus on the mathematical theory and therefore adopt the notation used by chryssomalakos and okon @xcite instead of the notation used by ahluwalia - khalilova @xcite , with the exception of the part of section 5 .
the clifford algebra @xmath0 is a 16 dimensional associative algebra with a basis consisting of one scalar @xmath19 , four vectors @xmath20 , six bivectors @xmath21 , three trivectors @xmath22 and a pseudoscalar @xmath23 which for brevity we write @xmath24 , together with the clifford product ( geometric product ) used to multiply elements of the algebra .
duals of elements of the clifford algebra are found by multiplying through by the pseudoscalar @xmath25 .
trivectors are thus dual to vectors . instead of writing trivectors as @xmath22 , we write them as @xmath26 for reasons which will become clear later .
as noted earlier , we do not have to concern ourselves with the issue of stability when using a clifford algebra .
all clifford algebras are stable because the diagonal entries of the metric @xmath27 are non - zero and therefore under a small perturbation , the signature remains unchanged .
we are looking for a clifford representation of the stabilized poincare - heisenberg algebra .
this means that @xmath28 , @xmath29 , @xmath30 and @xmath31 are to be represented by elements of a real clifford algebra .
the stabilized poincare - heisenberg algebra is a lie algebra so that we will require the lie product to be a commutator in the clifford algebra , i.e. @xmath32=xy - yx$ ] .
the scalar element of the clifford algebra commutes with every element leaving us with 15 generators
. we will start with @xmath33 , @xmath34 , @xmath35 and @xmath36 , where @xmath37 , and calculate their commutators .
@xmath38&=&2\omega_{3}^2(\eta_{\nu\rho}e_{\mu\sigma}+\eta_{\mu\sigma}e_{\nu\rho}-\eta_{\mu\rho}e_{\nu\sigma}-\eta_{\nu\sigma}e_{\mu\rho})\\ \notag[\omega_3 e_{\mu\nu},\omega_2 ee_{\rho}]&=&2\omega_2 \omega_3(\eta_{\nu\rho}ee_{\mu}-\eta_{\mu\rho}ee_{\nu})\\ & = & 2\omega_3(\eta_{\nu\rho}(\omega_2 ee_{\mu})-\eta_{\mu\rho}(\omega_2 ee_{\nu}))\\ \notag[\omega_3 e_{\mu\nu},\omega_1 e_{\rho}]&=&2\omega_1 \omega_3(\eta_{\nu\rho}e_{\mu}-\eta_{\mu\rho}e_{\nu})\\ & = & 2\omega_2(\eta_{\nu\rho}(\omega_1 e_{\mu})-\eta_{\mu\rho}(\omega_1 e_{\nu}))\end{aligned}\ ] ] @xmath39&=&2\omega_{2}^2e_{\mu\nu}\\ & = & 2\frac{\omega_{2}^2}{\omega_{3}}(\omega_3e_{\mu\nu})\\ \notag[\omega_1e_{\mu},\omega_1e_{\nu}]&=&2\omega_{1}^2e_{\mu\nu}\\ & = & 2\frac{\omega_{1}^2}{\omega_{3}}(\omega_3e_{\mu\nu})\\ \notag[\omega_2ee_{\mu } , \omega_1e_{\nu}]&=&2\omega_1\omega_2\eta_{\mu\nu}e\\ & = & 2\eta_{\mu\nu}\frac{\omega_1\omega_2}{\omega_4}(\omega_4e)\\ \notag[\omega_2ee_{\mu},\omega_4e]&=&2\omega_2\omega_4e_{\mu}\\ & = & 2\frac{\omega_2\omega_4}{\omega_1}(\omega_1e_{\mu})\\ \notag[\omega_1e_{\mu } , \omega_4e]&=&-2\omega_1\omega_4ee_{\mu}\\ & = & -2\frac{\omega_1\omega_4}{\omega_2}(\omega_2ee_{\mu})\\ \label{last}[\omega_3e_{\mu\nu},\omega_4e]&=&0\end{aligned}\ ] ] by defining the 15 operators @xmath40 @xmath41 @xmath42 @xmath43 we obtain the stabilized poincare - heisenberg algebra for the special case where @xmath5 is equal to zero .
( we discuss this point in the next section . )
we call @xmath28 , @xmath29 , @xmath30 and @xmath31 above the clifford generators of the stabilised poincare - heisenberg algebra . as in @xcite , we interpret @xmath44 to be the position vectors , @xmath45 to be the momenta and @xmath30 to be the rotations and boosts .
in this section we will transform the clifford representation in such a way that the transformed clifford generators will generate the entire stabilized poincare - heisenberg algebra rather than just the special case where @xmath5 is equal to zero .
the physical interpretation of this transformation will be discussed in the following section and gives rise to a new concept in physics .
we start by defining by redefining @xmath44 as @xmath47 giving @xmath48 and therefore @xmath49=2(a^2+b^2)\,e_{\mu\nu},\qquad \mu\neq\nu\end{aligned}\ ] ] what we require is ; @xmath50=i\,q\,\alpha_2\,j_{\mu\nu}\ ] ] to get this we must thus define ; @xmath51 similarly , define @xmath29 as @xmath52 giving @xmath53 and @xmath54=2(c^2+d^2)\,e_{\mu\nu},\qquad \mu\neq\nu\end{aligned}\ ] ] we want this to be equal to @xmath55=i\,q\,\alpha_1\,j_{\mu\nu}\ ] ] which requires @xmath56 note that consistency in the definition for @xmath57 implies that @xmath58 the above transformed definitions for @xmath44 and @xmath45 can be written in the form @xmath59&= & \left[\begin{matrix}a&b\\c&d\end{matrix}\right]\ , \left[\begin{matrix}e_\mu\\ee_\mu\end{matrix}\right]\end{aligned}\ ] ] and therefore @xmath60&=&\frac{1}{ad - bc } \left[\begin{matrix}d&-b\\-c&a\end{matrix}\right]\ , \left[\begin{matrix}x_\mu\\p_\mu\end{matrix}\right]\end{aligned}\ ] ] or @xmath61 where @xmath62 is the determinant of the matrix . working out the commutator @xmath63 $ ] with
the new transformed expressions for @xmath44 and @xmath45 , we get @xmath64 while @xmath65 so , if @xmath66 , @xmath67&=&p_\mu x_\mu - x_\mu p_\mu\\ & = & ( ac+bd)\eta_{\mu\nu}+(ad - bc)\eta_{\mu\nu}\,e\\ & & -(ac+bd)\eta_{\mu\nu}+(ad - bc)\eta_{\mu\nu}\,e\\ & = & 2(ad - bc)\eta_{\mu\nu}\,e\end{aligned}\ ] ] while for @xmath68 , @xmath67=2(ac+bd)e_{\mu\nu}\end{aligned}\ ] ] we want this to be equal to @xmath69=i\,q\eta_{\mu\nu}m+i\,q\alpha_3j_{\mu\nu}\ ] ] so that we have @xmath70 and @xmath71 i.e. @xmath72 and @xmath73 note that consistency now requires @xmath74 next consider @xmath75=\alpha_3p_\mu-\alpha_1x_\mu$ ] .
we have @xmath76&=&\left[dee_\mu+ce_\mu,\frac{2(ad - bc)}{q}\,e\right]\\ \notag&=&\frac{2(ad - bc)}{q}\left[dee_\mu+ce_\mu , e\right]\\ \notag&=&\frac{2(ad - bc)}{q}(dee_\mu e+ce_\mu e - de^2e_\mu - cee_\mu)\\ \notag&=&\frac{2(ad - bc)}{q}(de_\mu - ce_\mu e - de_\mu - cee_\mu)\\ \notag&=&\frac{4(ad - bc)}{q}(de_\mu - cee_\mu)\\ \notag&=&\frac{4\delta}{q}\left(\frac{d}{\delta}(dx_\mu - bp_\mu)-\frac{c}{\delta}(-cx_\mu+ap_\mu)\right)\\ \notag&=&\frac{4}{q}(d^2x_\mu - bdp_\mu+c^2x_\mu - acp_\mu)\\ & = & -\frac{4}{q}(ac+bd)p_\mu+\frac{4}{q}(c^2+d^2)x_\mu\end{aligned}\ ] ] for this to be equal to @xmath77=\alpha_3 p_\mu-\alpha_1 x_\mu\ ] ] we must have @xmath78 and the consistency conditions must now be extended to read @xmath79 we also need to check @xmath80=\alpha_2p_\mu-\alpha_3x_\mu$ ] .
we have @xmath81&=&\left[ae_\mu+bee_\mu,\frac{2(ad - bc)}{q}\,e\right]\\ \notag&=&\frac{2(ad - bc)}{q}(ae_\mu e+bee_\mu e - aee_\mu - be^2e_\mu)\\ \notag&=&\frac{2(ad - bc)}{q}(-2aee_\mu+2be_\mu)\\ \notag&=&\frac{4\delta}{q}(be_\mu - aee_\mu)\\ \notag&=&\frac{4\delta}{q}\left(\frac{b}{\delta}(dx_\mu - bp_\mu)-\frac{a}{\delta}(-cx_\mu+ap_\mu)\right)\\ \notag&=&\frac{4}{q}(bdx_\mu - b^2p_\mu+acx_\mu - a^2p_\mu)\\ & = & -\frac{4}{q}(a^2+b^2)p_\mu+\frac{4}{q}(ac+bd)x_\mu\end{aligned}\ ] ] so we need @xmath82 which hold by the consistency equations . to satisfy the consistency equation ( [ consistency ] )
, we may define @xmath83 and @xmath84 and since also @xmath85 , we obtain @xmath86 comparing these results with those of ahluwalia - khalilova @xcite , we find that ) above , there are two more conversions , one between @xmath87 and @xmath88 and one between @xmath5 and @xmath4 .
a comparison reveals that consistency between @xcite and @xcite requires that ; @xmath89 ] @xmath90 this in turn tells us that @xmath91 where @xmath92 . in this section
we have transformed the clifford generators of the representation to obtain a representation where @xmath5 is not equal to zero .
this is going in the opposite direction to chryssomalakos and okon @xcite who begin with a representation where @xmath5 is not necessarily zero and then show that there always exists a representation in the @xmath93-@xmath94 plane with @xmath5 equal to zero by performing a linear redefinition of the generators .
the clifford algebra approach is thus consistent with @xcite .
chryssomalakos and okon , @xcite , comment that physicists in particular may frown upon the idea of working with arbitrary linear combinations of momenta and positions . for this reason it is important that we interpret what the transformation in the previous section might mean physically .
the transformation @xmath95 and @xmath96 is equivalent to adding some momentum to the position vector and vice versa .
the magnitude of the vector is invariant .
the transformation looks like a rotation in the position - momentum plane however @xmath97and @xmath98 are rotated by different angles . what is the physical interpretation of this transformation ? in a newtonian mindset we consider time and space to be disjoint and one can determine the absolute time and position of an event .
switching to a relativistic mindset , we know that we can no longer treat space and time separately but that in fact we have to consider them together in what we call spacetime .
it is no longer possible to determine the time and position of an event absolutely .
similarly , in a newtonian mindset , we can treat position and momentum separately and thus the above transformation may seem unphysical . in a quantum mechanical frame of mind however , we can not measure position @xmath44 and momentum @xmath45 with absolute certainty . the more accurately we know one , the less accurately we know the other as is described by the heisenberg uncertainty relationship @xmath99 .
this suggests that we can not just think about position or momentum without considering the other .
furthermore , making a measurement of the position of some particle will in itself affect the position .
the photon used to measure the particle s position will give the particle some momentum . on the quantum scale
therefore a linear combination of position and momentum does make sense and in fact treating position and momentum separately as in newtonian physics may no longer be desirable . at the interface of gr and
qm one combines quantum mechanics and relativity and therefore should consider spacetime - momentum instead of spacetime alone @xcite .
the authors of this paper wish to thank and acknowledge dharamvir ahluwalia - khalilova , benjamin martin and adam gillard for the many discussions . in particular we wish to thank dharamvir ahluwalia - khalilova for suggesting that at the interface of gr and qm one should consider spacetime - momentum instead of spacetime alone . c. n. yang , `` on quantized space - time , '' phys . rev .
* 72 * ( 1947 ) 874 .
r. vilela mendes , `` deformations , stable theories and fundamental constants , '' j. phys .
a * 27 * ( 1994 ) 8091 .
l. d. faddeev 1989 , `` mathematician s view on the development of physics , frontiers in physics : high technology and mathematics '' ed h. a. cerdeira and s. lundqvist ( singapore : word scientific , 1989 ) pp .
238 - 46 c. chryssomalakos and e. okon , `` generalized quantum relativistic kinematics : a stability point of view , '' int .
j. mod .
d * 13 * ( 2004 ) 2003 [ arxiv : hep - th/0410212 ] . c. chryssomalakos and e. okon , `` linear form of 3-scale relativity algebra and the relevance of stability , '' int
. j. mod .
d * 13 * ( 2004 ) 1817 [ arxiv : hep - th/0407080 ] .
d. v. ahluwalia - khalilova , `` a freely falling frame at the interface of gravitational and quantum realms , '' class .
quant . grav .
* 22 * ( 2005 ) 1433 [ arxiv : hep - th/0503141 ] .
d. v. ahluwalia - khalilova , `` minimal spatio - temporal extent of events , neutrinos , and the cosmological constant problem , '' int .
j. mod .
d * 14 * ( 2005 ) 2151 [ arxiv : hep - th/0505124 ] .
d. v. ahluwalia , `` quantum measurements , gravitation , and locality , '' phys .
b * 339 * ( 1994 ) 301 [ arxiv : gr - qc/9308007 ] .
a. nijenhuis and r. w. richardson 1967 `` deformations of lie algebra structures '' j. math .
@xmath100 89 v. v. khruschev and a. n. leznov , `` relativistically invariant lie algebras for kinematic observables in quantum space - time , '' grav .
cosmol .
* 9 * ( 2003 ) 159 [ arxiv : hep - th/0207082 ] . c. doran and a. lasenby 2003 `` geometric algebra for physicists '' ( cambridge university press ) p. butler and l. mcaven 1998 `` space : the anti - euclidean metric from the structure of rotations '' , in proceedings of the xii international colloquium on group theoretical methods in physics , 494498 ( international press ) g. amelino - camelia , `` proposal of a second generation of quantum - gravity - motivated lorentz - symmetry tests : sensitivity to effects suppressed quadratically by the planck scale , '' int .
j. mod .
d * 12 * ( 2003 ) 1633 [ arxiv : gr - qc/0305057 ] .
j. kowalski - glikman and l. smolin , `` triply special relativity , '' phys .
d * 70 * ( 2004 ) 065020 [ arxiv : hep - th/0406276 ] .
s. judes and m. visser , `` conservation laws in doubly special relativity , '' phys .
d * 68 * ( 2003 ) 045001 d. grumiller , w. kummer and d. v. vassilevich , `` a note on the triviality of kappa - deformations of gravity , '' ukr .
j. phys .
* 48 * ( 2003 ) 329 [ arxiv : hep - th/0301061 ] .
r. schutzhold and w. g. unruh , `` problems of doubly special relativity with variable speed of light , '' jetp lett .
* 78 * ( 2003 ) 431 [ pisma zh .
* 78 * ( 2003 ) 899 ] [ arxiv : gr - qc/0308049 ] .
j. christian , `` absolute being vs relative becoming , '' arxiv : gr - qc/0610049 .
d. v. ahluwalia - khalilova 2006 _ private communications _ | the stabilized poincare - heisenberg algebra ( spha ) is the lie algebra of quantum relativistic kinematics generated by fifteen generators .
it is obtained from imposing stability conditions after attempting to combine the lie algebras of quantum mechanics and relativity which by themselves are stable , however not when combined . in this paper
we show how the sixteen dimensional clifford algebra @xmath0 can be used to generate the spha .
the clifford algebra path to the spha avoids the traditional stability considerations , relying instead on the fact that @xmath0 is a semi - simple algebra and therefore stable .
it is therefore conceptually easier and more straightforward to work with a clifford algebra .
the clifford algebra path suggests the next evolutionary step toward a theory of physics at the interface of gr and qm might be to depart from working in space - time and instead to work in space - time - momentum . |
nuclear rainbow scattering , which is observed under incomplete absorption , can uniquely determine the interaction potential family up to the internal region without ambiguity @xcite .
the interaction potential for the @xmath0o+@xmath0o system has been most thoroughly investigated both experimentally and theoretically .
although a shallow potential had been used in heavy ion scattering and reactions for many years @xcite , the observation of a nuclear rainbow in @xmath0o+@xmath0o scattering at @xmath2=350 mev finally showed that a global interaction potential for this system is deep @xcite .
it was shown in ref.@xcite that the global deep potential determined in nuclear rainbow scattering can describe in a unified way not only the prerainbows and the airy structure in the 90@xmath4 excitation function , but also the molecular resonances in the low energy region and the band structure with the @xmath0o+@xmath0o cluster structure .
it was also found @xcite that the highest order airy structure @xcite evolves into molecular resonances with the @xmath0o+@xmath0o cluster structure in @xmath10s as the incident energy decreases .
the gross structures in the 90@xmath4 excitation function in rainbow scattering separated by the airy minima have been visually interpreted as panchdermous airy elephants in ref .
@xcite . rainbow scattering and interaction potentials for the asymmetric @xmath0o+@xmath7c system have been studied systematically at @xmath2=63 - 260 mev @xcite and @xmath2=608 - 1503 mev @xcite and a global potential was determined .
the global potential could explain not only the rainbows and prerainbows @xcite but also the molecular resonances in the low energy region and the superdeformation with the @xmath0o+@xmath7c cluster structure in a unified way @xcite .
however , in order to explain the airy minimum observed at much larger angles at around @xmath2=300 mev @xcite , which was impossible to reproduce in the optical model calculations with the global potential , a deeper family potential was needed . in
ref.@xcite the order of the airy minimum was reassigned systematically to be one higher than that in previous literature in refs .
@xcite . for example
, the airy minimum at @xmath3=82@xmath4 at @xmath2=132 mev was assigned @xmath6 instead of @xmath5 . very recently
this dilemma was rescued @xcite by noticing that the airy minimum at the large angle is a new kind of airy minimum caused dynamically by the coupling to an excited state of @xmath7c and it was found that the experimental angular distributions are reproduced by the coupled channel calculations with a global extended folding potential derived from the microscopic wave functions for @xmath7c and @xmath0o . the @xmath0o+@xmath1c system is situated between @xmath0o+@xmath0o and @xmath0o+@xmath7c .
@xcite measured rainbow scattering for the @xmath0o+@xmath1c system at 132 , 281 and 382.2 mev .
@xcite investigated the airy structure and concluded that the order of the airy minimum at @xmath3=76@xmath4 in the angular distribution at @xmath2=132 mev is @xmath6 , which is similar to the airy minimum @xmath6 at @xmath3=82@xmath4 in the angular distribution of @xmath0o+@xmath7c at @xmath2=132 mev claimed with a deeper family potential in ref.@xcite .
the purpose of this paper is to study the airy structure of rainbow scattering for the @xmath0o+@xmath1c system with the extended double folding model used successfully in ref.@xcite for the @xmath0o+@xmath7c system and to determine the order of the airy minimum from the energy evolution of the airy minimum over a wide range of incident energies .
it is shown that the airy minimum at @xmath3=76@xmath4 in the angular distribution at @xmath2=132 mev is @xmath5 .
this is different from the previous assignment in ref.@xcite .
the evolution of the airy structure into the molecular resonances and the cluster structure in the low energy region is discussed in comparison with typical system such as @xmath0o+@xmath0o .
we study rainbow scattering for @xmath0o+@xmath1c with an extended double folding ( edf ) model that describes all the diagonal and off - diagonal coupling potentials derived from the microscopic wave functions for @xmath0o using a density - dependent nucleon - nucleon force .
the diagonal and coupling potentials for the @xmath0o+@xmath1c system are calculated using the edf model and are given as follows : @xmath11 where @xmath12 represents the diagonal nucleon density of the ground state of @xmath1c , which is obtained by the convolution of the proton size from the charge density distribution taken from ref.@xcite .
@xmath13 is the diagonal ( @xmath14 ) or transition ( @xmath15 ) nucleon density of @xmath0o taken from the microscopic @xmath16+@xmath7c cluster model wave functions calculated in the orthogonality condition model ( ocm ) in ref.@xcite , which uses a realistic size parameter both for the @xmath16 particle and @xmath7c .
this is an extended version of ref.@xcite , which well reproduces almost all the energy levels up to @xmath17@xmath1813 mev and the electric transition probabilities in @xmath0o .
the wave functions have been successfully used for the systematic analysis of elastic and inelastic scattering over a wide range of incident energies @xcite .
we take into account the important transition densities available in ref.@xcite , i.e. , g.s @xmath19 @xmath20 ( 6.13 mev ) and @xmath21 ( 6.92 mev ) in addition to all the diagonal potentials . for the effective interaction @xmath22 we use the ddm3y - fr interaction @xcite , which takes into account the finite - range nucleon exchange effect . in the calculations we introduce the normalization factor @xmath23 for the real part of the double folding potential @xcite .
an imaginary potential with a woods - saxon volume - type form factor ( nondeformed ) is introduced phenomenologically to take into account the effect of absorption due to other channels .
in fig . 1 the angular distributions in elastic @xmath0o+@xmath1c scattering calculated using the single channel double folding ( df ) model potential are compared with the experimental data @xcite at @xmath2= 132 , 281 and 382.2 mev . the normalization factor and volume integral per nucleon pair , @xmath24 , for the real folding potential , and the imaginary potential parameters used are given in table i. the experimental angular distributions are well reproduced by the single channel calculations .
the calculated cross sections are decomposed into the farside ( dashed line ) and nearside ( dash - dotted line ) components .
the nearside component decreases rapidly beyond the diffraction region and the farside component dominates toward the intermediate angular region .
thus the broad structure of the angular distribution is the airy structure of the nuclear rainbow caused by refractive scattering .
[ t ] ( color online ) comparison of the single channel df potential model calculations ( blue solid line ) with the experimental angular distributions of cross sections ( ratio to rutherford scattering ) ( points ) @xcite in elastic @xmath0o+@xmath1c scattering at @xmath2= 132 mev ( a ) , 281 mev ( b ) , and 382.2 mev ( c ) .
the dashed ( green ) and dash - dotted ( grey ) lines display the calculated farside and nearside components , respectively .
the angular distributions calculated by switching off the imaginary potential ( w=0 ) are displayed by the dotted ( pink ) lines .
, title="fig:",width=328 ] .
[ table i ] the volume integral per nucleon pair @xmath24 of the the ground state diagonal potential ( in units of mev@xmath25 ) and the imaginary potential parameters used in the single channel double folding calculations in fig . 1 and coupled channel calculations with edf in fig . 2 .
@xmath23=1 is used except @xmath260.95 at 132 mev ( single channel ) and 281 mev ( coupled channel ) . [
cols="^,^,^,^,^,^,^,^,^,^ " , ] [ table1 ] the order of the airy minimum is determined by calculating the angular distribution by switching off the imaginary potential at the highest energy @xmath2=382.2 mev in fig .
the fall - off of the cross sections in the angular distribution , i.e. , the darkside of the rainbow , starts beyond @xmath3=@xmath27 , which means that the minimum at @xmath28 is the first order airy minimum @xmath29 . at @xmath2=281 mev in fig .
1(b ) the second order airy minimum @xmath5 is seen at @xmath28 in addition to the @xmath29 at @xmath30 . in order to determine the order of the airy minimum at @xmath3=@xmath31 at the lowest energy @xmath32132 mev in fig .
1(a ) without ambiguity , the energy evolution of the airy structure of the angular distribution between @xmath2= 281 mev and 116 mev is calculated using the single channel double folding potential by switching off the imaginary potential .
this is displayed in fig .
the value of @xmath23 was interpolated or extrapolated from those at @xmath2=281 mev ( @xmath23=1.0 ) and 132 mev ( @xmath23=0.95 ) .
the energy dependence of the df potential comes mostly from the ddm3y - fr effective two body interaction . at @xmath2=140
mev the @xmath29 airy minimum is clearly seen at @xmath33 .
thus the airy minimum at @xmath31 at @xmath2=132 mev is found to be @xmath5 .
this assignment of the @xmath5 airy minimum at @xmath31 for the @xmath0o+@xmath1c system at 132 mev corresponds well to the @xmath5 assignment of the airy minimum at @xmath34 for the @xmath0o+@xmath7c system at the same @xmath2=132 mev in refs.@xcite .
the energy evolution of the airy minimum seems to support the interpretation that the minimum ( not visible in fig .
1 ) at around @xmath35 at @xmath2=132 mev is a remnant of the airy minimum @xmath29 . in fact
, the calculated angular distribution beyond this angle turns into diffraction - like high - frequency oscillations rising toward the extreme backward angle @xmath36 . at @xmath32132 mev in fig .
1(a ) @xmath6 is observed at @xmath37 .
it is worth mentioning that coupled reaction channel ( crc ) calculations for the @xmath0o+@xmath1c system at @xmath2=132 and 281 mev in ref.@xcite show that the potential scattering dominates at angles smaller than @xmath38 and the contribution of the two proton cluster transfer reactions dominates at large angles .
the minimum at around @xmath3=@xmath35 in @xmath0o+@xmath1c scattering at @xmath2=132 mev visible in the farside component in fig . 2 could be seen only as a remnant of the airy minimum @xmath29 in experiment .
we note that the wrong @xmath6 assignment to the airy minimum at @xmath31 at 132 mev in ref.@xcite , which should be @xmath5 due to the luneberg lens @xcite of the mean field potential , was done simply based on the similarity of the shape of the angular distributions and the airy minimum between @xmath0o+@xmath1c and @xmath0o+@xmath7c scatterings at the same energy .
( color online ) the energy evolution of the airy structure in the angular distributions of cross sections ( ratio to rutherford scattering ) in @xmath0o+@xmath1c scattering calculated using the single channel df potential by switching off the imaginary potential is shown by the solid lines .
the dashed lines at 116 and 132 mev are the farside component of the calculated cross sections . , width=325 ] ( color online ) comparison of the coupled channel calculations ( blue solid line ) with experimental angular distributions of cross sections ( ratio to rutherford scattering ) ( points ) @xcite in @xmath0o+@xmath1c scattering at @xmath2= 132 mev ( a ) , 281 mev ( b ) , and 382.2 mev ( c ) .
the dashed ( green ) and dash - dotted ( grey ) lines display the calculated farside and nearside components , respectively .
the angular distributions calculated by switching off the imaginary potential ( w=0 ) are displayed with dotted ( pink ) lines .
note that the upper horizon scale is for ( a ) and ( b ) and the lower horizontal scale is for ( c ) .
, width=328 ] in fig .
3 the angular distributions calculated using the coupled channel method are compared with the experimental data .
the potential parameters used are given in table i. the experimental angular distributions are well reproduced by the coupled channel calculations .
there is little difference between the coupled channel and the single channel calculations in fig . 1 at the higher energies , 382.2 and 281 mev , although the airy minimum @xmath29 is slightly shifted forward at 132 mev compared with the single channel calculation .
essentially , the effect of the channel coupling on the airy structure is not important and the angular distributions are well described by the mean field df potential .
a dynamical secondary rainbow due to the coupling to the excited state of @xmath7c observed in @xmath0o+@xmath7c rainbow scattering is not seen in the calculated angular distributions . in this sense , @xmath0o+@xmath1c rainbow scattering is similar to the @xmath0o+@xmath0o system @xcite rather than the @xmath0o+@xmath7c system @xcite in the way that @xmath16+@xmath1c scattering @xcite is similar to @xmath16+@xmath0o scattering @xcite .
( color online ) the observed minima ( filled circles ) in the 90@xmath4 excitation functions for the @xmath0o+@xmath0o @xcite , @xmath7c+@xmath7c @xcite , and @xmath0o+@xmath7c @xcite systems are shown as a function of the reduced mass .
the predicted airy minima for the @xmath0o+@xmath1c system , @xmath29 and @xmath5 , are indicated by open squares .
the line is to guide the eye . , width=328 ] ( color online ) the values of the volume integrals per nucleon pair of the real potential , @xmath24 , for @xmath0o+@xmath1c rainbow scattering ( red filled circle ) are shown in comparison with those for @xmath0o+@xmath7c ( pink open square @xcite , pink filled diamonds @xcite ) , @xmath0o+@xmath0o ( blue upper triangle ) @xcite and @xmath7c+@xmath7c ( black down triangle ) @xcite rainbow scattering .
the line is to guide the eye .
, width=328 ]
airy elephants in the @xmath38 excitation function in heavy - ion rainbow scattering has been a continuing interest @xcite since their famous discovery in the @xmath7c+@xmath7c excitation function @xcite .
the existence of the airy elephants and their numbers can be determined calculating the airy minima that cross the @xmath38 excitation function . to determine the airy minima theoretically , the global interaction potential that describes rainbow scattering over a wide range of incident energies has to be determined uniquely .
the energy of the @xmath29 minimum in the @xmath38 excitation function can be determined using the global potential . in the @xmath0o+@xmath0o system , which has been most thoroughly investigated ,
its unique global potential has made it possible to understand airy structure @xcite , molecular resonances and cluster structure with the @xmath0o+@xmath0o configuration at low energy in a unified way @xcite .
as seen in fig . 4 , the @xmath29 airy minimum in the @xmath38 excitation function appears at around @xmath8=95 mev and other higher order airy minima @xmath5 , @xmath6 , @xmath39 , @xmath40 and @xmath41 appear at around @xmath8= 75 , 62 , 47 , 40 and 32 mev , respectively @xcite .
the highest order of the airy minimum is @xmath41 for the @xmath0o+@xmath0o system . at the lower energies below @xmath8=32 mev
, the gross structure of the airy structure evolves into the gross structure of the molecular resonances with the @xmath0o+@xmath0o structure as was shown in ref.@xcite . in fig .
4 , the observed @xmath29 airy minimum and the highest order airy minimum for the @xmath7c+@xmath7c system determined from refs.@xcite and those for the @xmath0o+@xmath7c system from ref.@xcite are also displayed . for the @xmath0o+@xmath7c system the energy of the @xmath29 airy
minimum was interpolated from the experimental result at @xmath2=170 mev ( @xmath8=72.9 mev ) and 132 mev ( @xmath8=56.6 mev ) in ref.@xcite . in fig .
4 there exist four airy minima for @xmath7c+@xmath7c system and three airy minima for the @xmath0o+@xmath7c system . for the @xmath0o+@xmath1c system , although there is no experimental data , the energy evolution of the airy minimum in fig .
2 predicts that the @xmath29 minimum appears at @xmath38 at @xmath2=158 mev and the @xmath5 minimum appears at @xmath2=116 mev .
we see in fig . 4 that the @xmath0o+@xmath1c system is situated between the two identical systems , @xmath0o+@xmath0o and @xmath7c+@xmath7c .
the energy between the @xmath29 and @xmath5 minima of about 20 mev is similar to that of the @xmath0o+@xmath0o system rather than the @xmath0o+@xmath7c system . from this similarity , airy minima with orders higher than @xmath6 are likely to exist below @xmath2=115 mev before the transition into the molecular resonances with the @xmath0o+@xmath1c structure that have been observed in the @xmath8=30 mev region @xcite . in fig .
5 the energy evolution of the volume integral of the real potential for the @xmath0o+@xmath1c system is compared with the systematic data for the @xmath0o+@xmath0o , @xmath7c+@xmath7c and @xmath0o+@xmath7c systems .
the volume integrals for the @xmath0o+@xmath1c system are consistent with the behavior of the other systems in the energy region where experimental data are available .
it seems that the number of the airy minima for identical systems is larger than that for the asymmetric systems .
it is highly desired to observe the airy minima @xmath5 and @xmath6 for the @xmath0o+@xmath1c system experimentally .
the lowest energy ( highest order ) airy structure , airy elephant , will evolve into the molecular resonances with the @xmath0o+@xmath1c cluster structure in the lower energy region similar to the @xmath0o+@xmath0o system @xcite .
the molecular resonance with the @xmath0o+@xmath1c structure has been studied theoretically @xcite and experimentally @xcite . the existence of the molecular resonances with 18@xmath42 , @xmath43 and @xmath44 ( or 20@xmath42 , @xmath44 and @xmath45 ) at @xmath8=23.4 , 27.4 and 31.05 mev , respectively ,
have been reported by freeman _
et al._@xcite .
_ reported the existence of the molecular resonances with @xmath46 = 11 , 13 , 17 and ( 15 ) at @xmath8 = 18.2 , 19.1 , 22.9 and 23.8 mev , respectively @xcite .
therefore it is expected that the gross structure , airy elephant , corresponding to the fourth or fifth airy minimum in the 90@xmath4 excitation function evolves into molecular resonance at around @xmath8=30 mev .
the experimental study of @xmath0o+@xmath1c elastic scattering in the energy region below @xmath2=132 mev and above 65 mev is highly desired to connect the airy structures ( airy elephants ) and the molecular resonances with the @xmath0o+@xmath1c configuration .
to summarize , we studied the airy structure in @xmath0o+@xmath1c rainbow scattering with an extended double folding ( edf ) model that describes all the diagonal and off - diagonal coupling potentials derived from the microscopic wave functions for @xmath0o using a density - dependent nucleon - nucleon force . the experimental angular distributions at @xmath2=132 , 281 and 382.2 mev were analyzed and well reproduced by the theoretical calculations .
the airy minimum at @xmath3=76@xmath4 in the angular distribution at @xmath2=132 mev was found to be a second order airy minimum @xmath5 .
the number of the airy minima in the 90@xmath4 excitation function was investigated in comparison with the typical identical @xmath0o+@xmath0o and @xmath7c+@xmath7c systems and at least two airy minima , airy elephants , are predicted to exist above @xmath2=110 mev ( @xmath8=51 mev ) .
the evolution of the airy minima in the 90@xmath4 excitation function related to the airy elephants into molecular resonances with the @xmath0o+@xmath1c cluster structure in the low energy region around @xmath8=30 mev is discussed .
aa d. t. khoa , w. von oertzen , h.g .
bohlen , s. ohkubo , j. phys . *
g 34 * , r111 ( 2007 ) .
p. e. hodgson , _ nuclear heavy ion reactions _ ( clarendon press , oxford 1978);r . bock _ heavy ion collisions _ , vol . 1 ( north - holland publishing , amsterdam 1979 ) and vol.2 ( 1980 ) .
s. ohkubo and k. yamashita , phys .
c * 66 * , 021301(r ) ( 2002 ) .
f. michel , g. reidemeister , s. ohkubo , phys .
63 * , 034620 ( 2001 ) ; f. michel , f. brau , g. reidemeister , and s. ohkubo , phys .
* 85 * , 1823 ( 2000 ) .
k. w. mcvoy and m. e. brandan , nucl . phys . * a 542 * , 295 ( 1992 )
. s. szilner _
et al . _ ,
c * 64 * , 064614 ( 2001 ) .
a. a. ogloblin _
c * 57 * , 1797 ( 1998 ) .
a. a. ogloblin _
et al . _ , phys .
c * 62 * , 044601 ( 2000 ) .
m. p. nicoli _
et al . _ ,
c * 61 * , 034609 ( 2000 ) .
d. t. khoa , w. von oertzen , and h. g. bohlen , phys .
c * 49 * , 1652 ( 1994 ) .
s. ohkubo and k. yamashita , phys . lett . *
b578 * , 304 ( 2004 ) .
a. a. ogloblin _
et al . _ ,
phys . at .
nucl . * 66 * , 1478 ( 2003 ) .
s. ohkubo and y. hirabayashi , phys .
* c 89 * , 051601(r ) ( 2014 ) .
a. glukhov , v. p. rudakov , k.p.artemov , a. s.demyanova , a. a. ogloblin , s. a. goncharov , and a. izadpanakh , phys . at .
* 70 * , 1 ( 2007 ) . h. de vries , c. w. de jager , and c. de vries , at . data and nucl . data tables * 36 * , 495 ( 1987 ) ; f. j. kline , h. crannell , j. t. obrien , j. mccarthy , r. r. whitney , nucl .
phys . * a 209 * , 381 ( 1973 ) .
s. okabe , _ tours symposium on nuclear physics ii _ , edited by h. utsunomiya _
( world scientific , singapore , 1995 ) , p. 112 ; private communication . y. suzuki , prog .
. phys . * 55 * , 1751 ( 1976 ) ; prog .
* 56 * , 111 ( 1976 ) . y. hirabayashi and s. ohkubo , phys .
c * 88 * , 014314 ( 2013 ) . s. ohkubo and y. hirabayashi , phys .
rev . * c 89 * , 061601(r ) ( 2014 ) .
s. ohkubo , y. hirabayashi , a. a. ogloblin , yu . a. gloukhov , a. s. demyanova , and w. h. trzaska , phys . rev .
* c 90 * , 064617 ( 2014 ) . a. m. kobos _
et al . _ ,
a384 * , 65 ( 1982 ) ; a. m. kobos _
et al . _ ,
a425 * , 205 ( 1984 ) .
g. r. satchler and w. g. love , phys . rep .
* 55 * , 183 ( 1979 ) .
m. e. brandan and g. r. satchler , phys .
rep . * 285 * , 143 ( 1997 ) .
a. t. rudchik _
et al . _ ,
j. * a 47 * , 50 ( 2011 ) .
f. michel , g. reidemeister , s. ohkubo , phys .
* 89 * , 152701 ( 2002 ) .
dao t. khoa , w. von oertzen , h. g. bohlen , f. nuoffer , nucl .
* a672 * , 387 ( 2000 ) .
f. michel and g. reidemeister , phys . rev .
c * 47 * , r1486 ( 1993 ) .
f. michel , s. ohkubo , g. reidemeister , prog .
. suppl . * 132 * , 7 ( 1998 ) .
a. s. demyanova _
et al . _ , nucl . phys . * a 834 * , 473c ( 2010 ) ; a. s. demyanova _
et al . _ , _ international symposium on exotic nuclei _ , sochi ( russia ) , 2009 ( exon2009 ) , aip conference proceedings no.1224 ( 2010 ) ( eds .
penionzhkevich , s. m. lukyanov ) , p.82 .
m. p. nicoli , f. haas , r. m. freeman , n. aissaoui , c. beck , a. elanique , r. nouicer , a. morsad , s. szilner , z. basrak , m. e. brandan , and g. r. satchler , phys .
c * 60 * , 064608 ( 1999 ) .
m. l. halbert _
et al . _ ,
. lett . * 51b * , 341 ( 1974 ) . f. michel and s. ohkubo , eur .
j. * a 19 * , 333 ( 2004 ) .
w. reilly , r. wieland , a. gobbi , m. w. sachs , j. maher , r. h. siemssen , d. mingay , and d. a. bromley , nuovo cimento a , * 13 * , 913 ( 1973 ) .
u. abbondanno , f. demanins , g. vannini , l. vannucci , p. boccaccio , r. dona , r .a .
ricci , m. bo@xmath47 in , and n. cindro , j. phys .
g * 16 * , 1517 ( 1990 ) .
freeman , z. basrak , f. haas , a. haehem , g.a .
monnehan , and m. youlal , z.phys . * a 341 * , 175 ( 1992 ) .
heenen and d. baye , phys . lett . *
b81 * , 295 ( 1979 ) . | the airy structure in @xmath0o+@xmath1c rainbow scattering is studied with an extended double folding ( edf ) model that describes all the diagonal and off - diagonal coupling potentials derived from the microscopic realistic wave functions for @xmath0o using a density - dependent nucleon - nucleon force .
the experimental angular distributions at @xmath2=132 , 281 and 382.2 mev are well reproduced by the calculations . by studying the energy evolution of the airy structure , the airy minimum at around @xmath3=76@xmath4 in the angular distribution at @xmath2=132 mev
is assigned as the second order airy minimum @xmath5 in contrast to the recent literature which assigns it as the third order @xmath6 .
the airy minima in the 90@xmath4 excitation function is investigated in comparison with well - known @xmath0o+@xmath0o and @xmath7c+@xmath7c systems .
evolution of the airy structure into the molecular resonances with the @xmath0o+@xmath1c cluster structure in the low energy region around @xmath8=30 mev is discussed .
it is predicted theoretically for the first time for a non-@xmath9 @xmath0o+@xmath1c system that airy elephants in the 90@xmath4 excitation function are present . |
observations of star formation within the milky way and other nearby metal - rich galaxies show that the surface density of star formation correlates better with the surface density of molecular gas than with the total surface density or the surface density of atomic gas ( @xcite ; although see also @xcite ) .
it is natural to assume that this correlation arises because the presence of molecular gas is a necessary prerequisite for star formation , but recent studies by @xcite , @xcite and @xcite have shown that this interpretation is incorrect .
numerical simulations of star formation in dense gas clouds have been used to demonstrate that the star formation rate in these clouds is largely insensitive to the composition of the gas .
if molecule formation and molecular cooling are artificially suppressed , the predicted star formation rate that one obtains is very similar to that coming from models that include molecular cooling @xcite .
these models suggest that the observed correlation between molecular gas and star formation comes about because the conditions that favour star formation also favour the formation of molecules .
furthermore , they predict that this correlation will break down at low metallicities @xcite .
however , these studies focussed on the thermal physics of individual gas clouds that were assumed to have already been assembled from the diffuse interstellar medium ( ism ) .
they did not address the issue of whether molecular gas can play an important role at an earlier stage of the process , prior to , or during the assembly of , these dense , cold clouds .
this question has been examined to some extent by a few previous numerical studies .
@xcite performed simulations of protogalaxy formation using a simplified chemical model that accounted for h@xmath1 formation and destruction , as well as the ionization and recombination of c , o and si , and showed that under these conditions , h@xmath1 cooling dominates over metal line cooling at low densities ( @xmath3 and below ) for metallicities @xmath4 .
however , their study did not include the effects of a pre - existing background radiation field , which is arguably a reasonable approach when considering the formation of the earliest protogalaxies , but which is unlikely to be valid for low metallicity galaxies in the local universe .
more recently , @xcite reported that in their galaxy formation simulations , h@xmath1 cooling dominates over metal - line cooling at temperatures of a few thousand kelvin and below for simulations with various different gas - phase metallicities and ultraviolet ( uv ) field strengths .
however , the limited spatial resolution of their model means that they have to include a clumping factor into their h@xmath1 formation rate in order to correct for unresolved small - scale density fluctuations .
this correction factor is easy to justify within regions representing dense atomic or molecular clouds , which are dominated by turbulence and therefore have significant density substructure , but it is less obvious whether including a correction of this form is appropriate for regions corresponding to lower density , warm patches of the ism , which are dominated by thermal motions and hence have much less turbulent substructure .
another important recent study is that of @xcite .
they performed a series of simulations of a small protogalaxy ( @xmath5 ) and explored the effects of varying the initial metallicity of the gas and the strength of the ambient radiation field .
they found that at metallicities below @xmath6 , their model protogalaxies were cooled primarily by h@xmath1 and that cooling in these low metallicity systems was very sensitive to the strength of the radiation field , being largely suppressed for field strengths @xmath7 in units of the @xcite field .
one drawback of all of these studies is that their computational cost limits the number of simulations that can be performed and makes it difficult to explore a wide parameter space of densities , metallicities etc .
their results suggest that there are situations in which h@xmath1 cooling is important in the low density ism , but these simulations do not , by themselves , allow us to identify the full range of physical conditions for which this is the case . for this reason ,
it is useful to study the interplay between chemistry and thermodynamics in the low density ism using simpler methods that do not have a large computational cost . in this paper , we carry out such a study .
we use a simple one - zone model of the ism , in which the chemical and thermal evolution of gas that is initially hot and ionized is followed in detail , but where the gas density is held constant . with this setup
, we can follow the evolution over long periods of time for minimal computational cost , allowing us to explore the behaviour of the gas for a wide range of different densities , metallicities and uv field strengths .
our aim is to establish the physical conditions in which gas that is initially in a warm , low - density state can cool significantly within a dynamical time , which we take to be a necessary prerequisite for the formation of molecular clouds and , ultimately , stars . in order to do this , we make use of a very simple one - zone model of the thermal and chemical evolution of the ism . in this model , we keep the gas density fixed , and simply track the evolution of the gas and dust temperatures , as well as the chemical composition of the gas . we note that although gas that is able to cool will likely also increase its density either by gravitational collapse or due to the effects of isobaric compression from surrounding warmer material
gas that is unable to cool is unlikely to change its density significantly unless perturbed by some external force .
our model is therefore well - suited to identifying which gas can cool and which can not , but will provide an incomplete picture of the behaviour of gas that does manage to cool .
we model the chemical and thermal evolution of the gas in the ism using a simplified chemical network coupled with a detailed atomic and molecular cooling function .
our chemical model is based on three main sources .
we take our treatment of the collisional gas - phase chemistry of hydrogen , helium and deuterium from @xcite , and our treatment of the collisional gas - phase chemistry of carbon , oxygen and silicon from @xcite .
rates for the key photochemical reactions ( e.g. h@xmath8 photodetachment , h@xmath1 photodissociation ) are taken from @xcite .
our model tracks the abundances of atomic and ionized carbon , oxygen and silicon , but does not include the formation of molecules containing these elements ( e.g. co , water ) .
our justification for ignoring these heavy molecules is that at the densities and temperatures considered in this study , the cooling they provide per molecule is no more than a factor of a few larger than the cooling per atom provided by atomic fine structure cooling @xcite .
therefore , cooling from these heavy molecules will be important only when a significant fraction of the available carbon and oxygen has been incorporated into molecules .
we know that in the local ism , this occurs only in dense , well - shielded gas , and this gas cools rapidly , in much less than a free - fall time , even when cooling from these molecules is neglected @xcite . at lower metallicity ,
the dust extinction in the ism will be lower and we therefore expect these heavy molecules to be even less important for determining whether the gas can cool .
we therefore do not expect that this simplification will significantly affect our conclusions .
we account for the formation of h@xmath1 on dust grains using the standard @xcite prescription , but do not include any other grain - surface chemistry in our model .
heating in our model comes primarily from the photoelectric effect and cosmic ray heating , with additional minor contributions from effects such as h@xmath1 formation heating or ultraviolet pumping of excited vibrational states of h@xmath1 .
cooling is provided by the electronic excitation of atomic h , he and he@xmath9 , rotational and vibrational emission from h@xmath1 and hd , fine structure and metastable line emission from c , c@xmath9 , o , si and si@xmath9 , and thermal emission from dust .
metastable line emission represents a new addition to our cooling function and is treated using the data given in table 9 of @xcite .
the collisional rate coefficients that we adopt for the various chemical reactions in our model are for the most part the same as in @xcite ( for the metals ) and @xcite ( for the h , he and d chemistry ) , but in a few cases we have updated the values , in light of new experimental or theoretical data . for the associative detachment reaction @xmath10 we now use the new rate coefficient measured by @xcite in place of the older value referenced in @xcite . for the mutual neutralization reaction
@xmath11 we use the rate coefficient given in @xcite , which agrees well with recent theoretical @xcite and experimental @xcite determinations .
we have also updated several of the rate coefficients used in our cooling function . to model the contribution to the h@xmath1 cooling rate made by h@xmath1-proton collisions ,
we now make use of the excitation rates recently calculated by @xcite for the transitions for which these are available , supplementing them with data from @xcite and @xcite for those transitions for which newer data is not available .
the contribution of h@xmath1-electron collisions is now modelled using the excitation rate coefficient data given in @xcite , in place of the much older data used in our previous treatment ( see * ? ? ?
finally , in our treatment of fine structure cooling now makes use of the new excitation rates for c - h and o - h collisions computed by @xcite in place of the older values used in @xcite .
we consider gas that is initially warm ( @xmath12 ) and fully ionized , although in section [ res : ion ] below we discuss the effects of starting with a reduced level of ionization .
we explore the influence of three main free parameters : the gas density , the metallicity , and the strength of the interstellar radiation field ( isrf ) .
we can describe the gas density in terms of the number density of hydrogen nuclei @xmath13 , which is related to the mass density by @xmath14 , where @xmath15 is the proton mass , and we have assumed a 10:1 ratio of hydrogen to helium , as is appropriate for the local ism . to parameterize the metallicity
, we assume that at solar metallicity , the elemental abundances of c , o and si relative to hydrogen are the same as those measured in the warm neutral medium ( see e.g. * ? ? ? * ) . at other metallicities , we assume that the elemental abundances simply scale linearly with the total metallicity z. we therefore have @xmath16 where @xmath17 , @xmath18 , and @xmath19 are the total fractional abundances of these three elements .
we also assume in most of our models that the dust - to - gas ratio , @xmath20 , scales linearly with metallicity , although in section [ res : dust ] below we report the results of runs in which we adopted a steeper dependence of @xmath20 on z. to model the isrf , we adopt the spectral shape described in @xcite in the ultraviolet , and that from @xcite at longer wavelengths .
we assume that the shape of the spectrum does not change as we change the normalization , @xmath21 , where setting @xmath22 gives us the original draine and mathis et al
. normalizations . in some of our models
, we include an approximate treatment of the effects of h@xmath1 self - shielding .
this is modelled using a self - shielding function based on the work of @xcite , but modified according to the prescription in @xcite in order to more accurately represent the effects of self - shielding in warm gas . as input to the self - shielding function , we need to provide an h@xmath1 column density .
we compute this as @xmath23 , where @xmath24 is the current h@xmath1 number density and @xmath25 is a characteristic self - shielding scale length . in our default set - up , we set @xmath26 , so that there is no self - shielding .
we explore the effects of adopting a non - zero value of @xmath25 in section [ res : h2ss ] .
we adopt a cosmic ray ionization rate for atomic hydrogen given by @xmath27 .
we include a dependence on @xmath21 to reflect the fact that the cosmic rays ionization rate and the uv photodissociation rate would both be expected to increase as the star formation rate increases .
the ionization rates for our other chemical species ( h@xmath1 , he , c , etc . )
relative to @xmath28 are scaled as described in @xcite .
we begin our study by examining to what extent the ism can cool in the absence of h@xmath1 cooling . to do this
, we use the one - zone model described in the previous section to explore the thermal evolution of the gas for a wide range of different densities , @xmath29 , and isrf strengths , @xmath30 , at eight different metallicities : @xmath31 and @xmath32 . for each value of z , @xmath21 and @xmath13 , we run the model for a single gravitational free - fall time , @xmath33 , and examine the temperature at the end of this time period . our choice of the free - fall time here is motivated by the classic rees - ostriker criterion for dynamical fragmentation @xcite .
however , we have verified that our results are not significantly different if we allow the gas to evolve for e.g. two free - fall times .
our results are plotted in figure [ no - h2 ] . at solar metallicity
, we see that there are two main temperature regimes . for values of @xmath34
greater than around 0.3 , the gas remains hot , with a temperature of around 80009000 k. for lower values of @xmath34 , however , the gas cools significantly , reaching temperatures of order a few hundred k when @xmath35 , and temperatures as low as 20 k when @xmath36 .
we can understand this behaviour as a consequence of the fact that over a wide range in temperatures , the cooling of the neutral ism is dominated by fine structure line emission , while the heating is dominated by photoelectric heating ( see e.g. * ? ? ?
* ; * ? ? ?
if we write the cooling rate as @xmath37 where @xmath38 depends only weakly on @xmath39 for @xmath40 k , and the heating rate as @xmath41 where @xmath42 is also a weak function of @xmath39 , then it is easy to show that in solar metallicity gas , @xmath43 since the temperature dependence of both the photoelectric heating rate and the fine structure cooling rate is weak in the temperature range of interest , it is the size of the factor @xmath34 that primarily determines whether heating or cooling dominates , and hence whether the gas remains close to its starting temperature , or cools until it reaches a temperature at which further fine structure cooling becomes ineffective .
if we now decrease the metallicity of the gas by an order of magnitude , we see that the main change which occurs is a slight increase in the size of the region in which cooling is ineffective .
the critical value of @xmath34 decreases from 0.3 to around 0.1 , but the behaviour of the gas for @xmath44 or @xmath45 does not significantly change . again
, this can be understood as a consequence of the cooling rate being dominated by fine structure emission and the heating rate by photoelectric emission from dust . since we have assumed that the dust - to - gas ratio is proportional to the metallicity , both the heating and the cooling rates scale linearly with the metallicity , and the resulting equilibrium temperature is largely independent of metallicity .
the fact that it is not completely independent is largely due to the fact that the cosmic ray heating rate does not decrease when the metallicity decreases , meaning that cosmic ray heating becomes increasingly dominant as we move to lower metallicities .
if we decrease the metallicity further , to @xmath46 , we find that there is a qualitative change in the behaviour of the gas . at densities @xmath47
, the gas behaves largely as before , but at lower densities , the gas remains warm regardless of the value of @xmath21 .
this change in behaviour is not due to a change in the equilibrium temperature of the gas , which remains low when @xmath21 is low .
instead , it reflects the fact that at low densities in this simulation , the gas does not have time to reach its equilibrium temperature . in this regime ,
the cooling time @xmath48 becomes larger than the free - fall time @xmath49 before the temperature of the gas can reach equilibrium .
for example , consider the case of gas with @xmath50 k at a number density @xmath51 . for these conditions , and for @xmath46 , the total cooling rate due to
fine structure emission from c@xmath9 , si@xmath9 and o is approximately @xmath52 .
the cooling time of this gas is therefore at least @xmath53 , regardless of the value of @xmath21 .
this is significantly longer than the free - fall time , which for gas at this density is approximately @xmath54 , and hence the gas is unable to cool below around 6000 k within a free - fall time . continuing to decrease the metallicity beyond this point
makes it even harder for the gas to cool within a free - fall time , as its cooling time scales inversely with metallicity while its free - fall time remains fixed . by the time we reach a metallicity of @xmath55
, we see that none of the gas in the region of g@xmath56@xmath13 parameter space examined here can cool very much within a free - fall time .
it is also easy to see how these results would change if we examined the gas after a longer period of time . since the limiting factor in the gas with low density and low @xmath21 is the time available for cooling , increasing the time period considered from @xmath49 to e.g.@xmath57 would shift the density at which the transition from hot gas to cold gas occurs , in this case by a factor of 4 .
our precise results therefore depend on our choice of @xmath49 as the moment to examine the gas temperature . however , to completely offset the effects of a substantial drop in metallicity
, one would have to consider a much longer time period ; for example , a decrease in @xmath58 by a factor of ten could be mitigated by looking at the gas after ten free - fall times . in practice , it is unlikely that the gas would remain in an undisturbed state for such a long time period , and even if it did so , gas which can cool and collapse only on a timescale @xmath59 is unlikely to form stars efficiently . in summary , we see that for the majority of the metallicities considered here , the main factor that limits the temperature to which the gas can cool within a free - fall time is not the influence of photoelectric and/or cosmic ray heating , it is simply the fact that the cooling time is long compared to the free - fall time .
we next look at how the ability of the ism to cool changes once we include the effects of h@xmath1 and hd cooling . in figure [ with - h2 ] , we show the results of a similar set of simulations to those in the previous section .
the only difference in these simulations is that in this case , we account for molecular cooling .
we see immediately that at solar metallicity , there is essentially no difference in the outcome .
indeed , if we quantitatively compare the temperatures reached at each point in @xmath21density space in this run with those in the model without h@xmath1 cooling , we find that the maximum difference in the final temperature is less than 1% .
we have also looked in detail at the fraction of the total cooling provided by h@xmath1 throughout the lifetime of each of the runs , and find that this is never greater than around 10% .
we can therefore immediately conclude that in metal - rich gas , and in the absence of self - shielding , h@xmath1 cooling does not play a significant role in regulating the temperature .
this is consistent with the idea that in galaxies like our own milky way , the observed correlation between molecular gas and star formation is a consequence of the fact that both molecules and stars form preferentially in regions with high column density and high volume density , rather than an indication that molecular cooling is required for star formation ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . at lower metallicities , however , we start to see significant differences between the two sets of runs .
we have already seen that in the runs without molecular cooling , as we lower the metallicity , the the question of whether a given parcel of gas can cool within a free - fall time becomes determined almost entirely by the gas density , with @xmath21 having little or no influence on whether the gas can cool .
this is not the case in the runs that include molecular cooling . in these runs
, @xmath21 plays an important role in determining the outcome of the simulations for all of the metallicities that we examine . for metallicities in the range @xmath60 , the boundary dividing those regions in @xmath21density space that cool from those that do not roughly follows a line along which @xmath61 is constant , while at lower metallicities , the dividing line is better described as a line of constant @xmath34 .
we see also that the temperature reached by the cooling gas increases significantly as we decrease the metallicity , from @xmath6220 k at solar metallicity to @xmath63300 k at @xmath64 , although we caution that these numbers should be treated with care , as in a realistic system , it is likely that the gas would not remain at constant density as it cools .
the difference between the runs with and without molecular cooling is particularly apparent at metallicities below @xmath6 . at these metallicities ,
little cooling occurs in the runs without h@xmath1 , whereas in the runs with h@xmath1 , cooling remains efficient over a large portion of the parameter space that we examine . in order to understand the behaviour that we find in these runs , and in particular why we recover the scalings that we do , it is useful to look at how the fractional abundance of h@xmath1 evolves with time in gas . at metallicities @xmath65 ,
the final h@xmath1 fraction is set by the balance between h@xmath1 formation on dust and h@xmath1 photodissociation @xcite .
we can write the h@xmath1 formation rate as @xmath66 where @xmath67 is the reaction rate coefficient for h@xmath1 formation on grain surfaces ( see e.g. * ? ? ?
* ; * ? ? ?
* ) , @xmath13 is the number density of hydrogen nuclei and @xmath68 is the number density of atomic hydrogen . for the h@xmath1 photodissociation rate
, we have @xmath69 where @xmath70 is the photodissociation rate per h@xmath1 molecule , which for our adopted radiation field is given by @xmath71 . combining these equations ,
it is easy to show that in chemical equilibrium , the h@xmath1 number density is simply @xmath72 this equilibrium state is reached on a timescale @xmath73 , and hence @xmath74 for the majority of the densities and uv field strengths that we examine . at densities far below the h@xmath1 critical density @xmath75 , the cooling time of the gas due to h@xmath1 line cooling can be written as @xmath76 where @xmath77 is the h@xmath1 cooling rate expressed in units of @xmath78 , which is in general a strong function of temperature . at a fixed gas temperature
, @xmath79 therefore scales with the h@xmath1 number density as @xmath80 if we now set @xmath81 , then since @xmath82 , it is easy to demonstrate that our required h@xmath1 number density varies as @xmath83 as we change @xmath13 . comparing this with equation [ nh2_eq_hz ]
, we see that if for some combination of @xmath13 and @xmath21 we have @xmath84 ( i.e. our equilibrium h@xmath1 fraction is just enough to cool the gas within a free - fall time ) , then the same condition will hold for other values of @xmath21 and @xmath13 only if @xmath85 in agreement with the behaviour that we see in figure [ with - h2 ] .
, plotted as a function of time , in gas with density @xmath86 , illuminated by an isrf with strength @xmath87 , and with metallicity @xmath88 ( solid line ) or @xmath55 ( dashed line ) .
[ time - evol],scaledwidth=50.0% ] below @xmath89 , the behaviour of the gas changes , owing to the increasing influence of the gas - phase processes @xmath90 the behaviour of the gas in this regime has been analyzed in detail by @xcite , who show that the ability of the gas to cool is determined by the size of the ratio @xmath34 .
when @xmath34 is very small ( @xmath91 ) , the average uv flux in the lyman - werner bands in units of @xmath92 . to convert approximately between these two different normalization schemes , we can use the fact that @xmath22 corresponds to @xmath93 . ] , the most important h@xmath0 destruction process is the charge transfer reaction @xmath94 some of the h@xmath95 ions produced by this reaction reform h@xmath0 via the inverse reaction @xmath96 but many are instead destroyed by dissociative recombination , @xmath97 and the end result is a net loss of h@xmath0 from the gas . at high temperatures ,
this process strongly limits the equilibrium h@xmath1 abundance .
however , as the gas cools , the rate of reaction [ h2_dest_ct ] falls off exponentially , allowing more h@xmath1 to form ( and more cooling to occur ) , until the h@xmath0 fractional abundance freezes out at a value of approximately @xmath98 @xcite , owing to the ongoing loss of electrons from the gas . when @xmath99 , on the other hand , the behaviour of the h@xmath1 fraction is quite different .
initially , it increases steadily with time until it reaches an equilibrium set by the balance between h@xmath0 formation via h@xmath8 and h@xmath1 destruction by charge transfer and by photodissociation . following this , however , it begins to decrease again as the electron fraction drops and the gas recombines .
the falling electron fraction leads to a drop in the h@xmath1 formation rate , and although the destruction rate due to charge transfer also falls , the photodissociation rate does not , and so the equilibrium h@xmath1 fraction decreases . this behaviour is illustrated in figure [ time - evol ] , where we plot the time evolution of the h@xmath1 fraction for two different metallicities , @xmath88 and @xmath55 , for gas with @xmath86 and @xmath87 . in the higher metallicity run ,
the h@xmath1 fraction increases steadily with time until it reaches a peak value of around @xmath98 at @xmath100 . at this point ,
the fractional ionization of the gas is around @xmath101 , h@xmath1 formation is dominated by the gas - phase reaction chain , and the equilibrium h@xmath1 fraction is around @xmath98 . at later times
, the equilibrium h@xmath1 fraction decreases as the gas recombines .
however , once the fractional ionization has decreased by a further order of magnitude , h@xmath1 formation on grain surfaces takes over as the most important process , and the h@xmath1 fraction tends to a roughly constant equilibrium value of @xmath102 . in the lower metallicity run ,
the initial behaviour is very similar . however , in this case , the much lower dust abundance means that h@xmath1 formation on grain surfaces is far less effective and gas - phase formation of h@xmath1 remains the dominant h@xmath0 formation process throughout the period shown in the figure . as a result , the h@xmath0 fraction is far more sensitive to the falling fractional ionization of the gas , and by the end of the period plotted has decreased by almost two orders of magnitude in comparison to its peak value .
the significant fall - off with time of the h@xmath1 fraction that occurs in the low metallicity run , together with the fact that the decreasing ionization also leads to a decreasing cooling rate per h@xmath0 molecule , implies that most of the cooling that occurs in this case takes place at early times , prior to or at the point at which the h@xmath1 abundance reaches its peak value .
consequently , the important timescale in this case is not the free - fall time , but rather is the time taken for the h@xmath1 abundance to reach a maximum , which is simply the photodissociation time @xmath103 .
if @xmath104 , then the gas will cool , but if @xmath105 , it will remain warm . as before , we know that @xmath106 and it is also easy to show that the photodissociation time scales as @xmath107 independent of the value of the h@xmath1 abundance
. therefore , the condition that @xmath108 implies that @xmath109 when @xmath34 is large , this scales with @xmath13 and @xmath21 as @xmath110 and hence our condition that @xmath108 is implies that @xmath111 and hence that @xmath112 in agreement with our simulation results .
so far , we have assumed , for simplicity , that the gas is initially fully ionized . however , we know that in practice this is an overestimate , as the typical ionization fraction in the warm neutral medium is closer to @xmath113 @xcite .
it is therefore important to establish whether this simplification significantly affects our results .
we have therefore run a series of models that include both metal and h@xmath1 cooling , but that start with fractional ionizations given by the following expression @xmath114,\ ] ] where @xmath28 is the cosmic ray ionization rate of atomic hydrogen .
this expression yields values for @xmath115 that are close to the equilibrium value that we would obtain if cosmic ray ionization and radiative recombination were the only processes acting to change the ionization state of the gas .
we find that in general , the change in the initial ionization fraction has only a minor effect on the ability of the gas to cool .
as an example , we show in figure [ lowx ] the results that we obtain for a run with @xmath55 and a low initial fractional ionization .
if we compare this with the corresponding panel in figure [ with - h2 ] , we see that there are only a few differences the region that is unable to cool has become slightly larger , and the final temperature reached by gas with high @xmath13 and low @xmath21 is also somewhat larger , but on the whole , the behaviour is very similar to that in the high ionization fraction case . , but for a set of simulations with metallicity @xmath55 that was run using a much lower value for the initial fractional ionization of the gas .
[ lowx],scaledwidth=40.0% ] in the regions of parameter space dominated by metal - line cooling , it is easy to understand why changing the initial fractional ionization has so little effect . in these simulations ,
the cooling rate is sensitive to the fractional ionization only when the latter is very large @xcite .
therefore , although the cooling rate in the high ionization simulations is initially larger than that in the low ionization simulations , the cooling rates converge as the gas recombines and @xmath116 falls towards its equilibrium value .
it is slightly more surprising that changing @xmath115 has such a limited effect in the case where h@xmath1 cooling dominates , since as we have seen , the h@xmath1 formation rate in these runs depends on the fractional ionization . however , in practice , the dependence of the h@xmath1 formation rate on @xmath116 becomes quite weak for @xmath117 , owing to the influence of the mutual neutralization reaction @xmath118 when @xmath116 is large , most of the h@xmath8 ions formed by reaction [ ra ] are destroyed by this reaction and do not survive for long enough to form h@xmath0 .
consequently , within this regime , further increasing @xmath116 has only a minor effect on the h@xmath1 formation rate and the peak h@xmath1 abundance . for this reason , the effect that decreasing @xmath115 has on the peak h@xmath1 abundance is much smaller than one might initially expect ,
thereby explaining why this change has only a limited effect on the ability of the gas to cool . in most of our models , we assume that the dust - to - gas ratio , @xmath20 , scales linearly with metallicity .
this appears to be a good assumption for galaxies with metallicities close to that of the milky way @xcite . however , there are observational indications that at metallicities below around @xmath119 , @xmath20 falls off more rapidly with decreasing metallicity than predicted by a simple linear scaling @xcite .
we have therefore examined the effect of adopting a scaling for @xmath20 that better matches the observational data , namely @xmath120 we find that changing the dust scaling in this fashion has little effect on our results . obviously , for metallicities @xmath121 , there is no change in @xmath20 from the values used with our default set - up , and hence no change in behaviour .
if we drop the metallicity to @xmath88 , we do see a slight change in behaviour , as illustrated in figure [ new - dust ] : the gas in the simulation with less dust cools more readily than the gas in the simulation with more dust .
this is a consequence of the fact that photoelectric heating is still an important heat source in the diffuse ism at these metallicities . by reducing the dust abundance ,
we reduce the photoelectric heating rate , making it easier for the gas to cool .
nevertheless , the difference between the two runs is small , and grows smaller still if we further reduce the metallicity .
we can therefore be confident that the main results of this study are robust against uncertainties in the nature of the dust in low metallicity galaxies .
up to this point , we have neglected the effects of h@xmath0 self - shielding . however , it is not immediately obvious that this is a good approximation .
the h@xmath1 fractions that we find in the diffuse gas are small , but even h@xmath1-poor gas can produce a significant amount of self - shielding if its total column density is large enough . in the real ism , the amount of self - shielding that we have at any given point depends on the distribution of h@xmath1 column densities around that point , as well as the velocity field of the gas . our simple one - zone model does not allow us to accurately account for these effects , and so to investigate the influence of self - shielding we instead adopt the approximation discussed in section [ sec : num ] : we assume that the h@xmath1 column density is directly related to the h@xmath1 number density via @xmath122 in figure [ h2-shield ] , we show the results we obtain from a set of runs performed using this approximation , with @xmath123 . comparing the results plotted in figures [ with - h2 ] and [ h2-shield ]
, we see that at metallicities @xmath124 , the inclusion of h@xmath1 self - shielding has little effect on the final gas temperature . in these simulations , the regions in @xmath21density parameter space where cooling is inefficient also correspond to regions where the peak h@xmath1 fraction is very small .
the h@xmath1 column density in these regions therefore never becomes large enough to produce significant self - shielding of h@xmath1 , and the final result is therefore the same in simulations with and without this effect included . at higher densities and lower values of g@xmath56 ,
the peak h@xmath0 fraction is much higher and the gas does become able to self - shield , but this occurs in regions that can already cool efficiently , and so once again the inclusion of h@xmath1 self - shielding has little effect on the outcome .
the only significant difference between the two sets of runs is in the minimum temperature reached by the gas with high density and low g@xmath56 : the inclusion of h@xmath1 self - shielding allows this gas to cool to slightly lower temperatures than before .
however , we caution that this result should be treated with caution as it is likely that in reality this strongly cooling gas will change its density in response to the loss of thermal support , rather than remaining at constant density as assumed in our model . in the lower metallicity runs , @xmath125 , the inclusion of h@xmath1 self - shielding has a more pronounced effect .
the boundary in g@xmath56density space between regions that can and can not cool becomes much sharper and also slightly steeper , corresponding to a line where @xmath126 , rather than to a line of constant @xmath34 as in the runs without self - shielding .
the fact that the transition becomes sharper is easy to understand : it is simply a consequence of the sudden jump in the equilibrium abundance of h@xmath1 that occurs once the gas becomes able to self - shield .
this jump occurs because the onset of self - shielding leads to a drop in the photodissociation rate , which leads to an increase in the equilibrium h@xmath1 abundance , which leads to an increase in the h@xmath1 column density and hence a further drop in the photodissociation rate , etc .
this positive feedback drives the equilibrium h@xmath1 abundance sharply upwards once we enter the self - shielding regime .
the steepening of the transition is a consequence of the fact that by including self - shielding , we are including another source of density dependence : the amount by which self - shielding reduces the h@xmath1 photodissociation rate depends on n@xmath127 , which in turn depends on the density via equation [ lss ] .
therefore , in gas with any given value of @xmath34 , self - shielding , and hence h@xmath1 cooling , will always be more effective at large @xmath13 ( and hence large @xmath21 ) than at small @xmath13 .
we have also examined the effect of varying @xmath25 .
we find , as one might expect , that decreasing @xmath25 leads to an increase in the size of the region that does not cool , whereas increasing @xmath25 has the opposite effect .
however , in both cases the effect is fairly small .
the reason for this is that @xmath128 varies strongly as a function of @xmath21 and @xmath13 , as we have already seen , and hence the value of @xmath129 at a given point in @xmath21density space depends far more on the value of @xmath128 than on the value of @xmath25 .
our main results are therefore robust against changes in @xmath25 .
at first sight , it might appear that our results are in conflict with previous studies of star formation in metal - poor systems that typically find that metal - line cooling dominates over h@xmath1 cooling at metallicities above a critical value of around @xmath130 ( see e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , and that dust cooling can dominate at even lower metallicities @xcite .
however , these studies were primarily concerned with higher density gas than is treated in this paper . for example , @xcite derive their constraints on the metallicity required to provide effective fine - structure cooling by considering the behaviour of gas at a density of @xmath131 and a temperature of 200 k without addressing how the gas gets to this density and temperature in the first place .
similarly , studies of dust cooling in very metal - poor systems find that dust dominates the cooling only at very high densities @xcite . in order to reach the densities at which metal cooling begins to dominate in these metal - poor systems , which are orders of magnitude higher than typical ism densities , some other form of cooling
is required .
one possibility is lyman-@xmath132 cooling from atomic hydrogen , but this maintains the gas temperature at close to @xmath133 , resulting in a very large jeans mass ( @xmath134 ) .
this will therefore be an important process only in relatively massive systems , such as protogalaxies illuminated by strong uv radiation fields ( see e.g. * ? ? ?
what our results demonstrate is that when the ambient radiation field is relatively weak , molecular hydrogen alone can cool the gas in less than a free - fall time , thereby enabling it to collapse to the higher densities required for efficient metal cooling .
it is also interesting to compare our results for solar metallicity gas with those presented by @xcite . in their simulations ,
@xcite find that the h@xmath0 cooling rate is almost an order of magnitude larger than the metal - line cooling rate at temperatures of around 1000 k , and hence conclude that h@xmath0 dominates the thermal evolution of the warm gas . on the other hand , in our solar metallicity simulations , we find that h@xmath1 never provides more than about 10% of the total cooling , even in runs in which we account for the effects of h@xmath1 self - shielding .
the reason for this discrepancy is not entirely clear .
our treatment of h@xmath1 cooling differs from that in @xcite owing to the changes we have made to our treatment of h@xmath1-proton and h@xmath1-electron collisions ( see section [ sec : num ] ) , but the overall change is at most a factor of three , far too little to account for the difference in our results .
@xcite also make use of a different treatment of metal - line cooling , based on @xcite and @xcite , but do not provide enough information on the ionization state of the gas in their models to allow us to judge how significantly their total metal - line cooling rate differs from ours . finally , it is possible that the initial conditions that we adopt for the warm gas in our present study do not properly reflect the conditions that @xcite find in their 3d hydrodynamical simulation . specifically , we assume that the gas is initially completely devoid of h@xmath1 .
this means that the amount of h@xmath0 available to participate in the cooling can not be larger than the amount that can form within a cooling time , even if the equilibrium h@xmath1 abundance is in principle much larger . in the @xcite simulations , some of their warm gas will have previously been cold , dense gas , and it may therefore retain some of the h@xmath1 that it formed during this period .
if so , then this would naturally lead to higher h@xmath1 cooling rates than we find in our simulations .
our results demonstrate that h@xmath1 cooling can potentially play an important role in regulating the temperature of the diffuse ism , but only if two important conditions are met .
first , the metallicity must be below @xmath2 , as at higher metallicities , metal line cooling dominates throughout the parameter space considered here , rendering the presence or absence of h@xmath1 irrelevant as far as the thermal evolution of the gas is concerned .
second , the ratio of the isrf strength to the gas density must also be low , typically @xmath135 , or else not enough h@xmath1 will survive to provide effective cooling . in the extreme case where @xmath136 , i.e. when there is no interstellar radiation field present , we predict that h@xmath1 will dominate over a wide range of physical conditions , in good agreement with the results of the previous study of @xcite .
on the other hand , when the radiation field is strong , our results show that h@xmath1 cooling will be ineffective unless the gas density is already large . in a highly turbulent system ( e.g. * ? ? ? * ) or in one which is undergoing large - scale gravitational collapse , such as a high - redshift protogalaxy ( e.g. * ? ? ?
* ; * ? ? ?
* ) , it is possible that the required densities could be achieved without the need for efficient cooling at @xmath137 , but in more quiescent systems , it is likely that the required value of @xmath34 can be achieved only when @xmath21 itself is small , i.e. when the interstellar radiation field is weak .
our results therefore suggest that although h@xmath1 cooling may enable star formation to occur in low metallicity systems that otherwise would be unable to cool , the effect will be strongly self - limiting : if too many stars form , the interstellar radiation field will become too strong , inhibiting further cooling and star formation .
the authors would like to thank f. bigiel , e. ostriker and s. walch for useful discussions on the topic of cooling in the ism .
they would also like to thank the anonymous referee for a useful and constructive report .
financial support for this work was provided by the deutsche forschungsgemeinschaft ( dfg ) via sfb 881``the milky way system '' ( sub - projects b1 and b2 ) , and from the baden - wrttemberg - stiftung by contract research via the programme internationale spitzenforschung ii ( grant p- ls - spii/18 ) .
pcc also acknowledges support from grant cl 463/2 - 1 , part of the dfg priority program 1573 `` physics of the interstellar medium '' . | we use a simple one - zone model of the thermal and chemical evolution of interstellar gas to study whether molecular hydrogen ( h@xmath0 ) is ever an important coolant of the warm , diffuse interstellar medium ( ism ) .
we demonstrate that at solar metallicity , h@xmath1 cooling is unimportant and the thermal evolution of the ism is dominated by metal line cooling . at metallicities below @xmath2
, however , metal line cooling of low density gas quickly becomes unimportant and h@xmath0 can become the dominant coolant , even though its abundance in the gas remains small .
we investigate the conditions required in order for h@xmath0 to dominate , and show that it provides significant cooling only when the ratio of the interstellar radiation field strength to the gas density is small .
finally , we demonstrate that our results are insensitive to changes in the initial fractional ionization of the gas or to uncertainties in the nature of the dust present in the low - metallicity ism .
galaxies : ism ism : clouds ism : molecules stars : formation |
the consideration of unicity properties is motivated by two perceptions .
usually , in a well - posed sudoku it is required that there exists exactly one solution , i.e. , there exists a solution and the solution is unique .
this assumption makes it reasonable to investigate this property in order to gain additional insight , which can be used in the development of algorithms . on the other side in optimization and approximation
theory it is a common approach to impose a unicity condition in order to obtain stronger results .
a classical example is the alternation theorem for haar subspaces in chebyshev approximation .
haar subspaces are a global unicity assumption and the corresponding theory can be found , e.g. , in the monograph of cheney @xcite .
this paper introduces three different types of unicity properties . in section [
s : constraint_sets ] constraint sets are defined .
these sets generalize the terms row , column and block of a classical sudoku puzzle and prepare the unicity statements . in section [ s : unicity_perm ]
we describe the solution set of the generalized sudoku problem in terms of permutations .
these permutations have specific properties on the constraint sets .
the unique solvability depends on the nonexistence of such a permutation . in section [ s : unicity_cells ] we introduce unicity cells , which is a local unicity condition . as a result
we prove that the generalized sudoku problem is uniquely solvable if and only if each cell is a unicity cell . in section [
s : unicity_rectangles ] we consider a generalized type of rectangles as a set of cells .
if the generalized sudoku problem admits a unique solution with a minimality condition on a rectangle , then this rectangle contains a given , i.e. , a populated cell .
unicity properties for the sudoku puzzle had been considered by provan @xcite and herzberg and murty @xcite .
provan showed that the complete solvability of a sudoku puzzle using the pigeon - hole rule implies the unique solvability of the sudoku puzzle .
herzberg and murty presented a necessary condition for the unique solvability . finally , we collect some basic terms and notations .
let @xmath0 denote the set of integers .
the @xmath1-times cartesian product of @xmath0 is indicated by a superscript @xmath1 , i.e. , @xmath2 .
the transpose of a vector or a matrix is indicated by a superscript @xmath3 .
the vectors @xmath4 respectively @xmath5 denote the zero respectively one vector , consisting of zeros respectively ones in each component .
the number of components is indicated by an index .
each vector is considered to be a column vector .
@xmath6 denotes the identity matrix .
the elements of a set are called distinct if each two elements of the set are distinct , i.e. , if and only if the elements are pairwise distinct . the sum over an empty index set
is considered to be zero .
the symbol @xmath7 denotes the number of elements ( cardinality ) of a finite set .
we treat sudoku problems on the basis of a general model introduced in @xcite and replicate the model in this section .
we restrict ourselves to the primal problem and do not consider the dual problem introduced in @xcite .
let @xmath8 be an integer with @xmath9 .
we define the sum @xmath10 and define a matrix @xmath11 with @xmath12 rows and @xmath8 columns inductively . for @xmath13 , let @xmath14 denote the empty matrix , i.e. , a matrix without entries
. assume the matrix @xmath15 had been defined with @xmath16 rows and @xmath17 columns .
we set @xmath18 we extend the matrix @xmath11 to a matrix @xmath19 with @xmath20 rows and @xmath21 columns .
the matrix @xmath19 consists in the main diagonal " of @xmath8 matrices @xmath11 and the remaining values are set to zero .
the matrix @xmath19 depends on the value @xmath8 , but we do not state this dependence explicitly . given the set @xmath22 , let @xmath23 be any permutation on this set , i.e. , @xmath24 be a permutation . we extend the notion of permutation to the matrix @xmath19 , i.e. , we define @xmath25 , where @xmath26 denotes the @xmath27 column of @xmath19 for @xmath28 . given a permutation @xmath23 on @xmath29 , we define the matrix @xmath30 , i.e. , we interchange the columns of @xmath19 according to the permutation @xmath23 . [
d : nonzero ] let @xmath31 .
for any point @xmath32 , we write @xmath33 if each component of @xmath34 is nonzero , i.e. , if @xmath35 for @xmath36 . this definition
should not be confused with the expression @xmath37 , where only one component of @xmath34 has to be nonzero .
given is @xmath38 , some permutations @xmath39 , @xmath40 , @xmath41 on @xmath29 , some @xmath42 , an index set @xmath43 , and givens @xmath44 with @xmath45 for @xmath46 .
the generalized sudoku problem is defined in the following way : @xmath47 we restrict ourselves to this mathematical model and do not refer directly to the classical sudoku puzzle . in particular , we will not investigate the relation of this model to the sudoku puzzle in detail .
this had been described in @xcite already .
from now on we denote the elements of @xmath29 in the definition of the generalized sudoku problem as cells .
we divide the set of all cells into @xmath8 subsets of @xmath8 cells .
[ d : constraintsets ] let @xmath23 be a permutation on @xmath29 .
the @xmath8 sets @xmath48 for @xmath49 are called the constraint sets of @xmath23 and are denoted by @xmath50 .
each of the sets @xmath51 contains exactly @xmath8 elements .
if the generalized sudoku problem describes a classical sudoku puzzle ( as defined in ( * ? ? ?
* section 3 ) ) , then the constraint sets of @xmath39 , @xmath40 respectively @xmath41 describe the indices of the rows , columns respectively blocks of the sudoku square . [ l:31 ] let @xmath23 be a permutation on @xmath29 .
the constraint sets of @xmath23 form a partition of the cells @xmath29 , i.e. , a. @xmath52 for @xmath53 , and b. @xmath54 .
\(i ) this is clear , because @xmath23 is a one - to - one mapping .
+ ( ii ) obviously , @xmath55 for @xmath56 .
the inclusion @xmath57 follows , because @xmath23 is onto .
let @xmath23 be a permutation on @xmath29 and let @xmath58 .
we adopt a notation of ( * ? ? ?
* section 2 ) and define @xmath59 the proof of the next lemma follows immediately from ( * ? ? ?
* lemma 3.2 ) .
[ l:32 ] let @xmath23 be a permutation on @xmath29 , let @xmath60 and let @xmath61 .
the following statements are equivalent : a. @xmath62
. b. @xmath63 are distinct . from ( * ? ? ?
* lemma 3.3 ) , we derive this lemma .
[ l:33 ] let @xmath23 be a permutation on @xmath29 , let @xmath60 and let @xmath61 .
the following statements are equivalent : a. @xmath62 + and @xmath64 for @xmath65 . b. @xmath66 .
the results on the subvectors of @xmath67 can be extended to the full length .
[ l:34 ] let @xmath23 be a permutation on @xmath29 and let @xmath61 .
the following statements are equivalent : a. @xmath68
. b. @xmath63 are distinct for @xmath49 . using ( * ? ? ?
* lemma 3.4 ( i ) ) , @xmath69 the claim follows from the diagonal structure of @xmath19 and lemma [ l:32 ] . from lemma [ l:34 ] and (
* lemma 3.1 ) , we derive immediately the next lemma .
[ l:35 ] let @xmath23 be a permutation on @xmath29 and let @xmath61 .
the following statements are equivalent : a. @xmath68 and @xmath70 for @xmath71
. b. @xmath72 for @xmath49 . in the sequel
we use from time to time the notation @xmath73 respectively @xmath74 for constraint sets , i.e. , we do not number them , but use an abbreviated version .
in this section , we consider the unique solvability of the generalized sudoku problem introduced in section [ s : model ] and we focus on a description of unicity using the dependance on permutations of @xmath29 .
[ d : pi - consistent ] let @xmath23 be a permutation on @xmath29 . a permutation @xmath75 on @xmath29 is called @xmath23-consistent if @xmath76 for @xmath49 .
a permutation is called @xmath23-consistent if it preserves the constraint sets of @xmath23 . in a classical sudoku puzzle
this definition states that @xmath75 permutes only values within a row , a column or a block ( depending on @xmath23 ) .
please note , @xmath75 is @xmath23-consistent if and only if @xmath77 is @xmath23-consistent .
the concatenation @xmath78 of two permutations @xmath23 and @xmath77 is a permutation again .
[ l:41 ] let @xmath23 and @xmath75 be permutations on @xmath29 and let @xmath61 .
the following statements are equivalent : a. @xmath79
. b. @xmath80 are distinct for @xmath81 we apply ( * ? ? ?
* lemma 3.4 ( i ) ) and obtain @xmath82 using lemma [ l:34 ] with @xmath78 ( in the role of @xmath23 ) gives the desired result .
[ l:42 ] let @xmath23 be a permutation on @xmath29 , let @xmath75 be a @xmath23-consistent permutation on @xmath29 and let @xmath58 .
the following statements hold : a. @xmath70 for @xmath71 if and only if @xmath83 + for @xmath71 . b. @xmath84 if and only if @xmath79 .
( i ) " this is clear , because @xmath75 is a permutation and does not modify the values of the components of @xmath67 . +
( ii ) " using lemma [ l:34 ] , @xmath68 if and only if @xmath85 by assumption , @xmath75 respectively @xmath77 are @xmath23-consistent , i.e. , @xmath84 if and only if @xmath86 the claim follows from lemma [ l:41 ] .
[ d : pi - x - consistent ] let @xmath23 be a permutation on @xmath29 and let @xmath61 . a permutation @xmath75 on @xmath29
is called @xmath23-@xmath67-consistent if @xmath87 for @xmath49 . in a classical sudoku puzzle
this definition states , that @xmath75 may permute not only values within a row , a column or a block , but more changes are allowed .
it is required , that after the permutation each row , column and block contains the same values as before .
each @xmath23-consistent permutation is @xmath23-@xmath67-consistent for each @xmath88 and each permutation @xmath23 on @xmath29 .
[ l:43 ] let @xmath23 and @xmath75 be permutations on @xmath29 and let @xmath61 , such that @xmath70 for @xmath71 . consider the statements : a. @xmath84 .
b. @xmath79 .
c. @xmath75 is @xmath23-@xmath67-consistent .
each two of the statements imply the third
. by lemma [ l:35 ] , ( i ) holds if and only if @xmath89 using lemma [ l:41 ] and ( * ? ? ?
* lemma 3.1 ) , ( ii ) holds if and only if @xmath90 substituting @xmath91 with @xmath92 , @xmath93 for @xmath49 , i.e. , ( ii ) holds if and only if @xmath94 the claim follows from the definition of @xmath23-@xmath67-consistent .
we define the solution set of the generalized sudoku problem by @xmath95 [ l:44 ] let @xmath96 and let @xmath97 .
there exists a permutation @xmath75 on @xmath29 with the following properties : a. @xmath98 , b. @xmath75 is @xmath99-consistent , c. @xmath75 is @xmath100-@xmath67-consistent for @xmath101 and d. @xmath102 for @xmath103 .
let @xmath104 and @xmath105 .
we apply lemma [ l:35 ] to @xmath99 , @xmath67 and @xmath34 , which implies @xmath106 for @xmath49 .
there exist permutations @xmath107 on @xmath108 , such that @xmath109 for each @xmath110 and @xmath56 . by lemma [ l:31 ] , the constraint sets form a partition of @xmath29 and we can combine the @xmath107 to a permutation @xmath75 on @xmath29 with @xmath111 here @xmath112
denotes the restriction of a mapping to a set .
this @xmath75 is also a permutation , since all @xmath113 are permutations and the constraint sets form a partition of @xmath29 .
( i ) " let @xmath114 be any cell
. there exists a ( uniquely determined ) index @xmath115 , such that @xmath110 , i.e. , @xmath91 is contained in the @xmath27 constraint set .
this implies @xmath116 .
since @xmath91 had been chosen arbitrarily , we obtain @xmath117 for @xmath114 and @xmath118 ( ii ) " obviously , @xmath75 is @xmath99-consistent , since each @xmath107 is a permutation on @xmath108 for @xmath56 .
( iii ) " let @xmath119 . since @xmath120 , @xmath121 and , by lemma [ l:43 ] , @xmath75 is @xmath100-@xmath67-consistent .
( iv ) " let @xmath122 .
there exists an index @xmath115 , such that @xmath123 .
then @xmath124 and @xmath125 by lemma [ l:34 ] , @xmath126 for each @xmath127 if @xmath128 .
this implies @xmath129 and proves the claim .
applied to a classical sudoku puzzle with @xmath130 the permutation @xmath75 describes a mapping , which changes cells within a row .
this permutation does not preserve the columns , but after the permutation we have distinct values in each column .
the cells with a given are not touched by the permutation @xmath75 .
now we prove two descriptions of the solution set of the generalized sudoku problem .
these descriptions are based on permutations .
[ t:41 ] let @xmath67 be a solution of the generalized sudoku problem , i.e. , @xmath131
. then @xmath132 @xmath133 " let @xmath120 . by lemma [ l:44 ] , there exist permutations @xmath134 , @xmath135 and @xmath136 on @xmath29 , such that @xmath137 for @xmath138 , @xmath139 is @xmath99-consistent for @xmath138 and @xmath140 for @xmath103 and @xmath141 . +
@xmath142 " let @xmath143 , let @xmath134 , @xmath135 and @xmath136 be permutations on @xmath29 , such that @xmath137 for @xmath138 , @xmath139 is @xmath99-consistent for @xmath138 and @xmath140 for @xmath103 and @xmath141 .
let @xmath97 and @xmath144 , then @xmath145 since @xmath139 is @xmath99-consistent and using lemma [ l:35 ] .
by lemma [ l:35 ] applied to @xmath34 , @xmath146 and @xmath147 for @xmath148 .
moreover @xmath149 for @xmath103 , i.e. , @xmath120 .
for each solution @xmath120 the permutations in theorem [ t:41 ] are uniquely determined as we see from the next lemma .
let @xmath23 be a permutation on @xmath29 , let @xmath88 such that @xmath84 and let @xmath143 .
let @xmath134 and @xmath135 be @xmath23-consistent permutations on @xmath29 , such that @xmath150 and @xmath151
. then @xmath152 .
let @xmath114 .
by lemma [ l:31 ] , there exists @xmath60 and a constraint set @xmath51 , such that @xmath153 . by assumption , @xmath154 , hence @xmath155 . since @xmath134 and @xmath135 are @xmath23-consistent , @xmath156 and @xmath157 .
by lemma [ l:34 ] , @xmath158 . since @xmath91 had been chosen arbitrarily , @xmath159 and we obtain @xmath160 .
the description in theorem [ t:41 ] uses the condition @xmath23-consistent for three permutations .
the next theorem describes the solution set by one permutation with the weaker condition @xmath23-@xmath67-consistent .
again by lemma [ l:45 ] , the permutation @xmath75 in theorem [ t:42 ] is also uniquely determined for each solution @xmath34 .
[ t:42 ] let @xmath67 be a solution of the generalized sudoku problem , i.e. , @xmath131 .
then @xmath161 @xmath133 " let @xmath120 .
using lemma [ l:44 ] , there exists a permutation @xmath75 with the desired properties . +
@xmath142 " let @xmath143 , @xmath75 be a permutation , such that @xmath162 , @xmath75 is @xmath39-consistent , @xmath40-@xmath67-consistent , @xmath41-@xmath67-consistent and @xmath102 for @xmath103 .
the components of @xmath34 are permutations of the components of @xmath67 , i.e. , @xmath147 for @xmath148 . by lemma [ l:43 ]
, @xmath163 moreover @xmath164 for @xmath103 and this shows @xmath120 . from this theorem
we can derive a characterization of uniquely solvable generalized sudoku problems .
[ t:43 ] let @xmath67 be a solution of the generalized sudoku problem , i.e. , @xmath131 .
then @xmath67 is the unique solution of the generalized sudoku problem if and only if there does not exist a permutation @xmath75 on @xmath29 , such that @xmath165 , @xmath75 is @xmath39-consistent , @xmath40-@xmath67-consistent , @xmath41-@xmath67-consistent and @xmath102 for @xmath103 .
in this section , we consider a unicity concept , which we call unicity cell .
this definition allows local " considerations of unicity , i.e. , we characterize , when a single cell contains a unique value .
[ d : unicity_cell ] a cell @xmath114 is called a unicity cell with unique value @xmath166 if all solutions @xmath167 of the generalized sudoku problem satisfy @xmath168 .
please note , this definition does not say anything on the solvability of the generalized sudoku problem . in an unsolvable generalized sudoku problem
each cell is a unicity cell for trivial reasons .
it is immediately clear from the definition , that a solvable generalized sudoku problem admits a unique solution if and only if each cell @xmath114 is a unicity cell .
we continue with the definition of a sudoku subproblem , which we call the reduced problem .
the condition @xmath169 for @xmath46 in the definition of the generalized sudoku problem can be written as @xmath170 with a suitable @xmath171 matrix @xmath172 ( describing the equality conditions ) and a right side @xmath173 .
roughly spoken , the matrix @xmath172 is a unit matrix , where some of the rows had been left .
we divide the set of all cells @xmath29 into the set of known cells @xmath174 ( determining @xmath172 and @xmath175 ) and the set of unknown cells @xmath176 ( see section [ s : model ] ) .
each known cell @xmath177 is a unicity cell with unique value @xmath178 for @xmath103 .
let @xmath179 be an integer and let @xmath180 be a set of cells with @xmath181 ( and possibly containing known and unknown cells ) .
we consider the projection @xmath182 defined by @xmath183 this projection induces a transformation @xmath184 defined by @xmath185 for @xmath186 .
the @xmath187-component of @xmath188 , @xmath189 for each @xmath190 and @xmath191 . for any @xmath192-matrix @xmath193 with @xmath194 rows , @xmath195 , and vectors @xmath196 for @xmath197 , we define @xmath198 we consider the projected matrices @xmath199 , @xmath200 , @xmath201 and @xmath202 and delete all zero rows
. additionally , we delete all rows from @xmath203 for @xmath141 with at most one nonzero component .
the resulting matrices are denoted by @xmath204 , @xmath205 , @xmath206 and @xmath207 .
we also delete the components of the point @xmath208 , which belong to deleted rows of @xmath202 and denote the new point by @xmath209 .
we define the reduced problem induced by @xmath210 in the following way : @xmath211 the reduced problem is built from the generalized sudoku problem by dropping cells .
simultaneously , all comparisons between dropped cells and between remaining cells and dropped cells are also dropped .
all givens in the dropped cells do not appear in the reduced problem .
the generalized sudoku problem and the reduced problem induced by @xmath210 are related .
[ l:51 ] ( i ) let @xmath67 be a solution of the generalized sudoku problem and let @xmath212 be a set of cells . then @xmath188 is a solution of the reduced problem induced by @xmath210 .
+ ( ii ) if @xmath213 , then the reduced problem induced by @xmath210 equals the generalized sudoku problem .
( i ) " this is clear , since all constraints of the reduced problem induced by @xmath210 are also constraints of the generalized sudoku problem . +
( ii ) " if @xmath213 , then @xmath214 and @xmath215 is the identical mapping .
both problems are identical .
obviously , the converse of lemma [ l:51 ] ( i ) is not true .
if @xmath188 is a solution of the reduced problem , there is no guarantee , that @xmath67 satisfies all constraints of the generalized sudoku problem .
there is an immediate consequence of lemma [ l:51 ] . if the generalized sudoku problem is solvable , then the reduced problem is also solvable .
[ d : unicity_cell_wrt ] let @xmath179 be an integer , let @xmath212 be a set of @xmath216 cells , let @xmath217 and let @xmath166 . the cell @xmath91 is called a unicity cell w.r.t .
the base set @xmath210 with unique value @xmath218 if all solutions @xmath219 of the reduced problem induced by @xmath210 satisfy @xmath220 .
note , the conditions in definition [ d : unicity_cell_wrt ] do not require the solvability of the reduced problem . we also do not impose any minimality condition on the base set .
using lemma [ l:51 ] ( ii ) , a cell @xmath91 is a unicity cell with unique value @xmath218 ( see definition [ d : unicity_cell ] ) if and only if @xmath91 is a unicity cell w.r.t .
the base set @xmath29 with unique value @xmath218 .
mainly , we are interested in small " base sets @xmath210 , i.e. , with a small number of elements . a possible solution method could try to determine a base set in each step and to solve the reduced problem induced by it .
[ e:51 ] we consider the example depicted in fig .
[ fig1 ] and the union @xmath210 of the sets @xmath221 ( the first block ) and @xmath222 ( the cells populated with @xmath223 s ) .
the cell @xmath224 ( row 3 , column 3 ) is a unicity cell w.r.t . @xmath210 with unique value @xmath223 .
all solutions of the reduced problem induced by @xmath210 contain a @xmath223 in the @xmath225 component ( which corresponds to cell @xmath224 ) .
the relation between unicity cell and unicity cell w.r.t .
a base set will be examined in the following theorem .
[ t:51 ] let @xmath114 be a cell and let @xmath166 .
the following statements are equivalent : a. the cell @xmath91 is a unicity cell with unique value @xmath218 .
b. there exists a base set @xmath226 , such that @xmath227 + and @xmath91 is a unicity cell w.r.t . @xmath210 with unique value @xmath218 .
( i)@xmath228(ii ) " define @xmath213 .
the claim follows from lemma [ l:51 ] ( ii ) . + ( ii)@xmath228(i )
" let @xmath167 and @xmath229 be solutions of the generalized sudoku problem .
by assumption ( ii ) , there exists a base set @xmath230 , such that @xmath217 and @xmath91 is a unicity cell w.r.t . @xmath210 with unique value @xmath218 .
we consider the reduced problem induced by @xmath210 and by lemma [ l:51 ] ( i ) , @xmath188 and @xmath231 are solutions of the reduced problem .
consequently , @xmath232 .
the strategy of humans to solve a uniquely solvable sudoku is to find a cell @xmath91 and a pretty small " base set @xmath210 .
the preceeding theorem suggests , it is not necessary to find some cell @xmath91 , but we can choose any cell @xmath91 and we still are able to find a base set @xmath210 . but the price for selecting an arbitrary cell @xmath91 may be a large base set @xmath210 , which results in a difficult to solve reduced problem . we illustrate this with the next example . in the next two examples we change the indexation to a more readable form @xmath233 .
[ e:52 ] we consider a continuation of the diabolical " example of mepham @xcite depicted in fig .
the term continuation means , we added some values to the puzzle , which were easy to find .
the cell @xmath234 ( row 2 , column 1 ) is a unicity cell w.r.t .
the base set @xmath210 consisting of the circled cells with unique value @xmath223 .
this can be seen by the following argumentation .
let @xmath67 be a solution of the reduced problem induced by @xmath210 .
in cell @xmath234 there can be only a @xmath223 or a @xmath235 .
suppose there is a @xmath235 , then @xmath236 , @xmath237 and @xmath238 @xmath239 and @xmath240 and @xmath241 and this is a contradiction .
this example of mepham had been analyzed by crook @xcite , too .
crook described an algorithm , which combined the deduction of values with a trial - and - error - method .
provan @xcite defined a pigeon - hole rule " , which is equivalent to the preemptive sets of crook .
the algorithm of provan consists of the consecutive application of this rule to a solvable sudoku puzzle .
if it is possible to apply this rule in each step , then the algorithm leads to a solution and the solution is unique ( * ? ? ?
* theorem 2 ) .
[ e:53 ] we consider an example of provan ( * ? ? ?
* table 2 ) depicted in fig .
the cell @xmath242 ( row 1 , column 1 ) is a unicity cell w.r.t .
the base set @xmath210 consisting of the circled cells with unique value @xmath243 .
this can be seen by the following argumentation .
let @xmath67 be a solution of the reduced problem induced by @xmath210 .
in cell @xmath242 there can be only a @xmath242 or @xmath243 .
suppose there is a @xmath242 , then @xmath244 and @xmath245 @xmath246 .
but this is a contradiction to the circled values in column 2 and row 8 .
in this section we introduce generalized rectangles , which are sets of cells with a rectangle - type shape and show the relation to unicity . [
d : rectangle ] let @xmath247 and @xmath248 . a set @xmath212 consisting of @xmath249 cells
is called a @xmath216-@xmath250-rectangle if there exist distinct indices @xmath251 , @xmath252 for @xmath253 and @xmath141 , such that @xmath254 for @xmath253 and @xmath141 .
each single cell is a @xmath242-@xmath242-rectangle .
the set of all cells @xmath255 is an @xmath8-@xmath8-rectangle . in a classical sudoku puzzle
a set @xmath210 is a @xmath216-@xmath250-rectangle if and only if there exist @xmath250 rows , @xmath250 columns and @xmath250 blocks and each of them contains @xmath216 elements of @xmath210 .
[ e:61 ] in fig .
[ fig4 ] the unpopulated cells form a @xmath256-@xmath257-rectangle .
the constraint sets are rows 2 , 4 , 5 , columns 4 , 5 , 8 and blocks 2 , 5 , 6 .
the possible values in this rectangle are @xmath258 and @xmath235 .
we can allocate these values in row 2 in two different ways and the values of the remaining cells are uniquely determined .
in particular this sudoku puzzle admits two solutions . if @xmath259 and @xmath210 is a @xmath216-@xmath250-rectangle , then the @xmath216 values @xmath260 , where @xmath261 are distinct .
consequently , the whole set @xmath262 contains at least @xmath216 values .
assume there exists a solution of the generalized sudoku problem which is minimal on a @xmath216-@xmath250-rectangle with @xmath264 .
then it is possible to interchange two values in @xmath265 for each @xmath1 and we obtain a new solution , i.e. , the generalized sudoku problem is not uniquely solvable .
[ t:61 ] let @xmath131 and let @xmath210 be a @xmath216-@xmath250-rectangle for some @xmath266 and @xmath248 .
assume @xmath267 , i.e. , @xmath210 does not contain any given .
if @xmath67 is minimal on @xmath210 , the generalized sudoku problem admits more than one solution .
let @xmath104 be a solution of the generalized sudoku problem , which is minimal on the @xmath216-@xmath250-rectangle @xmath210 .
there exist distinct points @xmath268 , such that @xmath269 for @xmath253 and @xmath141 .
since @xmath264 , there exist two distinct points @xmath270 .
we define @xmath271 and a new point @xmath272 with @xmath273 . by definition of @xmath274 @xmath275 for @xmath253 and @xmath141 . since @xmath276 for @xmath277 , @xmath278 for @xmath56 and @xmath141 and @xmath279 for @xmath103 . using lemma [ l:35 ]
, @xmath274 solves the generalized sudoku problem , i.e. , we have more than one solution .
[ e:62 ] we consider an example of herzberg and murty ( * ? ? ?
4 ) depicted in fig .
the point @xmath67 with @xmath280 is a solution .
the set @xmath281 is a @xmath256-@xmath256-rectangle , which does not contain a given and @xmath67 is minimal on @xmath210 . by theorem [ t:61 ] ,
@xmath67 is not the only solution .
another solution is @xmath34 with @xmath282 .
[ t:62 ] let @xmath210 be a @xmath216-@xmath250-rectangle for some @xmath266 and @xmath248 .
let @xmath67 be the unique solution of the generalized sudoku problem , such that @xmath67 is minimal on @xmath210 .
then @xmath210 contains a given , i.e. , @xmath283 .
the preceding theorem formalizes and generalizes an observation of herzberg and murty @xcite , who stated if in the solution to a sudoku puzzle , we have a configuration of a type indicated in figure 6 in the same vertical stack , then at least one of these entries must be included as a given in the initial puzzle , for otherwise , we would have two possible solutions to the initial puzzle simply by interchanging @xmath284 and @xmath285 in the configuration . " .
the term figure 6 " refers to a figure in their paper , which is recovered in fig .
[ fig5 ] , where the cells @xmath286 , @xmath287 , @xmath288 and @xmath289 contain the values @xmath284 respectively @xmath285 .
cheney , _ introduction to approximation theory _
, mcgraw - hill , new york , 1966 .
crook , a pencil - and - paper algorithm for solving sudoku puzzles , _ notices amer .
_ 56 ( 2009 ) , 460 - 468 . t. fischer , a necessary solution condition for sudoku , arxiv:1210.6343 [ math.co ] , 2012 .
t. fischer , duality for sudoku , arxiv:1301.0790 [ math.co ] , 2013 .
a.m. herzberg and m.r .
murty , sudoku squares and chromatic polynomials , _ notices amer . math .
_ 54 ( 2007 ) , 708 - 717 . m. mepham , solving sudoku , crosswords ltd . ,
frome , england , 2005 , retrieved oct 2 , 2014 , from http://www.sudoku.org.uk/pdf/solving_sudoku.pdf .
provan , sudoku : strategy versus structure , _ amer . math .
monthly _ 116 ( 2009 ) , 702 - 707 . | this paper deals with a generalized sudoku problem and investigates the unicity of a given solution .
we introduce constraint sets , which is a generalization of the rows , columns and blocks of a classical sudoku puzzle .
the unicity property is characterized by three different properties .
we describe unicity by permutations , by unicity cells and by rectangles .
these terms are defined in this paper and are illustrated with examples . throughout this paper
we are not concerned with the existence of a solution . |
the study of rectangular subdivisions of rectangles is motivated by vlsi floorplan design @xcite and cartographic visualization @xcite .
the rich combinatorial structure of rectangular floorplans has also attracted theoretical research @xcite .
combinatorial properties lead to efficient algorithms for the recognition and reconstruction of the _ rectangular graphs _ induced by the corners of the rectangles in a floorplan @xcite , the _ contact graphs _ of the rectangles @xcite , and the contact graphs of the horizontal and vertical _ line segments _ that separate the rectangles @xcite .
the number of combinatorially different floorplans with @xmath0 rectangles is known to be @xmath3 , the @xmath0th baxter number @xcite .
rectangular subdivisions in the presence of points have also been studied in the literature .
given a finite set @xmath4 of points in the interior of an axis - aligned rectangle @xmath5 , a _
rectangulation _ of @xmath6 is a subdivision of @xmath5 into rectangles by pairwise noncrossing axis - parallel line segments such that every point in @xmath4 lies in the relative interior of a segment ( see figure [ fig : example ] ) . finding a rectangulation of minimum total edge length has attracted attention @xcite due to its applications in vlsi design and stock cutting in the presence of material defects .
this problem is known to be np - hard @xcite , however , its complexity is unknown when the points in @xmath4 are in general position in the sense that they have distinct @xmath7- and @xmath8-coordinates , that is , the points are _ noncorectilinear_. it is not hard to see that in this case the minimum edge - length rectangulation must consist of exactly @xmath0 line segments @xcite . for the rest of this paper
, we consider only noncorectilinear point sets @xmath4 , and rectangulations determined by @xmath9 line segments , one containing each point in @xmath4 .
the space of all the rectangulations of a point set @xmath4 ( within a rectangle @xmath5 ) can be explored using the following two elementary operators introduced in @xcite ( refer to figure [ fig : operations ] ) . of a set of 6 points , @xmath10 , and @xmath11 . ] * ( flip ) * [ def : flip ] let @xmath12 be a rectangulation of @xmath4 and let @xmath13 be a point such that the segment @xmath14 that contains @xmath15 does not contain any endpoints of other segments . the operation _
flip@xmath16 _ changes the orientation of @xmath14 from vertical to horizontal or vice versa . * ( rotate ) * [ def : rotate ] let @xmath12 be a rectangulation of @xmath4 .
let @xmath17 be a segment that contains @xmath13 .
let @xmath18 be a segment such that @xmath19 lies in @xmath20 and @xmath21 does not contain any endpoints of other segments .
the operation rotate@xmath22 shortens @xmath23 to @xmath24 , and extends @xmath25 beyond @xmath19 until it reaches another segment or the boundary of @xmath5 . for a finite set of noncorectilinear points @xmath26
, we denote by @xmath27 the _ graph of rectangulations of @xmath4 _ , where the vertex set is @xmath28 is a rectangulation of @xmath29 and the edge set is @xmath30 a single flip or rotate operation on @xmath31 produces @xmath32 . since both operations are reversible , @xmath33 is an undirected graph .
it is not hard to show that @xmath33 is connected @xcite , and there is a sequence of @xmath34 flip and rotate operations between any two rectangulations in @xmath33 when @xmath4 is a set of @xmath0 points in @xmath5 .
it is natural to ask for the diameter of @xmath33 , which we call the _ flip diameter _ of @xmath4 for short . in this paper , we show that the flip diameter of @xmath4 is @xmath1 for every @xmath0-element noncorectilinear point set @xmath4 ( section [ sec : rect ] ) , and it is @xmath35 for some @xmath0-element noncorectilinear point sets ( section [ sec : construction ] ) .
however , there are @xmath0-elements noncorectilinear points sets for which the flip diameter is @xmath2 ( section [ sec : diagonal ] ) . that is , the flip diameter is always between @xmath1 and @xmath36 , depending on the point configuration , and both bounds are the best possible .
we extend the flip and rotate operations and the notion of flip diameter to convex subdivisions ( section [ sec : convex ] ) .
a _ convex subdivision _ of a set @xmath37 of points is a subdivision of the plane into convex faces by pairwise noncrossing line segments , halflines , and lines , each of which contains exactly one point of @xmath4 .
we show that the flip diameter for the convex subdivisions of @xmath0 points is always @xmath1 and sometimes @xmath2 .
determining the exact number of rectangulations on @xmath0 noncorectilinear points remains an elusive open problem in enumerative combinatorics @xcite .
recently , felsner @xcite proved that every combinatorial floorplan with @xmath38 rooms can be embedded into every set of @xmath0 noncorectilinear points , hence every set of @xmath0 noncorectilinear points has at least @xmath3 rectangulations .
the currently best known upper bound , @xmath39 by ackerman @xcite , uses the so - called `` cross - graph '' charging scheme @xcite , originally developed for counting the number of ( geometric ) triangulations on @xmath0 points in the plane .
this method is based on elementary `` flip '' operations that transform one triangulation into another .
lawson @xcite proved that every triangulation on @xmath0 points in the plane in general position ( i.e. , no three on a line ) can be transformed into the delaunay triangulation with @xmath34 flips , and this bound is the best possible by a construction due to hurtado et al .
however , for @xmath0 points in convex position , @xmath40 flips are sufficient , due to a bijection with binary trees with @xmath41 internal nodes @xcite .
hence the flip diameter of every triangulation on @xmath0 points in the plane is always between @xmath2 and @xmath42 depending on the point configuration .
eppstein et al .
@xcite and buchin et al .
@xcite define two elementary flip operations on floorplans , in terms of the directed dual graph , and solve optimization problems on floorplans by traversing the flip graph .
generalizing simultaneous edge flips in a triangulation , meijer and rappaport @xcite considered _ simultaneous edge flips _ in convex decompositions of a point set @xmath4 .
a _ convex decomposition _ of @xmath4 is a subdivision of the convex hull of @xmath4 into convex polygons whose joint vertex set is @xmath4 ; and two convex decompositions of @xmath4 are related by a simultaneous edge flip if their union contains no crossing edges .
in this section , we show that for every set @xmath4 of @xmath0 noncorectilinear points in a rectangle @xmath5 , the diameter of @xmath33 is @xmath1 .
[ thm : rect ] for every noncorectilinear set @xmath4 of @xmath0 points in the plane , the diameter of the graph @xmath33 is @xmath1 . given a rectangulation @xmath12 of @xmath4 , we construct a sequence of @xmath1 operations that transforms @xmath12 into a rectangulation with all segments vertical ( a _ canonical _ rectangulation ) .
our method relies on the concept of independent sets , defined in terms of the bar visibility graph .
let @xmath12 be a rectangulation @xmath12 of @xmath4 .
the _ bar visibility graph _
@xcite on the horizontal segments of @xmath12 is defined as a graph @xmath43 , where the vertices correspond to the horizontal segments in @xmath12 ; and two horizontal segments @xmath23 and @xmath25 are adjacent in @xmath43 if and only if there are points @xmath44 and @xmath45 such that @xmath46 is a vertical segment ( not necessarily in @xmath12 ) that does not intersect any other horizontal segment in @xmath12 .
it is clear that the bar visibility graph is planar .
observe that we can always change the orientation of any line segment @xmath14 with @xmath47 operations : successively shorten @xmath14 using rotate operations until @xmath14 contains no other segment endpoints , and then flip @xmath14 .
this simple procedure is formulated in the following subroutine .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ` shorten&flip`@xmath48 .
let @xmath14 be a segment in a rectangulation @xmath12 .
assume @xmath49 and @xmath13 is in the relative interior of @xmath14 . while @xmath14 contains the endpoint of some other segment ,
let @xmath50 and @xmath51 be the endpoints of some other segments closest to @xmath52 and @xmath53 , respectively ( possibly @xmath54 ) . if @xmath55 , then apply rotate@xmath56 to shorten @xmath49 to @xmath57 .
else , apply rotate@xmath58 to shorten @xmath14 to @xmath59 .
when @xmath14 does not contain the endpoint of any other segment , apply flip@xmath16 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the proof of theorem [ thm : rect ] follows from a repeated invocation of the following lemma . [
lem : lin ] let @xmath12 be a rectangulation of a set of @xmath0 pairwise noncorectilinear points in a rectangle @xmath5 .
there is a sequence of @xmath47 flip and rotate operations that turns at least one quarter of the horizontal segments into vertical segments , and keeps vertical segments vertical . by the four color theorem @xcite
, @xmath43 has an independent set @xmath60 that contains at least one quarter of the horizontal segments in @xmath12 .
the total number of endpoints of vertical segments that lie on some horizontal segment in @xmath60 is @xmath47 .
successively call the subroutine ` shorten&flip`@xmath48 for all horizontal segment @xmath61 .
the horizontal segments in @xmath60 are shortened and flipped into vertical orientation .
all operations maintain the invariants that ( 1 ) the segments in @xmath60 are pairwise disjoint , and ( 2 ) the remaining horizontal segments in @xmath60 form an independent set in the bar visibility graph ( of all horizontal segments in the current rectangulation ) .
it follows that each operation either decreases the number of horizontal segments in @xmath60 ( flip ) , or decreases the number of segment endpoints that lie in the relative interior of a segment in @xmath60 ( rotate ) .
after @xmath47 operations , all segments in @xmath60 become vertical . since only the segments in @xmath60 are flipped ( once each ) , all vertical segments in @xmath12 remain vertical , as required .
let @xmath4 be a set of @xmath0 pairwise noncorectilinear points in a rectangle @xmath5 .
denote by @xmath62 the rectangulation that consists of @xmath0 vertical line segments , one passing through each point in @xmath4 .
we show that every rectangulation @xmath31 of @xmath4 can be transformed into @xmath62 by a sequence of @xmath1 flip and rotate operations . by lemma [ lem : lin ] , a sequence of @xmath47 operations can decrease the number of horizontal segments by a factor of at least @xmath63 .
after at most @xmath64 invocations of lemma [ lem : lin ] , the number of horizontal segments drops below 1 , that is , all segments become vertical and we obtain @xmath62 .
[ [ remark . ] ] remark .
+ + + + + + + the proof of theorem [ thm : rect ] is constructive , and provides an algorithm for transforming a rectangulation on @xmath0 noncorectilinear points to the rectangulation with @xmath0 vertical segments .
if the rectangulations are maintained in a doubly connected edge list ( dcel ) data structure , then a flip or rotate operation can be implemented in @xmath65 time , and the bar - visibility graph can be computed in @xmath47 time . the bottleneck of the overall running time is the 4-coloring of the bar - visibility graph . the current best algorithm for 4-coloring an @xmath66-vertex planar graph runs in @xmath67 time @xcite , and a repeated call to this algorithm to exponentially decaying bar - visibility graphs takes @xmath34 time .
if we use an @xmath68-time 5-coloring algorithm @xcite , up to @xmath69 times , the overall running time improves to @xmath47 .
we show that the diameter of the graph @xmath33 is @xmath35 when @xmath4 is an @xmath0-element _ bit - reversal point set _ ( alternatively , halton - hammersley point set ) ( * ? ? ?
* section 2.2 ) . for every integer @xmath70
, we define a point set @xmath71 of size @xmath72 with integer coordinates lying in the square @xmath73 ^ 2 $ ] . for an integer @xmath66 , @xmath74 , with binary representation @xmath75 ,
the bit - reversal gives @xmath76 .
the bit - reversal point set of size @xmath72 is @xmath77 . by construction , no two points in @xmath71 are corectilinear .
we establish a lower bound of @xmath78 for the diameter of @xmath79 using a charging scheme .
we define @xmath80 empty rectangles ( called _ boxes _ ) spanned by @xmath71 , and charge one unit for `` saturating '' a box with vertical segments ( as defined below ) .
we show that when a rectangulation with all horizontal segments is transformed into one with all segments vertical , each box becomes saturated .
we also show that each rotate ( resp .
, flip ) operation contributes a total of at most 2 ( resp . , 4 ) units to the saturation of various boxes in our set .
it follows that at least @xmath81 operations are required to saturate all @xmath80 boxes .
consider the point set @xmath71 for some @xmath82 .
we say that a rectangle @xmath83 ^ 2 $ ] is _ spanned _ by @xmath71 if two opposite corners of @xmath84 are in @xmath71 ; and @xmath84 is _ empty _ if its interior is disjoint from @xmath71 .
let @xmath85 be the set of closed rectangular boxes spanned by point pairs in @xmath71 whose corresponding binary representation @xmath86 differ in exactly one bit .
see figure [ fig : hh1 ] , @xmath87 , and @xmath88 .
the edges connect points whose binary representations differ in a single bit , showing that @xmath71 is a projection of a @xmath89-dimensional hypercube .
the gray rectangles are spanned by point pairs whose binary representations differ exactly in the last coordinate . ] for examples .
each point in @xmath71 is incident to @xmath89 boxes in @xmath85 , since there are @xmath89 bits .
every box in @xmath85 is spanned by two points of @xmath71 , thus @xmath90 .
each point is incident to @xmath89 boxes of sizes @xmath91 for @xmath92 , since changing the @xmath93th bit @xmath94 incurs an @xmath95 change in the @xmath7-coordinate and an @xmath96 change in the @xmath8-coordinate .
however , if we change several bits successively , then either the @xmath7-coordinate changes by more than @xmath95 or the @xmath8-coordinate changes by more than @xmath96 , and so every box in @xmath85 is empty . note also that the boxes of the same size are pairwise disjoint .
we now define the `` saturation '' of each box @xmath97 with respect to a rectangulation of @xmath71 .
let @xmath97 and let @xmath12 be a rectangulation of @xmath71 .
the vertical _ extent _ of @xmath84 is the orthogonal projection of @xmath84 into the @xmath8-axis .
consider the vertical segments of @xmath12 clipped in @xmath84 ( i.e. , the segments @xmath98 for all vertical segments @xmath14 in @xmath12 ) .
the _ saturation of @xmath84 _ with respect to @xmath12 is the percentage of the vertical extent of @xmath84 covered by projections of vertical segments of @xmath12 clipped in @xmath84 .
see figure [ fig : hh2 ] .
the saturation of box @xmath99 is @xmath100 , 0 , 1 , 0 , 1 , 1 , 1 , and 1 , respectively . ] for examples . by definition
, the saturation of @xmath84 is a real number in @xmath101 $ ] . for every @xmath97 , we have that the saturation of @xmath84 is 0 when @xmath12 is a rectangulation with all horizontal segments , and it is 1 when @xmath12 consists of all segments vertical . if we transform an all - horizontal rectangulation into an all - vertical one by a sequence of operations , the total saturation of all @xmath80 boxes in @xmath85 increases from @xmath102 to @xmath80 .
the key observation is that a single operation increases the total saturation of all boxes in @xmath85 by at most a constant .
it remains to bound the impact of a single operation on the saturation of a box in @xmath85 .
consider first an operation rotate@xmath22 that increases the saturation of some box @xmath97 .
a rotate operation shortens a segment @xmath23 and extends an orthogonal segment @xmath25 .
the saturation of a box @xmath84 can increase only if a vertical segment grows , so we may assume that @xmath23 is horizontal and @xmath25 is vertical .
denote by @xmath14 the newly inserted portion of @xmath25 .
note that @xmath14 lies in a single face of the rectangulation @xmath12 .
similarly , if an operation flip@xmath16 increases the saturation of a box in @xmath84 , then it replaces a horizontal segment by a vertical segment passing through @xmath15 .
the new vertical segment lies in _ two _ adjacent faces of @xmath12 , separated by the original horizontal segment through @xmath15 .
we represent the new vertical segment as the union of two collinear vertical segments @xmath103 that meet at point @xmath15 . in summary ,
an operation rotate@xmath22 inserts one vertical segment @xmath14 that lies in the interior of a face of @xmath12 , and an operation flip@xmath16 inserts two such segments .
we show now that if such a new vertical segment @xmath14 increases the saturation of some box in @xmath97 , then @xmath14 must lie in @xmath84 .
[ lem : ce ] suppose that an operation inserts a vertical segment @xmath14 that lies in a face @xmath104 of the rectangulation @xmath12 .
if the insertion of @xmath14 increases the saturation of a box @xmath97 , then @xmath105 .
suppose , to the contrary , that @xmath106 .
let @xmath107 denote the two opposite corners of points that span @xmath84 , such that @xmath15 is the upper left or upper right corner of @xmath84 , and @xmath108 is the opposite corner of @xmath84 .
since @xmath14 increases the saturation of @xmath84 , it must intersect @xmath84 .
hence @xmath109 .
since @xmath106 , at least one of the endpoints of @xmath14 lies in the exterior of @xmath84 .
assume , without loss of generality , that the upper endpoint of @xmath14 lies outside @xmath84 , and @xmath15 is the upper left corner of @xmath84 .
then , the top side of @xmath104 is strictly above the top side of @xmath84 . since point @xmath15 can not be in the interior of @xmath104 , the left side of the face @xmath104 intersects the top side of @xmath84 .
note that @xmath14 and the left side of @xmath104 have the same orthogonal projection on the @xmath8-axis .
therefore , the insertion of @xmath14 can not increase the saturation of @xmath84 , contradicting our assumption .
we conclude that both endpoints of @xmath14 lie in @xmath84 , and @xmath105 .
[ lem : constant ] a rotate ( resp .
, flip ) operation increases the total saturation of all boxes in @xmath85 by at most 2 ( resp . , 4 ) .
suppose that an operation rotate@xmath22 inserts a vertical segment @xmath14 , or an operation flip@xmath16 inserts two collinear vertical segments @xmath103 that meet at @xmath15 .
by lemma [ lem : ce ] , the insertion of @xmath14 increases the saturation of a box @xmath97 of height @xmath110 by @xmath111 if @xmath112 , and does not affect the saturation of boxes of height @xmath113 . recall that the boxes in @xmath85 have only @xmath89 different sizes , @xmath91 for @xmath114 , and the boxes of the same size are pairwise disjoint .
let @xmath115 be the largest index such that @xmath116 .
for @xmath117 , segment @xmath14 increases the saturation of at most one box of height @xmath118 , and the increase is at most @xmath119 .
so @xmath14 increases the total saturation of all boxes in @xmath85 by at most @xmath120 , as required .
[ thm : construction ] for every @xmath121 , there is an @xmath0-element point set @xmath122 ^ 2 $ ] such that the diameter of @xmath33 is @xmath35 .
first assume that @xmath72 for some @xmath123 .
we have defined a set @xmath71 of @xmath72 points and a set @xmath85 of @xmath124 boxes spanned by @xmath71 .
the total saturation of all boxes in @xmath85 is 0 in the rectangulation with horizontal segments , and @xmath125 in the one with all segments vertical . by lemma [ lem : constant ] ,
a single flip or rotate operation increases the total saturation by at most 4 .
therefore , at least @xmath126 operations are required to transform the horizontal rectangulation to the vertical one , and the diameter of @xmath79 is at least @xmath126 .
if @xmath0 is not a power of two , then put @xmath127 and let @xmath122 ^ 2 $ ] be the union of @xmath71 and @xmath128 arbitrary ( noncorectilinear ) points in @xmath129 ^ 2 $ ] .
all axis - parallel segments containing the points in @xmath130 are in the exterior of @xmath131 ^ 2 $ ] .
therefore @xmath132 operations are required when all segments containing the points in @xmath133 change from horizontal to vertical .
we say that a point set @xmath4 is _ diagonal _ if all points in @xmath4 lie on the graph of a strictly increasing function ( e.g. , @xmath134 ) . in this section
we show that the flip diameter is @xmath47 for any @xmath0-element diagonal set .
we denote by @xmath135 the segment that contains @xmath136 .
[ thm : diagonal ] for every @xmath121 , the diameter of @xmath33 is at most @xmath137 when @xmath4 is a diagonal set of @xmath0 points . without loss of generality
, we may assume that the diagonal set is @xmath138 , where @xmath139 .
given a rectangulation @xmath12 for @xmath4 , we construct a sequence of at most @xmath140 flip and rotate operations that transforms @xmath12 into a rectangulation with all segments vertical .
our algorithm consists of the following four phases .
* phase 1 : no three consecutive segments are parallel . *
we describe an algorithm that , given a rectangulation @xmath12 of @xmath4 , applies less than @xmath141 operations and returns a rectangulation in which no three consecutive points are contained in parallel segments . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ` nothreeparallel`@xmath142 .
while there are three consecutive points in @xmath4 contained in parallel segments , let @xmath143 be arbitrary points such that the segments @xmath144 , @xmath135 , and @xmath145 are parallel , and call ` shorten&flip`@xmath146 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ note that ` shorten&flip`@xmath146 changes the orientation of the middle segment @xmath135 only . hence , after one invocation of
` shorten&flip`@xmath146 , segments @xmath144 , @xmath135 , and @xmath145 have alternating orientations , and will never be in the middle of a consecutive parallel triple .
this implies that ` nothreeparallel`@xmath142 terminates after changing the orientation of at most @xmath147 segments , and no three consecutive points are in parallel segments in the resulting rectangulation .
it remains to bound the number of flip and rotation operations : we have already seen that there are at most @xmath147 flip operations , and we shall charge each rotation operation to the segment extended by that operation . note that after an invocation of ` shorten&flip`@xmath146 , the two endpoints of @xmath135 lie on @xmath144 and @xmath145 , respectively , and so segment @xmath135 is not extended after a flip in phase 1 .
thus , each segment can be extended in two possible directions ( left / right or up / down ) .
we claim that each segment is extended at most once in each direction .
suppose to the contrary that a segment @xmath148 is extended twice in the same direction ( refer to figure [ fig : no3 ] , , @xmath136 , and @xmath149 are parallel , we perform ` shorten&flip`@xmath150 .
right : if a rotation extends a vertical segment @xmath148 up from @xmath135 to @xmath151 , then @xmath145 can not be horizontal . ]
right ) . by symmetry
, we may assume that @xmath148 is vertical , and it is extended upward in two rotate operations involving some horizontal segments @xmath135 and @xmath151 ( each of which is sandwiched between two horizontal segments ) . in the first rotation , @xmath148
is extended from its intersection with @xmath135 to an intersection with @xmath151 .
hence , the order of the corresponding three points is @xmath152 .
this step creates a rectangle bounded by @xmath135 on the top , @xmath151 on the bottom , and @xmath148 on the left .
the right boundary of this rectangle can not be @xmath145 , since it is horizontal ; and it can not be any segment @xmath153 , @xmath154 , otherwise the rectangle would contain @xmath149 .
the contradiction confirms the claim .
we conclude that this phase involves at most @xmath155 rotations and at most @xmath147 flip operations .
* phase 2 : creating a staircase . *
we define a _ staircase _ in a rectangulation of @xmath4 as a monotone increasing path along the segments that does not skip two or more consecutive points ( refer to figure [ fig : staircase ] ) . in the second phase of our algorithm
, we transform a rectangulation @xmath12 with no three consecutive parallel segments into a rectangulation @xmath12 that contains a staircase from the lower left to the upper right corner of @xmath5 .
we construct a staircase @xmath156 incrementally , starting from the lower left corner . in each step
we successively extend @xmath156 with one horizontal and one vertical segment to a point of @xmath4 or to the upper right corner of @xmath5 .
let @xmath156 be a current staircase from the lower left corner to a point @xmath157 .
denote by @xmath53 and @xmath19 the next two points of the diagonal set @xmath4 . if there is a monotone path from @xmath52 to @xmath53 or from @xmath52 to @xmath19 , then append this path to @xmath156 to obtain a longer staircase .
suppose there is no such path .
then the segments @xmath158 and @xmath159 must be parallel , otherwise there would be a monotone path from @xmath52 to @xmath53
. the segment @xmath160 must be perpendicular to the segment @xmath158 since there are no three consecutive parallel segments .
we would like to perform a rotation at the intersection of segments @xmath159 and @xmath160 , and then we can extend @xmath156 from @xmath52 to @xmath19 ( along @xmath158 and @xmath160 ) .
however , other operations may be necessary before we can rotate at @xmath161 .
successively rotate all intersection points of @xmath159 with other segments that are to the right of @xmath53 if @xmath159 is horizontal , and that are above @xmath53 if @xmath159 is vertical ( rotating at @xmath162 eventually ) .
we claim that this phase requires at most @xmath0 rotations .
let @xmath52 , @xmath53 , and @xmath19 be defined as before , such that there is no path from @xmath52 to either @xmath53 or @xmath19 .
assume , by symmetry , that @xmath52 and @xmath53 are horizontal and @xmath19 is vertical .
let @xmath163 be segment endpoints on @xmath159 to the right of @xmath53 .
the bottom endpoints of the vertical segments through @xmath164 are now below @xmath53 , and so these endpoints will never be involved in another rotation in our incremental procedure .
similarly , we perform a rotation at the left endpoint of each horizontal segment at most once . therefore , the total number of rotations is bounded by the number of segments , which is @xmath0 .
* phase 3 : extending all vertical segments of the staircase to full length . *
the third phase of our algorithm returns a rectangulation in which every vertical segment of the staircase extends to the top and bottom sides of the bounding box @xmath5 .
the staircase subdivides @xmath5 into two regions , one above and one below ( see figure [ fig : more_vert ] ) .
we handle the two regions independently without modifying the staircase . to handle the region above
, we sweep from top to bottom and successively shorten each horizontal segment we encounter until further shortening is impossible or would modify the staircase .
if we encounter a horizontal segment @xmath14 that is not part of the staircase , we use subroutine ` shorten&flip`@xmath48 to shorten it as much as possible and then flip it to vertical position .
similarly , we sweep the region below the staircase bottom - up .
each flip or rotate operation in phase 3 increases the number of segment endpoints lying on the top or bottom side of @xmath5 .
therefore , the number of operations is bounded by the total number of segment endpoints , which is @xmath155 .
* phase 4 : final steps .
* the vertical segments of the staircase extend to the top and bottom of the bounding box and partition it into vertical strips ( figure [ fig : more_vert ] , right ) .
each strip contains a horizontal segment of the staircase , including a point in @xmath4 . because the staircase does not skip two or more consecutive points , there are at most three points in each strip .
the total number of points in all strips combined is less than @xmath0 .
if there is just one point in a strip , then we flip the horizontal segment that contains it .
if there are two points in a strip , then one lies on a horizontal and one on a vertical segment . in this case , perform a rotation so the vertical segment reaches maximum length , and then flip the horizontal segment . if there are three points in a strip , then the middle point lies on a horizontal segment and the other two lie on vertical segments above and below the staircase , respectively . in this case
, we perform two rotations so that both vertical segments reach full length , and then flip the horizontal segment . similarly to phase 3 , each operation in this phase increases the number of segment endpoints lying on the top or bottom side of @xmath5 .
therefore , @xmath155 is an upper bound on the number of operations in phases 3 and 4 combined .
given a set @xmath4 of @xmath0 points in the plane , a _
convex subdivision _ for @xmath4 is a subdivision of the plane into convex cells by @xmath0 pairwise noncrossing line segments ( possibly lines or half - lines ) , such that each segment contains exactly one point of @xmath4 and no three segments have a point in common .
the flip and rotate operations can be interpreted for convex subdivisions of a point set @xmath4 , as well ( see figure [ fig : operations2 ] ) . of 6 points , @xmath165 , and @xmath11 .
] the definition of the operation rotate@xmath22 is identical to the rectilinear version .
the operation flip@xmath16 requires more attention , since a segment may have infinitely many possible orientations .
[ def : flip+ ] let @xmath12 be a convex subdivision of @xmath4 , let @xmath13 be a point such that the segment @xmath14 containing @xmath15 does not contain any endpoints of other segments , and let @xmath166 be a unit vector .
the operation _ flip@xmath167 _ replaces @xmath14 by a segment of direction @xmath168 containing @xmath15 .
similarly to the graph of rectangulations @xmath33 , we define the _ graph of convex subdivisions of @xmath4 _ , @xmath169 , where the vertex set is @xmath28 is a convex subdivision of @xmath29 and the edge set is @xmath30 a single flip or rotate operation on @xmath31 produces @xmath32 .
our main result in this section is that even though @xmath170 is an infinite graph , its diameter is @xmath1 , where @xmath171 .
[ thm : subd ] for set @xmath4 of @xmath0 points , the graph @xmath170 is connected and its diameter is @xmath1 . we show that any convex subdivision can be transformed into a subdivision with all segments vertical through a sequence of @xmath1 operations .
subroutine ` shorten&flip`@xmath48 from section [ sec : rect ] can be adapted almost verbatim : for a unit vector @xmath166 , subroutine ` shorten&flip`@xmath172 shortens segment @xmath14 maximally by rotate operations , and then flips it to direction @xmath168 .
[ lem : dir ] let @xmath12 be a convex subdivision of a set of @xmath0 points in the plane with distinct @xmath7-coordinates .
there is a sequence of @xmath47 flip and rotate operations that turns at least a @xmath173 fraction of the nonvertical segments vertical , and keeps all vertical segments vertical . before proving lemma [
lem : dir ] , we need to introduce a few technical terms . consider a convex subdivision @xmath12 of a set of @xmath0 points with distinct @xmath7-coordinates .
we say that a segment @xmath23 _ hits _ another segment @xmath25 if an endpoint of @xmath23 lies in the relative interior of @xmath25 .
extension of @xmath23 beyond @xmath25 hits @xmath174 _ if @xmath23 hits @xmath25 and @xmath23 is contained in a segment @xmath175 , such that @xmath175 hits @xmath174 and @xmath175 crosses at most one segment ( namely , @xmath25 ) .
we define the _ extension visibility digraph _
@xmath176 on all segments in @xmath12 , where the vertices correspond to the segments in @xmath12 , and we have a directed edge @xmath177 if @xmath25 hits @xmath174 or there is a segment @xmath23 such that an extension of @xmath23 beyond @xmath25 hits @xmath174 .
the graph @xmath176 is not necessarily planar : it is not difficult to construct a convex subdivision @xmath12 of a set of @xmath178 points where @xmath176 is isomorphic to the complete graph @xmath179 ( figure [ fig : k5 ] ) . of a set of 15 points .
the five bold segments induce @xmath180 in the extension visibility graph @xmath176 . ]
note that the number of edges in @xmath176 is at most @xmath181 , since each segment hits at most two other segments , but some segments extend to infinity ; and the extension of each segment beyond each of its endpoints hits at most one other segment .
if @xmath4 contains @xmath0 points , the average degree in @xmath176 is less than @xmath182 . therefore , @xmath176 has an independent set of size at least @xmath183 ( obtained by successively choosing minimum - degree vertices @xcite ) .
let @xmath12 be a convex subdivision of a set of @xmath0 points with distinct @xmath7-coordinates .
let @xmath184 be an independent set in the extension visibility graph @xmath176 induced by all nonvertical segments . as noted above , @xmath184 contains at least @xmath185 of the nonvertical segments in @xmath12 .
let @xmath186 be an independent set in the bar visibility graph of the segments in @xmath184 ( two nonvertical segments in @xmath184 are mutually visible if there is a vertical segment between them that does not cross any segment of the subdivision ) .
since the bar visibility graph is planar , we have @xmath187 , and so @xmath188 contains at least a @xmath189 fraction of the nonvertical segments in @xmath12 .
the total number of segment endpoints that lie in the relative interior of segments in @xmath188 is @xmath47 .
an invocation of subroutine ` shorten&flip`@xmath172 for each segment @xmath190 changes their orientation to vertical .
the operations maintain the invariants that ( 1 ) the segments in @xmath188 are pairwise disjoint ; and ( 2 ) the nonvertical segments in @xmath188 form an independent set in both @xmath191 and the bar visibility graph of all segments in @xmath184 .
it follows that each operation either decreases the number of nonvertical segments in @xmath188 ( flip ) , or decreases the number of segment endpoints that lie in the relative interior of a nonvertical segment in @xmath188 ( rotate ) .
after performing @xmath47 operations , all segments in @xmath188 become vertical . since only the segments in @xmath188 change orientation ( each of them is flipped to become vertical ) , all vertical segments in @xmath12 remain vertical , as required .
let @xmath4 be a set of @xmath0 points in a bounding box .
we may assume , by rotating the point set if necessary , that the points in @xmath4 have distinct @xmath7-coordinates .
denote by @xmath62 the convex subdivision given by @xmath0 vertical line segments , one passing through each point in @xmath4 .
consider a convex subdivision @xmath31 of @xmath4 . by lemma [ lem : lin ]
, @xmath47 operations can decrease the number of nonhorizontal segments by a factor of at least @xmath192 .
after at most @xmath193 invocations of lemma [ lem : lin ] , the number of nonvertical segments drops below 1 , that is , all segments become vertical and we obtain @xmath62 , as claimed .
we show that the bound @xmath1 on the diameter of the flip graph @xmath170 from theorem [ thm : subd ] can be improved to @xmath47 for some simple point configurations .
[ thm : collinear ] for every @xmath121 , the diameter of @xmath170 is @xmath47 when @xmath4 is a set of @xmath0 collinear points .
we may assume that @xmath194 , @xmath195 .
let @xmath12 be a convex subdivision of @xmath4 .
no segment in @xmath12 is horizontal , since each segment contains a unique point .
we show that there is a sequence of @xmath47 operations that transforms @xmath12 into a convex subdivision with all segments vertical .
suppose that not all segments in @xmath12 are vertical , and let @xmath196 be the minimum absolute value of the slope of a nonvertical segment .
our algorithm proceeds in two phases .
* phase 1 : building a staircase . * in this phase , we transform @xmath12 into a convex subdivision in which every @xmath135 is a nonvertical ray with a left endpoint at infinity , and has slope @xmath197 when @xmath93 is odd and @xmath198 when @xmath93 is even .
we process the points @xmath199 successively in this order .
we maintain the following invariant ( refer to figure [ fig : collinear1 ] ) : points lying on the @xmath7-axis into rays of alternating slopes @xmath197 and @xmath198 . ] when we start processing @xmath136 , segments @xmath200 are already rays with the desired properties , and they do not contain the endpoints of any segment @xmath201 in their relative interiors .
suppose that all points @xmath202 , @xmath203 , have already been processed .
we use the subroutine ` shorten&flip`@xmath204 to shorten @xmath135 and flip it into the desired orientation ( of slope @xmath197 or @xmath198 depending on the parity of @xmath93 ) . for @xmath205 ,
this already guarantees that segment @xmath206 is a ray with a left endpoint at infinity . if @xmath207 , however , the left endpoint of @xmath135 lies on the ray @xmath144 , whose slope is different from that of @xmath135 . due to the invariant
, @xmath144 contains no segment endpoints to the right of the intersection point @xmath208 .
we can now apply a rotation at @xmath208 .
since the slopes of all segments are at least @xmath197 in absolute value , @xmath209 extends to infinity , and our invariant is established for segments @xmath210 .
we argue that phase 1 uses @xmath47 operations . for each point @xmath136 , we perform one flip followed by at most one rotation .
it is enough to bound the number of rotations performed in the invocations of subroutines ` shorten&flip ` .
we claim that these invocations extend each segment at most once in each direction .
suppose , to the contrary , that a segment @xmath148 is extended in the same direction twice ( when processing points @xmath136 and @xmath202 ) .
since the points are processed from left to right , we have @xmath211 . consider the step in which a rotation extends @xmath148 from its intersection with @xmath135 to an intersection with @xmath151 .
then , the @xmath7-axis and segment @xmath148 intersect @xmath135 and @xmath151 in different orders .
hence , @xmath135 and @xmath151 must cross each other , contradicting the fact that they are part of a convex subdivision .
this proves our claim .
it follows that phase 1 uses at most @xmath212 operations .
* phase 2 : orienting all lines to be vertical . *
refer to figure [ fig : collinear2 ] .
we are given a convex subdivision in which every @xmath135 is a ray with a left endpoint at infinity , and has slope @xmath197 when @xmath93 is odd and @xmath198 when @xmath93 is even .
note that only consecutive segments intersect .
( this is still not a `` canonical '' subdivision , as @xmath66 depends on the initial convex subdivision . ) in phase 2 , we transform all segments into vertical lines .
we make five passes over the odd or even points : ( 1 ) for every odd point @xmath136 from left to right , rotate ray @xmath135 into a line of slope @xmath197 . as a result ,
the lines through the odd points separate the even points . ( 2 ) we can now flip the segments through the even points independently into vertical segments .
( 3 ) for every even point @xmath136 from left to right , rotate the top endpoints to infinity .
( 4 ) for every even point from right to left , rotate the bottom endpoint to infinity .
we obtain vertical lines through the even points , that separate the odd points .
( 5 ) finally , flip the segments through the odd points independently into vertical lines .
we made three passes over even segments and two passes over odd segments , so the total number of operations in phase 2 is @xmath213 .
the two phases combined use at most @xmath214 operations .
we have shown that the diameter of the flip graph @xmath33 is between @xmath36 and @xmath1 for every @xmath0-element point set @xmath4 , and these bounds can not be improved .
the diameter is @xmath2 for diagonal point sets , and @xmath215 for the bit - reversal point set .
the flip graph @xmath33 of a noncorectilinear set @xmath4 is uniquely determined by the permutation of the @xmath7- and @xmath8-coordinates of the points @xcite ( e.g. , diagonal point sets correspond to the identity permutation ) .
it is an open problem to find the average diameter of @xmath33 over all @xmath0-element permutations .
it would already be of interest to find broader families of point sets with linear diameter : is the diameter of @xmath33 linear if @xmath4 is in convex position , unimodal , or corresponds to a separable permutation ( see @xcite ) ?
we have shown that the diameter of the flip graph is also @xmath1 for the convex subdivisions of @xmath0 points in the plane .
we do not know whether this bound is tight .
it is possible that the flip diameter is @xmath35 for the bit - reversal point set defined in section [ sec : construction ] , but our proof of theorem [ thm : construction ] heavily relies on axis - aligned boxes and does not seem to extend to convex subdivisions . of 6 points .
middle : operation @xmath216 performed as a continuous deformation of the segment containing @xmath15 .
right : operation @xmath217 performed as a continuous deformation of part of the segment containing @xmath19 . ]
flip@xmath167 rotates the segment containing @xmath15 continuously to position @xmath168 ; and rotate@xmath22 rotates continuously a portion of the segment containing @xmath19 into the extension of the segment that currently ends at @xmath19 .
the _ weight _ of an operation can be defined as the number of vertices swept during this continuous deformation . by theorem [ thm : subd ]
, a sequence of @xmath1 operations can transform any convex subdivision to any other convex subdivision on @xmath0 points . a single operation , however , may have @xmath36 weight .
we conjecture that the weighted diameter of the graph @xmath170 is also @xmath1 for every @xmath0-elements point set @xmath4 .
research on this paper was initiated at the workshop on _ counting and enumerating of plane graphs _ held in march 2013 at schloss dagstuhl .
we thank sonia chauhan , michael hoffmann , and andr schulz for insightful comments and stimulating conversations about the problems discussed in this paper . , _ counting problems for geometric structures : rectangulations , floorplans , and quasi - planar graphs _
thesis , dept . of computer science ,
israel inst . of technology , haifa , israel ,
july 2006 .
williams , a linear algorithm for colouring planar graphs with five colours , _ computer journal _ , 28 ( 1985 ) , 7881 .
wismath , characterizing bar line - of - sight graphs , _ proc .
1st acm symp . on computational geometry _ ,
baltimore , md , 1985 , 147152 . | we study the configuration space of rectangulations and convex subdivisions of @xmath0 points in the plane .
it is shown that a sequence of @xmath1 elementary _ flip _ and _ rotate _ operations can transform any rectangulation to any other rectangulation on the same set of @xmath0 points .
this bound is the best possible for some point sets , while @xmath2 operations are sufficient and necessary for others .
some of our bounds generalize to convex subdivisions of @xmath0 points in the plane . |
jakob bernoulli s _ ars conjectandi _ established the field of probability theory , and founded a long and remarkable mathematical development of deducing patterns to be observed in sequences of random events .
the theory of statistical inference works in the opposite direction , attempting to solve the inverse problem of deducing plausible models from a given set of observations .
laplace pioneered the study of this inverse problem , and indeed he referred to his method as that of inverse probability .
the likelihood function , introduced by @xcite , puts this inversion front and centre , by writing the probability model as a function of unknown parameters in the model .
this simple , almost trivial , change in point of view has profoundly influenced the development of statistical theory and methods . in the early days ,
computing data summaries based on the likelihood function could be computationally difficult , and various _ ad hoc _ simplifications were proposed and studied . by the late 1970s , however , the widespread availability of computing enabled a parallel development of widespread implementation of likelihood - based inference .
the development of simulation and approximation methods that followed meant that both bayesian and non - bayesian inferences based on the likelihood function could be readily obtained . as a result
, construction of the likelihood function , and various summaries derived from it , is now a nearly ubiquitous starting point for a great many application areas .
this has a unifying effect on the field of applied statistics , by providing a widely accepted standard as a starting point for inference . with the explosion of data collection in recent decades
, realistic probability models have continued to grow in complexity , and the calculation of the likelihood function can again be computationally very difficult .
several lines of research in active development concern methods to compute approximations to the likelihood function , or inference functions with some of the properties of likelihood functions , in these very complex settings . in the following section
, i will summarize the standard methods for inference based on the likelihood function , to establish notation , and then in section [ sec3 ] describe some aspects of more accurate inference , also based on the likelihood function . in section [ sec4
] , i describe some extensions of the likelihood function that have been proposed for models with complex dependence structure , with particular emphasis on composite likelihood .
suppose we have a probability model for an observable random vector @xmath0 of the form @xmath1 , where @xmath2 is a vector of unknown parameters in the model , and @xmath3 is a density function with respect to a dominating measure , usually lebesgue measure or counting measure , depending on whether our observations are discrete or continuous .
typical models used in applications assume that @xmath2 could potentially be any value in a set @xmath4 ; sometimes @xmath4 is infinite - dimensional , but more usually @xmath5 .
the inverse problem mentioned in section [ sec1 ] is to construct inference about the value or values of @xmath6 that could plausibly have generated an observed value @xmath7 .
this is a considerable abstraction from realistic applied settings ; in most scientific work such a problem will not be isolated from a series of investigations , but we can address at least some of the main issues in this setting . the likelihood function is simply @xmath8 i.e. , there is an equivalence class of likelihood functions @xmath9 , and only relative ratios @xmath10 are uniquely determined . from a mathematical point of view , ( [ likelihood ] ) is a trivial re - expression of the model @xmath1 ; the re - ordering of the arguments is simply to emphasize in the notation that we are more interested in the @xmath11-section for fixed @xmath12 than in the @xmath12-section for fixed @xmath2 .
used directly with a given observation @xmath13 , @xmath14 provides a ranking of relative plausibility of various values of @xmath2 , in light of the observed data
. a form of direct inference can be obtained by plotting the likelihood function , if the parameter space is one- or two - dimensional , and several writers , including fisher , have suggested declaring values of @xmath2 in ranges determined by likelihood ratios as plausible , or implausible ; for example , @xcite suggested that values of @xmath2 for which @xmath15 , be declared ` implausible ' , where @xmath16 is the maximum likelihood estimate of @xmath2 , i.e. , the value for which the likelihood function is maximized , over @xmath2 , for a given @xmath12 .
in general study of statistical theory and methods we are usually interested in properties of our statistical methods , in repeated sampling from the model @xmath17 , where @xmath18 is the notional ` true ' value of @xmath2 that generated the data .
this requires considering the distribution of @xmath19 , or relative ratios such as @xmath20 . to this end
, some standard summary functions of @xmath19 are defined .
writing @xmath21 we define the _ score function _
@xmath22 , and the observed and expected fisher information functions : @xmath23 if the components of @xmath24 are independent , then @xmath25 is a sum of independent random variables , as is @xmath26 , and under some conditions on the model the central limit theorem for @xmath27 leads to the following asymptotic results , as @xmath28 : @xmath29 where we suppress the dependence of each derived quantity on @xmath24 ( and on @xmath30 ) for notational convenience .
these results hold under the model @xmath1 ; a more precise statement would use the true value @xmath18 in @xmath31 , @xmath32 , and @xmath33 above , and the model @xmath17 .
however , the quantities @xmath34 , @xmath35 and @xmath36 , considered as functions of both @xmath2 and @xmath24 , are approximate _ pivotal quantities _ , i.e. , they have a known distribution , at least approximately . for @xmath37 we could plot , for example , @xmath38 as a function of @xmath2 , where @xmath39 is the standard normal distribution function , and obtain approximate @xmath40-values for testing any value of @xmath37 for fixed @xmath12 .
the approach to inference based on these pivotal quantities avoids the somewhat artificial distinction between point estimation and hypothesis testing .
when @xmath37 , an approximately standard normal pivotal quantity can be obtained from ( [ lrt ] ) as @xmath41^{1/2 } { \mathop{\longrightarrow}^{\mathcal{l}}}n(0 , 1).\ ] ] the likelihood function is also the starting point for bayesian inference ; if we model the unknown parameter as a random quantity with a postulated prior probability density function @xmath42 , then inference given an observed value @xmath43 is based on the posterior distribution , with density @xmath44 bayesian inference is conceptually straightforward , given a prior density , and computational methods for estimating the integral in the denominator of ( [ bayes ] ) , and associated integrals for marginal densities of components , or low - dimensional functions of @xmath2 , have enabled the application of bayesian inference in models of considerable complexity .
two very useful methods include laplace approximation of the relevant integrals , and markov chain monte carlo simulation from the posterior .
difficulties with bayesian inference include the specification of a prior density , and the meaning of probabilities for parameters of a mathematical model .
one way to assess the influence of the prior is to evaluate the properties of the resulting inference under the sampling model , and under regularity conditions similar to those needed to obtain ( [ score ] ) , ( [ mle ] ) and ( [ lrt ] ) , a normal approximation to the posterior density can be derived : @xmath45 implying that inferences based on the posterior are asymptotically equivalent to those based on @xmath46 .
this simple result underlines the fact that bayesian inference will in large samples give approximately correct inference under the model , and also that to distinguish between bayesian and non - bayesian approaches we need to consider the next order of approximation .
if @xmath47 , then ( [ score])([lrt ] ) can be used to construct confidence regions , or to test simple hypotheses of the form @xmath48 , but in many settings @xmath2 can usefully be separated into a parameter of interest @xmath49 , and a nuisance parameter @xmath50 , and analogous versions of the above limiting results in this context are @xmath51 where @xmath52 is the profile log - likelihood function , @xmath53 is the constrained maximum likelihood estimate of the nuisance parameter @xmath50 when @xmath49 is fixed , @xmath54 is the dimension of @xmath49 , and @xmath55 is the fisher information function based on the profile log - likelihood function .
the third result ( [ lrt2 ] ) can be used for model assessment among nested models ; for example , the exponential distribution is nested within both the gamma and weibull models , and a test based on @xmath56 of , say , a gamma model with unconstrained shape parameter , and one with the shape parameter set equal to 1 , is a test of fit of the exponential model to the data ; the rate parameter is the nuisance parameter @xmath50 .
the use of the log - likelihood ratio to compare two non - nested models , for example a log - normal model to a gamma model , requires a different asymptotic theory ( @xcite , @xcite , ch .
a related approach to model selection is based on the akaike information criterion , @xmath57 where @xmath58 is the dimension of @xmath2 .
just as only differences in log - likelihoods are relevant , so are differences in @xmath59 : for a sequence of model fits the one with the smallest value of @xmath59 is preferred .
the @xmath59 criterion was developed in the context of prediction in time series , but can be motivated as an estimate of the kullback - leibler divergence between a fitted model and a notional ` true ' model .
the statistical properties of @xmath59 as a model selection criterion depend on the context ; for example for choosing among a sequence of regression models of the same form , model selection using @xmath59 is not consistent ( @xcite , @xcite , ch .
several related versions of model selection criterion have been suggested , including modifications to @xmath59 , and a version motivated by bayesian arguments , @xmath60 where @xmath30 is the sample size for the model with @xmath58 parameters .
the approximate inference suggested by the approximate pivotal quantities ( [ score2 ] ) , ( [ mle2 ] ) and ( [ lrt2 ] ) is obtained by treating the profile log - likelihood function as if it were a genuine log - likelihood function , i.e. as if the true value of @xmath50 were @xmath61 .
this can be misleading , because it does not account for the fact that the nuisance parameter has been estimated .
one familiar example is inference for the variance in a normal theory linear regression model ; the maximum likelihood estimate is @xmath62 which has expectation @xmath63 , where @xmath64 is the dimension of @xmath65 .
although this estimator is consistent as @xmath66 with @xmath64 fixed , it can be a poor estimate for finite samples , especially if @xmath64 is large relative to @xmath30 , and the divisor @xmath67 is used in practice .
one way to motivate this is to note that @xmath68 is unbiased for @xmath69 ; an argument that generalizes more readily is to note that the likelihood function @xmath70 can be expressed as @xmath71 where @xmath72 is proportional to the density of @xmath73 and @xmath74 is the marginal density of @xmath75 or equivalently @xmath76 .
the unbiased estimate of @xmath69 maximizes the second component @xmath74 , which is known as the restricted likelihood , and estimators based on it often called `` reml '' estimators .
higher order asymptotic theory for likelihood inference has proved to be very useful for generalizing these ideas , by refining the profile log - likelihood to take better account of the nuisance parameter , and has also provided more accurate distribution approximations to pivotal quantities .
perhaps most importantly , for statistical theory , higher order asymptotic theory helps to clarify the role of the likelihood function and the prior in the calibration of bayesian inference .
these three goals have turned out to be very intertwined . to illustrate some aspects of this , consider the marginal posterior density for @xmath49 , where @xmath77 : @xmath78 laplace approximation to the numerator and denominator integrals leads to @xmath79 where @xmath80 is the block of the observed fisher information function corresponding to the nuisance parameter @xmath50
, @xmath81 has been computed using the partitioned form to give the second expression in ( [ laplace2 ] ) , and in the third expression @xmath82 when renormalized to integrate to one , this laplace approximation has relative error @xmath83 in independent sampling from a model that satisfies various regularity conditions similar to those needed to show the asymptotic normality of the posterior @xcite .
these expressions show that an adjustment for estimation of the nuisance parameter is captured in @xmath84 , and this adjustment can be included in the profile log - likelihood function , as in the third expression in ( [ laplace2 ] ) , or tacked onto it , as in the second expression .
the effect of the prior is isolated from this nuisance parameter adjustment effect , so , for example , if @xmath85 , and the priors for @xmath49 and @xmath50 are independent , then the form of the prior for @xmath50 given @xmath49 does not affect the approximation .
the adjusted profile log - likelihood function @xmath86 is the simplest of a number of modified profile log - likelihood functions suggested in the literature for improved frequentist inference in the presence of nuisance parameters , and was suggested for general use in @xcite , after reparametrizing the model to make @xmath49 and @xmath50 orthogonal with respect to expected fisher information , i.e. , @xmath87 .
this reparameterization makes it at least more plausible that @xmath49 and @xmath50 could be modelled as _ a priori _ independent , and also ensures that @xmath88 , rather than the usual @xmath89 .
a number of related , but more precise , adjustments to the profile log - likelihood function have been developed from asymptotic expansions for frequentist inference , and take the form @xmath90 where @xmath91 ; see , for example , @xcite ( @xcite ) and @xcite .
the change from @xmath92 to @xmath93 is related to the orthogonality conditions ; in ( [ mpl ] ) orthogonality of parameters is not needed , as the expression is parameterization invariant .
inferential statements based on approximations from ( [ score2])([lrt2 ] ) , with @xmath86 or @xmath94 substituting for the profile log - likelihood function , are still valid and are more accurate in finite samples , as they adjust for errors due to estimation of @xmath50 .
they are still first - order approximations , although often quite good ones .
one motivation for these modified profile log - likelihood functions , and inference based on them , is that they approximate marginal or conditional likelihoods , when these exist .
for example , if the model is such that @xmath95 then inference for @xmath49 can be based on the marginal likelihood for @xmath49 based on @xmath96 , and the theory outlined above applies directly .
this factorization is fairly special ; more common is a factorization of the form @xmath97 : in that case to base our inference on the likelihood for @xmath49 from @xmath96 would require further checking that little information is lost in ignoring the second term
. arguments like these , applied to special classes of model families , were used to derive the modified profile log - likelihood inference outlined above .
a related development is the improvement of the distributional approximation to the approximate pivotal quantity ( [ root ] ) .
the laplace approximation ( [ laplace2 ] ) can be used to obtain the bayesian pivotal , for scalar @xmath49 , @xmath98 where @xmath99^{1/2 } , \\
\label{qb}q_b(\psi ) & = & -\ell_{\mathrm{p}}'(\psi ) j_{\mathrm{p}}^{-1/2}(\hat\psi ) \biggl\{\frac{|j_{\lambda\lambda}(\psi,\hat\lambda_\psi ) |}{|j_{\lambda\lambda}(\hat\psi,\hat\lambda)| } \biggr \}^{1/2}\frac{\pi(\hat\psi,\hat\lambda)}{\pi(\psi,\hat\lambda_\psi)}\end{aligned}\ ] ] and the approximation in ( [ rstarb ] ) is to the posterior distribution of @xmath100 , given @xmath12 , and is accurate to @xmath83 .
there is a frequentist version of this pivotal that has the same form : @xmath101 where @xmath102 is given by ( [ root2 ] ) , but the expression for @xmath103 requires additional notation , and indeed an additional likelihood component .
in the special case of no nuisance parameters @xmath104 in ( [ qf1 ] ) , we have assumed that there is a one - to - one transformation from @xmath12 to @xmath105 , and that we can write the log - likelihood function in terms of @xmath106 and then differentiate it with respect to @xmath107 , for fixed @xmath108 .
expression ( [ qf2 ] ) is equivalent , but expresses this sample space differentiation through a data - dependent reparameterization @xmath109 , where the derivative with respect to @xmath110 is a directional derivative to be determined .
the details are somewhat cumbersome , and even more so for the case of nuisance parameters , but the resulting @xmath111 approximate pivotal quantity is readily calculated in a wide range of models for independent observations @xmath112 .
detailed accounts are given in @xcite , @xcite , @xcite , @xcite and @xcite ( @xcite , ch .
8.6 ) ; the last emphasizes implementation in a number of practical settings , including generalized linear models , nonlinear regression with normal errors , linear regression with non - normal errors , and a number of more specialized models . from a theoretical point of view ,
an important distinction between @xmath113 and @xmath111 is that the latter requires differentiation of the log - likelihood function on the sample space , whereas the former depends only on the observed log - likelihood function , along with the prior .
the similarity of the two expressions suggests that it might be possible to develop prior densities for which the posterior probability bounds are guaranteed to be valid under the model , at least to a higher order of approximation than implied by ( [ bayes2 ] ) , and there is a long line of research on the development of these so - called `` matching priors '' ; see , for example , @xcite .
while the asymptotic results of the last section provide very accurate inferences , they are not as straightforward to apply as the first order results , especially in models with complex dependence . they do shed light on many aspects of theory , including the precise points of difference , asymptotically , between bayesian and nonbayesian inference . and
the techniques used to derive them , saddlepoint and laplace approximations in the main , have found application in complex models in certain settings , such as the integrated nested laplace approximation of @xcite . a glance at any number of papers motivated by specific applications , though , will confirm that likelihood summaries , and in particular computation of the maximum likelihood estimator , are often the inferential goal , even as the models become increasingly high - dimensional .
this is perhaps a natural consequence of the emphasis on developing probability models that could plausibly generate , or at least describe , the observed responses , as the likelihood function is directly obtained from the probability model .
but more than this , inference based on the likelihood function provides a standard set of tools , whose properties are generally well - known , and avoids the construction of _ ad hoc _ inferential techniques for each new application .
for example , @xcite write `` the likelihood framework is an efficient way to extract information from a neural spike train we believe that greater use of the likelihood based approaches and goodness - of - fit measures can help improve the quality of neuroscience data analysis '' . a number of inference functions based on the likelihood function , or meant to have some of the key properties of the likelihood function ,
have been developed in the context of particular applications or particular model families . in some cases the goal is to find ` reasonably reliable ' estimates of a parameter , along with an estimated standard error ; in other cases the goal is to use approximate pivotal quantities like those outlined in section [ sec2 ] in settings where the likelihood is difficult to compute .
the goal of obtaining reliable likelihood - based inference in the presence of nuisance parameters was addressed in section [ sec3 ] . in some settings ,
families of parametric models are too restrictive , and the aim is to obtain likelihood - type results for inference in semi - parametric and non - parametric settings . in many applications with longitudinal , clustered , or spatial data , the starting point is a generalized linear model with a linear predictor of the form @xmath114 , where @xmath115 and @xmath116 are @xmath117 and @xmath118 , respectively , matrices of predictors , and @xmath119 is a @xmath46-vector of random effects .
the marginal distribution of the responses requires integrating over the distribution of the random effects @xmath119 , and this is often computationally infeasible .
many approximations have been suggested : one approach is to approximate the integral by laplace s method @xcite , leading to what is commonly called penalized quasi - likelihood , although this is different from the penalized versions of composite likelihood discussed below .
the term quasi - likelihood in the context of generalized linear models refers to the specification of the model through the mean function and variance function only , without specifying a full joint density for the observations .
this was first suggested by @xcite , and extended to longitudinal data in @xcite and later work , leading to the methodology of generalized estimating equations , or gee .
@xcite compared penalized quasi - likelihood to pairwise likelihood , discussed in section [ sec4.3 ] , in simulations of multivariate probit models for binary data with random effects . in general
penalized quasi - likelihood led to estimates with larger bias and variance than pairwise likelihood .
a different approach to generalized linear mixed models has been developed by lee and nelder ; see , for example , @xcite and @xcite , under the name of @xmath120-likelihood .
this addresses some of the failings of the penalized quasi - likelihood method by modelling the mean parameters and dispersion parameters separately .
the @xmath120-likelihood for the dispersion parameters is motivated by reml - type arguments not unrelated to the higher order asymptotic theory outlined in the previous section .
there are also connections to work on prediction using likelihood methods @xcite .
likelihood approaches to prediction have proved to be somewhat elusive , at least in part because the ` parameter ' to be predicted is a random variable , although bayesian approaches are straightforward as no distinction is made between parameters and random variables .
composite likelihood is one approach to combining the advantages of likelihood with computational feasibility ; more precisely it is a collection of approaches .
the general principle is to simplify complex dependence relationships by computing marginal or conditional distributions of some subsets of the responses , and multiplying these together to form an inference function . as an _ ad hoc _ solution it has emerged in several versions and in several contexts in the statistical literature ;
an important example is the pseudo - likelihood for spatial processes proposed in @xcite ( @xcite ) . in studies of large networks ,
computational complexity can be reduced by ignoring links between distant nodes , effectively treating sub - networks as independent . in gaussian process models with high - dimensional covariance matrices , assuming
sparsity in the covariance matrix is effectively assuming subsets of variables are independent .
the term composite likelihood was proposed in @xcite , where the theoretical properties of composite likelihood estimation were studied in some generality .
we suppose a vector response of length @xmath46 is modelled by @xmath121 . given a set of events @xmath122 , the composite likelihood function
is defined as @xmath123 and the composite log - likelihood function is @xmath124 because each component in the sum is the log of a density function , the resulting score function @xmath125 has expected value @xmath126 , so has at least one of the properties of a genuine log - likelihood function . relatively simple and widely used examples of composite likelihoods include independence composite likelihood , @xmath127 pairwise composite likelihood @xmath128 and pairwise conditional composite likelihood @xmath129 where @xmath130 and @xmath131 are the marginal densities for a single component and a pair of components of the vector observation , and the density in ( [ clcond ] ) is the conditional density of one component , given the remainder .
many similar types of composite likelihood can be constructed , appropriate to time series , or spatial data , or repeated measures , and so on , and the definition is usually further extended by allowing each component event to have an associated weight @xmath132 . indeed one of the difficulties of studying the theory of composite likelihood is the generality of the definition . inference based on composite likelihood is constructed from analogues to the asymptotic results for genuine likelihood functions .
assuming we have a sample @xmath133 of independent observations of @xmath12 , the composite score function , @xmath134 is used as an estimating function to obtain the maximum composite likelihood estimator @xmath135 , and under regularity conditions on the full model , with @xmath136 and fixed @xmath137 , we have , for example , @xmath138 where @xmath139 is the @xmath140 godambe information matrix , and @xmath141 are the variability and sensitivity matrix associated with @xmath142 .
the analogue of ( [ lrt ] ) is @xmath143 where @xmath144 are the eigenvalues of @xmath145 .
neither of these results is quite as convenient as the full likelihood versions , and in particular contexts it may be difficult to estimate @xmath146 accurately , but there are a number of practical settings where these results are much more easily implemented than the full likelihood results , and the efficiency of the methods can be quite good . a number of applied contexts are surveyed in @xcite . as just one example , developed subsequently , @xcite investigate pairwise composite likelihood for max - stable processes , developed to model extreme values recorded at a number @xmath147 of spatially correlated sites .
although the form of the @xmath147-dimensional density is known , it is not computable for @xmath148 , although expressions are available for the joint density at each pair of sites .
composite likelihood seems to be particularly important for various types of spatial models , and many variations of it have been suggested for these settings . in some applications , particularly for time
series , but also for space - time data , a sample of independent observations is not available , and the relevant asymptotic theory is for @xmath149 , where @xmath46 is the dimension of the single response .
the asymptotic results outlined above will require some conditions on the decay of the dependence among components as the ` distance ' between them increases .
asymptotic theory for pairwise likelihood is investigated in @xcite for linear time series , and in @xcite for max - stable processes in space and time .
composite likelihood can also be used for model selection , with an expression analogous to @xmath59 , and for bayesian inference , after adjustment to accommodate result ( [ clrt ] ) .
_ statistica sinica _ * 21 * , # 1 is a special issue devoted to composite likelihood , and more recent research is summarized in the report on a workshop at the banff international research station @xcite . in some applications
, a flexible class of models can be constructed in which the nuisance ` parameter ' is an unknown function .
the most widely - known example is the proportional hazards model of @xcite for censored survival data ; but semi - parametric regression models are also widely used , where the particular covariates of interest are modelled with a low - dimensional regression parameter , and other features expected to influence the response are modelled as ` smooth ' functions .
@xcite developed inference based on a partial likelihood , which ignored the aspects of the likelihood bearing on the timing of failure events , and subsequent theory based on asymptotics for counting processes established the validity of this approach .
in fact , @xcite s partial likelihood can be viewed as an example of composite likelihood as described above , although the theory for general semi - parametric models seems more natural .
@xcite showed that partial likelihood can be viewed as a profile likelihood , maximized over the nuisance function , and discussed a class of semi - parametric models for which the profile likelihood continues to have the same asymptotic properties as the usual parametric profile likelihood ; the contributions to the discussion of their results provide further insight and references to the extensive literature on semi- and non - parametric likelihoods .
there is , however , no guarantee that asymptotic theory will lead to accurate approximation for finite samples ; it would presumably have at least the same drawbacks as profile likelihood in the parametric setting .
improvements via modifications to the profile likelihood , as described above in the parametric case , do not seem to be available in these more general settings .
some semi - parametric models are in effect converted to high - dimensional parametric models through the use of linear combinations of basis functions ; thus the linear predictor associated with a component @xmath150 might be @xmath151 , or @xmath152 .
the log - likelihood function for models such as these is often regularized , so that @xmath153 is replaced by @xmath154 , where @xmath155 is a penalty function such as @xmath156 or @xmath157 , and @xmath50 a tuning parameter .
many of these extensions , and the asymptotic theory associated with them , are discussed in @xcite ( @xcite , ch .
penalized likelihood using squared error is reviewed in @xcite ; the @xmath72 penalty has been suggested as a means of combining likelihood inference with variable selection ; see , for example , @xcite . penalized composite likelihoods
have been proposed for applications in spatial analysis ( @xcite , @xcite ; @xcite , @xcite ; @xcite , @xcite ) , gaussian graphical models @xcite , and clustered longitudinal data @xcite .
the difference between semi - parametric likelihoods and nonparametric likelihoods is somewhat blurred ; both have an effectively infinite - dimensional parameter space , and as discussed in @xcite and the discussion , conditions on the model to ensure that likelihood - type asymptotics still hold can be quite technical .
empirical likelihood is a rather different approach to non - parametric models first proposed by @xcite ; a recent discussion is @xcite .
empirical likelihood assumes the existence of a finite - dimensional parameter of interest , defined as a functional of the distribution function for the data , and constructs a profile likelihood by maximizing the joint probability of the data , under the constraint that this parameter is fixed .
this construction is particularly natural in survey sampling , where the parameter is often a property of the population ( @xcite , @xcite ; @xcite , @xcite ) .
distribution theory for empirical likelihood more closely follows that for usual parametric likelihoods .
simulation of the posterior density by markov chain monte carlo methods is widely used for bayesian inference , and there is an enormous literature on various methods and their properties . some of these methods can be adapted for use when the likelihood function itself can not be computed , but it is possible to simulate observations from the stochastic model ; many examples arise in statistical genetics .
simulation methods for maximum likelihood estimation in genetics was proposed in @xcite ; more recently sequential monte carlo methods ( see , for example , @xcite ) and abc ( approximate bayesian computation ) methods ( @xcite , @xcite ; @xcite , @xcite ) are being investigated as computational tools .
a reviewer of an earlier draft suggested that a great many applications , especially involving very large and/or complex datasets , take more algorithmic approaches , often using techniques designed to develop sparse solutions , such as wavelet or thresholding techniques , and that likelihood methods may not be relevant for these application areas .
certainly a likelihood - based approach depends on a statistical model for the data , and for many applications under the general rubric of machine learning these may not be as important as developing fast and reliable approaches to prediction ; recommender systems are one such example .
there are however many applications of ` big data ' methods where statistical models do provide some structure , and in these settings , as in the more classical application areas , likelihood methods provide a unifying basis for inference .
this research was partially supported by the natural sciences and engineering research council .
thanks are due to two reviewers for helpful comments on an earlier version . | i review the classical theory of likelihood based inference and consider how it is being extended and developed for use in complex models and sampling schemes . |
recently shukla and eliasson published in physical review letters @xcite a calculation where they applied linearized quantum hydrodynamics in a simple way to obtain the effective potential between classical protons in a dense hydrogen plasma containing degenerate electrons .
their result indicated a novel attractive force between two protons . if well - founded , this is not only an interesting results , but the relatively simple calculation methodology is also potentially one of considerable importance as a theorist s tool .
it is , therefore , important to discuss the implications of these results for real physical systems , such as dense quantum plasmas , and to rigorously verify the validity and limitations of the method against some more fundamental calculation .
however , presenting such a result can not be adequately done within the space limits of a comment in physical review letters , so it is being presented here as a full paper .
shukla and eliasson ( se hereafter ) claim @xcite to have calculated their potential as that between classical protons in a dense hydrogen plasma containing degenerate electrons .
the hydrogen plasma is considered in equilibrium ( essentially in the ground state , @xmath0 ) , and the authors stress that the observed `` novel attractive force '' is different from friedel oscillations . furthermore , from the existence of a `` hard core negative potential '' , eq .
( [ eq : phi_general ] ) , they predict the formation of bound states , a proton lattice , as well as `` critical points and phase transitions at nanoscales '' .
if correct , these results could have far - reaching consequences for hydrogen and dense quantum plasmas , in general .
it is , therefore , of interest to consider the results of shukla and eliasson more in detail , which is the goal of this paper .
we first summarize their approach
linearized quantum hydrodynamics ( lqhd ) .
wen then re - evaluate their final formulas transforming to common length and energy scales .
we then compare their screened proton potential to results from _ ab initio _ density functional ( dft ) simulations and observe qualitative deviations . from this
, we have to conclude that the lqhd results of ref .
@xcite are not applicable to dense quantum plasmas ad do not provide evidence for attractive interactions between protons .
we conclude by analyzing the origin of the deviations of the results of ref .
@xcite from dft which are traced to violation of the applicability range of linearized quantum hydrodynamics
. dense correlated quantum plasmas are presently of high interest in many fields including condensed matter , astrophysics and laser plasmas , e.g. @xcite . despite remarkable theoretical and experimental progress over recent decades , even for the simplest plasma system hydrogen
still interesting questions remain unsolved , including details of the phase diagram and the behavior under high compression , e.g. @xcite .
the modification of the pair interactions between the ions by the surrounding plasma is of prime importance for the theoretical understanding of these systems .
in contrast to the familiar coulomb potential , @xmath1 , of an ion observed in vacuo , in a plasma , correlation and quantum effects cause screening . at weak non - ideality
this gives rise to an isotropic yukawa - type potential , @xmath2 , where @xmath3 is the screening length given , in the limit of a classical high - temperature plasma , by the debye radius or , in a high - density quantum plasma , by the thomas - fermi length @xmath4 .
more general screened potentials are successfully computed using linear response theory to obtain the dynamic dielectric function , @xmath5 ( longitudinal density response ) , giving rise to @xcite @xmath6 where the longitudinal dielectric function derives from the dielectric tensor via @xmath7 .
the potential ( [ eq : phi_general ] ) typically decays slower than the exponential yukawa potential .
moreover , screening effects are known to give rise to an attractive region of the potential ( `` anti - screening '' ) in nonequilibrium situations such as a charge embedded into a streaming flow of oppositely charged particles , which is well studied theoretically @xcite , and the resulting attractive force has been confirmed e.g. in dusty plasma experiments with streaming ions .
no attractive forces have been seen in simulations in equilibrium plasmas with two exceptions : i ) the formation of bound states ( such as hydrogen molecules or molecular ions ) evidently corresponds to a net attractive force between the constituents ( hydrogen atoms , protons ) .
ii ) oscillatory potentials with ( very shallow ) negative parts ( friedel oscillations ) emerge in a strongly degenerate fermi gas as a consequence of the fermi edge singularity ( essentially due to the step character of the zero temperature momentum distribution of an ideal fermi gas ) , e.g. @xcite .
friedel oscillations have been observed in experiments probing surface states at very low temperature , e.g. @xcite .
@xcite se compute the potential ( [ eq : phi_general ] ) of a proton in a hydrogen plasma by 1 .
using zero temperature classical hydrodynamic equations for the electron density and mean velocity , coupled to the poisson equation for the electrostatic potential , 2 . adding , in the momentum equation , quantum diffraction effects via the bohm potential @xmath8 ( quantum hydrodynamics , qhd ) , 3 .
adding the pressure of the ideal fermi gas at zero temperature , 4 . adding exchange and correlation effects via an additional potential @xmath9 using a simple parametrization , ref .
33 in @xcite , 5 .
neglecting dynamic effects , @xmath10 , 6 . solving the resulting hydrodynamic equations in linear response for the dielectric function @xmath11 ( @xmath12 , in their notation ) , which parametrically depends on the coupling parameter ( brueckner parameter ) @xmath13 that completely defines the plasma state at zero temperature .
[ here we introduced the mean interparticle distance , @xmath14 , which is related to the unperturbed density by @xmath15 .
for the explicit form of the used lqhd equations we refer to ref .
@xcite , for the discussion of @xmath9 and earlier related work , see ref . @xcite . in the following we
restrict ourselves to the linearized equations used by se , i.e. to the lqhd .
the full qhd is beyond the scope of this paper , for references , see ref . @xcite . versus brueckner parameter @xmath16 .
an attractive proton potential is predicted in the density interval @xmath17 at zero temperature .
the shaded area denotes the range of @xmath16 where the plasmon energy is smaller than the fermi energy , @xmath18 ( weak coupling ) . ] using lqhd , se evaluate the static dielectric function , @xmath19 , and screened proton potential , @xmath20 , at zero temperature , as a function of a single parameter @xmath21 .
we plot @xmath21 as a function of the brueckner parameter in fig .
[ fig : alpha ] .
se observe that the screened potential of a proton , eq .
( [ eq : phi_general ] ) , develops a negative minimum for @xmath22 . as can be seen in the figure
, @xmath21 exceeds the value @xmath23 , in a range of coupling parameters @xmath24 . and [ fig : phi_min ] , respectively .
thin vertical lines indicate the equilibrium nearest neighbor distance of two protons , cf .
[ fig : r_min ] .
the black arrow marks the range of values of @xmath25 shown in fig .
[ fig : dft - electrostatic_potential ] . ]
let us now analyze the results for the screened proton potential in the whole range of @xmath16 values where an attractive hydrogen interaction is observed , as done in ref .
@xcite . in fig .
[ fig : pot_se ] we plot the potential energy corresponding to the screened proton potential in units of ev for three densities corresponding to @xmath26 . at low and high densities ( i.e. for cases where @xmath27 )
, we confirm that the potential is purely repulsive . in agreement with fig .
[ fig : alpha ] , for @xmath28 a negative minimum develops at a distance of several bohr radii from the proton .
the position of the minimum is plotted versus @xmath16 in fig .
[ fig : r_min ] .
there exists a minimum of this distance of about @xmath29 for @xmath30 with a slow ( rapid ) increase for lower ( higher ) densities .
finally , we analyze the depth of the attractive potential minimum .
these values are plotted in fig .
[ fig : phi_min ] for @xmath16 between @xmath31 and @xmath32 below which the minimum vanishes .
the deepest minimum is observed around @xmath33 and amounts to about @xmath34mev corresponding to a temperature of about @xmath35k . ,
compared to the mean interparticle distance .
note that the ground state nearest neighbor distance of two protons in the molecular phase ( @xmath36 ) is in the range of @xmath37 . ]
corresponding to @xmath38 , cf .
[ fig : alpha ] . ]
we now compare the lqhd results of se to what is known about dense hydrogen .
there have been numerous studies of hydrogen plasmas based on analytical approximations and first - principle simulations as well , e.g. @xcite .
there is hardly any question that the low - density ground state consists of @xmath39-molecules ( either in a gas , liquid or solid phase ) .
when the density is increased , the molecular binding is reduced until , eventually around @xmath40 , molecules break up .
even though the detailed scenario maybe quite complex , e.g. @xcite , involve intermediate states ( including @xmath41-atoms ) and vary with temperature it is clear that , around some critical density @xmath42 ( mott density ) , all bound states will break up into electrons and protons , e.g. @xcite .
this pressure ionization is a gradual process and a reasonable estimate for @xmath42 corresponds to @xmath43 . upon further density increase one
expects formation of a liquid and , eventually , solid phase of protons embedded into a fermi gas of electrons , e.g. @xcite .
while there remain interesting questions about the precise values of the different phase boundaries @xcite , the general structure of the hydrogen phase diagram at high pressures is well understood , e.g. @xcite . as an illustration , and for direct comparison with se , we present results for the screened proton potential and for the interaction of hydrogen atoms embedded in a jellium background , as obtained from density - functional total energy calculations .
the _ ab initio _ scheme of dft rests on firm theoretical grounds and has been thoroughly tested in recent decades , in particular in application to dense plasmas , see e.g. refs .
it provides a fully consistent treatment of quantum effects and uses the fundamental coulomb interaction between charged particles as input .
the only uncontrolled approximation contained in dft total - energy calculations is the approximation applied to the exchange - correlation energy functional , but the accuracy and limitations of the different approximations available have been investigated in detail for many applications in solid state physics and molecular chemistry @xcite .
therefore , dft total - energy calculations can serve as a benchmark for the linearized quantum hydrodynamic model of se . before proceeding with our own calculations we point out that detailed dft calculations have been carried through for h and h@xmath44 in jellium before , see , e. g. , the work by almbladh _
et al . _ ,
@xcite , by bonev and ashcroft , ref .
@xcite , and by song , ref . @xcite .
our results concur with their findings .
for other recent dft results for dense hydrogen , see refs .
@xcite and references therein . in the following
we restrict ourselves to spin - unpolarized simulations .
it is a well - known artefact of the usually applied approximations to the exchange - correlation energy functional that in vacuo the hydrogen molecule undergoes a transition from a spin - unpolarized to a spin - polarized ground state when the separation of the two protons is increased @xcite .
however , in their spin - polarized dft - calculations for hydrogen atoms immersed in jellium , nazarov _ et al . _
@xcite have found a spin - unpolarized ground - state for @xmath45 .
for this reason , and as no spin - polarization is to be expected at even higher electron densities , we have carried through only spin - unpolarized relaxations for @xmath46 and @xmath47 .
this is consistent with ref .
@xcite . is the distance from the proton .
dft data have been calculated in a cubic box ( size @xmath48 ) using the pbe - gga for the exchange - correlation energy - functional , a single k - point , and a plane - wave cutoff - energy of 200 ry ( @xmath46 ) or 300 ry ( @xmath49 and @xmath50 ) .
the dft data ( symbols ) are compared to the electrostatic potential from lqhd of shukla and eliasson , ref .
@xcite ( lines ) .
the black arrow marks the voltage range shown in fig .
[ fig : pot_se ] .
, width=321 ] a minimum around the hydrogen molecule bond length in vacuo ( @xmath51 bohr ) is observed , for @xmath52 , the molecular bound state is unstable .
the dft data have been calculated in a cubic box ( with a size as noted in the inset ) using pbe - gga or lda for the exchange - correlation energy - functional , a single or 6 k - points in the irreducible part of the brillouin zone , and a plane - wave cutoff - energy of 100 ry .
, width=321 ] the data in figs .
[ fig : dft - electrostatic_potential ] and [ fig : dft - interaction_energy ] have been computed with the fhimd code @xcite .
a @xmath53 coulomb potential ( i.e. no pseudopotential ) has been taken for the proton electron interaction .
the cut - off energy of the plane - wave basis - set has been chosen at 100 ry for the calculation of the h
h interaction energy , while larger values have been used for the calculation of the screened potential of an h - atom , or proton , in jellium .
the two hydrogen atoms forming the h@xmath44 molecule are put into a cubic super - cell with the length of the edge equal to 25 bohr ( 15 bohr ) , in case of @xmath47 ( @xmath46 ) .
the super - cell is repeated periodically in all three directions .
calculations have been carried through using only a single special k - point .
additional computations with 6 special k - points derived from a @xmath54 back - folding @xcite allow an error estimate , as can be seen in fig .
[ fig : dft - interaction_energy ] .
the smearing of the fermi distribution is small , @xmath55 ev or 0.1 ev , and the total energy has been extrapolated to zero temperature under the usual assumption of a linear heat capacity @xcite .
the generalized gradient approximation by perdew and ernzerhof ( pbe96 ) @xcite and , for comparison , the local - density approximation , have been applied to the exchange - correlation energy functional .
let us now consider the results .
as a typical low - density case , we show in fig .
[ fig : dft - electrostatic_potential ] the electrostatic potential around a proton in jellium for @xmath50 .
it exhibits a clear negative minimum around a distance of @xmath56 . the minimum becomes shallower with increasing electron density , cf .
curve for @xmath57 , and vanishes for @xmath52 .
there are further friedel oscillations at larger distances , which , however , are often not well resolved due to computational limitations and the small size of the super - cell , but they are of no relevance for hydrogen at these conditions , see below .
our data can now be directly compared to the lqhd results of se , shown by lines in fig .
[ fig : dft - electrostatic_potential ] . in transferring the shukla - eliasson results from fig .
[ fig : pot_se ] to fig .
[ fig : dft - electrostatic_potential ] one should note the large changes in scale and range . the horizontal axis in fig .
[ fig : dft - electrostatic_potential ] extends out only to @xmath58 , while that of fig .
[ fig : pot_se ] extends more than three times further .
also the vertical scale of the potential energy in fig .
[ fig : pot_se ] is roughly six times smaller than that of fig .
[ fig : dft - electrostatic_potential ] .
thus , on the scale of fig .
[ fig : dft - electrostatic_potential ] , the lqhd minima , which only appear clearly in the inset of fig .
[ fig : pot_se ] , involve such small potential changes on the scale of fig .
[ fig : dft - electrostatic_potential ] that they can not be seen .
the small dip in the lqhd potential for @xmath52 is at about @xmath59 near the right hand edge of the @xmath25-axis , while the dip for @xmath50 is even further away , far beyond the right hand edge . in any case , in the region of the dft minimum , the lqhd potential is monotonically decreasing , i.e. without minima , and far from that of the full calculation , as denoted by the symbols . in part
we ascribe the minimum of the electrostatic potential from the dft calculation in fig .
[ fig : dft - electrostatic_potential ] to the formation of an h@xmath60 ion , i. e. a bound state which almbladh _ et al .
_ @xcite have found for @xmath61 . as pointed out by bonev and ashcroft @xcite
, this formation of a negative ion sheds doubts on the adequateness of jellium - type ( or one - component plasma ) models for the description of a nonideal hydrogen plasma .
the recent ab initio simulations of hydrogen by morales et al .
@xcite and bonev et al .
@xcite are in fact based on dft - born - oppenheimer molecular dynamics simulations of ensembles or monte carlo simulations and thus fully account for the dynamics of the protons .
the interaction energy of two hydrogen atoms immersed in a jellium background with @xmath50 or @xmath52 is displayed in fig .
[ fig : dft - interaction_energy ] .
in fact , the h h interaction energy has been computed by bonev and ashcroft @xcite before , using the dft total - energy program vasp .
they point out that linear response theory would be inadequate for this purpose .
furthermore , they describe the fate of the hydrogen bond when the electron density is increased : for @xmath62 a local energy minimum at a h h separation larger , but still close to , the bond length of the hydrogen molecule in vacuum , and an energy barrier towards h@xmath44 association are observed @xcite .
our dft results summarized in fig . [ fig : dft - interaction_energy ] show the same behavior : for @xmath47 we obtain a stable h@xmath44-bond at a bond length comparable to the bond length of a hydrogen molecule in vacuo , but with a much smaller binding energy .
this result is not sensitive with respect to the approximation applied to the exchange - correlation functional .
also the more technical effect of the restricted k - point sampling can be read from the figure . comparing again to the lqhd results of shukla and eliasson , it is obvious that the interaction energy minimum found in the _ ab initio _ data is not included in their approach .
thus , the se screened proton potential has , for small densities ( e.g. @xmath50 ) , a qualitatively incorrect shape , in particular it completely misses the molecular bound state . for larger density , i. e. for @xmath52 , fig .
[ fig : dft - interaction_energy ] shows that the minimum in our simulations becomes very shallow or vanishes at all , within the accuracy of the numerical dft computations .
similarly , we do not resolve extremely shallow ( below @xmath63mev ) binding potentials at larger distances from the proton , such as the ones related to friedel oscillations .
these oscillations could be reproduced by dft using a substantial computational effort .
however , there is no need for this since such extremely small features of the potential are expected to be of no relevance for all dense plasma applications since they always encounter a finite electron temperature of at least one ev .
the approach by se in ref .
@xcite starts from the assumption that the ions are immobile and embedded into a neutralizing background ( see eq .
( 3 ) in ref .
we have thus used dft , together with the lda or the pbe - gga @xcite applied to the exchange - correlation energy functional , to calculate the interaction energy of two protons immersed in a jellium background , without any further approximation applied to the quantum - mechanical problem . for the reliability of the dft as compared to quantum monte carlo simulations we refer e.g. to ref .
thus the dft results can serve as a reference to evaluate the accuracy of the interaction potential as derived from the lqhd approach by se @xcite . as shown from the comparison to ab initio dft simulations
, the se potential completely misses the bound states of protons in low temperature hydrogen .
when the density is increased the deviations between lqhd and dft potentials become smaller but are still noticeable . yet even at higher densities where no molecules exist the se potential is qualitatively wrong because it does not show an attractive minimum for @xmath64 at all .
dielectric theories of the electron gas and electron liquid including correlations in the frame of local field corrections , however , confirm that friedel oscillations persist up to high densities , e.g. @xcite . the present lqhd model of se also misses the friedel oscillations , as the authors themselves underline . furthermore , we consider two limiting cases of the lqhd potential discussed by se @xcite : first , in the limit @xmath65 se recover a yukawa potential with the screening length @xmath4 .
this is correct for high densities , @xmath66 , but not for low densities , @xmath67 where @xmath65 as well , cf .
[ fig : alpha ] .
second , se recover , for @xmath68 , the exponential cosine - screened coulomb potential @xcite .
this limit is questionable since @xmath21 can not exceed @xmath69 , cf .
[ fig : alpha ] .
these failures rule out any reliable predictions such as novel potential minima . besides , the extremely low value of the associated binding energy ( cf .
[ fig : phi_min ] ) would require a particularly accurate theory and careful verification .
finally , let us analyze possible reasons for the failure of the linearized qhd model in application to dense two - component plasmas .
i. : : as any fluid theory , lqhd can not resolve distances below some cut - off @xmath70 . this has been pointed out by manfredi and co - workers @xcite who performed a test of a 1d version of nonlinear qhd for thin metal films @xcite .
ii . : : shukla and eliasson note @xcite that qhd is valid only for @xmath71 , i.e. for weak coupling .
this range is indicated by the shaded area in fig . [
fig : alpha ] .
since the present model of lqhd contains exchange and correlation contributions , this range may actually be larger @xcite .
iii . : : linearized qhd fails when the linearization conditions are violated , i.e. when the perturbed density is comparable with the unperturbed density which is common in the description of strong attractive interactions .
while item ( ii ) is presently open and requires further analysis , our dft simulation results allow us to directly verify ( i. ) . in a hydrodynamic description of classical plasmas
the smallest length scale that can be resolved is the debye length .
similarly , in the quantum case , this cut - off is expected to be of the order of the thomas - fermi length , @xmath72 .
in fact , re - examining the dft and lqhd data for the screened proton potential in fig .
[ fig : dft - electrostatic_potential ] we observe that the most dramatic deviations occur on length scales smaller than @xmath73 .
thus we confirm that lqhd can not , by construction , yield potentials with atomic - scale resolution in a dense quantum plasma @xcite . yet
the most severe limitation is , apparantly , related to the linearization of the qhd equations , item ( iii ) .
in fact , similar linear response calculations for a proton in a degenerate electron gas have been done long ago .
almbladh et al .
@xcite have performed density functional calculations and compared them to a linearized version of dft ( which , by construction , is essentially more accurate than any hydrodynamic approach ) .
they observed a complete failure of the linear theory in the prediction of the electron density screening the proton .
not only is the electron density at the proton a factor @xmath74 to @xmath75 ( for @xmath76 and @xmath77 , respectively ) too small , but linear theory generally predicted friedel oscillations to occur at much too high distances from the proton .
thus , a possible explanation for the potential minimum observed in the present linearized qhd , cf . fig . [ fig : pot_se ] , is that it is a trace of the friedel oscillations ( the first minimum ) that is displaced to higher distances from the proton due to the linearization , cf . also fig . [ fig : r_min ] .
it would , therefore , be interesting to test this hypothesis by comparing the lqhd results of se to full nonlinear qhd .
let us assume for a moment that the attractive potential of se would be a real effect and consider its implications for proton crystallization @xcite . to this end ,
we plot in fig .
[ fig : r_min ] the location of the potential minimum and compare it to the mean interparticle distance for a given density .
it is obvious that there is a striking mismatch .
even if the system would be at zero temperature protons could not occupy the minimum locations simply because their density is several orders of magnitude too high ( see also fig .
[ fig : pot_se ] where the mean interparticle distance is marked on the se potential ) , making such a state energetically impossible .
we thus have to conclude , based on dft simulations and general considerations about the location and low depth of the potential minimum , that the predictions of se of novel attractive forces , novel ion lattices , atoms , molecules , critical points and phase transitions in dense hydrogen are invalid .
this work was supported by the deutsche forschungsgemeinschaft via sfb - tr 24 project a5 .
we are grateful fo the referees for suggestions that have helped improve the clarity of the manuscript . for an overview , see e.g. m. korth , s. grimme , j. chem .
theory comput .
* 5 * , 993 ( 2009 ) and c. fiolhais , f. nogueira , m. marques , _ a primer in density functional theory _ , lecture notes in physics * 620 * , springer - verlag ( berlin , 2003 ) .
for completeness we mention that for distances larger than @xmath70 there are still noticeable deviations of the lqhd potential from the dft result that are significantly larger than the minimum of the lqhd potential , cf .
[ fig : phi_min ] . | in a recent letter [ p.k . shukla and b. eliasson , phys .
rev . lett . * 108 * , 165007 ( 2012 ) ] the discovery of a new attractive force between protons in a hydrogen plasma was reported that would be responsible for the formation of molecules and of a proton lattice . here
we show , based on ab initio density functional calculations and general considerations , that these predictions are wrong and caused by using linearized quantum hydrodynamics beyond the limits of its applicability . |
markov chain monte carlo ( mcmc ) algorithms @xcite are often used to generate samples distributed according to non - trivial densities in high dimensional spaces .
many algorithms have been developed that allow mcmcs to produce samples @xmath0 from an unnormalized target density @xmath1 : @xmath2 in many applications , it is desirable or even necessary to be able to normalize the target density
. i.e. , to calculate @xmath3 where @xmath4 is the support of @xmath5 . this integral can be computationally very costly or impossible to perform with standard techniques if the volume where the target @xmath5 is non - negligible occupies a very small part of the total volume of @xmath6 .
an important area where such integration is necessary is for bayesian data analysis @xcite .
bayes formula reads , for a given model @xmath7 , @xmath8 where here @xmath9 are the parameters of the model and the data are used to extract probabilities for possible values of @xmath9 .
the denominator is usually expanded using the law of total probability and written in the form @xmath10 and goes by the names ` evidence ' , or ` marginal likelihood ' , and is the type of integral that we want to be able to calculate ( here the data are fixed and @xmath11 ) .
an example use of @xmath12 is for the calculation of bayes factors in the comparison of two models : @xmath13 another application where the calculation of a normalization can be very important is in the parallelization of the mcmc algorithm @xcite . while the mcmc approach has very attractive features , it is often slow in its execution due to the nature of the algorithm .
a goal is therefore to parallelize the computations needed to map out the target density .
this looks at first sight difficult since the mcmc algorithms are by construction serial .
a parallelization of the calculations can however be achieved via a partitioning of the support .
i.e. , we partition @xmath6 into sub volumes @xmath14 with @xmath15 and we run a separate mcmc sampling for each sub volume @xmath14 . in order to have a final set of samples representing the target density over the full support
, we need to know the relative probabilities for the different sub volumes .
i.e , we need @xmath16 the samples in the different regions are then given weights @xmath17 with @xmath18 the number of samples from @xmath1 in @xmath14 and @xmath19 .
a variety of techniques to calculate the evidence in bayesian calculations have been successfully developed .
a summary can be found in @xcite , where a number of mcmc related techniques are reviewed , including laplace s method @xcite , harmonic mean estimation @xcite , chib s method @xcite , annealed importance sampling techniques @xcite , nested sampling @xcite and thermodynamic integration methods @xcite .
we are here specifically interested in testing techniques directly applicable in an mcmc setting , and which is independent of the specific mcmc algorithm .
we assume that the mcmc algorithm has been successfully run to extract samples according to the target density , and the goal is to provide an algorithm for calculating the normalization ( or evidence ) .
given our requirements , only arithmetic mean estimation ( ame ) , harmonic mean estimation ( hme ) and laplace methods are directly applicable . using ame and hme methods
directly is known to fail in many situations , and the laplace method is only applicable if the target density is gaussian .
we introduce the use of a reduced integration volume and normalization using the mcmc output to improve the ame and hme performance . after a description of the techniques
, we report on numerical investigations of the different approaches using samples from the mcmc code bat @xcite . assuming the mcmc has been successfully run to extract samples according to @xmath20 , one of the quantities directly retrievable from the mcmc output
is an estimate of the parameter values at the global mode : @xmath21 is in the neighborhood of @xmath22 .
i.e. , we know approximately where the integrand in eq .
[ eq : integral ] has its maximum .
we note that @xmath23 with @xmath24 a sub support of @xmath6 is directly estimated from the mcmc output by counting the fraction of samples falling within @xmath25 , @xmath26 ( the reason for this notation will become clear below ) .
i.e. , the task of evaluating @xmath27 reduces to integrating the function @xmath1 over a well - chosen region - presumably a small region around @xmath22 and dividing by @xmath28 .
this integral can be much simpler to evaluate than the integral over the full support . in the following , we use a simple hypercube for our integration region .
from the mcmc samples , we can construct the marginalized distributions along each of the @xmath9 dimensions .
we define an interval along each dimension centered at @xmath21 with width which is a multiple of the standard deviation ( we use the symbol @xmath29 to represent this factor ) .
the optimum value of @xmath29 depends on the dimensionality of the problem as described below .
another option would be to produce a covariance matrix of the @xmath0 for sampling using a multivariate normal distribution if desired , but this was not found necessary in the examples we have studied .
the integral in the numerator in eq .
[ eq : scheme ] can presumably be determined in a straightforward way since now we are focusing on a small volume with significant mass .
the standard importance sampling approximation is given by @xmath30 where our sampling probability density is given by @xmath31 .
@xmath32 is the number of samples used in the calculation .
if we choose for @xmath31 a uniform distribution in the hypercube , then we have the well - known sample mean result @xmath33 with @xmath34 the volume of the hypercube .
our estimator for @xmath27 is then @xmath35 we will use this simplest version of the estimator for our examples below . assuming unbiased gaussian distributions for @xmath36 and @xmath28 about their true values
, we can estimate the uncertainty for @xmath27 with @xmath37 where @xmath38 the effective sample size @xcite is defined here as @xmath39 with the autocorrelation function at @xmath40 defined for our mcmc sample as @xmath41 in these equations , the subscript @xmath42 labels the component of @xmath0 , while the index @xmath40 labels the iteration in the mcmc .
the uncertainty from the sample mean integration is estimated by separating the sample mean calculation of @xmath36 into @xmath43 batches and looking at the variance of these calculations : @xmath44 with these definitions , we are able to report both an estimate for our integral and an uncertainty . these will be compared to accurately calculated values for the chosen examples in the following sections . the hme @xcite value for @xmath27 can be calculated as follows :
@xmath45_{\hat{f}(\lambda ) } & = & \int_{\omega } \frac{1}{f(\lambda ) } \cdot \frac{f(\lambda)}{i } d\lambda \\ & = & \frac{v}{i } \end{aligned}\ ] ] where @xmath46 is the normalised target density and @xmath47 is the total volume of the support .
the hme estimator is then @xmath48 this calculation is performed directly from the mcmc output from which the samples @xmath0 as well as @xmath49 are available , and does not require an extra sample mean calculation as in the ame scheme . however , it can be unstable because of samples occurring ( or missing ) in regions where @xmath1 is small ( relative to other regions ) .
we can improve the estimation , as originally noted in @xcite , by limiting ourselves to a small volume around the mode . using the same notation as above ,
we can write @xmath50 where now only the samples in the restricted support @xmath24 are used . the uncertainty in the estimate is calculated by separating the mcmc samples included in our integration region into batches and looking at the variation of these estimates . in this approach
, the target distribution is assumed to be represented by a ( multivariate ) gaussian distribution .
the estimator for the normalization is then @xmath51 where the target density is evaluated at the mode returned from the mcmc and @xmath52 is the determinant of the covariance matrix evaluated numerically from the samples @xmath0 .
this method is clearly only expected to work in cases where the assumption of normality is valid .
we start with a simple example - the target density is the product of a number of gauss functions depending on only one parameter - to describe our testing procedures in detail .
we then move on to more complicated examples in multivariate spaces , including functions with degenerate modes .
all mcmc calculations were performed using the bat program , with samples from the target density taken after convergence of the mcmc algorithm .
we start with the following target function : @xmath53 this type of function could , e.g. , be the likelihood function constructed for producing an estimate of a quantity , @xmath54 , given @xmath55 measurements , @xmath56 , with a sampling distribution modeled by a gaussian probability distribution of fixed width @xmath57 .
the normalization integral for the target can be performed analytically assuming the volume of interest extends well beyond the extreme values of the @xmath58 . for the more general case of the product of @xmath55 @xmath59-dimensional uncorrelated gaussian functions with known variances @xmath60 , the integral
is given by @xmath61 d\vec{\mu } \\ & = & \frac{1}{(2\pi)^{d\cdot ( m-1)/2}\vert\sigma\vert^{(m-1)/2 } m^{d/2 } } \exp\left(-\sum_{j=1}^d \frac{{\rm var}[x_j]}{2\sigma_j^2 } \right ) \end{aligned}\ ] ] where @xmath62 is the ( diagonal ) covariance matrix .
for our concrete example , we take @xmath63 and generate random values of @xmath64 from a gauss distribution of mean zero and unit standard deviation , and we find for the generated values @xmath65 . in evaluating the integral , we take for the support @xmath66 .
we then use @xmath67 samples from the mcmc output to find an estimate for the mode of @xmath68 and to calculate the standard deviation for @xmath54 .
the distribution of samples from @xmath68 from the mcmc are displayed in fig .
[ fig : mcmc1d](left ) .
the mode of the samples is found at @xmath69 and the standard deviation is found to be @xmath70 .
the effective sample size for this set of samples is @xmath71 . the dependence of @xmath28 on the chosen value of @xmath29 is also shown in fig .
[ fig : mcmc1d](right ) for 500 values of @xmath29 ranging from @xmath72 to @xmath73 in steps of @xmath72 . for a one - dimensional gaussian target density , which is what we have here
, the expectation is that 68 % of mcmc samples occur within @xmath74 and 95 % occur within @xmath75 , and this is indeed what is found . as a function of the value of @xmath29 ( in units of the standard deviation of the distribution).,title="fig:",width=264 ] as a function of the value of @xmath29 ( in units of the standard deviation of the distribution).,title="fig:",width=264 ]
we then perform a sample mean calculation with @xmath76 samples for each of the different choices of @xmath29 . for each calculation
, we extract a value of @xmath77 as described in section [ sec : sm ] as well as an estimate of the uncertainty .
the extracted values of @xmath77 ( divided by the true value ) are shown as a function of @xmath29 in fig .
[ fig : evidence1d](left ) .
the error bars are the estimated one standard deviation uncertainties .
we observe small systematic deviations of the results for small values of @xmath29 resulting from the inaccurate determination of @xmath78 from the mcmc samples ( note that the mcmc was only run once , so that the @xmath28 values are correlated ) . as a function of @xmath29 , scaled by the true value .
the error bars correspond to the estimated uncertainty .
right ) the actual error @xmath79 ( red ) , the estimated uncertainty @xmath80 ( black ) and the total estimated uncertainty @xmath81 ( blue ) as a function of @xmath29 .
, title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
the error bars correspond to the estimated uncertainty .
right ) the actual error @xmath79 ( red ) , the estimated uncertainty @xmath80 ( black ) and the total estimated uncertainty @xmath81 ( blue ) as a function of @xmath29 .
, title="fig:",width=264 ] to study the uncertainty estimation , we compare @xmath81 to @xmath82 at each value of @xmath29 .
the results are shown in fig .
[ fig : evidence1d](right ) . in this figure ,
the red points indicate the absolute value of @xmath82 , the black points the estimated uncertainty coming from the sample mean calculation , @xmath83 , and the blue points the total estimated uncertainty , @xmath81 .
we observe that our estimated uncertainty is accurate , and that there is a minimum of the uncertainty around @xmath75 .
the location of the minimum clearly depends on the number of samples chosen for the mcmc and sample mean calculations , but it is important that we can accurately estimate the uncertainty . in this case , the arithmetic mean calculation is quite accurate even at large values of @xmath29 since we are only working in one - dimension .
we now evaluate the harmonic mean estimate for @xmath27 as described in section [ sec : hme ] .
the estimate @xmath84 as well as the absolute deviation from @xmath85 as a function of @xmath29 are shown in fig .
[ fig : hme ] .
we see for this example that the hme technique works well , and that accuracies of a fraction of 1 % are possible from the hme estimation at @xmath86 . as @xmath29
is increased , the hme estimation worsens since , although more of the mcmc samples are included , reducing the binomial uncertainty on @xmath26 , imperfect sampling in the tails of the distribution plays a large role and we see the importance of limiting the range of the integration region for the hme calculation already with this simple one - dimensional example .
the uncertainty is somewhat worse than what was found for the ame calculation , but probably adequate for the majority of applications .
also , the calculation did not require the extra step of performing a sample mean calculation .
scaled by the true value as a function of @xmath29 .
the error bars correspond to the estimated uncertainty .
right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
, title="fig:",width=264 ] scaled by the true value as a function of @xmath29 .
the error bars correspond to the estimated uncertainty .
right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
, title="fig:",width=264 ] as seen in fig .
[ fig : mcmc1d ] , the target density is gaussian and therefore the laplace method is expected to work well .
indeed , the laplace method yields an estimate within @xmath88 % of the true value in this example : @xmath89 .
we now move to a target density composed of a product of ten dimensional gaussian distributions with non - diagonal covariance matrix .
the target function in this case is : @xmath90 where @xmath62 is the covariance matrix , assumed to be known , and @xmath91 .
the target function is ten - dimensional and has significant correlations among the ten parameters .
the values of @xmath92 were chosen by generating random vectors using @xmath93 and the following covariance matrix @xmath94 and again could represent a type of situation found in a data analysis setting .
the integration region for @xmath27 was taken as a @xmath95d hypercube of side length @xmath96 centered on @xmath97 .
the value for @xmath85 can again be evaluated analytically by finding the similarity transformation that diagonalizes the covariance matrix .
the expression of the integral in this case is @xmath98 \right ) .\ ] ] where @xmath99 and @xmath100 with @xmath101 a diagonal matrix .
the true value of the integral for randomly generated data was evaluated using this expression and yielded @xmath102 .
the mcmc program bat was used to sample from the target density with @xmath103 samples stored post - convergence ( yielding @xmath104 ) .
the value of @xmath28 is given as a function of @xmath29 in fig .
[ fig:10destimator ] . as a function of the value of @xmath29 ( in units of the standard deviation of the marginalized distribution ) for the product of ten - dimensional correlated gauss functions.,width=340 ]
the arithmetic mean calculation was performed at each of @xmath105 values of @xmath29 as in the one - dimensional case , with @xmath106 samples in each ame run .
the results are shown in fig .
[ fig : evidence10d ] .
as is seen , for values of @xmath29 around @xmath107 , the uncertainty is about 1 % .
the method does not show any systematic biases for @xmath108 , and the estimated uncertainty is again a good estimator for the error . at small @xmath29 , where a small number of mcmc samples are used
, the correlation between the mcmc samples produces some systematic errors in the evaluation of @xmath27 . as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ] the results for the hme estimator are also shown in fig . [
fig : evidence10d ] .
we see that accuracies of a few tens of % are achieved , but only in a narrow @xmath29 range . for @xmath109 ,
the error is more than @xmath110 % and the hme estimate is no longer useful .
also , the estimated uncertainty is too low and does not provide a reliable estimate of the true error .
the hme method is clearly already running into trouble at this level of complexity .
the target density is again a multivariate gaussian , and the laplace method works well , yielding @xmath111 .
we now move beyond simple unimodal gaussian type target densities and consider a function in @xmath59 dimensions with degenerate modes lying on a @xmath112 dimensional surface of fixed radius , a gaussian shell : @xmath113 this function is centered at @xmath114 with degenerate modes along a surface of radius @xmath78 .
the value of the function decreases away from the modal surface along a radius according to a gaussian shape with standard deviation @xmath57 .
the integral of this function can be evaluated using spherical coordinates centered at @xmath114 , where @xmath115 is the radial coordinate in the space , so that @xmath116 the volume element , integrated over the angular coordinates , is @xmath117 with @xmath118 , so that we have @xmath119 we are left with a one - dimensional integral that can be easily calculated numerically to high precision .
note that we have assumed that the integral in the region outside @xmath120 ( the corners in the hypercube ) is vanishingly small .
this is the case for the examples considered in this article .
for the three examples below , we use the following settings : radius @xmath121 , width @xmath122 and @xmath123 .
the integration region extends from @xmath124 in each dimension .
the parameter values result in @xmath125 .
we use the bat code to produce @xmath106 mcmc samples from the target density , yielding an effective sample size @xmath126 . the sample distribution from the mcmc as well as the estimate of @xmath28 as a function of @xmath29
are shown in fig .
[ fig:2dshell ] .
standard deviation ranges .
right ) fraction of mcmc samples falling within the hypercube of side length @xmath127 as a function of the value of @xmath29 ( in units of the standard deviation of the distribution).,title="fig:",width=264 ] standard deviation ranges .
right ) fraction of mcmc samples falling within the hypercube of side length @xmath127 as a function of the value of @xmath29 ( in units of the standard deviation of the distribution).,title="fig:",width=264 ] as can be seen in the figure , the mcmc has produced a reasonable sample distribution .
the location of the mode from the posterior samples happens to be close to @xmath128 and is indicated in the figure ( note that @xmath129 in the figure ) .
the lack of a single mode is not a problem for the ame and hme algorithms , but we no longer expect the laplace method to give sensible results .
the mean values of @xmath130 are very close to @xmath131 and the standard deviation in each direction is about @xmath73 units .
the hypercube centered at the mode found from the mcmc samples and with @xmath74 contains about @xmath132 % of the samples , and the hypercube with @xmath75 contains about @xmath133 % of the samples .
we again use @xmath106 samples for our sample mean calculations at each of the values of @xmath29 .
the results for @xmath77 are shown in the top plots in fig .
[ fig : evidence2dshell ] , and we see that there is no difficulty in achieving a good result for the integral despite not having a simple mode for the target distribution .
the accuracy of the calculation is good , and the uncertainty is better than 1 % for a wide range of @xmath29 , despite the rather small number of samples in the mcmc and ame calculations .
we again find that our estimated uncertainty gives a good reproduction of the actual error .
the hme evaluations are also given in figs .
[ fig : evidence2dshell ] . here
we find good performance ( few % level accuracy ) up to @xmath75 , at which point the hme calculation starts to systematically deviate from the correct value . in this case , the estimated uncertainty does not give a reliable indication of the actual error for @xmath109 and in fact the uncertainty is grossly underestimated .
this is a result of the missing mcmc samples at very small @xmath134 .
the volume term in the numerator in eq . [ eq : hme ] grows as @xmath29 is increased , but is not properly compensated by large terms that should appear in the denominator from small values of @xmath134 .
the inability to diagnose this behavior implies that the hme is unreliable . as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as expected , the laplace method does not work for the gaussian shell situation .
for the two - dimensional example considered here , @xmath135 . here
in a first calculation , we use the bat code to produce @xmath106 mcmc samples from the target density , yielding an effective sample size @xmath137 and calculate the evidence .
we again use @xmath106 samples for our sample mean calculations .
the results for the ame and hme evaluations are given in fig .
[ fig : evidence10dshell ] . for the arithmetic mean calculation
, we see the same pattern as in the previous examples . for small values of @xmath29 , the uncertainty coming from the small number of mcmc samples dominates .
however , sub % errors are possible for @xmath138 , which corresponds to @xmath139 . as @xmath29 increases
, the uncertainties from the sample mean calculation dominate since we move to regions of the space that do not contain significant probability mass .
the estimated uncertainty is again accurate and can be used as a guide to choose the optimal value of @xmath29 as we discuss below .
the hme estimate achieves few % accuracy at a somewhat smaller value of @xmath29 than the optimal for the sample mean calculation .
the estimated uncertainty is again tends too small at larger @xmath29 and is not reliable .
as expected , the laplace method does not work well and yields @xmath140 . as a check that these results are not due to small mcmc sample size
, the calculations were redone for @xmath141 mcmc samples .
the optimal value of @xmath29 changes somewhat for the sample mean calculation , but otherwise all results are basically as before .
the systematic behavior of the @xmath87 is the same as for the smaller mcmc sample size ; no significant improvement in performance was found with the 10 times large mcmc sample size . as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty.,title="fig:",width=264 ] as an extreme example , we considered a 50-dimensional gaussian shell . here
the modal surface is a 49-dimensional hypersphere and @xmath142 .
the bat code was used to initially produce @xmath106 mcmc samples from the target density , yielding an effective sample size @xmath137 .
the values of @xmath28 increase rapidly from @xmath143 at @xmath144 to @xmath145 at @xmath146 .
the standard deviations in each dimension is about @xmath147 units , so that @xmath148 approximately covers the full support defined for the function .
the results for the ame and hme evaluations are given in fig .
[ fig : evidence50dshell ] . the best result for the sample
mean calculation gives about @xmath95 % accuracy , whereas the hme calculation is within @xmath110 % of the correct result for a small range of @xmath29 where @xmath28 starts to increase .
we used @xmath106 samples for our sample mean calculations , although this is clearly too small a number for such a large dimensional volume . the error from the sample mean calculation increases rapidly as we increase @xmath29 , and becomes completely unreliable for @xmath109 .
for such a large volume , the vast majority of sample mean evaluations are in regions where the target density is vanishingly small and the uncertainty grossly underestimates the true error . in the next section ,
we discuss a choice of settings for the sample mean calculation and redo the calculation shown here .
as expected , the laplace method does not work well and yields @xmath149 .
we again checked that these results are not due to small mcmc sample size , the calculations were redone for @xmath150 mcmc samples .
the optimal location of @xmath29 changes to smaller values for the sample mean calculation and few % level accuracy is reached . for the hme calculation , a small improvement is also observed , but otherwise all results are basically as before . as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ] as a function of @xmath29 , scaled by the true value .
top right ) the actual error @xmath79 ( red ) , the estimated uncertainty from the sample mean calculation ( black ) and the total estimated uncertainty ( blue ) as a function of @xmath29 .
bottom left ) @xmath84 scaled by the true value as a function of @xmath29 .
bottom right ) the actual error @xmath87 and the estimated uncertainty as a function of @xmath29 .
the error bars in the left plots correspond to the estimated uncertainty .
, title="fig:",width=264 ]
based on the results in the previous sections , we discuss now a procedure for choosing the value of @xmath29 for both the sample mean and harmonic mean estimators .
as was seen in our examples , the uncertainty in the calculation for the ame estimator comes from two sources - the approximately binomial fluctuations in the number of mcmc samples included in our region of interest specified by @xmath29 , and the uncertainty coming from the sample mean calculation .
the first uncertainty can be estimated from the mcmc output , and can be used to define a value of @xmath28 by specifying that this source of uncertainty should contribute half of the final uncertainty .
i.e. , we find the value of @xmath28 such that ( see eq .
[ eq : uncertainty ] ) @xmath151 where @xmath152 is the target uncertainty .
we will use @xmath153 for our discussion below except for the fifty - dimensional gaussian shell example , where we take @xmath154 .
once we have fixed @xmath28 in this way , we then find the corresponding value of @xmath29 and use this to calculate sample mean integrals with for a batch of samples , requiring a minimum of @xmath95 batches .
we use the variance of these @xmath95 calculations to determine how many batches will be needed to get the desired uncertainty ; i.e. , @xmath155 the results for the examples given in the previous sections using this procedure for fixing the parameters of the algorithm is given in table [ tab : summary ] .
as is seen , the range of values for @xmath29 is relatively narrow and only grows slowly with the complexity of the target function .
the number of sample mean calculations however depends strongly on the complexity of the problem , and is also inversely dependent on the accuracy specified and on the size of the mcmc sample . for a given specified accuracy
, @xmath29 is reduced as @xmath156 is increased , and this reduces the number of sample mean calculations necessary .
we find that the ame algorithm gives a reliable estimate of the uncertainty for the examples chosen if the required number of sample mean calculations is not too large .
we conclude that the ame calculation of the integral of the target density using a reduced volume around the mode of the target works well for the types of cases we have studied .
.summary of the results on different target functions for the ame estimator of the normalizing integral .
@xmath156 is the number of posterior samples from the mcmc , @xmath157 is the effective sample size , @xmath152 is the specified accuracy for the integral calculation , @xmath29 is the multiplier of the standard deviation along each dimension chosen by the algorithm , @xmath32 is the number of samplings of the function used in the sample mean calculation , @xmath85 is the true value of the integral , @xmath158 is the fractional error made in the calculation and @xmath159 is the estimated fractional uncertainty from the calculation . [
cols="^,^,^,^,^,>,<,^,^",options="header " , ] as can be seen from the table , and as discussed earlier , the hme calculation works well for the simple target functions considered , but does not produce good results for the more complicated target functions . in particular , the estimated uncertainty does not provide a good estimate of the actual error , so that it is not possible to diagnose that the calcuclation is not performing well .
we therefore do not recommend the use of the hme estimator to calculate the normalization integral for anything but the simplest low - dimensional target densities , the laplace estimation works well in cases where the target density is well approximated by a ( multivariate ) gaussian distribution .
if this is known to be the case , then this approximation is easily calculated and can be used .
however , it should be avoided if the shape of the target distribution is not well known .
we have investigated techniques for the integration of the target density in cases where a mcmc algorithm has successfully run . we do not attempt to modify the sampling of the target density , but only to provide a post - processor for an mcmc algorithm . from the mcmc
, we have an estimate of the global mode and also the variance of the samples marginalized along each parameter dimension .
we use this information to define a hypercube centered on the global model and having side lengths proportional to the standard deviation along these directions , and then calculate the integral of the target function in the reduced volume using either an arithmetic mean or harmonic mean approach .
the fraction of mcmc samples within the reduced volume was used to estimate the integral of the target density over the full volume of interest .
this technique was tried on a variety of examples and also compared to a laplace estimator .
the key elements of the methods studied are : * given the mcmc has been run successfully , the evaluation of the normalization of the target function can be performed using any sub support of the support of the target function ; * from the mcmc , we can find a point near the maximum of the target function , and we can perform the integration in a region which is in some ways optimal by centering the sub support on this point ; * it is possible to also calculate an estimated accuracy for the integral .
our conclusions are that the arithmetic mean calculation performed in a hypercube centered on the observed mode works well and provides a technique for calculating the normalization of the target density with a reliable uncertainty estimate . on the other hand , the harmonic mean estimator only works well in situations where the range of values from the target density does not vary too widely , and
the laplace estimator is restricted for use on gaussian shaped target distributions .
the authors would like to thank frederik beaujean , daniel greenwald , stephan jahn and kevin krninger for many fruitful discussions .
9 see e.g. , c. robert and g. casella , ` monte carlo statistical methods ' , 2@xmath160 edition , springer ( 2004 ) . h. jeffreys , ` theory of probability ' , 3@xmath161 ed .
, claredon press , oxford , mr0187257 ( 1961 ) .
e. t. jaynes , ` probability theory : the logic of science ' , cambridge university press , cambridge , mr1992316 ( 2003 ) .
d. n. vanderwerken , s. c. schmidler , ` parallel markov chain monte carlo ' , arxiv:1312.7479v1 n. friel and j. wyse , ` estimating the evidence - a review ' , stat .
* 66 * ( 2012 ) 2800 .
l. tierney and j. b. kadane , ` accurate approximations for posterior moments and marginal densities ' , journal of the american statistical associations , * 81 * ( 1986 ) 82 .
m. a. newton and a. e. raftery , ` approximate bayesian inference with the weighted likelihood bootstrap ' , journal of the royal statistical society , series b*56 * ( 1994 ) 3 .
s. chib , i. jeliazkov , ` marginal likelihood from the metropolis - hastings output ' , journal of the american statistical association * 96 * ( 2001 ) 270 .
c. p. robert and d. wraith , ` computational methods for bayesian model choice ' , bayesian inference and maximum entropy methods in science and engineering : the 29@xmath162 international workshop on bayesian inference and maximum entropy methods in science and engineering ( aip conference proceedings ) , vol . 1193 ( 2009 ) 251 .
j. skilling , ` nested sampling for general bayesian computation ' , bayesian analysis * 1 * ( 2006 ) 833 .
a. gelman and x. l. meng , ` simulating normalizing constants : from importance sampling to bridge sampling to path sampling ' , statistical science * 13 * ( 1998 ) 163 .
n. friel and a. n. pettitt , ` marginal likelihood estimation via power posteriors ' , journal of the royal statistical society , series b*70 * ( 2008 ) 589 .
a. caldwell , d. kollar , k. krninger , ` bat - the bayesian analysis toolkit ' , comput .
commun . 180
( 2009 ) 2197 - 2209 .
r. e. kass , b. p. carlin , a. gelman , and r. neal , ` markov chain monte carlo in practice : a roundtable discussion ' , the american statistician , * 52 * ( 1998 ) 93 .
a. e. gelfand and d. k. dey , ` bayesian model choice : asymptotics and exact calculations ' , journal of the royal statistical society , b * 56 * ( 1994 ) 501 - 514 . | techniques for evaluating the normalization integral of the target density for markov chain monte carlo algorithms are described and tested numerically .
it is assumed that the markov chain algorithm has converged to the target distribution and produced a set of samples from the density .
these are used to evaluate sample mean , harmonic mean and laplace algorithms for the calculation of the integral of the target density
. a clear preference for the sample mean algorithm applied to a reduced support region is found , and guidelines are given for implementation . |
a proper understanding of the basic mechanism of the electroweak symmetry breaking is one of the central problems in particle physics .
if the fundamental particles are to remain weakly interacting up to high energies the symmetry breaking sector of the model should contain one or more scalar higgs bosons with masses of the order of weak scale @xmath2 .
the simplest mechanism for the breaking of the electroweak symmetry is realized in the sm where one scalar field remains in the spectrum , manifesting itself as the physical higgs particle @xmath3 .
though the scalar sector of the sm is simple enough to predict just one higgs boson , it has been criticized from various theoretical aspects , the most common of which is the quadratic divergence of the higgs mass .
this naturalness problem has triggered the study of supersymmetric models .
mssm , being the simplest of such models , lacks an explanation for the origin and scale of the bilinear higgs coupling mass , which is commonly known as the @xmath4 problem . in connection with this last point ,
extended supersymmetric models , in which bilinear higgs coupling mass @xmath4 is related to the vacuum expectation value of some sm gauge singlet , have been proposed . in these extended models ,
one extends either merely the higgs sector ( nmssm ) @xcite , or both higgs sector and the sm gauge group ( @xmath0 models ) @xcite , and these generally lead to a larger higgs spectrum.thus , in accelerator searches of the higgs , even if detected , it will still be a challenging issue to determine what kind of model it implies .
consequently , it is desirable to investigate the possibility of discriminating between these models given that some spectrum of higgs particles are observed at some ( @xmath1 ) collider .
high precision electroweak data indicate a preference for higgs boson to have a mass within a factor of 2 or so of 140 gev @xcite .
however , weak ( logarithmic ) dependence of the theoretical predictions on the higgs boson mass as well as the intrinsic uncertainities in the electroweak observables prevent to derive stringent predictions for the higgs mass , and literally almost the entire sm higgs boson range ( up to and above @xmath5 1 tev ) must be swept for the higgs search @xcite . from the direct searches ,
four lep experiments put the lower bound of 95 gev for the higgs mass @xcite .
the dominant production mechanism for higgs boson within the lep2 reach is the higgs strahlung process , @xmath6 , in which higgs boson is emitted from a virtual @xmath7 line .
this process has already been analyzed in sm @xcite and mssm @xcite , including the radiative corrections .
one notes that for @xmath8 and @xmath1 in the final state , fusion process where higgs boson is formed in @xmath9 and @xmath10 t - channel collisions will interfere with the higgs- strahlung amplitude . in this work
we analyze nmssm and @xmath0 models comparatively with particular emphasis on their cp even higgs spectra . to have direct effects of these particles at the weak scale one demands the gauge singlet scalar to have a vacuum expectation value ( vev ) around the weak scale .
such a breaking scheme leading relatively light particles ( including the @xmath0 boson itself ) has already been shown to exist in @xmath0 models @xcite .
this occurs when the trilinear coupling of the higgs doublets and the singlet in the soft supersymmetry breaking terms becomes large compared to the soft masses , and leads to appoximately equal vev s for both doublets and the singlet .
this type of vacuum state satisfies the existing constraints on the mixing between the @xmath11 and @xmath0 , and produces a spectrum of scalars at the weak scale .
however , in nmssm such higgs trilinear coupling driven minima , though can exist , are not required by some phenomenological requirement @xcite as in the @xmath0 models @xcite . in @xmath0 models for singlet to have large vev
one needs the singlet mass - squared to be large negative . in nmssm , in addition to this mechanism , one can allow for the singlet trilinear coupling to be large negative to have a large singlet vev .
the opposite limit , namely , the requirement of having both singlet and the doublet vev s to be around the weak scale could be satisfied by choosing the mass - squareds of the higgs fields appropriately , in particular , letting them be sufficiently larger than both singlet cubic coupling and higgs trilinear coupling , in absolute magnitude . however , this requirement can also be met by choosing higgs trilinear coupling large enough compared to soft masses and singlet cubic coupling . in this work
we will perform a comparative study of nmssm and @xmath0 models in the same kind of minimum induced by the relatively large values of the higgs trilinear coupling . in particular , we dicuss the cp even spectra , and investigate their collider signature , both at tree and loop levels , by analyzing the higgs production through the bjorken processes . in sec.2 we derive the couplings and masses of the scalars , and specify all of the relevant properties of nmssm and @xmath0 models .
we also discuss vacuum stability and scalar spectrum against radiative corrections . in sec.3 we give a comparative discussion of the electron - positron annihilation to four - fermion final states through higgs - strahlung in both nmssm and @xmath0 models with reference to the corresponding sm expressions . in sec .
4 we conclude the work .
we consider a supersymmetric model whose scalar sector consists of two higgs doublets @xmath12 , @xmath13 , and a sm - singlet @xmath14 .
nmssm is such a supersymmetric model whose scalar sector is spanned by these fields , and gauge symmetry is exactly that of the mssm .
in fact , it is specified by the superpotential @xmath15 where contributions of all fermions but the top quark are neglected as they are much lighter . here
@xmath16 and @xmath17 are left - handed su(2 ) doublet and singlet quark superfields .
the cubic term in ( 1 ) is necessary to avoid the unwanted peccei - quinn symmetry .
however , superpotential has still a @xmath18 symmetry which , when spontaneously broken , causes serious problems about domain walls @xcite .
this problem can be avoided by the addition of non - renormalizable terms of the form @xmath19 if gravity violates the @xmath18 symmetry @xcite . despite these problems ,
the superpotential ( 1 ) is sufficient to understand the basic implications of nmssm for particle physics applications . besides nmssm , @xmath0 models , which are those models having a low - energy supersymmetric extra u(1 ) factor , have also an extended scalar sector compared to mssm . unlike nmssm ,
these models are devoid of such cosmological problems , and are predicted in string compactifications and @xmath20 gut s @xcite . in @xmath0 models , not only the scalar sector but also the gauge sector is extended by an extra @xmath21 factor with the coupling @xmath22 .
consequently , all fields are charged under this group , and we make the charge assignment @xmath23 , @xmath24 , @xmath25 , @xmath26 and @xmath27 for @xmath12 , @xmath13 , @xmath14 , @xmath16 , and @xmath17 , respectively . in this case , the superpotential is given by @xmath28 where a cubic term is forbidden due to the extra abelian group factor .
formally , one can obtain this superpotential by setting @xmath29 in ( 1 ) .
although the superpotentials ( 1 ) and ( 2 ) formally differ only by the cubic term , resulting scalar potentials reveal explicitly the difference between the two models , through f- and d- terms .
in fact , the most general representation for the scalar potential is given by @xmath30 where the dimensionless @xmath31 coefficients are listed in table i for both models .
in writing this potential we suppressed the contribution of the squarks for simplicity , however , when discussing the loop effects we will explicitly take them into account .
for later use we parametrise the vacuum expectation values of the higgs fields as follows : @xmath32 with real @xmath33 and @xmath34 .
all physical quantities of interest can be expressed in terms of parameters of the potential and these vev s . before specializing to a particular minimum of the potential we derive vector boson - fermion couplings as they are essentially independent of the scalar sector of the model .
since the vector boson sector of the @xmath0 model is larger we analyze it in detail first , and then infer the necessary formulae for nmssm .
there are two massive neutral vector bosons in @xmath0 models , the usual @xmath7 of the sm gauge group , and @xmath0 of the extra @xmath21 group , which mix through the mass - squared matrix : @xmath35 whose entries are given by @xmath36 where @xmath37 . @xmath38 mixing angle , which is one of the most important parameters in @xmath0 models ,
is defined by @xmath39 diagonalization of @xmath40 leads to mass - eigenstates @xmath41 and @xmath42 with masses @xmath43 and @xmath44 , respectively . the @xmath38 mixing angle forms the mere sign of @xmath0 models in lep z - pole data and , in fact , in this way it is constrained to be @xmath45 @xcite
actually the way @xmath46 enters the lep z - pole observables can be seen through the @xmath47 couplings . for a fermion @xmath48 , we define @xmath49 where @xmath50 is the third component of the weak isospin , and @xmath51 is the electric charge .
then coupling of @xmath52 to a fermion line is given by the lagrangean @xmath53 where vector coupling @xmath54 and axial coupling @xmath55 are listed in table ii for @xmath0 models .
one notices that @xmath54 depends on the sum of the extra u(1 ) charges of the left
handed fermion doublet and right handed su(2 ) singlet , that is , @xmath56 . due to the gauge invariance of the superpotential
this sum equals minus the extra u(1 ) charge of the associated higgs doublet .
hence it is not possible to implement the leptophobic @xmath0 models though it is required by the high energy precision data .
nevertheless , this problem can be sidestepped if lepton charges under extra u(1 ) group vanish , and there is no trilinear mass term in the superpotential for leptons , that is , they acquire masses through the non - renormalizable interactions as all light quarks are supposed to do @xcite .
thus we rectrict our attention particularly to the top yukawa coupling given in the superpotential ( 2 ) . unlike the vector couplings , axial couplings depend on the difference between the extra u(1 ) charges of the doublet and su(2)-singlet fermions .
needless to say , in the limit of small mixing angles , @xmath47 couplings approach the corresponding sm ones . in nmssm , one has only the standard @xmath7 boson whose couplings can be obtained by setting @xmath57 , and @xmath58 in table ii .
nmssm couplings can be compared with the results of @xcite .
after obtaining the couplings of vector bosons to fermions , now we turn to the discussion of cp - even higgs - vector boson couplings which are of fundamental importance in discussing the higgs production through the higgs - strahlung type processes .
evaluation of the couplings of higgs bosons to vector bosons requires the minimization of the potential ( 3 ) after which one obtains the physical particle spectrum together with the necessary diagonalizing matrices .
each of the models under concern has its own phenomenological constraints to be satisfied . as discussed in the introduction , in the case of nmssm
, one should prevent the creation of a pseudoscalar goldstone mode , which can easily be satisfied for non - vanishing @xmath59 . in the case of @xmath0 models , however , one has stringent constraints on the possible vacuum state , that is , the @xmath38 mixing angle ( 9 ) should be @xmath60 @xcite , and @xmath0 mass must be @xmath61 600 gev , as required by the recent tevatron direct search with leptonic final states @xcite . as long as one considers the case of leptophobic @xmath0 in accord with the lep z - pole data , this latter condition can be relaxed . however , the former one is a non - trivial condition on the possible vacua of @xmath0 models . in the next section
we shall analyze the relevant vacuum state of the two models in detail .
for the models under concern , to have spectacular effects at lep2 and nlc energies , the breaking scale of the supersymmetry is expected to be around the weak scale . indeed , when the supersymmetry is broken above @xmath5 tev , at the weak scale one ends up with an effective 2hdm @xcite which carries only indirect information about the underlying model .
thus , in what follows we assume supersymmetry to be broken around the weak scale for both models under concern .
this observation requires all three vev s to be around the weak scale . in particular , @xmath62 is prohibited in both models as otherwise sm - singlet @xmath63 may acquire a large vev . in nmssm , in addition to this , one has to prevent latexmath:[$|k_{s}a_{k}| > > @xmath63 through the cubic soft term .
we first discuss @xmath0 models , and following it we turn to the discussion of nmssm . in analyzing the @xmath0 models
the basic quantity of interest is the @xmath65 mixing angle ( 9 ) which has to be small to satisfy the present phenomenological bounds @xcite .
an observation on ( 9 ) reveals that @xmath65 mixing angle @xmath46 can be made small either by choosing @xmath66 , or by forcing @xmath67 itself to be small without constraining @xmath68 .
while the former one requires sm - singlet vev to be much larger than the doublet vev s , the latter allows all three vev s to be of the same order of magnitude , as required by the discussions at the end of the last subsection .
one can realize small @xmath67 when the charges of the higgs doublets under extra u(1 ) are equal , and higgs trilinear coupling @xmath69 is larger than the other mass parameters @xcite .
we name that minimum of the potential ( 3 ) for which @xmath69 is larger than the other mass parameters as higgs trilinear coupling driven minimum ( htcdm ) from now on .
in fact , the potential possesses a htcdm when @xmath70 exceeds the critical point @xmath71 where @xmath72 . as this formula indicates , when @xmath73 , @xmath74 exists , and passage of the potential from small to large trilinear coupling regime is a first order phase transition , namely , all vev s are discontinuous at @xmath74 .
on the other hand , when @xmath75 , there is no critical point at all ; transition is exclusively second order @xcite .
however , independent of the sign of @xmath76 and the order of the transition , in the limit of large @xmath70 , all vev s converge the solution @xmath77 with which one can fix @xmath78 by using the @xmath79 mass .
when the potential ( 3 ) possesses a htcdm , in the basis @xmath80 , re[h_{2}^{0 } ] , re[s^{0}])^{t}$ ] , cp - even higgs mass - squared matrix is given by @xmath81 where @xmath82 and @xmath83 , because @xmath84 as assumed above , when discussing htcdm .
diagonalization of the cp - even higgs mass - squared matrix gives the following mass spectrum @xmath85 in the increasing order .
the physical mass eigenstates @xmath86 are related to the basis vector @xmath80 , re[h_{2}^{0 } ] , re[s^{0}])^{t}$ ] via the diagonalizing matrix @xmath87 having matrix @xmath88 at hand , one can easily compute the coupling strengths of @xmath89 to @xmath90 for @xmath91 and @xmath92 .
in fact , table iii gives a list of these couplings for all possible cases .
as one notices this table has important implications for the mechanism of higgs production through the higgs - strahlung type processes .
first of all , the sum of the squared @xmath93 couplings equals the square of the sm @xmath94 coupling .
thus , just like mssm , mixings in the scalar sector result in a reduction of the coupling strength , implying a smaller production cross section than that of the sm .
next , one observes that some couplings vanish .
while @xmath95 and @xmath96 have only diagonal couplings in the form @xmath97 , @xmath98 has only @xmath99 type coupling .
hence , in a transition of the form @xmath100 , if the initial and final vector bosons are identical then only lightest and heaviest higgs bosons could be radiated off . unlike this , if vector bosons are not identical , then the radiated scalar can only be next - to - lightest higgs . this well defined spectrum of the produced scalars may be important in a particular collider search , say lep2 or nlc .
this completes the discussion of the scalar sector of the @xmath0 models at the tree level .
when we discuss the loop effects we shall reanalyze some quantities derived in this section .
having discussed the scalar sector of the @xmath0 models , we now start analyzing nmssm scalar potential ( 3 ) in reference to table i. as mentioned at the beginning of this section , we require supersymmetry be broken around the weak scale , and thus , higgs vev s are of the same order of magnitude .
restrictions on @xmath0 models do not have any analogue in nmssm , and one is generally free to realize any kind of minimum as long as vev s do not have too big splittings among them .
in fact , all relevant portions of the nmssm parameter space have already been analyzed in @xcite whose results will not be reproduced here . for a comparative and parallel study of the @xmath0 models and nmssm , it would be convenient to discuss the latter one in that portion of the parameter space required by the former as it is severely constrained by the lep z - pole data . thus , differently than @xcite , we shall discuss nmssm also in the limit of large higgs trilinear coupling .
actually , nmssm scalar potential ( 3 ) would possess a htcdm provided that @xmath101 exceeds the critical value @xmath102 where @xmath103 , and @xmath72 .
in nmssm , instead of @xmath104 , one has @xmath101 characterizing the minimum of the potential .
at this point one should bare in mind that @xmath105 coupling @xmath106 singles out @xmath63 and its large values automatically creates @xmath107 , rather than @xmath108 .
consequently , if one wishes to obtain a htcdm for the potential ( 3 ) , @xmath106 must be much less than @xmath104 so that @xmath109 approximately applies to @xmath74 .
this is an approximate statement because one can not ignore @xmath106 completely due to the axion problem mentioned in the introduction . just like the @xmath0 models ,
type of the transition is sensitive to the sign of @xmath76 , however , independent of this , for large enough @xmath104 , all vev s converge to the same value given by @xmath110 which mainly follows @xmath69 since @xmath111 .
an analysis of the @xmath0 models reveals that cp - even higgs mass - squared matrix is highly sensitive to gauge and yukawa couplings as can be seen from ( 14 ) . in nmssm ,
the prescription @xmath112 implies two distinct cases to be analyzed in detail : @xmath113 where we named the two cases as nmssm1 and nmssm2 for later use .
it is convenient to discuss the implications of these two cases seperately .
@xmath114 in this case one can replace @xmath115 by @xmath116 so that @xmath117 . furthermore , neglecting @xmath118 in comparison with @xmath119 , one obtains @xmath120 which , when sufficiently exceeded by @xmath70 , implies the vev s @xmath121 with these vev s , in the basis @xmath80 , re[h_{2}^{0 } ] , re[s^{0}])^{t}$ ] , mass - squared matrix for cp - even scalars turns out to be @xmath122 the diagonalization of which yields higgs spectrum with masses @xmath123 in the increasing order .
the physical mass eigenstates @xmath86 are related to the basis vector @xmath80 , re[h_{2}^{0 } ] , re[s^{0}])^{t}$ ] via the diagonalizing matrix @xmath124 as this diagonalizing matrix shows , the heaviest higgs gets contributions only from @xmath125 $ ] , and the lighter higgs particles get contributions only from the neutral cp - even components of @xmath126 and @xmath127 . in this sense
doublets and the singlet decouple , and produce their higgs spectra .
one recalls that the situation in @xmath0 models was different ; there it was only the next - to - lightest higgs that was independent of @xmath125 $ ] .
this forms a clear distinction between the two models . with @xmath88 matrix at hand , it is easy to compute the strength of the coupling between a @xmath7 line and a cp - even higgs as already listed in the first column of table iv . as the table shows only next - to - lightest higgs is radiated off a @xmath7 line , and the lightest and the heaviest higgs scalars can not be produced . in this way
one concludes that higgs - strahlung type processes can lead to the production of next - to - lightest higgs only .
moreover , in terms of @xmath41 couplings , @xmath0 models and nmssm are complementary to each other .
this completes the discussion of the nmssm1 in terms of its particle spectrum and implications for bjorken production of the higgs particles .
@xmath128 in this case one can replace @xmath118 by @xmath70 and neglect @xmath129 in comparison with @xmath130 so that @xmath131
. then , in exact similarity with the @xmath0 models , one gets @xmath132 which , when sufficiently exceeded by @xmath70 , implies the vev s @xmath133 with these vev s , in the basis @xmath80 , re[h_{2}^{0 } ] , re[s^{0}])^{t}$ ] , mass - squared matrix for cp - even scalars turns out to be @xmath134 the diagonalization of which yields the particle spectrum @xmath135 in the increasing order .
the physical mass eigenstates @xmath86 are related to the basis vector @xmath80 , re[h_{2}^{0 } ] , re[s^{0}])^{t}$ ] via the diagonalizing matrix @xmath136 in comparison with nmssm1 , here one encounters some novel aspects .
though the cp - even higgs mass - squared matrices ( 27 ) and ( 22 ) have the same form their dependences on @xmath137 are not identical because of the fact that @xmath129 plays different roles in two cases
. however , one still expects doublets and the singlet to decouple due to the form of the mass - squared matrix ( 27 ) .
indeed , as the diagonalizing matrix ( 29 ) shows this time it is the next - to - lightest higgs that is a pure singlet state as opposed to nmssm1 where the heaviest higgs was a pure singlet .
one notices that since @xmath138 , @xmath95 and @xmath98 are nearly degenerate in mass .
a comparison of the diagonalizing matrices ( 24 ) and ( 29 ) implies the cyclic interchange @xmath139 , @xmath140 , and @xmath141 .
thus , in nmssm2 @xmath142 is a pure singlet , and the heaviest and the lightest higgs particles get contribution from only the neutral cp - even parts of the higgs doublets . in this case ,
higgs - strahlung type processes support the production of the lightest and the heaviest higgs scalars only . with the diagonalizing matrix ( 29 )
, one can compute the couplings of higgs scalars to a @xmath7 line .
in fact , the second column of table iv indicates these couplings . as we see
, only the lightest higgs @xmath143 can be produced in a higgs - strahlung type process in nmssm2 .
the heavier higgs scalars can not be obtained in such processes .
the discussions above show that , in the limit of large higgs trilinear couplings , both @xmath0 models and nmssm approach a htcdm in which all vacuum expectation values are of the same order .
however , for a proper analysis of the particle spectrum it is necessary to investigate the effects of the radiative corrections on the stability of the htcdm .
effects of the radiative corrections have already been discussed for @xmath0 models in @xcite , and nmssm in @xcite .
the radiative effects of the entire particle spectrum on the higgs potential could be parametrized by using the effective potential approximation . to one - loop accuracy
the effective potential has the usual coleman - weinberg form @xmath144 where @xmath145 is the higgs field dependent mass - matrices of the fields entering the supertrace @xmath146 .
here @xmath147 is the renormalization scale . generally ,
unless supersymmetry is broken below @xmath5 tev the loop effects as well as the direct production of the supersymmetric particles are hard to detect in present and near future experiments .
however , then the logarithms in the coleman - weinberg effective potential may not be satisfactorily large . in spite of this fact , this formula is still satisfactory for large enough yukawa couplings ( see the second reference in @xcite ) . among all supersymmetric particle spectrum especially top quark and
top squarks are important due to the large value of top yukawa coupling , @xmath148 .
then , taking only the dominant stop and top quark contributions into account , among all potential parameters only the following ones get significantly corrected : @xmath149\nonumber\\ \hat{\lambda}_{1s}&=&\lambda_{1s}+\beta_{h_{t } } s_{\tilde{t}q } h_{s}^{2}\\ \hat{\lambda}_{2}&=&\lambda_{2}+\beta_{h_{t}}s_{\tilde{t}t}h_{t}^{2}\nonumber\end{aligned}\ ] ] with @xmath150 where the last two quantities represent the splitting between stops and the scale @xmath147 , and top quark mass , respectively .
@xmath151 is the top trilinear coupling coming from the superpotential ( 1 ) or ( 2 ) , and the the remaining mass parameter @xmath152 is the sum of the soft mass - squareds of top squarks ; @xmath153 . in deriving these one - loop corrections
we assumed @xmath154 in accordance with the fcnc constraints @xcite , and expanded the stationarity conditions in powers of stop splitting @xmath155 .
detailed expressions for stop mass - squared matrix can be found in @xcite .
one recalls from the discussions of the tree - level potential that the existence of the higgs trilinear coupling driven minimum of the potential can be characterized by the threshold value @xmath74 of @xmath119 .
this threshold value of @xmath119 is highly precise if the sum of the soft mass - squareds of the higgs fields are positive , and irrespective of the order of the transition potential possesses a htcdm if @xmath119 is sufficiently large compared to @xmath74 .
thus , it is convenient to analyze the effects of the radiative corrections on the vacuum structure starting with a htcdm at the tree - level .
if @xmath156 is the critical value of @xmath119 in the presence of the radiative corrections , one has @xmath157 where @xmath158 represents the effects of the radiative corrections , and can be expressed solely in terms of the higgs- and top - trilinear couplings and the yukawa couplings : @xmath159\nonumber\\&+ & \frac{1}{18 } \beta_{h_{t}}(1-s_{\tilde{t}q})\big[16 \left(\frac{a_{t}^{crit}}{a_{s}^{crit}}\right)^{2}+3 \left(\frac{a_{t}}{a_{s}^{crit}}\right)^{2}\big]\nonumber\\&- & \frac{1}{3}\beta_{h_{t } } s_{\tilde{t}t}\left(\frac{h_{t}}{h_{s}}\right)^{2 } \end{aligned}\ ] ] where , following the discussion of ccb minima in mssm in @xcite , we introduced the critical value of @xmath151 via the relation @xmath160 which represents the threshold value of @xmath151 for which color and/or charge breaking just starts taking place , for a given @xmath161 . in these formuale , @xmath162 and @xmath163
if the vacuum state under concern is a non - ccb htcdm .
the exact value of @xmath158 requires a full specification of the scalar sector of the theory including the squark trilinear mass terms in superpotentials ( 1 ) or ( 2 ) .
the criterium induced by @xmath158 has a severe dependence on its sign : if @xmath164 , @xmath165 , and thus tree - level htcdm might disappear if the actual value of @xmath119 were close to @xmath74 at the tree level . in the opposite case of @xmath166 , @xmath167 and
htcdm is supported by the radiative corrections .
since the higgs vev s are of the same order , satisfying the cdf value of the top - quark requires @xmath148 , so that effects of the radiative corrections are no way negligable .
moreover , for the reasonable set of parameters , @xmath168 and @xmath169 , @xmath158 turns out to be negative so that radiative corrections prefer @xmath165 which can destabilize the vacuum state unless @xmath170 at the tree level .
radiative corrections modify mass matrices of all scalars , in particular that of the cp - even higgs scalars by addition of the following matrix @xmath171 which is obtained by assuming @xmath172 , and a small stop splitting .
in fact , one can write @xmath173 using the definition of @xmath174 ( 35 ) .
as @xmath175 is the natural renormalization scale , this approximation is reasonable .
the entries of the tree - level mass matrices ( 14 ) of @xmath0 models , and ( 22 ) and ( 27 ) of nmssm are now modified with the addition of ( 36 ) .
let us note an important difference between @xmath0 models and nmssm in terms of the behaviour of their mass matrices under radiative corrections . in @xmath0 models , ( 14 ) , with its all elements are non - vanishing , gets non - vanishing contributions to each of its elements through the radiative corrections , preserving its form . on the other hand , in nmssm , tree - level mass matrices , ( 22 ) and ( 27 ) , already have some of their elements vanishing so that there is no mixing between the doublet and singlet contributions .
however , with the addition of the radiative corrections ( 36 ) , this simple tree - level picture is destroyed , and now singlet and doublet sectors do mix . even in the limit of small @xmath151 , @xmath176re[s^{0}]$ ]
type mixing can not be avoided .
if @xmath70 is sufficiently large compared to the radiatively corrected critical point @xmath177 ( 33 ) then potential ( 3 ) will definitely have a htcdm in which the vev s are proportional to @xmath70 and their values are eventually fixed by the fermi scale @xmath178 .
thus one can safely take vev s equal in ( 36 ) .
factoring out @xmath179 from ( 36 ) , one observes that @xmath180 and @xmath181 terms contribute by @xmath182 and @xmath183 , respectively .
thus , especially the top quark contribution is important .
although one can analytically obtain the effects of ( 36 ) on the mass spectrum and couplings , the results will be algebraically involved ; and thus , we will numerically analyze the consequences of radiative corrections in the next section .
the search for the higgs particles is one of the most important issues at lep2 and nlc .
though two - fermion processes are still of interest , the true novelty at high energy @xmath1 colliders will come from four - fermion processes , among which the @xmath184-channel higgs - strahlung process @xmath6 is the most important one for higgs production .
in fact , other processes in which higgs is formed in @xmath9 and @xmath10 @xmath185- channel collisions have smaller cross sections at lep2 energies , and their interference with the higgs- strahlung type processes could be avoided by preventing @xmath186 and @xmath187 from the final products @xcite .
above all , we should analyze four - fermion ( 4f ) processes because intermediate state higgs particles show up as resonances when their mass and the invariant mass flow to that channel coincide . in the much simpler process @xmath6 one can not trigger the multi - higgs structure of the models under concern as the products in a specific collision process are fixed . with these constraints in mind , we analyze the following four - fermion process @xmath188 where , depending on the @xmath1 cm energy @xmath189 , vector bosons and higgs bosons above may come to the physical shell .
computation of the cross section for 4f processes like this is a highly complicated problem due to the phase space integration @xcite , and one mostly resorts to numerical techniques @xcite . despite this , however , we can extract the necessary information about the higgs structure of the underlying models without performing a full calculation if we take the ratio of the differential cross section to that of the sm . before going into the details of such a calculation
it is convenient to describe the properties of @xmath0 models due to its complicated higgs - vector boson sector . for @xmath0 models , in htcdm ,
@xmath38 mixing angle ( 9 ) is vanishingly small , and we assume it remains small also when radiative corrections are included . assuming further @xmath0 be leptophobic , it is seen that @xmath190 does not contribute to ( 37 ) , so an analysis of ( 37 ) with @xmath191 is sufficient . to calculate the total cross section @xmath192 one needs to integrate over the phase space of the final particles , which is not possible by hand @xcite . moreover , in the case of polarized electron - positron beams task will be much more complicated .
the amplitude for the scattering process ( 37 ) in the models under concern differs from that in the sm by the multitude of higgs scalars reflecting them in @xmath193 and @xmath194 couplings .
this observation greatly simplifies the analysis if one integrates over first @xmath195 and @xmath196 phase spaces using the momentum conservation , and identifies the remaining task to be done with that of the sm process .
then the ratio of the resulting cross section to that of the sm depends only on the invariant mass @xmath197 of the @xmath198 system : @xmath199 where @xmath200 are @xmath193 couplings , listed in tables iii and iv , in units of corresponding sm coupling @xmath201 .
higgs - fermion couplings are parametrized by @xmath202 , which can be 1 and 2 for down- and up - type fermions , respectively . furthermore ,
@xmath203 and @xmath204 designate the total widths of @xmath89 and @xmath3 , respectively .
@xmath197 , invariant mass flowing into the higgs branch , has the kinematical range of @xmath205 to @xmath206 . as @xmath197 varies in this range @xmath207
will be sharply peaked at each @xmath208 , as long as @xmath208 is kinematically accessible .
finally , one notices that @xmath207 is independent of the beam polarizations . for the compactness of the presentation
, we introduce the dimensionless parameter @xmath209 which equals unity when the sm higgs mass resonance is encountered . in what follows ,
we take @xmath210 equal to the lightest higgs boson mass in the particular model under concern . in this sense , at @xmath211 lightest higgs resonance occurs . as @xmath212 gets higher and higher values other higgs scalars with non - vanishing couplings
will be excited in the order of increasing mass . for trilinear higgs yukawa coupling @xmath130 , we assume @xmath213 .
actually , @xmath130 can be choosen as low as @xmath214 at which the present lep lower bound of the lightest higgs mass is exceeded at the tree level . for @xmath169 ,
one gets @xmath215 , which is a reasonable scale .
on the other hand , for @xmath22 we assume the usual gut constraint of @xmath216 .
the @xmath105 yukawa coupling @xmath129 is either @xmath217 ( nmssm1 ) , or @xmath218 ( nmssm2 ) . for @xmath219 , as is necessary for obtaining htcdm , @xmath118 dependence of the masses and the couplings cancel , as was illustrated before .
in addition to these , we take @xmath220 , so that @xmath151 dependence of the radiative corrections to the cp - even higgs mass - squared matrix ( 36 ) can be neglected .
finally , we approximate @xmath204 by @xmath221 , and assume all scalars have the same width . this last assumption is falsified especially when the higgs mass exceeds @xmath222 , but it does not affect the main conclusion of the work because the width of a particular resonance is not of central concern . at the final state we take @xmath223 and @xmath224 , which is convenient for detection purposes at lep2 @xcite . with the above mentioned values of parameters , one can calculate all couplings and masses .
in fact , tree- and loop- level values of the lightest higgs mass are tabulated in table v , for @xmath0 models , and nmssm for both cases in ( 19 ) .
an analysis of @xmath225 ( 38 ) reveals that it is necessary to have @xmath226 gev for all higgs scalars be excited in both models .
this value of @xmath189 is pretty much above the maximum value of @xmath227 gev aimed at lep2 , and thus one is to wait for nlc operation for the experimental realization of these scalar spectra , if any . in fig .
1 we show @xmath225 for @xmath0 models , at @xmath228 gev , with and without the radiative corrections . here
dashed curve represents the tree - level analysis , and as expected there is a single resonance curve at @xmath229 corresponding to the heaviest higgs whose coupling is given in table ii . in this graph , the full curve shows @xmath225 when radiative corrections are included . under radiative corrections none of the couplings remains vanishing , and therefore , effects of the scalars which were inert at the tree level show up . in this sense
, the first resonance curve at @xmath230 corresponds to the next - to - lightest higgs which was absent in the tree- level @xmath225 . that this resonance is much narrower than that of the heaviest higgs located at @xmath231 is caused by the smallness of @xmath232 .
both tree- and loop - level @xmath225 has a non - negligable value at @xmath211 because of the fact that lightest higgs is observable at the tree - level and has still a large enough coupling at the loop level . in fig .
2 we present @xmath225 for nmssm1 described in equation ( 19 ) . at the tree - level , in accordance with table iv , only next - to - lightest higgs is observable as is evidenced by the resonance at @xmath233 .
both dashed and full curves have vanishingly small values at @xmath234 due to the fact that @xmath235 , which vanishes at the tree - level , is still small compared to @xmath232 .
when radiative corrections are included place of the next - to - lightest higgs resonance practically remains the same , and there is a tiny resonance curve at @xmath236 representing the heaviest higgs contribution .
such a narrow resonance is caused by the smallness of @xmath237 compared to @xmath232 . in fig .
3 @xmath225 for nmssm2 is depicted . at the tree - level , in accordance with table iv , only the lightest higgs observable so that the dashed line @xmath238 arises .
differently than the models analyzed in figs .
1 and 2 in this model radiative corrections cause important modifications in the spectrum .
this is mainly caused by the smallness of the elements of the mass - squared matrix compared to the two cases discussed above which make them more sensitive to the radiative corrections .
consequently , when the radiative corrections are included next - to - lightest higgs is seen to have dominant couplings compared to others , so that aymptotics of the full curve are entirely determined by @xmath98 .
in fact , @xmath225 has vanishingly small values at @xmath234 and @xmath239 due to this reason .
the resonance curves of @xmath98 and @xmath96 are close to each other , narrower than tree - level ones of the previous cases , and located at @xmath240 and @xmath241 , respectively .
the analysis carried out in this work primarily concentrates on the similarities and dissimilarities between the @xmath0 models and nmssm concerning their neutral higgs scalars .
the two models have been comparatively analyzed in that minimum of the potential characterized by the large values of trilinear soft mass coupling the sm singlet and the higgs doublets .
the issue of radiative corrections deals essentially with the stability of this minimum together with the corresponding higgs mass squared matrix .
the radiative corrections are conveniently parametrized using the effective potential approximation . as a case study we analyzed higgs production via bjorken process at an @xmath1 collider .
both models are comparatively analyzed referring basically to the variation of the cross section with the di - fermion invariant mass .
once a high energy collider is given there appear several scattering processes whose signature necessarily depend on the details of the underlying model .
this becomes especially clear after a comparative study of the sm and mssm , for example , at @xmath1 linear colliders @xcite . on the other hand , characteristic to all supersymmetric models ,
there are sfermions , charginos , neutralinos , and fermions in addition to the higgs particles whose distinguishability needs be studied .
the two minimal extensions of the mssm , that is , @xmath0 models and nmssm , are already already distinguishable from the mssm and 2hdm in terms of their particle spectrum
. moreover , there are also differences between @xmath0 models and nmssm as they have different number of neutral gauge bosons ( two for the former , one for the latter ) , psedoscalars ( one for the former , two for the latter ) , and neutral gauginos ( three for the former , two for the latter ) .
therefore , for processes with such signature , it can be easy to know what model is the underlying one .
for example , the neutralino pair production , @xmath242 or associated higgs production @xmath243 ( @xmath244 is a pseudoscalar ) are candidate processes . despite such differences , however , the two models have equal number of cp even higgs particles .
the primary goal of the future colliders is the discovery of the higgs particle whose most likely production mode is the bjorken process . that this process does not involve the pair production of the supersymmetric particles is especially appealing as it could be realized for moderate collider energies
due to the reasons mentioned above , a discussion of the distinguishability of the cp even higgs scalars of the @xmath0 models and nmssm is an important issue by itself .
we have concentrated here on htcdm ( where @xmath0 models are likely to have for satisfying the lep constraints ) in which the higgs potentials of the two model are similar apart from their respective parameters .
the results of the analysis are presented in the tabels and figures for both tree level and one loop higgs potentials .
nmssm , having no gauge extension at all , is particularly interesting in htcdm .
for example , a glance at table iv shows that these couplings are identical to those of the minimal model .
however , one
loop @xmath245 couplings for both nmssm1 and nmssm2 destroy this tree
level picture by causing other coupligs to have non vanishing values .
then , as suggested by figs . 2 and 3 there
appear new resonances allowing , eventually , for distinguishability .
this examplifying case particularly shows how important the radiative corrections are . at the tree level
, there is a certain hierarchy , in particular , some couplings vanish identically as shown in tables 3 and 4 .
however , the radiative corrections are strong enough to elevate other couplings to non - vanishing values .
this is particularly observed in the figures ( solid curves ) where there arise additional resonances compared to the tree
level results ( dashed curve ) corresponding to additional excited higgs bosons due to their raditively induced couplings . in sec .
iii we have considered only @xmath246 fermion processes in illustrating higgs search via the bjorken mechanism . the actual experimental search strategy @xcite is to look for @xmath247 and @xmath223 in eq . (
37 ) , which kills down the possibility of having a pseudoscalar boson coupling to @xmath248 pair .
such an event selection mechanism necessarly establishes the existence of a scalar or set of scalars as the collider energy increases .
however , when the collider energy gets higher and higher , pairs of lightest neutralinos and charginos will be produced in association with the higgs . in this case @xmath7 and @xmath244 couplings to @xmath248 pair compete for @xmath169 as has been assumed throughout the work . in spite of all such supersymmetric particle production modes occuring in association with a cp even higgs particle , the main search mecanishm for the cp
even higgs remains to be the bjorken process with @xmath247 and @xmath223 @xcite . as a final point
, it is obvious that @xmath207 described in the figures is not a directly measurable quantity .
however , it is highly useful in distinguishing the two models in terms of the di - fermion invariant mass .
indeed , @xmath207 could be a useful tool in monte carlo simulations of these models in the next linear collider @xcite .
both the simulation studies ( with more sophistcated numerical techniques ) and analysis of the experimental data can be guided by the results of this analysis .
the author is grateful to goran senjanovic for highly stimulating discussions and helpful comments .
he also acknowledges the discussions with alejandra melfo .
normalized cross section differential @xmath225 for @xmath0 models at tree- ( dotted curve ) and one - loop ( full curve ) levels
2 . same as in fig . 1 , but for nmssm1 .
3 . same as in fig . 1 , but for nmssm2 | two well - known extended supersymmetric models , @xmath0 models and nmssm , are comparatively analyzed in the limit of large trilinear higgs couplings .
the two models are found to have distinguishable higgs spectra at both tree- and loop- levels .
higgs production through bjorken processes at an @xmath1 collider is shown to discriminate between the two models . |
classical chaos is related to the extreme sensitivity of the dynamics of a system to its initial conditions , a concept that can be traced back to poincar.@xcite the main features of classical chaos can be illustrated by a dynamical billiard , which is an idealized billiard table with no friction where a particle reflects elastically from boundaries which can have any shape .
the motion of the particle is represented in phase space by a trajectory whose evolution is restricted to a surface of constant energy . depending on the shape of the boundaries , the system may be chaotic , which means that two trajectories whose initial conditions are very close will diverge exponentially in time .
the rate of this separation is characterized by the lyapunov exponent.@xcite the trajectories may also become ergodic , which implies that after a long time the particle will have visited the entire surface of constant energy .
equivalently , we may say that after a long time , the particle is equally likely to be found in any point of the accessible phase space . for quantum systems
the notion of phase - space trajectories loses its meaning as can be seen from the heisenberg uncertainty principle .
nevertheless , because classical physics is a limit of quantum physics , it is natural to search for quantum signatures of classical chaos .
the relation between quantum mechanics and classical chaos is the subject of quantum chaos.@xcite quantum chaos is a very broad field .
our intention is not to review the entire field , but to focus on a particular aspect of it , namely , the spectral statistics and structures of eigenstates of quantum many - body systems as they transition from the regular to the chaotic domain . as a preparation for our goal , we first give a brief idea of some of the important topics of the field ( more details are in refs . and in the references
to be cited ) .
quantum chaos encompasses semiclassical theories aiming at establishing direct links between the trajectories of classically chaotic systems and the properties of these systems in the quantum domain .
the main advances in this area were achieved by periodic - orbit theory , which provides a way to calculate the spectrum of a quantum system from its classical periodic orbits ( trajectories that repeat themselves after a certain time ) .
an excellent introduction to the subject was given by gutzwiller.@xcite gutzwiller s method allowed for the development of a theory of `` scars,''@xcite which refer to the structure of eigenstates that concentrate along the classical periodic orbits of chaotic systems .
scars were systematically studied by heller,@xcite and the first experimental observations occurred in quantum billiards . @xcite quantum billiards obey the laws of quantum mechanics and correspond to miniature versions of dynamical billiards . another important step in the understanding of quantum chaos came from the verification that the distribution of the spacings between neighboring energy levels of a quantum billiard depend on the billiard s classical counterpart.@xcite if the latter is chaotic , the energy levels are highly correlated and repel each other ; if it is regular ( integrable ) , the energy levels are uncorrelated , randomly distributed , and can cross .
level repulsion is one of the main features of quantum chaos .
it has been observed in other quantum systems , such as atoms in strong magnetic fields @xcite and systems of coupled particles,@xcite .
interestingly , in the context of number theory , level repulsion has been associated also with distributions of prime numbers.@xcite in this article we focus on quantum many - body systems . a common approach when dealing with such systems
is to ignore the details of the interactions and treat them statistically with random matrices .
the idea is that when the interactions are strong and the behavior of the system is sufficiently complex , generic properties should emerge .
this approach was taken by wigner @xcite to describe the spectrum of heavy nuclei .
he employed matrices with random elements whose only constraint was to satisfy the symmetries of the system .
the level spacing distributions of these matrices showed level repulsion and agreed surprisingly well with the data from actual nuclei spectra . when level repulsion was later verified in billiards , a connection between quantum chaos and random matrices became established .
soon after the introduction of random matrices in nuclear physics , they were employed in the analysis of the spectrum of other quantum many - body systems , such as atoms , molecules , and quantum dots.@xcite the application of random matrix theory is not restricted to the statistics of eigenvalues , but accommodates also studies of eigenstates .
eigenstates of random matrices are pseudo - random vectors ; that is , their amplitudes are random variables.@xcite all the eigenstates are statistically similar , they spread through all basis vectors with no preferences and are therefore ergodic . despite the success of random matrix theory in describing spectral statistical properties
, it can not capture the details of real quantum many - body systems .
the fact that random matrices are completely filled with statistically independent elements implies infinite - range interactions and the simultaneous interaction of many particles .
real systems have few - body ( most commonly only two - body ) interactions which are usually finite range . a better picture of systems with finite - range interactions is provided by banded random matrices , which were also studied by wigner . @xcite
their off - diagonal elements are random and statistically independent , but are non - vanishing only up to a fixed distance from the diagonal .
there are also ensembles of random matrices that take into account the restriction to few body interactions , so that only the elements associated with those interactions are nonzero ; an example is the two - body - random - ensemble@xcite ( see reviews in refs . ) .
other models which describe systems with short - range and few - body interactions do not include random elements , such as nuclear shell models,@xcite and the systems of interacting spins which we consider in this article .
all the matrices we have mentioned can lead to level repulsion , but differences are observed .
for instance , eigenstates of random matrices are completely spread ( delocalized ) in any basis , whereas the eigenstates of systems with few - body interactions delocalize only in the middle of the spectrum.@xcite in this paper we study a one - dimensional system of interacting spins 1/2 .
the system involves only nearest - neighbor interactions , and in some cases , also next - nearest - neighbor interactions . depending on the strength of the couplings , the system may develop chaos , which is identified by calculating the level spacing distribution .
we also compare the level of delocalization of the eigenstates in the integrable and chaotic domains .
it is significantly larger in the latter case , where the most delocalized states are found in the middle of the spectrum .
the paper is organized as follows .
section ii provides a detailed description of the hamiltonian of a spin 1/2 chain .
section iii explains how to compute the level spacing distribution and how to quantify the level of delocalization of the eigenstates .
section iv shows how the mixing of symmetries may erase level repulsion even when the system is chaotic .
final remarks are given in sec .
-0.1 cm we study a one - dimensional spin 1/2 system ( a spin 1/2 chain ) described by the hamiltonian [ hamall ] @xmath0.\label{ham3}\end{aligned}\ ] ] we have set @xmath1 equal to 1 , @xmath2 is the number of sites , @xmath3 are the spin operators at site @xmath4 , and @xmath5 are the pauli matrices . the term @xmath6 gives the zeeman splitting of each spin @xmath4 , as determined by a static magnetic field in the @xmath7 direction .
all sites are assumed to have the same energy splitting @xmath8 , except a single site @xmath9 , whose energy splitting @xmath10 is caused by a magnetic field slightly larger than the field applied on the other sites .
this site is referred to as a defect .
a spin in the positive @xmath7 direction ( up ) is indicated by @xmath11 or by the vector @xmath12 ; a spin in the negative @xmath7 direction ( down ) is represented by @xmath13 or @xmath14 .
an up spin on site @xmath4 has energy @xmath15 , and a down spin has energy @xmath16 .
a spin up corresponds to an excitation .
the second term , @xmath17 , is known as the xxz hamiltonian .
it describes the couplings between nearest - neighbor ( nn ) spins ; @xmath18 is the strength of the flip - flop term @xmath19 , and @xmath20 is the strength of the ising interaction @xmath21 .
the flip - flop term exchanges the position of neighboring up and down spins according to @xmath22 or , equivalently , it moves the excitations through the chain .
we have assumed open boundary conditions as indicated by the sum in @xmath17 which goes from @xmath23 to @xmath24 .
hence , an excitation in site 1 ( or @xmath2 ) can move only to site 2 ( or to site @xmath24 ) .
closed boundary conditions , where an excitation in site 1 can move also to site @xmath2 ( and vice - versa ) are mentioned briefly in sec .
the ising interaction implies that pairs of parallel spins have higher energy than pairs of anti - parallel spins , that is , @xmath25 and @xmath26 for the chain described by eqs .
( [ hamall ] ) the total spin in the @xmath7 direction , @xmath27 , is conserved , that is , @xmath28=0 $ ] .
this condition means that the total number of excitations is fixed ; the hamiltonian can not create or annihilate excitations , it can only move them through the chain .
to write the hamiltonian in matrix form and diagonalize it to find its eigenvalues and eigenstates , we need to choose a basis .
the natural choice corresponds to arrays of up and down spins in the @xmath7 direction , as in eqs .
( [ flipflop ] ) , ( [ upup ] ) and ( [ updown ] ) .
we refer to it as the site basis . in this basis , @xmath6 and the ising interaction contribute to the diagonal elements of the matrix , and the flip - flop term leads to the off - diagonal elements . in the absence of the ising interaction ,
the excitations move freely through the chain . in this case
the eigenvalues and eigenstates can be found analytically . the existence of an analytical method to find the spectrum of a system guarantees its integrability .
the addition of the ising interaction eventually leads to the onset of quantum chaos .
the source of chaos is the interplay between the ising interaction and the defect.@xcite to bring the system to the chaotic regime , we set @xmath29 ( arbitrary units ) , choose @xmath30 ( arbitrary units ) , and place the defect on site @xmath31 , where @xmath32 stands for the largest integer less than or equal to @xmath33 .
these choices are based on the following factors .
( a ) the strength of the ising interaction can not be much larger than @xmath18 , because if it were , basis vectors with different numbers of pairs of parallel spins would have very different energies and @xmath18 would not be able to effectively couple them .
for example , the energy difference between @xmath34 and @xmath35 is @xmath36 ; if we had @xmath37 , the matrix element @xmath38 coupling these two basis vectors would become ineffective . as a result
, the eigenstates would involve only a small portion of the basis vectors , which would mean localized eigenstates and therefore non - chaotic systems .
( b ) the defect can not be placed on the edges of the chain , because it has been shown @xcite that in this case an analytical solution exists and the system is therefore still integrable .
( c ) we can not have @xmath39 , because this would break the chain in two ; that is , an excitation on one side of the chain would not have enough energy to overcome the defect and reach the other side of the chain . in effect
, we would be dealing with two independent chains described by the xxz model , which is integrable.@xcite
we use the level spacing distribution to identify when the system becomes chaotic .
we analyze also what happens to the structure of the eigenstates once level repulsion occurs .
the distribution @xmath40 of the spacings , @xmath41 , of neighboring energy levels differs depending on the regime of the system.@xcite the energy levels of integrable systems are not correlated , and are not prohibited from crossing , so the distribution is poissonian ( p ) , @xmath42 in chaotic systems the eigenvalues become correlated and crossings are avoided .
there is level repulsion , and @xmath40 is given by the wigner - dyson ( wd ) distribution , as predicted by random matrix theory .
the form of the wigner - dyson distribution depends on the symmetry properties of the hamiltonian .
systems with time reversal invariance are described by a gaussian orthogonal ensemble , which corresponds to an ensemble of real symmetric matrices , whose elements @xmath43 are independent random numbers chosen from a gaussian distribution .
the average of the elements and the variance satisfy @xmath44 and @xmath45 .
the level spacing distribution of a gaussian orthogonal ensemble is given by @xmath46 the hamiltonian in eqs .
( [ hamall ] ) is also real and symmetric , so the distribution achieved in the chaotic limit is the same as @xmath47 .
however , our system has only short - range - two - body interactions , so it can not reproduce all features of the gaussian orthogonal ensemble . the density of states , for instance , is gaussian for eqs .
( [ hamall ] ) , but for the gaussian orthogonal ensemble it has the shape of half of a circle .
interested readers can obtain these forms using the codes provided.@xcite another difference is the structure of the eigenstates , as discussed in sec .
[ princomp ] . for gaussian orthogonal ensembles ,
all the eigenstates are statistically similar and highly delocalized , whereas for the hamiltonian in eqs .
( [ hamall ] ) , delocalization is restricted to the middle of the spectrum . to obtain the level spacing distribution
, we first need to separate the eigenvalues according to their symmetry sectors ( subspaces ) .
if we mix eigenvalues from different symmetry sectors , we can not achieve a wigner - dyson distribution even if the system is chaotic because eigenvalues from different subspaces are independent and therefore uncorrelated , so do not repel each other .
we discuss further the danger of mixing eigenvalues from different symmetry sectors in sec .
[ symm ] . for the moment
, we need to remember only that the hamiltonian in eqs .
( [ hamall ] ) conserves @xmath48 .
this symmetry implies that the hamiltonian matrix is separated into uncoupled blocks , corresponding to subspaces with a fixed number of spins in the up direction . in the following ,
we select a particular subspace with @xmath49 up spins , whose hamiltonian matrix has therefore dimension @xmath50 $ ] .
the second essential step before computing the spacing distribution is to unfold the spectrum .
the procedure consists of locally rescaling the energies , so that the local density of states of the renormalized eigenvalues is 1 .
this rescaling allows for the comparison of spectra obtained for different parameters and for different systems .
there are different ways to unfold the spectrum .
a simple and commonly used procedure is to order the spectrum in increasing values of energy ; separate it into several smaller sets of eigenvalues ; and divide each eigenvalue by the mean level spacing of its particular set .
the mean level spacing of the new set of renormalized energies becomes 1 .
because the density of states is the number of states in an interval of energy , that is , the reciprocal of the mean level spacing , this procedure also ensures that the local density of states is unity .
this procedure is the one we used.@xcite given the unfolded spacings of neighboring levels , the histogram can now be computed . to compare it with the theoretical curves , the distribution needs to be normalized , so that its total area is equal to 1 .
figure [ fig : ps ] shows the level spacing distribution when the defect is placed on site 1 and on site @xmath51 .
the first case corresponds to an integrable model and the distribution is a poisson ; the second case is a chaotic system , so the distribution is wigner - dyson .
we now investigate how the transition from a poisson to a wigner - dyson distribution affects the structure of the eigenstates .
in particular , we study how delocalized they are in both regimes . to determine the spreading of the eigenstates in a particular basis , we look at their components .
consider an eigenstate @xmath52 written in the basis vectors @xmath53 as @xmath54 .
it will be localized if it has the participation of few basis vectors , that is , if a few @xmath55 make significant contributions .
it will be delocalized if many @xmath55 participate with similar values . to quantify this criterion
, we use the sum of the square of the probabilities , @xmath56 ( the sum of the probabilities would not be a good choice , because normalization implies @xmath57 ) , and define the number of principal components of eigenstate @xmath4 as@xcite @xmath58 the number of principal components gives the number of basis vectors which contribute to each eigenstate . it is small when the state is localized and large when the state is delocalized . for gaussian orthogonal ensembles , the eigenstates are random vectors , that is , the amplitudes @xmath59 are independent random variables .
these states are completely delocalized .
complete delocalization does not mean , however , that the number of principal components is equal to @xmath60 . because the weights @xmath55 fluctuate , the average over the ensemble gives @xmath61.@xcite to study the number of principal components for eqs .
( [ hamall ] ) , we need to choose a basis .
this choice depends on the question we want to address .
we consider two bases , the site- and mean - field basis .
the site - basis is appropriate when analyzing the spatial delocalization of the system .
to separate regular from chaotic behavior , a more appropriate basis consists of the eigenstates of the integrable limit of the model , which is known as the mean - field basis.@xcite in our case the integrable limit corresponds to eqs .
( [ hamall ] ) with @xmath62 , @xmath63 , and @xmath64 .
we start by writing the hamiltonian in the site - basis .
let us denote these basis vectors by @xmath65 . in the absence of the ising interaction
, the diagonalization of the hamiltonian leads to the mean - field basis vectors .
they are given by @xmath66 .
the diagonalization of the complete matrix , including the ising interaction , gives the eigenstates in the site - basis , @xmath67 .
if we use the relation between @xmath65 and @xmath68 , we may also write the eigenstates of the total hamiltonian in eqs .
( [ hamall ] ) in the mean - field basis as @xmath69 shows the number of principal components for the eigenstates in the site - basis [ ( a ) , ( b ) ] and in the mean - field basis [ ( c ) , ( d ) ] for the cases where the defect is placed on site 1 [ ( a ) , ( c ) ] and on site @xmath51 [ ( b ) , ( d ) ] .
the level of delocalization increases significantly in the chaotic regime .
however , contrary to random matrices , the largest values are restricted to the middle of the spectrum , the states at the edges being more localized .
this property is a consequence of the gaussian shape of the density of states of systems with two - body interactions .
the highest concentration of states appears in the middle of the spectrum , where the strong mixing of states can occur leading to widely distributed eigenstates . ) versus energy ; @xmath70 , 6 spins up , @xmath71 , @xmath72 , @xmath29 , and @xmath73 ( arbitrary units ) .
( a ) and ( b ) site - basis ; ( c ) and ( d ) mean - field basis ; ( a ) and ( c ) defect on site @xmath74 ; ( b ) and ( d ) defect on site @xmath75.,scaledwidth=40.0% ] an interesting difference between the integrable and chaotic regimes is the fluctuations of the number of principal components . for the regular system
the number of principal components shows large fluctuations .
in contrast , in the chaotic regime the number of principal components approaches a smooth function of energy .
chaotic eigenstates close in energy have similar structures and consequently similar values of the number of principal components .
the presence of a defect breaks symmetries of the system . in this section
we remove the defect and have a closer look at the symmetries .
we refer to the system in the absence of a defect ( @xmath76 ) as defect - free .
contrary to the case where @xmath77 , a defect - free spin-1/2 chain with nn couplings remains integrable even when the ising interaction is added .
this system can be analytically solved using the bethe ansatz.@xcite to drive the system to chaos , while keeping it defect free , we need to add further couplings . by considering couplings between next - nearest - neighbors ( nnns),@xcite the hamiltonian becomes @xmath78 where @xmath79.\ ] ] for sufficiently large @xmath80 ( @xmath81 for @xmath82 ) , there are various scenarios for which chaos can develop , which include the absence of ising interactions , @xmath83 ; the absence of the flip - flop term between next - nearest - neighbors , @xmath84 ; the absence of ising interaction between next - nearest - neighbors , @xmath85 ; and the presence of all four terms . depending on the parameters in eq .
( [ hamnnn ] ) , we might not obtain a wigner - dyson distribution even if the system is chaotic if not all symmetries of the system are taken into account.@xcite we have mentioned conservation of total spin in the @xmath7 direction . in the absence of a defect other symmetries of @xmath86 in eq .
( [ hamnnn ] ) include the following.@xcite _ parity_. parity may be understood by imagining a mirror in one edge of the chain . for eigenstates written in the site - basis , the probability of each basis vector is equal to that of its reflection .
for example , suppose we have @xmath87 and one excitation .
the eigenstates are given by @xmath88 .
the amplitudes are either @xmath89 and @xmath90 for even parity or @xmath91 and @xmath92 for odd parity . the level spacing distribution needs to be independently obtained for each parity .
_ spin reversal_. if the chain has an even number of sites and @xmath33 up spins , then @xmath93 . in this sector pairs of equivalent
basis vectors correspond to those which become equal if we rotate all the spins from one vector by 180@xmath94 .
for example , state @xmath95 pairs with state @xmath96 .
_ total spin_. if the system is isotropic , that is , @xmath97 and @xmath98 , the total spin , @xmath99 , is conserved . for closed boundary conditions
there is also momentum conservation .
the more symmetries the system has , the smaller the subspaces become for a given system size , which is not good for statistics .
for this reason we chose open boundary conditions . ) with @xmath100 , @xmath29 ; @xmath101 .
( a ) @xmath102 , 7 spins up , @xmath103 .
all eigenvalues of the subspace @xmath93 are considered .
( b ) @xmath104 , @xmath105 .
( c ) @xmath104 , @xmath106 .
( d ) @xmath84 , @xmath105 .
( e ) @xmath104 , @xmath107 , @xmath85 . for
( b)(e ) @xmath82 , 5 spins up .
the eigenvalues are separated according to the parity of the corresponding eigenstates .
@xmath40 is the average of the distributions of the two parity sectors.,scaledwidth=37.0% ] in fig .
[ fig : symmetry ] we show the level spacing distribution for the four chaotic systems we have described .
all the figures involve eigenvalues of a single selected @xmath48-sector . both figs .
[ fig : symmetry](a ) and [ fig : symmetry](b ) show results for a chaotic spin-1/2 chain with all interactions in eq . included . because fig .
[ fig : symmetry](a ) mixes the three symmetries we have discussed , @xmath40 becomes a poisson distribution , even though the system is in fact chaotic . in fig .
[ fig : symmetry](b ) , just as in figs .
[ fig : symmetry](c)-(e ) , we circumvent the @xmath93 subspace by avoiding chains with an even number of sites .
we also choose @xmath108 to avoid conservation of total spin . by doing so , the only remaining symmetry is parity , which we take into account .
the expected wigner - dyson distributions are then obtained .
the computer programs used to obtain the data for figs . [ fig : ps ] , [ fig : npc ] , and [ fig : symmetry ] are available.@xcite the reader will also find programs to study gaussian orthogonal ensembles and suggestions for other investigations .
spin 1/2 chains are excellent models for introducing students to some of the basic concepts of linear algebra , quantum mechanics , as well as to current areas of research .
in addition to the crossover from integrability to chaos , they can be used to introduce topics as diverse as the metal - insulator transition , quantum phase transition , entanglement @xcite , spintronics , and methods of quantum control @xcite .
they have been considered as models for quantum computers @xcite and magnetic compounds @xcite , and recently have been simulated in optical lattices.@xcite a. g. thanks stern college for women of yeshiva university for a summer fellowship and the kressel research scholarship for a one - year financial support .
this work is part of her thesis for the s. daniel abraham honors program .
l. f. s. thanks m. edelman , f. m. izrailev , a. small , and f. zypman for useful discussions .
a. v. sologubenko , e. felder , k. giann , h. r. ott , a. vietkine , and a. revcolevschi , `` thermal conductivity and specific heat of the linear chain cuprate sr@xmath109cuo@xmath110 : evidence for thermal transport via spinons , '' phys .
b * 62 * , r6108 - 14 ( 2001 ) .
s. trotzky , p. cheinet , s. flling , m. feld , u. schnorrberger , a. m. rey , a. polkovnikov , e. a. demler , m. d. lukin , and i. bloch , `` time - resolved observation and control of superexchange interactions with ultracold atoms in optical lattices , '' science * * , ( ) .
chen , s. nascimbne , m. aidelsburger , m. atala , s. trotzky , and i. bloch , `` controlling correlated tunneling and superexchange interactions with ac - driven optical lattices . ''
lett . * 107 * , 210405 ( 2011 ) . | we introduce aspects of quantum chaos by analyzing the eigenvalues and the eigenstates of quantum many - body systems .
the properties of quantum systems whose classical counterparts are chaotic differ from those whose classical counterparts are not chaotic .
the spectrum of the first exhibits repulsion of the energy levels .
this is one of the main signatures of quantum chaos .
we show how level repulsion develops in one - dimensional systems of interacting spins 1/2 which are devoid of random elements and involve only two - body interactions .
in addition to the statistics of the eigenvalues , we analyze how the structure of the eigenstates may indicate chaos . the programs used to obtain the data are available online . |
modelling of pedestrian dynamics is actual problem at present days .
different approaches from the social force model ( @xcite and references therein ) based on differential equations to stochastic ca models ( @xcite and references therein ) are developed .
they reproduce many collective properties including lane formation , oscillations of the direction at bottlenecks , the so - called `` faster - is - slower '' effect .
these are an important and remarkable basis for pedestrian modelling .
but there are still things to be done in order to reproduce individual pedestrian behavior more realistic and carefully .
the model presented takes its inspiration from stochastic floor field ( ff ) ca model @xcite . here
a static field is a map that pedestrian may use to orient in the space .
dynamic field is used to model herding behavior in panic situations .
it s known that regular situations imply that pedestrians analyze environment and choose their route more carefully ( see @xcite and reference therein ) .
pedestrians keep a certain distance from other people and obstacles .
the more hurried a pedestrian is and more tight crowd is this distance is smaller . we adopted a mathematical formalization of these points from @xcite .
pedestrians minimize efforts to reach their destinations : feel strong aversion to taking detours or moving opposite to their desired direction .
however , people normally choose the fastest rout but not the shortest .
this means that opportunity to wait ( to stay at present place ) has to be realized in the model .
( models @xcite ( and other ca models ) imply that people can stay at present place if there is no space to move only . )
we realize this point ( people patience ) in the algorithm . as well as it s necessary to take into account that some effects are more reside for certain regions .
for instance , clogging situations are more pronounced in the nearest to an exit areas .
this means that spatial adaptivity of correspondent model parameters to be introduced in the model .
all these changes and additions extend basis ff model towards emotional aspect and improve and make flexible decision making process . by this reason model obtained was named as _
intelligent ff model_.
as usual for ca models the space ( plane ) is sampled into cells @xmath0 ( it s an average space occupied by a pedestrian in a dense crowd @xcite ) which can either be empty or occupied by one pedestrian ( particle ) only . the von neumann neighborhood is used .
it implies that each particle can move to one of four its next - neighbor cells @xmath1 or to stay at the present cell at each discrete time step @xmath2 , e.i .
, @xmath3 .
( empirically the average velocity of a pedestrian is about @xmath4 .
so real time corresponding to one time step in the model is about @xmath5 .
) such movement is in accordance with certain transition probabilities that are explained below .
static ( @xmath6 ) and dynamic ( @xmath7 ) floor fields are introduced and discussed in @xcite .
for each cell @xmath1 values of @xmath8 and @xmath9 are given .
_ static floor field _
@xmath6 describes the shortest distance to an exit ( or other destination point that depends on a task ) . it does nt evolve with time and is nt changed by the presence of the particles .
the value of @xmath8 is set inversely proportional to the distance from the cell @xmath1 to the exit .
one can consider @xmath6 as a map that pedestrian can use to move to the target point , e.g. , exit .
_ dynamic floor field _
@xmath7 is a virtual trace left by the pedestrians similar to the pheromone in chemotaxis .
it is used to model a " long - ranged attractive interactions between the pedestrians , e.g. , herding behavior that is observed in panic situations .
dynamic floor field is time dependent . in each time
step each @xmath9 decays with probability @xmath10 and diffuses with probability @xmath11 $ ] to one of its four neighboring cells .
decay and diffusion lead to broadening , dilution , and finally vanishing of the trace . at @xmath12 for all cells @xmath13 .
@xmath14 just after particle left cell @xmath1 .
field @xmath6 and @xmath7 works in such a way that transition probability increases in the direction of higher fields @xmath6 and @xmath7 , i.e. motion in such direction is more feasible . to make next
step pedestrian observes surroundings .
let @xmath15 be maximum distance ( in cells ) at which pedestrian can see around , by @xmath16 denote the occupation number . if desired direction ( cell ) is occupied to minimize efforts to reach destination or to realize the fasters rout pedestrian has to a have an opportunity to stay at present cell or move to one of the rest unoccupied nearest cells .
keeping apart from other people and obstacles can be simulated by decreasing probability for such direction ( let @xmath17 be a name of the direction ) .
for present position of some pedestrian let cell @xmath1 be a next - neighbor in the direction @xmath17 . for cell
@xmath1 we will calculate a term @xcite @xmath18 where @xmath19 distance to a nearest obstacle ( wall , column , etc . , but not pedestrian ) in the direction @xmath17 starting from cell @xmath1 ( @xmath20 ) ; @xmath21 number of pedestrians that are in the direction @xmath17 starting from cell @xmath1 up to the nearest obstacle ( sum is over lines ( @xmath22 ) or over columns ( @xmath23 ) depending on direction @xmath17 ) , all cells behind obstacle are considered as occupied .
update rules for ca are the following : 1 . for each pedestrian the transition probability @xmath24 to move to cell @xmath1
one of four next - neighbors is @xmath25,\end{gathered}\ ] ] where @xmath26,@xmath27 sensitivity parameters ; @xmath28 inertia effect : @xmath29 for the direction of pedestrian s motion in the previous time step and @xmath30 for other cells ; normalization @xmath31 where the sum is over all possible target cells . 2 .
if @xmath32 then pedestrian stays at the present cell , otherwise pedestrian chooses randomly a target cell @xmath33 based on the transition probabilities determined by ( 1 ) .
if @xmath34 and @xmath35 ( i.e. , cell @xmath33 is occupied ) then pedestrian chooses randomly a target cell once again .
now target cell is chosen among the following candidates : rest next - neighbors available for moving ( i.e. , @xmath36 ) and the present cell . for the available next - neighbor cell @xmath1
new transition probability is @xmath37 and pedestrian can stay at present cell with probability @xmath38 ( where sum is over all available candidates at now ) .
obviously , if there is no available next - neighbors then particle stays at present cell .
+ in contrast to @xcite this step gives an opportunity for pedestrians not to move and wait when preferable direction will free .
4 . whenever two or more pedestrians have the same target cell , the movement of all involved pedestrians is denied with probability @xmath39 , i.e. all pedestrians remain at their old places @xcite .
one of the candidates moves to the desired cell with the probability @xmath40 .
pedestrian that is allowed to move has the largest probability among all candidates ( in this case probability is a measure of pedestrians physical strength ) .
the other probabilistic method @xcite can be used here as well .
pedestrians that are allowed to move perform their motion to the target cell .
@xmath7 at the origin cell @xmath1 of each moving particle is increased by one @xmath41 and therefor can take any non - negative integer value .
these rules are applied to all particles at the same time , i.e. , parallel update is used .
there are several parameters in the model obtained .
their values and physical meaning are presented below .
* @xmath15 maximal distance at which pedestrian can feel the surroundings .
people avoid to walk close to obstacles and other people . @xmath42 if there is no free space to move , @xmath43 if direction is free , @xmath44 for any other intermediate situation .
* @xmath45 sensitivity parameter that can be interpreted as the knowledge of the shortest way to the destination point , or as a wish to move in a certain direction .
@xmath46 means that pedestrian do nt use information from the field @xmath6 .
the higher @xmath26 is movement of the pedestrians is more directed .
* @xmath47 sensitivity parameter that can be interpreted as a rate of herding behavior .
it is known that people try to follow others particulary in panic situations @xcite .
@xmath48 means that pedestrian chooses a way of their own ignoring ways of others .
the higher @xmath27 is the herding behavior of the pedestrians is more pronounced .
* @xmath49 parameter that determines the strength of inertia which suppresses quick changes of the direction .
* @xmath50 : @xmath51 where @xmath52 $ ] friction parameter that controls the resolution of conflicts in clogging situations .
@xmath53 works as some kind of local pressure between the pedestrians .
the higher @xmath53 is pedestrians are more handicapped by others trying to reach the same target cell .
such situations are natural and well pronounced for nearest to exit ( destination point ) space . for other areas
it s not typical but it s possible .
so to realize it and make simulation of individuals realistic the coefficient @xmath54 is introduced ( in contrast with original ff model @xcite ) .
* @xmath10 , @xmath11 $ ] these constants control diffusion and decay of the dynamic floor field @xcite . `` it reflects the randomness of people s movement and the visible range of a person , respectively .
if the room is full of smoke , then @xmath10 takes large value due to the reduced visibility . through diffusion and
decay the trace is broadened , diluted and vanishes after some time . ''
model obtained is simple . for one time step
there are @xmath55 calculations if @xmath56 pedestrians are involved .
it gives advantage over continuous social - force model @xcite where each time step @xmath57 interaction terms have to be evaluated .
the discreteness of the model is advantage as well .
it allows for a very efficient implementation for large - scale computer simulations .
it s shown @xcite that original ff model reproduces variety of collective effects : clogging at large densities , lane formation in counterflow , oscillation in counterflow at bottlenecks , patterns at intersection , trail formation ,
faster - is - slower " and freezing - by - hearting " ( in panic ) .
all these effects are simulated by varying of the model parameters .
model obtained saves this opportunities .
modifications and improvements made here mainly concern the quality of pedestrians behavior reproducing .
they allow more realistically ( carefully ) simulate analysis that people accomplish while they choose a direction for moving .
idea of parameters adaptivity makes model more flexible and closer to real life .
at first let us consider components that determine probability @xmath24 .
in contrast to ff model @xcite one can distinguish two different types of terms in ( [ 1 ] ) .
term @xmath58 characterizes the physical possibility to move , @xmath59 $ ] .
it takes maximal value if movement conditions in the direction are favorable . and
@xmath42 if there is no free space to move .
term @xmath58 proportionally decreases with the advent and approaching of some obstacles ( people , wall , etc . ) in the direction .
other terms @xmath60 , @xmath61 , @xmath62 vary form @xmath63 to @xmath64 ( in general case ) and characterize style of people behavior
. minimal value of parameter ( @xmath65 , @xmath48 or @xmath30 ) means that correspondent feature of behavior is nt realized and term does nt affect the probability . if all three terms are minimal then pedestrians walk free . and in this case only term @xmath58 determines the transition probability for each next - neighbor cell @xmath1 in accordance with people features : keeping apart from other people and obstacles , patience . in ff model
@xcite pedestrians stay at present cell if there is no space to move only . here
we give pedestrians the opportunity to wait when preferable direction will free even if other directions are available for moving at this time .
such behavior is reside to low and middle densities . to realize it transition probabilities ( [ 1 ] )
do nt include a checking if cell @xmath1 occupied or not . in this case transition probabilities ( [ 1 ] ) can be considered as a rate of wish to go to the certain directions .
possibility not to leave present cell is realized in step 3 .
the idea of spatial adapted parameters is introduced here .
it s clear that conditions ca nt be equal for all people involved .
position of pedestrian in the space may determine some features of the behavior .
thus clogging situations are natural and well pronounced for nearest to exit ( destination point ) space . for other areas
it s not typical but it s possible . to realize it the coefficient @xmath66 for clogging parameter @xmath67 is introduced
in order to test our model a regular evacuation process was simulated .
this means that @xmath48 in examples presented . and we set @xmath30 , @xmath68 .
there was simulated evacuation of one person ( n=1 ) from a room @xmath69 ( @xmath70 cells @xmath71 @xmath70 cells ) with one exit ( @xmath72 ) in the middle of a wall .
recall that the space is sampled into cells of size @xmath0 which can either be empty or occupied by one pedestrian only .
static field @xmath6 was calculated in accordance with @xcite . stating position is a cell in a corner near wall opposite to the exit .
pedestrian moves towards the exit with @xmath73 . for
such sampled space minimal value of time steps that require to leave the room starting from initial position is @xmath74 .
different combinations of parameters @xmath26 and @xmath75 were considered .
total evacuation time and trajectories were investigated .
following table contains results over 500 experiments .
figures [ gist ] show total evacuation time distributions for some couples of the parameters from table [ tmo ] over 500 realizations . and
@xmath75 over 500 experiments.,title="fig : " ] and @xmath75 over 500 experiments.,title="fig : " ] + + and @xmath75 over 500 experiments.,title="fig : " ] and @xmath75 over 500 experiments.,title="fig : " ] + + and @xmath75 over 500 experiments.,title="fig : " ] and @xmath75 over 500 experiments.,title="fig : " ] + + figures [ treks ] present tracks of pedestrian ways over 500 realizations for some couples of the parameters from table [ tmo ] . and @xmath75 over 500 experiments.,title="fig : " ] and @xmath75 over 500 experiments.,title="fig : " ] + + and @xmath75 over 500 experiments.,title="fig : " ] and @xmath75 over 500 experiments.,title="fig : " ] + + and @xmath75 over 500 experiments.,title="fig : " ] and @xmath75 over 500 experiments.,title="fig : " ] + + different moving conditions are reproduced by these combinations .
they vary from @xmath76 , @xmath77 to @xmath78 , @xmath79 .
the former case can be interpreted as pedestrian moves in a very low visibility ( or by touch ) but he approximately knows the direction to destination point . in other words
one can say that pedestrian does nt see , knows , and wants not so much to go to destination point ( exit ) .
last one ( @xmath78 , @xmath79 ) describes situation when pedestrian sees , knows , and wants to go to destination point very much . note that if @xmath76 model presented corresponds to ff - model @xcite with the same other parameters . and
it s clear that the pedestrian patience that was introduced in the model does nt pronounced in one pedestrian case .
one can see that for small @xmath77 mode @xmath80 is very dependent on parameter @xmath75 .
the bigger parameter @xmath26 is an influence of @xmath75 to @xmath80 is less pronounced .
but the bigger parameter @xmath75 is more natural way pedestrian chooses tracks are more close to a line connecting starting point and exit ( non the less random component takes place ) . under @xmath81
a proximity of wall decreases the probability ( [ 1 ] ) to move in this direction .
thereby tracks are forced to tend to natural one .
thus parameter @xmath75 fulfils its role to simulate environment analysis here . for this collective experiment
the space was a room @xmath82 ( @xmath83 cells @xmath71 @xmath83 cells ) with one exit ( @xmath72 ) .
initial number of people is @xmath84 ( density @xmath85 ) .
initial positions are random and people start to move towards the exit with @xmath73 .
exit is in the middle of east wall .
figures [ 300people_ks1 ] , [ 300people_ks3 ] present typical stages of evacuation process for different @xmath26 and @xmath75 . + and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] + + + and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] + + and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] + + + and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] and different @xmath75.,title="fig : " ] + one can see that evacuation dynamics in case a ) differs from case b ) in both figures [ 300people_ks1 ] , [ 300people_ks3 ] .
the reason of it is different parameters @xmath75 .
if @xmath81 people avoid to approach to walls .
and while crowd density allows pedestrians try to follow more natural way to the exit .
the bigger @xmath75 is shape of crowd in front of exit is more diverse from the case of @xmath76 .
( note if @xmath76 we have ff - model with the same other parameters . )
one can notice that the closer to the exit pedestrians are circle shaped crowd in front of the exit is more unrealistic .
the problem comes from computational aspect of coefficient @xmath58 .
it s a positive that in the model obtained people avoid to approach to walls . but
this effect has to be less pronounced with approaching to wall ( walls as in a subsection example below ) surrounding exit .
thus parameter @xmath75 has to be spatial adaptive .
next table demonstrates numerical description of cases presented .
let @xmath86 be frequency to choose direction @xmath87 over all experiment for each couple of parameters . here
@xmath88 are north , east , sought , west , center ( stay at present place ) correspondingly , @xmath89 total number of movements including stayings at current position over all experiment . comparing cases 1 ) and 2 ) , 3 ) and 4 )
correspondingly one can notice that greater @xmath75 leads to significant redistribution of flow
. in the case of @xmath76 low @xmath90 says that people move as much as possible ( because people can stay at current position if all nearest cells are occupied only ) .
if @xmath91 the opportunity to wait is realized in cases 2 ) and 4 ) @xmath90 has the greatest value .
@xmath92 has the smallest value ( this direction is opposite to the exit ) , @xmath93 , @xmath94 are approximately equal because exit is in the middle of the wall .
thus in cases 2 ) and 4 ) model reproduces more natural decision - making process . at the same time
let us remark that increasing @xmath26 makes evacuation process more directed and reduces total evacuation time ( see cases 1 ) and 3 ) , 2 ) and 4 ) correspondingly ) . for this collective experiment
the space was a room @xmath95 ( @xmath70 cells @xmath71 @xmath96 cells ) with one exit ( @xmath72 ) .
initial @xmath97 ( density @xmath98 ) .
initial positions are random and people start to move towards the exit with @xmath73 .
spaces are presented in the figures [ 150people]a , [ 150people]b .
+ + two combinations of parameters @xmath26 and @xmath75 were considered . and
total evacuation time was investigated .
table [ tmoexits ] contains results over 100 experiments .
initial positions of people were the same for all experiments .
\a ) b ) one can notice once again that in the case a ) parameter @xmath75 does nt influence on total evacuation time under such @xmath26 .
but in other case increase of @xmath75 leads to significant delay of evacuation .
the reason of it is `` computational '' repulsion from the walls . and
as a result pedestrians do nt use corner between wall and exit , exit is not fully used , total evacuation time increases .
so this is one more example that shows necessity in at least parameter @xmath75 adaptivity .
in the paper the intelligent ff cellular automation model is presented .
modifications made are to improve realism of the individual pedestrian movement simulation .
the following features of people behavior are introduced : keeping apart from other people ( and obstacle ) , patience .
idea of spatial adaptation of model parameters is pronounced and one method is presented ( parameter @xmath53 ) .
model obtained saved opportunities to reproduce variety of collective effects of pedestrian movement from free walk to escape panic and took more flexibility .
the simulation made showed real improvements in decision - making process in comparison with basic ff model and pointed out some problems .
the following points seem to be very important for realistic pedestrian simulation and are under future investigations . in a case of emergency
appearing of clogging situation in front of exit often leads to appearing fallen or injured people ( or jam ) .
the physical interactions in such crowd add up and cause dangerous pressures up to @xmath99 @xcite which `` can bend steel barriers or push down brick walls '' .
fallen or injured people act as `` obstacles '' , and escape is further slowed .
continuous model can reproduce pushing and physical interactions among pedestrians .
ca model does nt allow to do it .
parameter @xmath53 works in the model as some kind of local pressure between the pedestrians ( the higher @xmath53 is pedestrians are more handicapped by others trying to reach the same target cell ) . but by means of @xmath53 fallen or injured people are not simulated .
our further intention is to produce method to evaluate common pressure to each pedestrian in ca model , and if pressure is over some barrier to indicate correspondent cell as new obstacle
. there are at least two reasons for parameters to be adaptive .
one of them is learning that is reside to people .
therefor at least parameters @xmath26 and @xmath27 need to be time adaptive and spatial dependent as well .
parameter @xmath75 needs to be spatial adaptive because of negative computational effects .
so methods to adapt model parameters are under further investigation .
1 burstedde , c. ; k. klauck ; a. schadschneider ; j. zittartz . 2001 .
`` simulation of pedestrian dynamics using a twodimensional cellular automaton . ''
_ physica a _ , no .
295 , 507525 .
helbing , d. 2001 .
`` traffic related self - driven many - particle systems . '' _ rev .
mod . phys .
73 _ , no .
4 . kirchner , a. ; a. schadschneider .
2002.``simulation of evacuation processes using a bionics - inspired cellular automation model for pedestrian dynamics . ''
_ physica a _ , no .
312 , 260276 .
malinetskiy , g.g . and m.e .
`` an application of cellular automation for people dynamics modelling . '' _ journal of computational mathematics and mathematical physics 44 _ , no . 11 , 21082112.(rus . ) nishinari , k. ; a. kirchner ; a. namazi ; a. schadschneider . `` extended floor field ca model for evacuation dynamics . ''
_ e - print cond - mat/0306262 _ yamamoto , k. , kokubo , s. , nishinari , k. 2007 .
`` simulation for pedestrian dynamics by real - coded cellular automata ( rca ) . ''
_ physica a _
, doi:10.1016/j.physa.2007.02.040 . | a stochastic cellular automata ( ca ) model for pedestrian dynamics is presented .
our goal is to simulate different types of pedestrian movement , from regular to panic .
but here we emphasize regular situations which imply that pedestrians analyze environment and choose their route more carefully .
and transition probabilities have to depict such effect .
the potentials of floor fields and environment analysis are combined in the model obtained .
people patience is included in the model .
this makes simulation of pedestrians movement more realistic .
some simulation results are presented and comparison with basic ff - model is made . |
probably the best understood adsorption system is the he monolayer on graphite @xcite .
experiments carried out at the university of washington ca .
1970 revealed _ for the first time _ behavior corresponding to a two dimensional ( 2d ) gas .
more dramatic was the appearance of a spectacular peak in the specific heat of @xmath0he near @xmath4 k. this peak , well described by the 3 state potts model , manifested a 2d transition from a high @xmath5 fluid to a low @xmath5 commensurate ( @xmath2 r30@xmath3 ) phase , providing a benchmark measure of coverage , not seen in previous adsorption experiments .
this ordered phase ( at density @xmath6@xmath7 ) corresponds to atoms localized on second nearest neighbor hexagons . at higher densities near completion of the first monolayer ( @xmath8@xmath7 ) an incommensurate
2d triangular solid phase is present ; the phase diagram at intermediate densities is not yet completely determined . a quantitative understanding of the he
graphite interaction was made possible by precise scattering measurements of surface bound states and band structures @xcite .
the availability of graphene ( gr ) and its derivatives like graphane ( gh ) @xcite and graphene fluoride ( gf ) @xcite offers the prospect of novel adsorption phenomena .
no special phenomenon is expected for he adsorbed on one side of gr because the interaction is similar to that on graphite .
the situation is different for gh and gf due to the modified symmetry of the adsorption potential .
we have developed a model adsorption potential for he on gf and gh . with exact quantum monte carlo methods
we have studied a single @xmath0he and @xmath1he atom on these substrates , as well as submonolayer films of @xmath0he at coverages similar to that ( @xmath9@xmath7 ) of the @xmath2 r30@xmath3 state on graphite .
graphane and graphene fluoride have a similar geometry ; half of the h ( f ) atoms are attached on one side of the graphene sheet to the carbon atoms forming one of the two sublattices of graphene .
the other half are attached on the other side to the c atoms forming the other sublattice .
the h ( f ) atoms are located on two planes ( see fig.[fig1]a ) ; one is an overlayer located at a distance h above the pristine graphene plane while the other is an underlayer at a distance h below the graphene plane .
in addition , there is a buckling of the c plane with the c atoms of one sublattice moving upward by a distance b while the other sublattice moves downward by the same amount . a he atom approaching gh ( gf ) from above will interact primarily with the h ( f ) overlayer , but it will interact also with the c atoms and the h ( f ) atoms of the underlayer . two schematic views of gf .
f ( c ) atoms are light ( dark ) gray .
positions of atoms are to scale but their sizes are arbitrary .
the black balls represent two adsorption sites for he , one of each kind .
gh is similar .
, title="fig:",width=340][a ] two schematic views of gf .
f ( c ) atoms are light ( dark ) gray .
positions of atoms are to scale but their sizes are arbitrary .
the black balls represent two adsorption sites for he , one of each kind .
gh is similar .
, width=188 ] we have adopted a traditional , semi empirical model to construct the potential energy @xmath10 of a single he atom at position @xmath11 near a surface @xcite .
the potential is written @xmath12 , a sum of a hartree fock repulsion derived from effective medium theory , and an attraction , @xmath13 , which is a sum of `` damped '' he atom van der waals ( vdw ) interactions and the polarization interaction with the surface electric field .
the first term is @xmath14 . here
@xmath15 ev
bohr@xmath1 is a value derived by several workers as the coefficient of proportionality between the repulsive interaction and the substrate s electronic charge density @xmath16 _ prior _ to adsorption .
the geometry of gh and gf , their electronic charge density and the electrostatic potential have been obtained using density functional theory with an all electron triple numerical plus polarization basis set with an orbital cutoff of 3.7 @xmath17 as implemented in the dmol3 code @xcite .
the exchange and correlation potential was treated in a generalized gradient approximation parametrized by perdew , burke , and ernzerhof @xcite .
we use a tetragonal unit cell containing four c atoms and four h ( f ) atoms for gh and gf , respectively .
the cell dimensions for gf are @xmath18 , @xmath19 , and @xmath20 , while for gh we use @xmath21 , @xmath22 , and @xmath20 .
the brillouin zone was sampled with a monkhorst
pack grid of @xmath23 * k * points in both cases .
the self consistent cycles were run until the energy difference was less than 10@xmath24 ev .
the atomic positions were relaxed until the forces on all atoms were lower than @xmath25 ev / . as a result , the c
f distance is @xmath26 , the c c distance @xmath27 , the c c distance projected on the @xmath28 plane is @xmath29 @xmath30 and the buckling displacement @xmath31 ; while in gh , the c h distance is @xmath32 , the c c distance @xmath33 , @xmath34 @xmath30 and @xmath35 .
the attraction is a sum of contributions ; for gh , @xmath36 the right most term is the induced dipole energy , where @xmath37 @xmath1 is the static polarizability of the he atom and @xmath38 is the electric field due to the substrate .
the three vdw terms for gh originate from the h overlayer , the graphene sheet ( we are neglecting in this term the small buckling of the graphene sheet ) and the h underlayer , respectively . the graphene term may be written @xmath39 , where @xmath40 is the he distance from the graphene sheet and @xmath41 ev @xmath0 . here @xmath42 mev
@xmath1 is the coefficient of proportionality entering the he_graphite _ vdw interaction and @xmath43 is the interlayer spacing of graphite .
the term @xmath44 , a sum of individual he h interactions , requires the use of damping because the he h separation can be small ; we have adopted the tang toennies damping procedure for this situation @xcite with the parameter @xmath45 @xmath46 .
also required are the vdw coefficients @xmath47 entering the asymptotic he
h interaction , @xmath48 . here
we have used the value @xmath49 ev
@xmath50 , derived from the _ ab initio _ he
h2 interaction of meyer , hariharan and kutzelnigg @xcite .
@xmath51 is treated similarly to @xmath44 but it is considered to be function of @xmath40 , with @xmath52 ev
@xmath0 , because of the larger he distance from the h- plane ( @xmath53 ) .
@xmath13 in the case of gf has an expression similar to ( 2 ) with @xmath54 and @xmath55 replaced by @xmath56 and by @xmath57 with the coefficient @xmath58 ev @xmath50 as given by frigo et al @xcite , @xmath59 @xmath46 and @xmath60 ev @xmath0 . with such model potentials the adsorption sites ( see fig.[fig1]b )
are above the centers of each triplets of h ( f ) atoms of the overlayer , forming a honeycomb lattice with the number of sites equal to the number of c atoms , twice as many as those on gr .
half of the sites are above h ( f ) of the underlayer but the difference between the well depths for the two kinds of adsorption sites is very small , below 1% . for gf
the well depth is 498 k and for gh it is 195 k. these values do not include the induced dipole energy which gives a contribution below 1% .
the inter
site energy barrier is 24 k for gf and 13 k for gh .
both values are significantly smaller than the barrier height 41k for graphite . in this last case
the energy barrier does not depend much on the direction in the @xmath28 plane whereas in the case of gf and gh the ratio between maximum and minimum barrier height in the @xmath28 plane is of order of 45 : the energy landscape of the two last substrates is characterized by a very large corrugation with narrow channels along which low potential barriers are present .
the motion of the he atom , especially in the case of gf , visits only these channels , as though the atom moves in a multiconnected space .
another significant difference is that the distance between two neighboring sites is @xmath61 @xmath30 for gf and @xmath62 @xmath30 for gh whereas it is @xmath63 @xmath30 for graphite and for gr .
prior to these studies , graphite was believed to be the most attractive surface for he , with a well depth a factor of 10 greater than that on the least attractive surface ( cs ) .
if correct , the present results reveal gf to replace graphite , since its well is a factor of 3 more attractive .
we verified that the electrostatic contribution to @xmath64 is small so we neglected it in this first computation .
we computed the exact ground state energy of one @xmath0he atom or one @xmath1he atom on gf and gh , see table [ table1 ] .
the binding energy on gh is similar to that on graphite , whereas that on gf is about three times that on graphite . in both cases the ground state
is delocalized over the full substrate and both kinds of adsorption sites are occupied with comparable probability .
our computation is based on the path integral ground state ( pigs ) method @xcite . with this method
we can compute quantum averages of the ground state of the system using the quantum evolution in imaginary
time @xmath65 of a trial wave function @xmath66 .
if @xmath66 is not orthogonal to the ground state , and @xmath65 is sufficiently long , the quantum evolution purges from @xmath66 the contributions of the excited states , yielding the ground state energy and wave function .
a valuable feature of the pigs method is that it is exact , in principle ; the results are independent of @xmath66 @xcite and systematic errors may be reduced below the statistical uncertainty .
an additional feature of pigs is that it can be used for a single particle as well as for many
body bose systems like the system in the study reported in sect .
3 . with the pigs method one can compute also imaginary time correlation functions like the density
density one in fourier space at imaginary time @xmath65 : @xmath67 , @xmath68 $ ] .
here @xmath69 is the position of the atom at imaginary time @xmath65 .
@xmath70 contains information on the excited states of the system .
the state of wave vector @xmath71 of the lowest energy band has a dominant role in @xmath70 and its energy can be estimated via laplace inversion using powerful inversion methods @xcite . the computed energy spectrum along the directions @xmath72k and @xmath72 m for he on gf and on gh is shown in fig.[fig2 ] .
these bands are represented rather accurately by a tight binding model with nearest and next nearest coupling @xcite . for comparison
we have computed with this same method the band energy for he on graphite finding substantial agreement with the carlos and cole result for the lowest band @xcite .
the bandwidths @xmath73 of he on these three substrates are given in table [ table1 ] . * 7l system & @xmath74 ( k ) & @xmath75 ( k)&@xmath76 ( k)&@xmath73 ( k ) @xmath0he+gf & 46.78(2 ) & -422.94(1 ) & -376.16(3 ) & 9.6(1 ) @xmath1he+gf & 51.08(1 ) & -413.41(1 ) & -362.33(1 ) &
13.7(1 ) @xmath0he+gh & 20.51(1 ) & -153.58(1 ) & -133.07(2 ) & 13.6(4 ) @xmath1he+gh & 22.53(1 ) & -149.50(1 ) & -126.97(3 ) & 19.4(4 ) @xmath0he+gr & 25.30(4 ) & -168.49(1 ) & -143.19(4 ) & 9.6(2 ) @xmath1he+gr & 27.05(2 ) & -162.87(1 ) & -135.82(2 ) & 15.7(4 ) panel ( a ) : the lowest energy band of @xmath1he and @xmath0he on gf along the directions @xmath72k , @xmath72 m and mk .
error bars represent uncertainty from the inversion procedure .
lines ( dashed line for @xmath1he and full line for @xmath0he ) are the fit with the tight binding model with parameters in the inset .
panel ( b ) : same as in panel ( a ) on gh .
, title="fig:"][a ] panel ( a ) : the lowest energy band of @xmath1he and @xmath0he on gf along the directions @xmath72k , @xmath72 m and mk .
error bars represent uncertainty from the inversion procedure . lines ( dashed line for @xmath1he and full line for @xmath0he ) are the fit with the tight binding model with parameters in the inset .
panel ( b ) : same as in panel ( a ) on gh . ]
the effective masses @xmath77 of the various systems reflect the varying corrugations of the potentials . for @xmath0he ( @xmath1he ) ,
the ratios of @xmath77 to the bare mass are 1.40 ( 1.25 ) , 1.10 ( 1.08 ) and 1.05 ( 1.01 ) on gf , graphite and gh , respectively .
the smaller mass enhancement of @xmath1he than @xmath0he reflects the smaller ratio of the corrugation potential to the translational zero point energy .
static structure factor at density @xmath78 ( panel ( a ) ) and off diagonal one body density matrix ( panel ( b ) ) for @xmath0he on gf at @xmath79 @xmath7 ( open circles ) and @xmath80 @xmath7 ( filled circles ) .
lines are guides to the eyes . , title="fig:"][a ] static structure factor at density @xmath78 ( panel ( a ) ) and off diagonal one body density matrix ( panel ( b ) ) for @xmath0he on gf at @xmath79 @xmath7 ( open circles ) and @xmath80 @xmath7 ( filled circles ) .
lines are guides to the eyes . ]
we have studied a @xmath0he submonolayer on gf . as he he interaction we have used an aziz potential @xcite .
the ground state has been computed for a number of @xmath0he atoms from 22 to about 100 spanning the density range @xmath81=0.040.09 @xmath7 . on graphite
the ground state is the commensurate @xmath2 r30@xmath3 state with filling factor 1/3 of the adsorption sites .
a similar state on gf is obtained by populating fourth neighbor sites ( this corresponds to second neighbors in one of the sublattices of the honeycomb at a distance 4.482 ) with a filling factor of the adsorption sites equal to @xmath82 and it corresponds to a density @xmath80 @xmath7 .
notice that this density is smaller than the @xmath6 @xmath7 on graphite due to the dilation of the c plane in gf .
we find that this ordered state is unstable : starting the simulation from an ordered configuration after a short monte carlo evolution the bragg peaks corresponding to the @xmath2 r30@xmath3 state disappear and the system evolves into a disordered fluid state modulated by the substrate potential .
@xmath83 at this density is plotted in fig.[fig3]a as function of @xmath84 and @xmath85 for two numbers @xmath86 of particles : the intensity of some of the peaks do not depend on @xmath86 so they are due to short range order , others scale roughly as @xmath86 and arise from the modulation of the density due to the adsorption potential .
the energy has a minimum value @xmath87 k per atom at the density @xmath88 @xmath7 .
this lies 1.5 k below the single particle energy , implying that the ground state is a self bound liquid . for comparison , we note that the strictly 2d cohesive energy of @xmath0he @xcite is just 0.84 k and the equilibrium density is @xmath89 @xmath7 .
we have computed the off diagonal one body density matrix @xmath90 . as can be seen in fig.[fig3]b @xmath91
reaches a plateau at large @xmath92 and the bose einstein condensate ( bec ) fraction is @xmath93 % at @xmath79 @xmath7 and @xmath94 % at @xmath95 ; the system is superfluid .
we reach a similar conclusion in the case of the gh substrate : the ground state is a liquid with density @xmath96 @xmath7 and @xmath97 k per atom and the bec fraction is @xmath98 % near the equilibrium density and @xmath99 % at @xmath100 @xmath7 .
note that this condensate fraction is significantly smaller than the value ( @xmath101 % ) for @xmath0he in 2d @xcite .
the smaller value is a consequence of the spatial order , albeit imperfect , induced by the substrate potential and of the smaller effective surface available to the atoms due to the strong channeling induced by that potential .
he adsorption on new substrate materials is valuable because of the fundamental importance of helium in many body physics , with a variety of phases seen in both 2d and 3d .
our results indicate that the gf substrate provides the strongest binding of any surface ( since the previous record was held by graphite ) .
moreover , the novel symmetry , the smaller intersite distance and large corrugation imply that quite novel properties may be anticipated for this system .
this is indeed the case .
when many @xmath0he atoms are adsorbed on gf and on gh the most striking result is that the ground state is a low density liquid modulated by the substrate potential and the system has bec , i.e. it is a superfluid .
this is qualitatively different from graphite for which the lowest energy state is the @xmath2 r30@xmath3 commensurate one with no bec @xcite .
we have verified that such an ordered state on gf and gh is unstable relative to the liquid phase . in subsequent work we will provide predictions concerning the phase diagrams and thermodynamic properties for
both he / gf and he / gh , hoping to stimulate experimental studies of these systems . it should be noticed that some of the parameters in the adsorption potential are not known with high precision or they have been adopted from other systems .
we have verified that even a change of parameters like @xmath102 , @xmath103 and @xmath104 by 1020% does not modify the qualitative behavior of the adsorbed he atoms even if there can be a sizable change in the value , for instance , of the adsorption energy .
measurement of thermodynamic properties and he atomic beam scattering experiments from gf and gh will be important to test the accuracy of our model potentials .
this work has been supported by regione lombardia and cilea consortium through a lisa initiative ( laboratory for interdisciplinary advanced simulation ) 2010 grant [ http://lisa.cilea.it ] .
chris billman and jorge sofo are partially supported by the donors of the petroleum research fund administrated by the american chemical society .
9 dash j g , schick m and vilches o e 1994 _ surf .
sci . _ * 299/300 * 405 - 414 cole m w , frankl d r and goodstein d l 1981 _ rev .
phys . _ * 53 * 199 ; boato g and cantini p 1983 _ adv . electronics and electron phys . _ * 60 * 95 sofo j o , chaudhari a s and barber g d 2007 _ phys .
_ b*75 * 153401 nain r r et al .
2010 _ small _ * 6 * 2877 ; zboril r et al .
2010 _ small _ * 6 * 2885 stott m j and zaremba e 1980 _ phys . rev . _
b*22 * 1564 ; cole m w and toigo f 1985 _ phys .
_ b*31 * 727 vidali g , cole m w and schwartz c 1979 _ surf .
sci . _ * 87 * l273 tang k t and toennies j p 1984 _ j. chem .
phys . _ * 80 * 3726 delley b 1990 _ j. chem . phys . _
* 92 * 508 and 2000 _ j. chem .
phys . _ * 113 * 7756 perdew j p , burke k and ernzerhof m 1996 _ phys .
lett . _ * 77 * 3865 meyer w , hariharan p c and kutzelniggw 1980 _ j. chem .
phys . _ * 73 * 1880 frigo a , toigo f , cole m w and goodman f o 1986 _ phys .
_ b * 33 * 4184 sarsa a , schmidt k e and magro w r 2000 _ j. chem . phys . _ * 113 * 1366 rossi m , nava m , reatto l and galli d e 2009 _ j. chem .
phys . _ * 131 * 154108 vitali e , rossi m , reatto l and galli d e 2010 _ phys . rev . _
b*82 * 174510 carlos w e and cole m w 1980 _ phys
_ b*21 * 3713 aziz r a et al .
1979 _ j. chem .
phys . _ * 70 * 4330 whitlock p a , chester g v and kalos m h 1988 _ phys .
_ b*38 * 2418 castro neto et al .
2009 _ rev .
phys . _ * 81 * 109 buzzacchi m , galli d e and reatto l 2002 _ j. low temp .
phys . _ * 126 * 205 | the discovery of fullerenes has stimulated extensive exploration of the resulting behavior of adsorbed films .
our study addresses the planar substrates graphene fluoride ( gf ) and graphane ( gh ) in comparison to graphene .
we present initial results concerning the potential energy , energy bands and low density behavior of @xmath0he and @xmath1he films on such different surfaces .
for example , while graphene presents an adsorption potential that is qualitatively similar to that on graphite , gf and gh yield potentials with different symmetry , a number of adsorption sites double that on graphene / graphite and a larger corrugation for the adatom . in the case of gf ,
the lowest energy band width is similar to that on graphite but the he atom has a significantly larger effective mass and the adsorption energy is about three time that on graphite .
implications concerning the monolayer phase diagram of @xmath0he are explored with the exact path integral ground state method .
a commensurate ordered state similar to the @xmath2 r30@xmath3 state on graphite is found the be unstable both on gf and on gh .
the ground states of submonolayer @xmath0he on both gf and gh are superfluids with a bose einstein condensate fraction of about 10% . |
since the beginning of thermodynamics in the first half of the 19.century its original motivation was the description of steam engines and the liquid to gas transition of water
. here water becomes inhomogeneous and develops a separation of the gas phase from the liquid , i.e. water boils .
a little later statistical mechanics was developed by boltzmann@xcite to explain the microscopic mechanical basis of thermodynamics .
up to now it is generally believed that this is given by the boltzmann - gibbs canonical statistics . as traditional canonical statistics
works only for homogeneous , infinite systems , phase separations remain outside of standard boltzmann - gibbs thermo - statistics , which , consequently , signal phase - transitions of first order by yang - lee singularities .
it is amusing that this fact that is essential for the original purpose of thermodynamics to describe steam engines was never treated completely in the past 150 years .
the system must be somewhat artificially split into ( still macroscopic and homogeneous ) pieces of each individual phase @xcite .
the most interesting configurations of two coexisting phases can not be described by a single canonical ensemble .
important inter - phase fluctuations remain outside , etc .
this is all hidden due to the restriction to homogeneous systems in the thermodynamic limit .
also the second law can rigorously be formulated only microcanonically : already clausius @xcite distinguished between external and internal entropy generating mechanisms .
the second law is only related to the latter mechanism @xcite , the internal entropy generation .
again , canonical boltzmann - gibbs statistics is insensitive to this important difference . for this purpose , and also to describe small systems like fragmenting nuclei or non - extensive ones like macroscopic systems at phase - separation , or even very large , self - gravitating , systems
, we need a new and deeper definition of statistical mechanics and as the heart of it : of entropy . for this purpose
it is crucial to avoid the thermodynamic limit .
as the main aspects of this new thermodynamics were published in @xcite i skip here to repeat all the arguments .
instead i will stress here only the fact that negative heat capacity and convex entropy can be seen at proper phase transitions of 1 .
order , i.e. at phase _ separation _ , in small as well in macroscopic systems independently whether they have long or short range interactions .
as there was a hot discussion at this conference about this point , it seems necessary to repeat the arguments here .
the argument is simple c.f.@xcite : at phase separation the weight @xmath1 of the configurations with energy e in the canonical partition sum @xmath2 becomes _ bimodal _ , at the transition temperature it has two peaks , the `` liquid '' and the `` gas '' configurations which are separated in energy by the latent heat .
consequently @xmath0 must be convex ( @xmath3 , like @xmath4 ) and the weight in ( [ canonicweight ] ) has a minimum at @xmath5 between the two pure phases .
of course , the minimum can only be seen in the microcanonical ensemble where the energy is controlled and its fluctuations forbidden .
otherwise , the system would fluctuate between the two pure phases by an , for macroscopic systems even macroscopic , energy @xmath6 of the order of the latent heat in clear contrast to the usual assumption of the fluctuations in the canonical ensemble @xmath7 .
the heat capacity is @xmath8 i.e. _ the convexity of @xmath0 and the negative microcanonical heat capacity are the generic and necessary signals of any phase - separation@xcite_. this `` convex intruder '' in @xmath0 with the depth @xmath9 has a direct physical significance : its depth is the surface entropy due to constraints by the existence of the inter - phase boundary between the droplets of the condensed phase and the gas phase and the corresponding correlation .
@xmath10 is directly related to the surface tension per surface atom ( with number @xmath11 ) of the droplets .
@xmath12 in my paper together with m.madjet @xcite
we have compared the values of @xmath10 calculated by monte - carlo using a realistic short range interaction with the values of the surface tension of the corresponding macroscopic system . in these calculations
we used the empirical liquid drop parameters for the ground - states energies of the different clusters as given by @xcite . &
@xmath13&@xmath14&@xmath15&@xmath16 & + & @xmath17$]&@xmath18&@xmath19&@xmath20 & + & @xmath21$]&@xmath22&@xmath23&@xmath24 & + & @xmath25&@xmath26&@xmath27&@xmath28 & + & @xmath29&@xmath30&@xmath31&@xmath32 & + & @xmath11&@xmath33&@xmath34&@xmath35 & + & @xmath36&@xmath37&@xmath38&@xmath39 & + roughly @xmath40 and one may argue that this will vanish compared to the ordinary leading volume term @xmath41 .
however , this is not so as @xmath42 at energies inside the phase - separation region ( the convex intruder ) is the of @xmath0 ( its slope gives the maxwell construction of the caloric curve @xmath43 ) .
it is a straight line and its curvature @xmath44 .
consequently for large @xmath13 @xmath45 and the depth of the intruder @xmath46 goes to infinity in the thermodynamic limit . of course
, the ubiquitous phenomena of phase separation exist only by this reason .
it determines the ( negative ) heat capacity as in eq.([negheat ] ) .
the physical ( quite surprising ) consequences are discussed in @xcite .
gross and j.f .
the microcanonical thermodynamics of finite systems : the microscopic origin of condensation and phase separations ; and the conditions for heat flow from lower to higher temperatures . , 122:224111 ; and http://xxx.lanl.gov/abs/condmat/0503604 , ( 2005 ) . d.h.e .
gross and m.e .
fragmentation phase transition in atomic clusters iv the relation of the fragmentation phase transition to the bulk liquid - gas transition .
, 104:541551 , 1997 ; and http://xxx.lanl.gov/abs/cond-mat/9707100 . | conventional thermo - statistics address infinite homogeneous systems within the canonical ensemble . however , some 170 years ago the original motivation of thermodynamics was the description of steam engines , i.e. boiling water .
its essential physics is the separation of the gas phase from the liquid . of course , boiling water is inhomogeneous and as such can not be treated by conventional thermo - statistics .
then it is not astonishing , that a phase transition of first order is signaled canonically by a yang - lee singularity .
thus it is only treated correctly by microcanonical boltzmann - planck statistics .
this was elaborated in the talk presented at this conference .
it turns out that the boltzmann - planck statistics is much richer and gives fundamental insight into statistical mechanics and especially into entropy .
this can be done to a far extend rigorously and analytically . the deep and essential difference between `` extensive '' and `` intensive '' control parameters , i.e. microcanonical and canonical statistics , was exemplified by rotating , self - gravitating systems . in the present paper the necessary appearance of a convex entropy @xmath0 and the negative heat capacity at phase separation in small as well macroscopic systems independently of the range of the force is pointed out .
foundation of classical thermodynamics , negative heat capacity , macroscopic systems |
we grew pbbr@xmath1 crystals using silica hydrogels prepared in a magnetic field of @xmath2 = 5 t as media of crystal growths @xcite .
aligned arrays of nanocrystallites are found in gels @xcite .
the magnetic field was applied during the preparation of the gels and the crystallographic axes of the crystallites were oriented along the direction of the magnetic field .
even if a magnetic field was applied during the crystal growth , it did not affect appreciably .
thus , it is anticipated that the magnetic field brought a structural anisotropy in the pb(ii)-doped silica hydorgels during the preparation .
identification of structural anisotropy has , so far , been one of our recent subjects .
there are a lot of potential uses of the silica gels with controlled structure
. therefore , control of the structure of silica gels has , so far , been studied .
the transport of materials in the gels depends on the structure .
aerogels can be used as media in column chromatography .
if pore size is highly controlled uniformly , the filtering of the column must be improved .
one can intuitively realize anisotropy in mechanical and/or thermomechanical properties due to the anisotropic structure . as described in a monograph @xcite and reported in literatures such as refs .
@xcite , one can control the structure of silica gels by the selection of starting materials , ph control of the solvent , solvent exchange during polymerization stage , and aging .
in addition , porosity of commercially available silica gels for column chromatography was modified by several method @xcite .
magnetic field applied during the gelation affects the network structure of polymer gels .
chemically cross - linked poly(@xmath3-isopropylacrylamid ) forms aligned network structure perpendicular to the magnetic field @xcite .
physically cross - linked agarose gels also exhibit the perpendicular alignment @xcite .
if side - chain groups prefer the parallel alignment , main chains align perpendicularly . on the other hand , in a case
that a magnetic moment is included in a group of the main chain , the main chains align parallel @xcite .
because silica is diamagnetic , one can not imagine the alignment of silica polymers due to direct interaction with a magnetic field . in a previous paper @xcite , basically base on the results of birefringence measurement and the discussions on a likely mechanism of interaction with the magnetic field
, we reached incorrectly at a conclusion that the silica gel network must extend perpendicular to the magnetic field .
a negative birefringence on the order of @xmath4 for 5 t samples , which was a result from a devised snarmont method as described in ref .
@xcite ( also , see ref .
@xcite for corrections of typographical errors in ref .
@xcite ) , was interpreted as follows .
there have been an interest in the structure of scale around several tens or a few hundreds of nanometers .
pores of such scale have been reported in silica aerogels @xcite and commercial gels for column chromatography @xcite . also , along with fractal nature , which were more often reported for silica xerogels ( such as in refs .
@xcite ) , such scales were observed in silica hydrogels @xcite .
it is natural to imagine the existence of closed loops of the same scale in the hydrogels .
let us remember the presence of pb@xmath5 ions and that the skeletons made of silicon and oxygen atoms possess a lot of dangling bonds in hydrogels .
we speculate ring currents along close loops due to complexes formed on such skeletons with pb@xmath5 ions through a mechanism similar to that of electric conduction of the conjugated polymers @xcite and the force tending to direct those rings perpendicular to the magnetic field . in this way , providing a positive intrinsic birefringence of silica , in ref .
@xcite we struggled to give an interpretation to the negative birefringence from this speculated structural anisotropy .
narrowing of the pore size distribution , which resulted from the scanning microscopic light scattering ( smils ) @xcite performed in ref .
@xcite , can be understood consistently , too .
it should be noted that , irrespective of the direction of magnetic alignment , certain characteristic size distributes narrowly if ordering occurs .
the aim of this paper is to overturn the conclusion of ref .
@xcite on the structural anisotropy in silica hydrogels prepared in a high magnetic field . to do so , a consideration is given relying on a previous result @xcite and a recent one @xcite with help of a result of the reinvestigated birefringence @xcite .
further analysis of smils result in ref .
@xcite is also given .
in ref . @xcite as well as in refs .
@xcite , the samples were prepared in the same way as described previously @xcite , except for the application of magnetic field .
the starting material was sodium metasilicate , the acid to maintain the solution strongly acidic was acetic acid , and the source of pb(ii ) was a pb(no@xmath6)@xmath1 aqueous solution . in ref .
@xcite , unlike previous studies @xcite , a lead ( ii ) acetate aqueous solution was employed as the source of pb(ii ) . after preparing silica hydrogels in a magnetic field of various strengths like in refs .
@xcite , the samples were dried in test tubes as prepared for a year .
then , electrons were irradiated to the samples in a transmission electron microscope ( tem ) environment .
we observed the samples by a tem ( hitachi h-9000nar ) .
we should note on the difference between the mechanisms for crystallization in those two kinds of experiments . in the former ,
crystallization is governed by the diffusion of br@xmath7 ions and reaction with pb@xmath5 ions thereafter . in the latter ,
such diffusion - reaction mechanism does not exist .
pb@xmath5 ions are reduced by irradiated electrons and then metallic pb clusters precipitate .
( a ) a tem image of pb(ii ) nanocrystallites formed inside macropores of a silica xerogel prepared in 10 t induced by electron irradiation @xcite ( this image is not a reuse of one in ref .
@xcite ) , and ( b ) a magnification of a part of ( a ) , indicated by a square .
sharpness has been reduced to recover the quality of magnified image . among many crystallites
two pairs of vertical lines have been added to two of them for guide for eyes in ( b).,scaledwidth=90.0% ] we have recently reported formation of pb(ii ) nanocrystallites induced by electron irradiation inside macropores of pb(ii)-doped silica xerogels prepared in a magnetic field of various strengths @xcite .
a tem image is shown in fig .
[ fig : ei ]
. nanocrystallites are observed in aggregates inside macropores . a part of fig .
[ fig : ei](a ) is taken up in fig .
[ fig : ei](b ) .
sharpness has been reduced and two pairs of vertical lines have been inserted auxiliary .
their sizes are a few tens of nanometers , which coincide to the length of short axis of pbbr@xmath1 nanocrystallites in fig .
3(a ) of ref .
@xcite [ carried into this paper as fig .
[ fig : jcg](a ) ] . it is noted that due to the reduction of sharpness we failed to put smaller crystallites in relief as indicated .
those sizes must be the length of short axes of pores elongated affected by the magnetic field in the hydrogels .
we guess that similar structures remain inside macropores in the xerogels . during drying process of hydrogels , macropores form .
therefore , the microstructure of xerogels also more or less differs from that of hydrogels .
nevertheless , there must remain gel network structure in macropores similar to that of hydrogels .
from the fact that the orders are the same , we speculate that the length of the short axis of pores in the hydrogels is also on the same order .
( a ) a tem image of an aligned array of pbbr@xmath8 nanocrystallites formed in a silica hydrogel [ fig .
3(a ) of ref .
@xcite ] , ( b ) white - black binary image of ( a ) , and ( c ) the particles pucked up by particle analysis of imagej software .
auxiliary lines are drawn beneath one of merged objects in ( c ) .
the direction of magnetic field is indicated by the arrow in ( a).,scaledwidth=95.0% ] from the same point of view , the long axes of pbbr@xmath1 nanocrytallites must correspond to the long axes of pores . in fig .
[ fig : jcg](a ) , nanocrystallites are elongated in the direction of magnetic field .
remember the difference in the mechanisms of crystallizations as pointed out in the preceding section .
in the crystal growth in hydrogels used as media , diffusion controls the growth .
moreover , nutrients are supplied by diffusion in hydrogels .
therefore , crystallites can grown along the elongated micropores . on the hand , in crystallization in xerogels induced by an electron irradiation diffusion of nutrient
does not govern the crystal growth .
unless coalescence occurs , the precipitates may grow in a spherical - like form .
it is speculated that order of the underlaying gel networks or that of self - similar network structure whose scale is less than that affects to the outer form of the precipitates must influences the crystallinity of precipitates ( discussion with regard to the fractal nature of the gel network on micropore scale will be given later ) .
it may result in faceting behavior ; however , fig .
[ fig : ei ] does not possess a sufficient resolution .
we do not wish to improve the resolution because it is not the present conc1ern .
a rough estimation is possible from fig .
[ fig : jcg](a ) .
dimension of the nanocrystallites was evaluated using imagej software . after making the tem image [ fig .
[ fig : jcg](a ) ] into a white - black binary image , particle analysis was carried out .
a binarized image is shown in fig .
[ fig : jcg](b ) and a resultant drawing of particle analysis in fig .
[ fig : jcg](c ) .
result of binarization depends on the threshold , and thus the result of particle analysis is not very unique .
the average lengths of major and minor axes were 80 and 40 nm in approximation by ellipses .
regarding the present tem image approximation by rectangles brought vague results due to merged objects made of a number of particles .
dotted lines are drawn to one of merged ones for guide , from which one can understand a tendency of overestimation even by the approximation by ellipses .
one can separate two particles contacting each other by selecting the threshold . with the threshold which can diminish the border of objects and
separate merged ones , the size of objects decreases .
this situation occurs in general . along with the fact that the pore in the network only give the upper limit of the precipitate dimension ,
the present estimation is underestimated for the pore dimension . a result of smils for the scattering angle of @xmath9 ( fig . 4 of ref .
@xcite ) . in the inset peaks of relaxation time distributions locate in the range of millisecond for both birefringent and no - birefringent samples .
the corresponding characteristic length is on the order of hundreds nanometers.,scaledwidth=78.0% ] a rough estimation for a previous smils experiment @xcite on a characteristic length gives a support .
figure 4 of ref .
@xcite is carried as fig .
[ fig : ml ] .
the inset is a logarithmic gaussian distribution spectra of the relaxation times ( @xmath10 ) for autocorrelation functions of the electric field of scattered light .
peaks appear in the rage of millisecond for both samples .
the diffusion coefficient @xmath11 of the corrective diffusion mode is usually estimated by fitting by @xmath12 , where @xmath13 denotes @xmath14 with @xmath15 being the scattering vector . since the working wavelength in this measurement was @xmath16 = 532 nm , @xmath13 = @xmath17 with @xmath18 being the scattering angle
is evaluated as @xmath19 m@xmath20 for @xmath18 = @xmath9 @xcite ( @xmath21 is the diffactive index of the media ) .
therefore , for the result of fig .
[ fig : ml ] one can estimate @xmath11 on the order of @xmath22 m@xmath23/sec . owing to the stokes - einstein relation
@xmath24 the corresponding characteristic length @xmath25 is roughly estimated as one hundred and several tens of nanometers or a few hundreds of nanometers , where @xmath26 is the temperature multiplied by boltzmann s constant and @xmath27 the viscosity coefficient of the solvent . taking into consideration a note that the estimate of 80 nm was underestimated ,
this characteristic length is speculated to correspond to a long axis of elongated pores .
the present concern has been limited to the direction of magnetic alignment in a scale of micropore .
pore size distribution has not been determined . if there is a fractal nature in the scale under consideration , to determine the distribution around a mean value is meaningless because of a scale - free property .
the dimension of nanocrystallites was not limited only by the compartment within the micropores .
even if one can determine the nanocrystallite size distribution , it unnecessarily coincide to the pore size distribution .
we conclude that the silica gel networks align along the magnetic field .
this conclusion is opposite to that of ref .
this is , however , consistent with a recent result on the sign of measure of birefringence of silica hydrogels prepared in a magnetic field of various strengths @xcite .
we have also overturned a conclusion on the sign of birefringence of ref .
@xcite . in ref .
@xcite , the vessels used as sample cells were cylindrical tube made of borosilicate glass . in ref .
@xcite , we performed the same measurements using square cross - sectional cells made of both borosilicate glass and quartz .
special cares were payed to selection of cells so that the optical path difference of empty cells was on the order of @xmath28 nm . nevertheless , obtained results were qualitatively different from each other .
the former indicted decreasing property of @xmath29 and negative birefringence was measured for @xmath2 @xmath30 5 t. a negative birefringence on the order of @xmath4 was not reproduced for 5 t samples . instead , negative birefringence on the order of @xmath31 was measured for 7 t and 10 t samples .
however , we infer that the resultant negative birefringence was attributed to the effect of the borosilicate glass cell wall @xcite .
it is known that borosilicate glasses include a lot of impurities , and thus the cell wall surfaces are not featureless .
therefore , alignment of silica chains is affected by the surface structure and/or other unexpected factors due to impurities .
on the other hand , a trend of positive birefringence was seen in the latter .
effect of cell wall on the alignment of silica chains is expected to be smaller for quartz glass than for borosilicate glass . in ref .
@xcite , we have reinvestigated the results for quartz cells and overturned the conclusion of ref .
@xcite on the sign of birefringence .
that is , the silica hydrogels prepared in a high magnetic field have a positive birefringence with the optic axis along the direction of magnetic field .
providing a positive intrinsic birefringence , one reaches to a conclusion that silica chains tend to align parallel to the magnetic field let us discuss the mechanism of magnetic alignment of the silica network in hydrogels . in ref .
@xcite , we assumed excitation of soliton type , such as for conducting polymers @xcite , on silica chain with presence on pb@xmath32 ions and this excitation plays a role of electric conduction carrier .
ring currents on loops in gel networks make magnetic moments perpendicular to the loops .
accordingly , the loops tend to align perpendicular to the magnetic field .
now , we point that excitation is not limited to such type ; excitation accompanying magnetic moments can arise .
in such case , silica chains align parallel to the magnetic field .
confirmation of this speculation is a subject of future researches .
we wish to conclude this paper with the following sentence .
a conclusion of a previous paper @xcite has been overturned : a silica gel network in pb(ii)-doped silica hydrogels aligns parallel to a magnetic field .
let us close this paper with a comment on a future research .
we have constructed a set up of the snarmont method with a mechanical rotating analyzer to reduce the measurement time @xcite . in previous studies
@xcite , the numbers of measurement were small because the measurements were extremely time - consuming .
we expect an improvement of statistics accuracy to support the conclusion on the sign , and to update the magnitude of the measure of birefringence of silica gels .
99 t. kaito , s .-
yanagiya , a. mori , m. kurumada , c. kaito , t. inoue , effect of magnetic field on the gel growth of pbbr@xmath1 , j. cryst .
growth 289 ( 2006 ) 275 - 277 .
t. kaito , s .-
yanagiya , a. mori , m. kurumada , c. kaito , t. inoue , characteristic nanocrystallite growth of pbbr@xmath1 in a magnetic field in gel , j. cryst .
growth 294 ( 2006 ) 407 - 410 .
r. k. ilre , the chemistry of silica ( new york , wiley - iterscience , 1979 ) .
r. takahashi , k. nakanishi , n. sogo , effects of aging and solvent exchange on pore structure of silica gels with interconnected macropores , j. non - cryst .
solids 189 ( 1995 ) 66 - 76 .
d. j. s. birch , c. d. geddes , sol - gel particle growth studied using fluorescence anisotropy : an alternative to scattering technique , phys .
b 62 ( 2000 )
2977 - 2980 . b. knoblich , th .
gerber , the arrangement of fractal cluster dependent on the ph value in silica gels from sodium silicate solutions , j. non - cryst .
solids 296 ( 2001 ) 81 - 87 .
h. hoffmann , m. meyer , i. zeitler , control of morphology inside the mesoporous gelstructure in silica - gels , coll .
a 291 ( 2006 ) 117 - 127 .
n. a. rahman , i. widhiana , s. r. huliastuti , h. setywan , synthesis of mesoporous silica with controlled pore structure from bagasse ash as a silica source , coll .
a 476 ( 2015 ) 1 - 7 .
r. leboda , e. mendyk , a. gierak , v. a. tertykh , hydrothermal modification of silica gels ( xerogels ) 1 .
effect of treatment temperature on their porous structure , coll .
a 105 ( 1995 ) 181 - 189 .
r. leboda , e. mendyk , a. gierak , v. a. tertykh , hydrothermal modification of silica gels ( xerogels ) 2 .
effect of the duration of treatment on their porous structure , coll .
a 105 ( 1995 ) 191 - 197 .
j. goworek , a. borwka , r. kusak , modification of the porosity of commercial silica gels by dissolution in alkaline mixtures , coll .
a 157 ( 1999 ) 127 - 135 .
r. leboda , j. skubiszewska - ziba , a. dabrowski , v. a. tertykh , effect of aerosil on the hydrothermal modification of porous structure of silica gel adsorbents , coll .
a 172 ( 2000 ) 69 - 77 . v. m. gunko , j. skubiszewka - ziba , r. leboda , v. v. turov , impact of thermal and hydrothermal treatments on structural characteristics of silice gel si-40 and carbon / silica gel adsorbents , coll .
a 235 ( 2004 ) 101 - 111 .
i. otsuki , h. abe , s. ozeki , magnetic field control of structure and function of poly(@xmath3-isopropylacrylamide ) gels , sci .
technol . adv .
( 2006 ) 327 - 331 .
i. yamamoto , s. saito , t. makino , m. yamaguchi , t. takamasu , the anisotropic properties of magnetically ordered gel , sci .
( 2006 ) 322 - 236 .
y. shgekura , y. m. chen , h. furukawa , t. kaneko , d. kaneki , y. osada , j. p. gong , anisotropic polyion - complex gels via template polymerization , adv .
17 ( 2005 ) 2695 - 2699 .
a. mori , t. kaito , h. furukawa , structural anisotropy of silica hydrogels prepared under magnetic field , mater .
62 ( 2008 ) 3459 - 3461 .
a. mori , t. kaito , h. furukawa , m. yamato , k. takahashi , birefringence of silica hydrogels prepared under high magnetic fields reinvestigated , mater .
exp . 1 ( 2014 ) 045202 .
j. wang , j. shen , b. zhou , z. deng , l. zhao , l. zhu , y. li , cluster structure of silica aerogel investigated by laser ablation , nano . struct . mater . 10 ( 1998 ) 909 - 916 .
w. stefaniak , j. goworek , b. biliski , pore size analysis by nitrogen adsorption and thermal desorption , coll . surf . a 214 ( 2003 ) 231 - 237 .
d. w. schaefer , k. d. keefer , fractal geometry of silica condensation polymers , phys .
53 ( 1984 ) 1383 - 1386 .
f. ferri , b. j. frisken , d. s. cannell , structure of silica gels , phys .
67 ( 1991 ) 3626 - 3629 .
j. a. heeger , s. kivelson , j. r. schrieffer , w .- p .
su , solitons in conducting polymers , rev .
60 ( 1988 ) 781 - 850 .
h. furukawa , k. horie , r. nozaki , m. okada , swelling - induced modulation of static and dynamic fluctuations in polyacrylamide gels observed by scanning microscopic light scattering , phys .
e 68 ( 2003 ) 031406 .
h. kusumoto , t. kaito , s .-
yanagiya , a. mori , t. inoue , growth of single crystals of pbbr@xmath1 in silica gel , j. cryst .
growth 277 ( 2005 ) 536 - 540 .
r. tomita , a. mori , m. yamato , h. futrukawa , k. takahashi , structural anisotropy of silica hydrogels prepared under magnetic field , world j. eng .
7 , suppl . 2 ( 2010 ) 518 - 519 .
t. kaito , a. mori , c. kaito , electron - irradiation induced nanocrystallization of pb(ii ) in silica gels prepared in high magnetic fields , j. chem .
chem . eng . 9
( 2015 ) 61 - 66 .
the scattering angle for fig . 3 of ref .
@xcite should be read as 60@xmath33 .
this does not affect the related conclusion therein .
a. mori , r. tomita , semi - automated snarmont method for measurement of small retardation , instr .
technol . , 43 ( 2015 ) 379 - 389 . | in a previous paper [ a. mori , t. kaito , h. furukawa , mater . lett .
62 ( 2008 ) 3459 - 3461 ] , we carried out birefringence measurements of pb(ii)-doped silica hydrogels prepared in a magnetic field ( @xmath0 ) .
for a 5 t sample , we observed a negative birefringence with the optic axis along @xmath0 . at that time , providing a positive intrinsic birefringence of silica , we speculated that in the birefringent gels the gel network extended perpendicular to @xmath0 .
the purpose of this paper is to reconsider this speculation on the basis of previous and recent results [ t. kaito , s .- i .
yanagiya , a. mori , m. kurumada , c. kaito , t. inoue , j. cryst .
growth 289 ( 2006 ) 275 - 277 ; t. kaito , a. mori , c. kaito , j. chem .
chem .
eng . , 9 ( 2015 )
61 - 66 ] . in the former ,
the silica gels were used as a medium of a crystal growth of pbbr@xmath1 and aligned arrays of crystallites with long axis parallel to @xmath0 were obtained . in the latter , pb(ii )
nanocrystallites were formed in silica xerogels by electron irradiation . both of the short axis of pbbr@xmath1 crystallites and the diameter of pb(ii )
crystallites were a few tens of nanometers .
this size must be a size of short axis of pores in the gel networks elongated affected by the magnetic field .
since the pbbr@xmath1 crystallites were elongated along the magnetic field , we conclude that the pb(ii)-doped silica gel networks aligned along the magnetic field .
silica gel ; magnetic field ; microstructure ; porosity ; anisotropy |
in the standard model ( sm ) of particle physics , triple and quartic interactions of the gauge bosons have not been determined with great accuracy .
the possible deviation of gauge boson self - interactions from the sm predictions will be an important sign of new physics beyond the sm .
assuming that the energy scale associated with the new physics is sufficiently high compared to the masses of the gauge bosons , we are motivated to use effective lagrangian method @xcite , which is model independent and can converge to the sm at lower energy scale .
the aqgcs are parameterized by effective operators which do not induce new trilinear gauge couplings .
therefore , we consider that aqgcs can be examined independently from the effects induced by trilinear gauge couplings . in the literature ,
the aqgc operators have been described by either linear or non - linear effective lagrangians .
while non - linear effective lagrangians have dimension 6 , linear effective lagrangians have dimension 8 .
however , the dimension 6 lagrangian only contains vertices with two photons . assuming that there is no higgs boson in the low energy spectrum we employ a nonlinear representation of the spontaneously broken @xmath14 gauge symmetry .
c and p conserving non - linear effective lagrangian for @xmath4 aqgcs are given by @xcite @xmath15 @xmath16 @xmath17 where @xmath18 is @xmath19 triplet , @xmath20 is the tensor for electromagnetic field strength , and @xmath21 and @xmath22 are the dimensionless .
@xmath23 is a mass - dimension parameter associated with the energy scale of the new degrees of freedom which have been integrated out .
in addition , the dimension 8 lagrangian induces quartic couplings among the neutral gauge bosons and it is obtained by using a linear representation of @xmath14 gauge symmetry that is broken by the conventional sm higgs mechanism . however , after the recent discovery of a new particle which is consistent with the sm higgs boson , it becomes important to examine aqgcs based on linear effective lagrangian . on the other hand , since many aqgc - related studies have been generally investigated using non - linear effective lagrangian , which we also choose in our aqgc study . in addition , our limits can compare with the results of the linear parameterizations which is published by cms collaboration @xcite .
for this reason , the non - linear effective lagrangian can be expressed by the linear effective lagrangian @xcite .
therefore , aqgc parameters obtained from non - linear operators can be easily translated into parameters of linear operator . in effective lagrangian approach ,
the cross sections of aqgc processes increase with the coupling strength , and hence unitarity is violated for sufficiently high energy collisions . as stated by ref .
@xcite , a plausible choice of form factor can really provide the unitarity requirement .
form factor formalism is commonly described by the formula @xcite @xmath24 where @xmath23 represents new physics scale . however , because of selection of form factor is arbitrary in the literature , we will examine aqgc parameters without using any form factor in this study . in the presence of the effective lagrangian in eqs.(2 - 3 ) , the vertex functions for @xmath25
are obtained respectively by @xmath26,\end{aligned}\ ] ] @xmath27.\end{aligned}\ ] ] @xmath4 aqgcs were examined in opal , d0 and cms experiments .
the @xmath8 c. l. limits on aqgcs @xmath9 and @xmath10 through the processes @xmath5 and @xmath28 by the opal collaboration at lep2 collider are given by @xmath29 tev@xmath12 , @xmath30 tev@xmath12 @xcite .
also , the current limits on aqgcs @xmath9 and @xmath10 have been provided through the processes @xmath31 by d0 collaboration at the tevatron @xcite and cms collaboration at the large hadron collider ( lhc ) @xcite .
the aqgc limits obtained by d0 collaboration are @xmath32 @xmath33 the most restrictive limits on the parameters of @xmath4 aqgcs were determined by cms collaboration @xmath34 @xmath35 in the literature , @xmath4 aqgcs at the linear colliders were studied through the processes @xmath36 @xcite , @xmath37 @xcite , @xmath38 @xcite , and @xmath39 @xcite , where @xmath40 or @xmath41 and @xmath42 or @xmath43 .
in addition , these couplings were examined at the lhc via the processes @xmath44 @xcite , @xmath45 @xcite , @xmath46 @xcite and @xmath47 @xcite . in this study
, we probe the @xmath4 aqgcs by analyzing three different processes @xmath5 , @xmath6 , and @xmath7 at the clic
. the lhc may not be the best platform to investigate genuine quartic gauge couplings due to remnants of usual @xmath48 deep inelastic processes .
on the other hand , since @xmath49 and @xmath50 are fundamental particles , lepton colliders can determine aqgc parameters with much higher precision measurements than hadron colliders .
the clic , which has high energy and luminosity , is one of the foreseen linear colliders .
it is designed to be running at three center - of - mass energy stages : @xmath1 , @xmath2 , and @xmath3 tev @xcite .
the fundamental parameters of the three energy stages are given in table i. in the clic s first energy stage , it provides good opportunities for the precise measurements of different observables and characteristics of the sm higgs boson , top quark and gauge bosons , etc .. the second phase will further open the window to the discovery of new physics beyond the sm . in addition , higgs boson properties such as the higgs self - coupling and rare higgs decay modes will be examined .
the third phase , which has a maximum energy of @xmath3 tev , is considered to be able to make the most precise measurements of the sm , and to directly determine the pair - production of new heavy particles with masses up to @xmath2 tev @xcite .
the linear @xmath0 collider also has @xmath51 and @xmath52 run modes to examine the new physics search . in @xmath51 and @xmath52
run modes , high energy real photons can be obtained by converting the incoming leptons beam into a photon beam via the compton backscattering mechanism @xcite .
in addition , the linear collider allows us to study photon - induced @xmath53 and @xmath54 reactions arising from almost real photons . here
, @xmath55 is emitted by any of the incoming lepton immediately after it collides with the other lepton .
hence , it is possible to investigate @xmath53 and @xmath54 collisions at the clic .
the photons in these processes have been described as a suitable framework by the weizsaecker - williams approximation ( wwa ) @xcite .
wwa is used to define processes of electro - production in the case of very small angle of @xmath49 or @xmath49 scattering . in this instance , since the virtuality of photon emitted from incoming leptons in wwa is very low , it is assumed to be on the mass shell .
it gives a possibility to reduce the process of electro - production to the photo - production one with an appropriate photon spectrum . as a result , @xmath53 and
@xmath54 processes are produced in a natural way from the @xmath0 process itself .
the diagrams defining these processes are presented in fig . @xmath56 and @xmath57 . in the literature ,
photon - induced reactions through the wwa have been widely examined at the lep , tevatron , and lhc @xcite .
the processes @xmath5 and the subprocesses @xmath58 , and @xmath59 at the clic are described by tree - level feynman diagrams in figs .
we can see from these figures that only one of these diagrams involves anomalous @xmath4 vertex while the others show the contributions arising from the sm .
we have used the comphep-4.5.1 program for numerical calculations in this study @xcite . in our study , only one aqgc coefficient is set to 0 at one time while the other one is set to non - zero .
the total cross sections as functions of @xmath9 and @xmath10 aqgcs for these processes at the clic with @xmath61 and @xmath3 tev are given in figs .
we observe from figs . @xmath62-@xmath63
that the cross section values induced by non - zero @xmath9 are larger than the values of the @xmath10 coupling .
hence , the obtained limits on the @xmath9 coupling are anticipated to be more restrictive than the limits on @xmath10 coupling .
during statistical analysis , we determined @xmath8 c.l .
limits on the aqgc parameters @xmath9 and @xmath10 using a simple one - parameter @xmath64 analysis .
the @xmath64 analysis is defined by the following formula @xmath65 where @xmath66 is the sm cross section , @xmath67 is the total cross section with aqgc , @xmath68 is the statistical error : @xmath69 is the number of events .
first , the number of events for the process @xmath5 is given by @xmath70 where @xmath71 is the integrated luminosity and @xmath66 is the sm cross section . the @xmath72 boson is heavy enough to decay both hadronically and leptonically .
it decays approximately @xmath73 of the time leptonically ( for electron or muon ) and @xmath74 of the time to hadrons @xcite .
so we consider one of the @xmath72 bosons decays leptonically and the other hadronically for the signal .
therefore , we consider that the branching ratio of the @xmath72 bosons pairs in the final state to be @xmath75 . in addition , we impose the acceptance cuts on the pseudorapidity @xmath76 and the transverse momentum @xmath77 gev for the final state photon . after applying these cuts ,
the sm background cross sections in the process @xmath5 , we obtain as @xmath78 pb at @xmath79 tev , @xmath80 pb at @xmath81 tev , and @xmath82 pb at @xmath83 tev . in table ii , we calculate limits on the aqgc parameters @xmath9 and @xmath10 for various integrated luminosities and center - of - mass energies of the process @xmath5 .
we observe from table ii that the obtained limits on the aqgcs via the process @xmath5 with @xmath79 tev are less restrictive than the best limits attained by the current experimental limits . on the other hand ,
our limits at @xmath2 and @xmath3 tev of center - of - mass energies of the same process are , even with low integrated luminosities , more restrictive than the experimental limits .
especially , we have found the limits of the @xmath9 aqgc as @xmath84 $ ] tev@xmath12 while the limits on @xmath10 aqgc as @xmath85 $ ] tev@xmath12 at the integrated luminosity of @xmath86 fb@xmath87 and center - of - mass energy of @xmath3 tev . in the second analysis ,
the number of events for the process @xmath6 is obtained as @xmath88 here , the @xmath72 boson can decay leptonically , but in this case , it will introduce a great uncertainty due to the production of two neutrinos in the final state of our process . for this reason , we consider the hadronic decay of the @xmath72 boson . also , we apply the cuts @xmath77 gev and @xmath76 for the photon in the final state . therefore , we calculate the sm cross sections as @xmath89 pb at @xmath79 tev , @xmath90 pb at @xmath81 tev , and @xmath91 pb at @xmath83 tev . we show the limits on the @xmath9 and @xmath10 aqgcs for the process @xmath6 at various integrated luminosities and center - of - mass energies in table iii . as we can see from this table , the best limits obtained on @xmath9 and @xmath10 aqgcs through process @xmath6 are found to be @xmath92 $ ] tev@xmath12 and @xmath93 $ ] tev@xmath12 , respectively for @xmath94 fb@xmath87 and @xmath83 tev at the clic .
finally , for the process @xmath7 , in order to get promising results the number of events is obtained as @xmath95 here ,
we find the sm cross sections to be @xmath96 pb for @xmath79 tev , @xmath97 pb for @xmath81 tev , and @xmath98 pb for @xmath83 tev . in table iv , we give the limits of the @xmath9 and @xmath9 aqgcs for @xmath7 at clic at @xmath99 and @xmath3 tev . as shown in all these tables , we realize that the sensitivities of @xmath9 and @xmath10 aqgcs are rapidly enhanced when the center - of - mass energy of the clic increases .
in addition , the combined limits obtained on @xmath9 and @xmath10 aqgcs through three different processes with the integrated luminosity of @xmath94 fb@xmath87 at the clic with @xmath83 tev are calculated as @xmath100 $ ] tev@xmath12 and @xmath101 $ ] tev@xmath12 , respectively .
the combined limits are worse than the limits derived in the process @xmath5 since the other two processes have worse limits .
the clic is a future linear collider at the planning stage .
having it running with high energy and high luminosity is quite important new physics research beyond the sm .
@xmath4 aqgcs are defined by dimension 6 effective lagrangians , which have very strong energy dependences .
hence , @xmath4 cross section with anomalous couplings interactions has a higher momentum dependence than the sm cross section .
we can easily understand that the contribution to the cross section of aqgcs quickly increases with center - of - mass energy .
the analyzed processes that have very high energy and clean experimental environment such as clic , are anticipated to be more sensitive to aqgcs .
therefore , we are motivated to study the the aqgc parameters @xmath9 and @xmath10 through the processes @xmath5 , @xmath6 , and @xmath7 at the clic .
we have shown that the best limits obtained on aqgcs @xmath9 and @xmath10 in these processes are at the order of @xmath11 and @xmath13 tev@xmath12 , respectively . since our limits
are derived based on the total cross section with only statistical uncertainty , they is significantly better than the limits derived in experimental studies .
99 w. buchmuller and d. wyler , nucl .
268 , 621 ( 1986 ) .
g. belanger and f. boudjema phys .
b 288 , 201 ( 1992 ) .
g. belanger and f. boudjema phys .
b 288 , 210 ( 1992 ) .
cms collaboration , phys .
d 90 032008 ( 2014 )
. m. baak @xmath102 , arxiv:1310.6708 [ hep - ph ] . c. degrande @xmath102 , arxiv:1309.7890 [ hep - ph ] .
eboli , m.c .
gonzalez - garcia and j.k .
mizukoshi , phys . rev .
d 74 073005 ( 2006 ) .
d. yang , y. mao , q. li , s. liu , z. xu , and k. ye , jhep 1304 108 ( 2013 ) .
o. j. p. eboli , m. c. gonzalez - garcia , and s. m. lietti , s. f. novaes phys . rev .
d 63 075008 ( 2001 ) .
v. m. abazov @xmath102 , d0 collaboration , phys .
b 695 67 ( 2011 ) .
g. abbiendi @xmath102 , opal collaboration , phys . rev .
d 70 032005 ( 2004 ) . v.
m. abazov @xmath102 , d0 collaboration , phys . rev .
d 88 012005 ( 2013 ) .
s. chatrchyan @xmath102 cms collaboration , jhep 1307 116 ( 2013 ) .
g. belanger , f. boudjema , y. kurihara , d. perret - gallix , and a. semenov , eur .
j. c 13 283 - 293 ( 2000 ) .
w. j. stirling and a. werthenbach , eur .
j. c 14 103 - 110 ( 2000 ) .
g. a. leil and w. j. stirling , j. phys .
g 21 517 - 524 ( 1995 ) .
dervan , a. signer , w.j .
stirling , and a. werthenbach , j. phys .
g 26 607 - 615 ( 2000 ) .
w.j . stirling and a. werthenbach ,
phys . lett .
b 466 369 - 374 ( 1999 ) .
s. atag and i. sahin , phys .
d 75 073003 ( 2007 ) .
o. j. p. eboli , m. c. gonzalez - garcia , and s. f. novaes , nucl .
b 411 381 - 396 ( 1994 ) .
o. j. p. eboli , m. b. magro , p. g. mercadante , and s. f. novaes , phys .
d 52 15 - 21 ( 1995 ) .
i. sahin , j. phys .
g 36 075007 ( 2009 ) .
t. pierzchala and k. piotrzkowski , nucl .
phys . proc .
179 - 180 257 ( 2008 ) .
e. chapon , c. royon and o. kepka , phys . rev .
d 81 074003 ( 2010 ) .
a. senol , arxiv:1311.1370 .
p. j. bell , eur .
j. c 64 25 - 33 ( 2009 ) .
o. j. p. eboli , m. c. gonzalez - garcia , and s. m. lietti , phys .
d 69 095005 ( 2004 ) .
d. dannheim @xmath102 , clic e+e- linear collider studies , arxiv:1305.5766v1 .
d. dannheim @xmath102 , clic @xmath103 linear collider studies , arxiv:1208.1402 . i. f. ginzburg , g. l. kotkin , v. g. serbo and v. i. telnov , nucl . instr . and meth .
205 47 ( 1983 ) .
i. f. ginzburg , g. l. kotkin , s. l. panfil , v. g. serbo and v. i. telnov , nucl .
instr . and meth .
219 5 ( 1984 ) . v.
i. telnov , nucl .
instr . and meth . a 294 72 ( 1990 ) .
d. v. soa @xmath102 , mod .
a 27 1250126 ( 2012 ) .
delahaye , mod .
a 26 2997 - 3023 ( 2011 ) .
budnev , i.f .
ginzburg , g.v .
meledin and v.g .
serbo , phys .
15 181 ( 1974 ) .
g. baur @xmath102 , phys .
364 359 ( 2002 ) .
e. yehudai , phys . rev .
d 41 33 ( 1990 ) . c. carimalo , p. kessler and j. parisi , phys .
d 20 1057 ( 1979 ) .
cieza montalvo , n. v. cortez , m. d. tonasse , phys.rev .
d 77 095015 ( 2008 ) .
r. nisius , phys . rept .
332 165 - 317 ( 2000 ) .
j. abdallah @xmath102 , delphi collaboration , eur .
j. c 35 159 ( 2004 ) .
a. abulencia @xmath102 , cdf collaboration , phys .
lett . 98 112001 ( 2007 )
. t. aaltonen @xmath102 , cdf collaboration , phys .
102 222002 ( 2009 )
. t. aaltonen @xmath102 , cdf collaboration , phys .
102 242001 ( 2009 ) .
s. chatrchyan @xmath102 , cms collaboration , jhep 1201 052 ( 2012 ) .
s. chatrchyan @xmath102 , cms collaboration , jhep 1211 080 ( 2012 ) .
cms collaboration , jhep 07 116 ( 2013 ) .
s. atag and a. billur , jhep 11 060 ( 2010 ) .
k. piotrzkowski , phys . rev .
d 63 071502 ( 2001 ) .
h. murayama and m. e. peskin , ann .
nucl . part .
46 533 - 608 ( 1996 ) .
s. atag , s. c. inan and i. ahin , phys . rev .
d 80 075009 ( 2009 ) .
i. ahin and s. c. inan , jhep 09 069 ( 2009 ) .
s. c. inan , phys .
d 81 115002 ( 2010 ) .
i. ahin and m. kksal , jhep 11 100 ( 2011 ) .
m. kksal and s. c. inan , arxiv:1305.7096 .
m. kksal and s. c. inan , ahep volume 2014 , article i d 315826 , 8 pages ( 2014 ) .
m. kksal , arxiv:1402.2915 .
a. a. billur and m. kksal , phys .
d 89 037301 ( 2014 ) .
a. a. billur and m. kksal , arxiv:1311.5326 .
a. senol , phys .
d 85 113015 ( 2012 ) .
i. ahin @xmath102 , phys.rev .
d 88 095016 ( 2013 ) .
s. c. inan and a. billur , phys . rev .
d 84 095002 ( 2011 ) .
r. s. gupta , phys .
d 85014006 ( 2012 ) .
i. ahin , phys . rev .
d 85 033002 ( 2012 ) .
l. n. epele @xmath102 , eur .
j. plus 127 60 ( 2012 ) .
i. ahin and b. ahin , phys .
d 86 115001 ( 2012 ) .
a. a. billur , europhys .
lett . 101 21001 ( 2013 ) .
k. nakamura @xmath102 , j. phys .
g 37 , 075021 ( 2010 ) .
a. pukhov @xmath102 , report no .
inp msu 98 - 41/542 ; arxiv : hep - ph/9908288 ; arxiv : hep - ph/0412191 .
.the fundamental parameters of the three energy options of the clic . here
@xmath104 is the center - of - mass energy , @xmath69 is the number of particles in bunch , @xmath105 is the total luminosity , @xmath106 are the average sizes of the bunches @xcite .
[ tab1 ] [ cols="^,^,^,^,^",options="header " , ] | the compact linear collider ( clic ) is one of the most popular linear colliders , planned to realize @xmath0 collisions in three energy stages of @xmath1 , @xmath2 , and @xmath3 tev .
it has an energy scale never reached by any existing lepton collider . in this study , we present the sensitivity studies of the @xmath4 anomalous quartic gauge boson coupling ( aqgc ) in the three different processes @xmath5 , @xmath6 , and @xmath7 at the clic .
we obtained @xmath8 confidence level limits on the aqgc parameters @xmath9 and @xmath10 at various integrated luminosities and center - of - mass energies .
the best limits obtained on the aqgcs @xmath9 and @xmath10 in these processes are at the order of @xmath11 tev@xmath12 and @xmath13 tev@xmath12 , respectively . |
recent high - resolution , multi - instrumental and multi - wavelength observations of chromospheric transients have re - introduced the question of how much do these affect the outer layers in terms of heating and mass loading .
rbes ( rapid blue - shifted excursions ) and rres ( rapid red - shifted excursions ) are small - scale transients observed in either the blue or the red far - wings of chromospheric lines @xcite .
the measured properties of rbes are similar to those reported for type ii spicules , first reported in ca h @xcite , and are therefore thought to be the h@xmath0 counterparts of these structures . in this work , the term rapid excursions ( res ) will be used to refer to both rbes and rres .
@xcite reported that rbes have signatures in transition region ( tr ) lines .
the observed fov was that of an active region displaying striking coronal loops .
furthermore , @xcite found a spatio - temporal match between h@xmath0 res and features in the iris tr passbands .
in particular , they found evidences that there are spectral signatures for res in iris c 1335 and 1336 and si 1394 and 1403 spectral lines .
however , in the lower activity network regions , they found only weak signals in the c and si spectra with none apparent in the corresponding slit - jaw imaging channels ( i.e. 133 and 140 nm ) that can be traced back to rbes .
this is despite an abundance of rbe detections along the slit .
furthermore , they observed a diffuse halo around the stronger network elements and network - jet like signals that match the rbes in the spectra .
the iris detections show that res have a heating impact up to 80 kk ( si ) , and at least in areas as quiet as network regions .
@xcite studied type ii spicules at the limb using hinode , iris , and sdo / aia .
they find matching heating signals in aia he 30.4 nm with a small spatial displacement from the chromospheric ca h signals ( i.e. continuing the path of ca h spicules , see their fig .
this offset indicates that spicules are being heated as they travel upwards , disappearing from the cooler channels , with signatures appearing in the hotter channels with matching signal found in lines as hot as si iv 140 nm . in a similar set of data , the temporal evolution of type ii spicules
was studied by @xcite who find stronger evidence for progressive heating .
the he 30.4 nm detections clearly shows that at least some type ii spicules reach temperatures as high as 200 kk .
@xcite detected large - scale jets with speeds of 80 - 250 , seen above network regions using iris slit - jaw images in the 133 nm si and 140 nm c passbands .
the latter were observed in bundles , on disk locations close to the limb above network bright - points .
@xcite argue that these " network jets , due to the observed temperature and velocity , likely do not fall back on the solar surface but instead contribute to the heating of the outer atmosphere and mass of the solar wind . @xcite used aia observations at the edge of a coronal hole at the limb , to study the property of propagating disturbances that they identify as `` jetlets '' .
they found that these jetlets have lower velocities than the iris network - jets reported by @xcite , a difference that they attribute to the lower activity of the region .
they also found a correspondence between some aia signatures and network - jets in slit - jaws images from iris , as well as some events without any correspondence .
similarly , the iris detections of @xcite also appear to be lower in velocity than those of @xcite .
however , @xcite find a clear relation between some of their res and the tr network - jets of @xcite . in active regions , jets with properties consistent with rbe s have been observed to be important for the heating of coronal loops @xcite .
more recently , based on in aia observations alone , further evidence has been put forward to suggest that `` plume '' or `` jetlet '' transients are the main source of heating in coronal loops @xcite .
models of the physical processes that could lead to chromospheric heating from transients can be found as early as 1982 @xcite .
evidence that this energy , deposited in the chromosphere , can make it s way into the corona has also emerged from simulations @xcite .
recently , @xcite proposed that the highly dynamical chromospheric jets could be unstable due to the kelvin - helmholtz instability produced by velocity discontinuities between the surface of the jet and surrounding media
. this could be responsible for the rapid heating , and hence the observed fast disappearance of res / type ii spicules in the chromosphere @xcite . while the detailed modeling of solar atmospheric heating based on transients events remains an open question , it is observationally clear that jet - like transients have an impact of at least the lower corona and tr .
what is not observationally clear is how big of an impact these transients have , especially for the most prominent regions on the sun such as coronal holes and quiet sun regions that do not display prominent loop structures . in this paper
we try to answer the question of how important rbes and rres are to the lower corona and transition region . for this
we look at a typical quiet sun region . our approach is to find coronal and transition region heating signatures and relate those signatures to transients observed in the chromosphere .
we use aia data due to its continuous sampling of the sun , due to the link with the previous results and , most importantly , due to the potential for large statistical significance .
sst data are used for the chromospheric observations providing high - spectral , spatial and temporal resolution in h@xmath0 .
limb observations are highly revealing but it is difficult to derive statistics from them due to multiple overlapping structures along the observer s line - of - sight . at disk centre this limitation does not exist and we use such advantage to automatically match structures in this work .
the observations presented in this paper were obtained between 09:06 and 09:35 ut on 2013 may 3 , and the target was a very quiet region located at disk centre ( fig.[quiet ] ) .
this compares to more active areas of the sun presented in the earlier work of @xcite .
two rosettes are visible , anchored in two sets of photospheric bright points and show remarkably little evolution throughout the time - series .
both rosettes show activity in h@xmath0 with the upper rosette showing torsional motions in alternating directions . some of these rotational motions translate to transverse displacements of res which have been analysed by @xcite .
this paper describes the ground - based observations and reduction , and hence here we only present a brief summary .
we used the crisp imaging spectropolarimeter instrument , at the 1-m swedish solar telescope ( sst ; * ? ? ? * ) .
adaptive optics , including an 85-electrode , were used ( an upgrade of the system described in * ? ? ? * ) .
all data were reconstructed with multi - object multi - frame blind deconvolution ( momfbd ; * ? ? ?
* ; * ? ? ?
* ) , using 51 karhunen - love modes sorted by order of atmospheric significance and @xmath1 pixel subfields . a prototype of the data reduction pipeline published by was used before and after momfbd .
this includes the method described by for alignment and destretching as in @xcite .
the spatial sampling is [email protected] pixel@xmath3 , with the spatial resolution reaching up to [email protected] in h@xmath0 over the field - of - view ( fov ) of 41@xmath441 mm ( fig.[rbe_context ] ) .
the h@xmath0 line scan consists of 7 positions ( -0.906 , -0.543 , -0.362 , 0.000 , 0.362 , 0.543 , + 0.906 from line core ) corresponding to a range of -41 to + 41 in velocity , and the temporal cadence of the full spectral scan was 1.34 s. we also use imaging data from the atmospheric imaging assembly ( aia ) , on board of the solar dynamics observatory @xcite in the he 304 and 171 pass - bands ( henceforth 304 and 171 , respectively ) .
the data were downloaded and processed to level 1.5 using solarsoft @xcite which includes co - alignment between the aia channels up to 2 .
we then aligned , de - rotated , and re - sampled the aia data to the sst h@xmath0 images .
all analysis is performed at the sst pixel - size for all data .
alignment using manual control points , solarsoft and standard idl routines were used .
the aia channels were co - aligned to better than one aia pixel ( 0.6 ) using manual comparison with blinking after the initial alignment and resampling to the sst scale .
sub - pixel shifting with cubic interpolation was used for all channels during the re - alignment procedures .
our goal is to create data cubes consisting of brightening detections from the 304 and 171 channels , and data cubes consisting of blue wing , red wing , rbe , and rre detections from h@xmath0 for comparison .
running - difference cubes are computed for all observations ( aia and sst ) comprising of the difference between the mean of the three central adjacent frames ( 30 seconds ) , and the mean of the two frames located 20 seconds before and 20 seconds after the central frames .
this is aimed high sensitivity to transients with life - times around those of re ( 50 seconds ) .
previous detection approaches have used differentiation with h@xmath0 frames close in time but in the photospheric wings ( with the sst * ? ? ? * ) or differentiation with the previous frame in a time - series of binned frames ( in aia and hinode * ? ? ?
our approach in this step is most similar to the latter . in all aforementioned works as well as
this , wavelengths just over 40 from h@xmath0 core are used .
binary maps are computed to locate the events by using thresholding and basic morphological operations . for 304 and 171 a threshold of the mean plus
1.5 standard deviations is applied .
then a second mask , the hit mask , selects the binary detections with at least one pixel above 1.9 standard deviations . for h@xmath0 a single mask with a threshold of the mean plus
1.1 standard deviations is used .
an erode , a dilate and another erode are performed on all h@xmath0 binary masks using disks of 1 , 4 , and 1 pixel diameter , respectively .
this was intended to work both as a morphological close and noise removal at the smallest scales . for 304 and 171 one single
dilate with a disk of one sst pixel ( 0058 ) was used to close regions adjacent at the vertices that would otherwise be disconnected .
each contiguous region was marked as one individual detection using a label operation .
standard idl routines were used for the morphological operators and labelling which should be available in any standard computer vision package .
the labelling step marks structures in the three - dimensional cubes thus tracing structures in space and time . relating with previous automated - detection work
, the morphological operations and labelling would be analogous to the distance thresholds used in @xcite to determine that two nearby structures in time and space are the same . unlike @xcite
, we do nt select for elongation as we often find events originating in groups , occasionally forming wide detection blobs that would be completely discarded .
we also find small blob - like rbes with little elongation , but very high contrast and a clear propagation direction , in the difference maps that would have been discarded had an elongation threshold been considered .
finally , all detections below a size of 500 sst pixels in time and space ( that is , in the three - dimensional data - cube ) were removed .
we further separate the h@xmath0 detections in `` blue '' and rbes as well as `` red '' and rres .
the difference between the re maps and the respective wings maps is that the re maps were zeroed where a signal was also picked up in the opposite wing . in fig.[movie1 ] , and the associated time - series , the contours of the detection maps obtained from 304 are shown plotted over 304 and 171 as well as the detection maps of the rbes , rres and 171 . statistics of the matches between the different detection maps are shown in table [ table : stats2 ] for matches with 304 , and table [ table:171 ] for matches with 171 .
we consider a detection to have a match with another wavelength map if it overlaps by at least 100 sst pixels , or @xmath5 , in time and space with detections from that other map .
if 100 pixels is less than 5% the size of the structure being matched , times the filling factor of the map it is being matched against , then the latter value is the threshold of overlap above which a match is considered .
this is done to prevent the largest structures from having a match with no statistical significance .
these thresholds are also computed in this way when finding the individual probabilities of a match ( @xmath6 ) due to randomness in section [ statistics ] .
the reasoning behind the 5% value is also explained in section .
[ statistics ] .
the selection of the threshold levels was made iteratively such that obvious events are always detected but separate from nearby events while , at the same time , keeping the total amount of pixels in a cube small .
however , small changes do not seem to generate significantly different patterns . while the number of removed small detections , detected noise ( in the aia case ) , and the size of the detected regions will change with different thresholds and filtering , our main aim is to make sure that the statistical significance of the results is high . for the thresholding of aia maps
this mainly means that the total amount of detected pixels should remain a small fraction of the whole fov , and that the overlap criteria remains comparatively high ( discussed in section [ statistics ] ) . [ cols="^,<,^,^,^,^,^ " , ] we choose as the null hypothesis ( h@xmath7 ) that the overlaps of the transient events across two passbands are due to pure coincidence or matches with noise , and assume that the transient detections can occur randomly at any point in time and space ( as noisy detections would ) .
every single labelled detection is taken , with its properties of a size in time and space , and the probability of that region satisfying our match criteria against the full pixel - space of the second passband , is computed .
the probability of a region being matched then follows a bernoulli distribution with the parameters being our pixel overlap criterion ( i.e , a minimum of 100 matching pixels or 100 bernoulli successes ) , the size of the region ( the number of bernoulli trials ) , and the probability of each trial being a success .
the latter is given by the total filling - factor of the detections in the second passband ( i.e. approximately 1.3% for rbes and 1% for rres ) .
we compute this probability for every single detection ( @xmath6 )
. with the individual probability of a match for each region known , finding the total probability of a certain number of matches between becomes a classic poisson - trial problem , where our random variable x is the sum of the outcomes of the several bernoulli variables with different probabilities . under certain circumstances ,
the probability distribution of x can be approximated by a gaussian with @xmath8 $ ] and @xmath9 $ ] as parameters . however
, an approximation by a gaussian may not be justifiable for this problem since the probabilities of each trial fluctuate between extremely low values , and values of the order of one due to the low filling factors of the h@xmath0 maps combined with the very different sizes of each 304 and 171 detection .
thus we chose to compute a stronger test statistic , without having to do any approximation or even having to know the exact distribution of x , by using chernoff bounds .
we use the upper chernoff bound ( see for e.g. * ? ? ?
* ) which provides an exact upper limit on the probability that our number of matches ( or number of bernoulli successes ) are higher than our measured value ( k ) for the null hypothesis : @xmath10 ) } \left(\frac{e[x]}{k}\right)^k , \label{eq:1}\ ] ] where e[x ] is the expected value for the number of successes given by @xmath11=\sum p_i$ ] .
this bound yields a probability lower than @xmath12 for h@xmath7 for every h@xmath7 except rbes to 171 which yields @xmath13 < 0.014 $ ] ( see table [ table : chernoff ] ) , which definitely excludes the null hypothesis of the matches occurring at random for every case .
this value is striking , but not surprising when we consider that we are dealing with small filling factors while requiring a broad overlap match ( minimum of 100 pixels ) . nor
is it surprising considering the actual number of matched structures and the degree of visual match presented in our figures , and associated time - series , when compared to what would be expected from chance alone . for completeness and familiarity
, we also compute the p - value for a z - test statistic , which assumes a gaussian distribution for x as described above .
we obtain a p - values of two digits for every statistic .
this means that our match statistics are tens of standard deviations away from the expected value of the null hypothesis assuming gaussian statistics .
we can therefore confidently reject the null hypothesis when assuming gaussian statistics to much better than 99% for all cases .
note that if our detection algorithms were picking mainly noise , then our test statistics would be necessarily close to those of the null hypothesis , at least within the confidence interval .
one could still argue , disregarding the very different band - passes and instruments used , that systematic errors in the detection method are generating noise at the same locations in both h@xmath0 and 304 . while the 304 data are undoubtedly noisy , the h@xmath0 detections follow very obvious , very high - contrast features in both h@xmath0 wings and are therefore easier to detect . additionally , from the time - series
, the 304 detections do seem to follow obvious intensity changes over the re time - scales . regarding the dependency of the test statistics on our specific choice of the match criteria ,
we note that chernoff bounds drop exponentially with the number of successes ( see eq.[eq:1 ] ) .
our criteria for matching structures do not decrease as fast due to the large number of detections that clearly overlap with those of h@xmath0 over thousands of pixels in space and time , as may be seen in the figures and associated time - series discussed in section [ discussion ] .
extensive testing with different overlap overlap criteria was done that confirms these dependencies .
further , these tests were used to select the final overlap criterium itself .
we essentially aimed at selection criteria that would be undisputably significant but not unecessarily harsh leading to underestimated matches .
this was achieved by simply selecting criteria that led to expected values for the null hypothesis ( @xmath11 $ ] for @xmath14 ) , of about 1% for the aia to sst cases . as mentioned in section .
[ analysis ] , this ended up being a minimum of 100 pixel overlap or 5% the size of the structure being matched , times the filling factor of the map the structure is being matched against , whichever is higher .
the 5% threshold mostly applied to unusually large 304 structures visible shooting from the the lower rosette ( e.g. fig.[lower ] ) .
everything considered , we successfully reject our null hypothesis of the computed matches in tables [ table : stats2 ] and [ table:171 ] being due to the random coincidences across all maps , for our selected criteria and using non distribution - dependent hypothesis testing .
we find that 11% of the 304 detected brightenings match rbes , with a 9% match between 304 and rres .
the @xmath15 lower confidence levels using gaussian statistics are 9% and 7% , respectively ( see table . [
table : stats2 ] ) .
computing the match between 304 and any re we obtain a value of 14% , which is not significantly higher than the rbe match value if one considers the respective increase in the expected value from noisy matches as measured by h@xmath7 ( table [ table : chernoff ] ) .
rbes and rres are often in the vicinity of each - other so it is not surprising that the match with res is not a sum of the matches from rbes and rres . for 171 ,
about 6% of the detections can be traced back to rbes and 5% to rres ( see table .
[ table:171 ] ) .
while these values may appear too low to claim that a relation exists , they are highly significant as we have completely invalidated the null hypothesis of this result being due to chance by using distribution independent chernoff bounds , and with expected matches due to chance of just @xmath16 for both 171 to rbes and 171 to rres . to our knowledge
, it is the first time that such a statistical match has been demonstrated between chromospheric jet - like features and coronal events .
further , to our knowledge , it is the first time an estimate for the relation between signatures in the corona in the 0.8 mk region and quiet - sun chromospheric events is proposed .
higher match values would be obtained by selecting lower detection thresholds for the 171 difference maps or by lowering the match - by - overlap criteria . however , any of the latter would translate into a higher expected value from random matches in our robust null - hypothesis testing . to put this into other words : the main effect of the amount of detections due to noise in 304 and 171 is to weigh on the selection of our match criteria following the procedure discussed in section [ statistics ] , and thus lower the computed number of detections , rather than lead to a lower significance of the result .
thus , we speculate that future higher - resolution , higher - aperture observations of the corona will reduce the noise of automated detection algorithms and thus lead to higher statistical estimates of the 304/171 to chromospheric re matches . finally , we computed the match percentage of 171 with any type of re for which we obtain a value of 10% . similarly to the 304 case , this is not significantly higher than the 171 to rbe match percentage ( 6% ) if one simply subtracts the respective increase in the expected value from noisy matches ( 3% ) .
we find that about 25% of all re features can be matched with both 304 and 171 signatures if one removes the expected values from chance alone from the measured matches ( see tables 1 through 3 ) . for the h@xmath0 detection maps the contrast is very high and thus , unlike for 304 and 171
, the absolute number of events detected is a relevant result .
we have detected 974 rbes and 864 rres .
a few of these structures may be part of the same feature but disconnected due to the common close proximity of rbes and rres , leading to a simple blue - wing detection to be split into two rbes because half way there is also overlapping red - wing signal .
the blue wing and red wing features detected were 867 and 759 respectively which we believe is a more reliable number for the actual amount of transients observed .
we generally find little difference between our wing detection - maps and the re maps . finally , this close proximity of rbes and rres , more visible in the lower rosette ( fig.[lower ] and time - series associated with fig .
[ movie1 ] ) , may be further evidence of the twist observed by @xcite and @xcite .
it is also further evidence ( together with the work of @xcite ) that rbes and rres are likely aspects of the same phenomena .
the total volume in time and space of the rbes and rres , divided over the total three - dimensional cube volume , gives a filling factor of 1.3% and 1% respectively .
this value should be useful for future studies on the importance of these structures for the quiet - sun dynamics .
the estimated filling factor is consistent with the filling factor of type ii spicules obtained by @xcite .
they assumed type ii spicules , observed with hinode in ca ii h filter , to be randomly distributed along the boundaries of circular supergranules with each spicule having a diameter of 0.1 mm , and obtained an area filling factor of 1.5 @xmath17 @xcite . for comparision , an estimate for the filling factor of type i spicules in the quiet sun chromosphere is 4 - 5@xmath17 @xcite .
@xcite , using high resolution h@xmath0 data , found an upper limit for the filling factor of chromospheric mottles , which are structures considered to be on - disk counterparts of type i spicules , of @xmath18 .
figures 4 through 7 depict selected regions of interest ( rois ) from the time - series associated with fig .
[ movie1 ] described in section [ analysis ] .
they display contours around the 304 detections , red features where rres were detected and blue features for rbe detections , with the intensity being proportional to the contrast of such detections . the time - stamps and the coordinates are in the same parameter space as the time - series associated with fig .
[ movie1 ] .
visually , the best matches between the h@xmath0 features and 304 seem to be at the extremities of structures . some matches , such as the one shown in fig .
[ lower ] at coordinates x=21and y=10 , are very striking with a progression of res into a large co - evolving 304 plume .
however , these visual matches have nearly no overlap and many were missed by our overlap criteria .
it is likely that the 11% matching value would be increased if we match borders of 304 regions with the tips or the h@xmath0 detections instead of matching overlaps over a large area .
such work , a computer vision problem , is being undertaken but the simple overlaps presented in this paper are significant and might indicate areas of multi - thermal gas or simply transition areas where the contribution for the opacity of h@xmath0 and 304 intensity happens to be significant .
( note that the formation of h@xmath0 fibrils is still largely unexplained , including formation temperatures ) .
similarly , we see more structures visually matched but not overlapping for rbes than rres . this observation may indicate that rres tend to have a more multi - thermal structure whereas rbes may become hotter `` sooner '' in the jet progression , which would then lead to higher overlaps between 304 and rres but more rbes disappearing into 304 detections .
however , the level of detections of rbes and rres are comparable in all aspects , with a slightly lower match of 304 to rres ( 2 percentage points lower ) but also slightly fewer rres detected ( leading to one percentage point lower matches expected due to randomness alone ) .
multiple 304 brightenings as well as res are visible quasi - periodically above and in the vicinity of the rosettes .
such is the case for the rois shown in figures [ lower ] and [ upper ] .
the roi in fig.[upper ] shows the last of a series of four pulses with only the latter showing re counterparts . these counterparts , however , have very high contrast and are centered inside the 304 contours
. the depicted pulse appears to have two lobes , with the lower one enveloping an rre and the upper an rbe .
both the rre and rbe are less than 0.6wide and would have been missed in lower resolution studies . as visible in the first 3 frames , the signatures in 304 appear before the res .
we observe several instances of this effect , even when the h@xmath0 feature seems to be propagating towards the 304 contour .
this , together with the quasi - periodic nature of the four 304 pulses , matched with the occasional re counterpart , at the same location and with the same propagating direction , suggests that the res and the 304 brightenings have a common generating mechanism .
furthermore , this mechanism may generate jets at temperatures higher than the h@xmath0 formation ranges at chromospheric heights .
brightenings also occur over areas that do not have any obvious relation with the rosettes nor any bright point concentrations .
figures [ norosette ] and [ awayright ] show three striking examples where 304 detections are matching rres and rbes .
this type of events have not , as far as we are aware , been detected before .
the first row of fig .
[ norosette ] shows a small rbe blob with a 304 counterpart of similar dimensions at the bottom left corner .
as may be seen from the intensity scale , the rbe has very high contrast and would have been detected even with very aggressive thresholding .
it also has a very small elongation even though it has a well defined propagation direction .
this example shows that any elongation criteria for rbe characterization would have removed features such as this .
it also demonstrates that we are capturing very small - scale events and that these are important to the visible features of the lower corona .
the second row of fig.[norosette ] shows a group of res with an adjacent and slight overlapping co - evolving 304 detection . in the second to final frame , 3 small 304 signals are visible at the tips of the central rbes .
for the roi in fig .
[ awayright ] , the match between h@xmath0 and 304 occurs at the final location of the h@xmath0 signature .
the h@xmath0 structure in fig . [ awayright ] , in the time - series , has the appearance of an arch as seen from above , and the composing features of the arch appear to converge from the extremities to the centre . in the centre of the arch , at the end of the lifetime of the structure , an overlapping 304
signature is visible .
this structure does not look like a jet but has spectral properties similar to those of an rbe as well as the transverse movements observed by @xcite .
note that the aia detected region is only 3 aia pixels wide and , with the time slices , barely large enough in our detection criteria to be counted as a match .
we believe that such features , located in unremarkable locations away from any network elements , would be ubiquitous in the sun .
the 171 detection map is noisier than the others . in this map
we focus on the most powerful events , and we observe multiple examples of matches across 304 , 171 and res .
an example of this is clear in fig.[movie1 ] at the coordinates x=40y=40as well as in the associated time - sequence .
the detections are elongated across all channels , along a diagonal in a very isolated region , and the 304 detection seems to appear first at frame t=08:44 followed by the rbe and the 171 signal that appear simultaneously . in the same location but at frame t=09:07 , we have again a detection in 304 and 171 but without an rbe counterpart . rather , this event seems to have a faint rre counterpart delayed in time that we tentatively attribute to coronal rain . examining the centre of the rosettes we can find more examples due to the quasi - periodic stream of detections across all channels .
one such example is visible in the lower rosette at t=9:19 of the time - series associated with fig.[movie1 ] ( coordinates x=30 y=10 ) which would correspond to looking straight - down at the centre of the fanning canopy .
this can easily be interpreted as the jet structure , directed at the observer , being visible at different heights as it is heated to coronal temperatures . around the same timeframe
one can see propagating events in the canopy , in all maps , at coordinates of x=20
at this location , detections in all channels seem to repeat throughout the time - series . together with quasi - periodic pulses shown in fig .
[ upper ] and discussed in section [ discussion ] , these multi - channel detections with a periodicity amount to evidence that jets exist that are generated by the same mechanisms but at a wide range of temperatures , including coronal temperatures , at very low geometrical heights .
0@div 0@div 0@div 0@div
we confirm statistically , for the first time , that there is a relation between strongly doppler - shifted h@xmath0 transients at small scales in the quiet sun and heating signatures in the tr and corona .
noise limits our measured match - levels by design , limited by statistic significance , thus our values should be taken as lower bounds .
it should be noted that a match implies overlapping in time as well as in space .
further , we find small features , matched in both h@xmath0 and in the tr , that occur over very quiet sun without any obvious bright photospheric - footpoints .
these two results show that frequent small - scale transient - flows in the chromosphere , including but perhaps not limited to type ii spicules , are indeed a significant contributor to the heating and dynamics of the upper atmospheric layers of the sun including the corona .
although this has been discussed before , we show for the first time that these chromospheric transients are continuously reaching the upper layers of the solar atmosphere at very unremarkable regions of the sun and provide values for the minimum amount of coronal events that can be traced back to chromospheric , quiet - sun res .
we further propose that res and coronal transients often share a common source that generates repeated events but at different temperatures in each pulse .
this is based on our direct observations of 304 signatures flowing out of a bright - point that sometimes , but not always , have a very clear h@xmath0 re counterpart .
such scenario would be consistent with jets of material being heated to low coronal temperatures at geometrical heights traditionally associated with the chromosphere , with some of these jets completely bypassing any meaningful intermediate signature at h@xmath0 formation temperatures .
this work was first presented at the hinode 9 conference , held at queen s university belfast in september 2015 .
we would like to thank peter stterlin for assisting with the observations .
we would also like to acknowledge support from robert ryans with computing infrastructure .
this work made use of an idl port of robert shine s routines made by tom berger .
solarsoft was used .
the swedish 1-m solar telescope is operated on the island of la palma by the institute for solar physics ( isp ) of stockholm university in the spanish observatorio del roque de los muchachos of the instituto de astrofsica de canarias .
this research was supported by the solarnet project ( www.solarnet-east.eu ) , funded by the european commissions fp7 capacities program under the grant agreement 312495 .
this work has been supported by the uk science and technology facilities council ( stfc ) .
would like to acknowledge funding from the european commissions seventh framework programme ( fp7/2007- 2013 ) under grant agreement no .
606862 ( f - chroma ) . , m. g. 2002 , in presented at the society of photo - optical instrumentation engineers ( spie ) conference , vol .
4792 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
j. bones , m. a. fiddy , & r. p. millane , 146155 , g. b. , bjelksjo , k. , korhonen , t. k. , lindberg , b. , & petterson , b. 2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4853 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
s. l. keil & s. v. avakyan , 341350 , g. b. , dettori , p. m. , lofdahl , m. g. , & shand , m. 2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
4853 , society of photo - optical instrumentation engineers ( spie ) conference series , ed . s. l. keil & s. v. avakyan , 370380 | rapid blue- and red - shifted excursions ( rbes and rres ) are likely to be the on - disk counterparts of type ii spicules .
recently , heating signatures from rbes / rres have been detected in iris slit - jaw images dominated by transition - region lines around network patches .
additionally , signatures of type ii spicules have been observed in aia diagnostics . the full - disk , ever - present nature of the aia diagnostics should provide us with sufficient statistics to directly determine how important rbes and rres are to the heating of the transition region and corona .
we find , with high statistical significance , that at least 11% of the low - coronal brightenings detected in a quiet - sun region in 304 , can be attributed to either rbes or rres as observed in h@xmath0 , and a 6% match of 171 detected events to rbes or rres with very similar statistics for both types of h@xmath0 features .
we took a statistical approach that allows for noisy detections in the coronal channels and provides us with a lower , but statistical significant , bound .
further , we consider matches based on overlapping features in both time and space , and find strong visual indications of further correspondence between coronal events and co - evolving but non - overlapping , rbes and rres . 1.0truecm |
in the limit of small electric and thermal bias away from the equilibrium temperature @xmath7 and chemical potential @xmath8 , the electric current @xmath9 and heat current @xmath10 flowing into the probe may be expressed as @xcite @xmath11 where @xmath12 are onsager linear - response coefficients with electrode labels @xmath13 and @xmath14 , and @xmath15 is the thermal conductance between the probe and the ambient environment @xcite .
( [ eq : three_terminal_currents ] ) is a _ completely general _ linear - response formula , and applies to macroscopic systems , mesoscopic systems , nanostructures , etc . , including electrons , phonons , and all other degrees of freedom , with arbitrary interactions between them . at sufficiently low temperatures or for sufficiently small systems , the electronic contribution to the coefficients @xmath12
may be calculated using elastic quantum transport theory @xcite @xmath16 where @xmath17 is the quantum mechanical transmission function @xcite describing the probability to propagate from electrode @xmath14 to electrode @xmath13 , and @xmath18 is the equilibrium fermi - dirac distribution . in the limit as the system temperature approaches absolute zero , eq .
( [ eq : lnu_elastic ] ) becomes @xmath19 then eqs . ( [ eq : equil0 ] ) and ( [ eq : three_terminal_currents ] ) may be solved to obtain bttiker s result @xcite for the voltage measured by the probe @xmath20 the question remains , how to generalize bttiker s result ( [ eq : mu_p_mb ] ) to systems at non - zero temperatures .
early on , engquist and anderson @xcite considered both voltage and temperature probes of quantum electron systems at finite temperature .
for the case of a voltage measurement , they assumed the entire system remains at ambient temperature @xmath21 , so that eqs .
( [ eq : equil0 ] ) and ( [ eq : three_terminal_currents ] ) imply @xmath22 however , substituting eq .
( [ eq : ea_voltage ] ) for the probe s chemical potential into eq .
( [ eq : three_terminal_currents ] ) gives @xmath23 which is generally non - zero at finite temperature .
this is a generic three - terminal thermoelectric effect occuring whenever the probe coupling to the source and drain electrodes ( through the system ) is unequal .
thus the voltage probe originally proposed by engquist and anderson is _ not in thermal equilibrium _ with the system . in the absence of thermal equilibrium ,
the identification of @xmath24 with the local electrochemical potential of the system is problematic , since any temperature differential between sample and probe will lead to a voltage differential through the seebeck effect .
moreover , the assumption that @xmath25 is inconsistent , given that @xmath26 , unless the thermal coupling of the probe to the environment is so large that the heat current flowing into the probe from the system can be neglected .
we define an _ ideal voltage measurement _ as one in which the probe is in _ both electrical and thermal equilibrium _ with the system . for an electrical bias @xmath27 applied between electrodes 1 and 2 ,
both held at ambient temperature ( @xmath28 ) , eqs .
( [ eq : equil0][eq : three_terminal_currents ] ) can be solved for the probe voltage of such an ideal measurement , yielding @xmath29 , where the thermoelectric correction to the voltage is @xmath30 @xmath31 is the thermopower of the probe - sample junction , and @xmath32 is the probe temperature satisfying @xmath33 where @xmath6 is given by eq .
( [ eq : ip1_ea ] ) , @xmath34 \label{eq : kappa_p}\ ] ] is the parallel thermal conductance from electrodes 1 and 2 into the probe , and @xmath35 is the thermal coupling of the probe to the environment at temperature @xmath7 .
in this section , we calculate the thermoelectric correction to the probe voltage for a prototypical ballistic quantum conductor , a graphene nanoribbon . however , we emphasize that the voltage error induced by thermoelectric effects is a generic phenomenon , and not material specific .
figure [ fig : cnr_voltage_voltage_error ] shows the computed voltage distribution for a zig - zag graphene nanoribbon with an electrical bias of 0.1v between the source and drain electrodes ( at right and left in the figure ) , which are held at the ambient temperature of @xmath36 .
the equilibrium chemical potential of the nanoribbon ( determined by doping and/or a backgate ) was taken as @xmath37=-57.5mev . in our calculations ,
the @xmath38-system of the graphene nanoribbon is described using a tight - binding model which has been shown to accurately reproduce the low - energy physics of this system @xcite .
the macroscopic electrodes are assumed to operate in the broad - band limit , where the electrode - nanoribbon coupling is independent of energy , with a per - orbital bonding strength of 2.5ev .
the voltage probe is modeled as an atomically - sharp pt tip scanned at a fixed height of 3 above the plane of the c nuclei ( tunneling regime ) .
the tunneling matrix elements between the probe atoms and the nanoribbon were determined using the methods outlined in ref . .
the linear - response coefficients were calculated using eq .
( [ eq : lnu_elastic ] ) following the methods of refs . @xcite .
additional details of our computational methods may be found in the supporting information .
the top panel of fig .
[ fig : cnr_voltage_voltage_error ] shows the engquist - anderson voltage @xmath39 computed from eq .
( [ eq : ea_voltage ] ) , while the bottom panel shows the thermoelectric correction @xmath40 to the probe voltage , computed from eqs .
( [ eq : delta_vp][eq : kappa_p ] ) . for this case , which is representative of various geometries
we have considered ( see supporting information ) , the thermoelectric correction to the measured voltage is @xmath0 of the maxiumum voltage and @xmath41 of the applied bias , highlighting the importance of thermoelectric effects on precision voltage measurements in quantum systems . as mentioned previously ,
this system is not unique and even larger corrections are expected for systems with larger thermoelectric responses .
the cause of the substantial thermoelectric correction to the voltage is elucidated in fig .
[ fig : cnr_current_temp_diff ] .
the top panel of fig .
[ fig : cnr_current_temp_diff ] shows the heat current @xmath6 flowing into the probe when its temperature is held fixed at @xmath7 , calculated using eq .
( [ eq : ip1_ea ] ) .
the peak values of @xmath42 may not be large in an absolute sense , but they correspond to a heat current density of @xmath43 through the apex atom at the tip of the probe , some 700 times the radiant energy flux at the surface of the sun ! clearly , the assumption that such a probe , whose voltage is given by eq .
( [ eq : ea_voltage ] ) , is in local equilibrium with the system is questionable . the bottom panel of fig .
[ fig : cnr_current_temp_diff ] shows the deviation of the temperature @xmath32 of an ideal thermoelectric probe from ambient temperature , calculated from eq .
( [ eq : tp ] ) .
the ideal probe is in local thermal equilibrium with the system , and as such , its temperature maps out the hot and cold regions of the system @xcite . the lower panel of fig .
[ fig : cnr_current_temp_diff ] shows clear evidence of peltier cooling / heating of up to @xmath44 within the system induced by the external electrical bias of 0.1v .
the large peltier effect in this system may be related to giant thermoelectric effects predicted in related @xmath38-conjugated systems @xcite , where quantum interference effects have been shown to strongly enhance thermoelectricity .
however , similar phenomena should occur in other ballistic quantum conductors .
flowing into the probe when it is held at the ambient temperature @xmath36 , calculated from eq .
( [ eq : ip1_ea ] ) .
bottom panel : the temperature @xmath32 of the probe when it is in both electrical and thermal equilibrium with the nonequilibrium electron system in the graphene nanoribbon , calculated from eq .
( [ eq : tp ] ) .
, width=336 ] to the probe voltage ( top panel ) and the deviation of the probe temperature from ambient temperature ( bottom panel ) for the same system shown in figs .
[ fig : cnr_voltage_voltage_error][fig : cnr_current_temp_diff ] , but with a finite thermal coupling @xmath45 of the probe to the environment , where @xmath46 ( 0.284nw / k at 300k ) is the thermal conductance quantum . eqs .
( [ eq : delta_vp ] ) and ( [ eq : tp ] ) indicate that the thermoelectric corrections for larger values of @xmath35 scale as @xmath47 . ,
width=336 ] let us now consider the effects of measurement non - ideality . the greatest source of error in a scanning thermoelectric measurement is likely to stem from the unavoidable coupling @xmath35 of the probe to the thermal background ( typically , the ambient environment ) @xcite . indeed
, state - of - the - art sthm still operates in the regime where the coupling of the probe to the thermal background is many times its thermal coupling to the system itself @xcite . while values of @xmath35 much less than the thermal conductance quantum @xmath46 ( 0.284nw / k at 300k ) @xcite are possible in principle for probes whose thermal coupling to the environment is predominantly radiative @xcite , current scanning probes @xcite have @xmath48 .
figure [ fig : cnr_voltage_temp_with_kappa_p0 ] shows the thermoelectric correction to the voltage ( upper panel ) and the probe temperature ( lower panel ) for @xmath49 . for this case , the thermal coupling of the probe to the environment exceeds its coupling to the system , so that the probe temperature is closer to ambient , and the thermoelectric correction to the voltage is reduced .
the reduction of the thermoelectric corrections is described analytically by eqs .
( [ eq : delta_vp ] ) and ( [ eq : tp ] ) . even for a thermal coupling of @xmath50 ,
typical of current state - of - the - art sthm @xcite , the voltage error would still be of order 1@xmath51v , well within the resolution of precision voltage measurements , which routinely obtain sub - ngstrom spatial resolution @xcite .
an ideal voltage measurement in a nonequilibrium quantum system was defined in terms of a floating thermoelectric probe that reaches both electrical and thermal equilibrium with a system via local ( e.g. , tunnel ) coupling .
this definition extends bttiker s quantum voltage probe paradigm @xcite to systems at finite temperature , where thermoelectric effects are important . as an example
, we developed a realistic model of a scanning potentiometer with atomic resolution and used it to investigate voltage measurement in a prototypical ballistic quantum conductor ( a graphene nanoribbon ) bonded to source and drain electrodes .
under ideal measurement conditions , we predict large thermoelectric voltage corrections ( @xmath5224% of the probe s peak voltage signal ) when the applied source - drain bias voltage is small .
we also derived expressions for the probe s voltage correction under non - ideal measurement conditions , finding that the voltage correction is reduced linearly as the probe - environment coupling is increased . in the graphene nanoribbon system
considered here , voltage corrections on the order of several @xmath51v persist even with strong environmental coupling . in summary , we predict a large thermoelectric correction to voltage measurement in quantum coherent conductors .
the origin of this correction is local peltier cooling / heating within the nonequilibrium quantum system , a generic three - terminal thermoelectric effect .
this finding has important implications for precision local electrical measurements : it implies that _ a precision voltage measurement requires a simultaneous precision temperature measurement_. this paper is dedicated to the memory of markus bttiker , whose ideas continue to inspire our quest to understand nonequilibrium quantum systems .
was supported by the department of energy
basic energy sciences grant no .
de - sc0006699 .
25ifxundefined [ 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/physrevb.24.1151 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.40.3409 [ * * , ( ) ] link:\doibase 10.1103/physrevb.75.035340 [ * * , ( ) ] @noop _ _ , vol .
( , ) link:\doibase 10.1103/physrevlett.57.1765 [ * * , ( ) ] link:\doibase 10.1103/physrevb.46.9648 [ * * , ( ) ] link:\doibase 10.1038/35075009 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.196802 [ * * , ( ) ] link:\doibase 10.1007/s10955 - 009 - 9697 - 1 [ * * , ( ) ] link:\doibase 10.1146/annurev.matsci.29.1.505 [ * * , ( ) ] link:\doibase 10.1021/nn300774n [ * * , ( ) ] link:\doibase 10.1021/nn401027u [ * * , ( ) ] @noop ( ) , link:\doibase 10.1021/nn2026325 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop _ _ , ed .
( , ) link:\doibase 10.1103/physrevb.90.035407 [ * * , ( ) ] @noop * * , ( ) | a generalization of bttiker s voltage probe concept for nonzero temperatures is an open third terminal of a quantum thermoelectric circuit .
an explicit analytic expression for the thermoelectric correction to an ideal quantum voltage measurement is derived , and interpreted in terms of local peltier cooling / heating within the nonequilibrium system .
the thermoelectric correction is found to be large ( up to @xmath0 of the peak voltage ) in a prototypical ballistic quantum conductor ( graphene nanoribbon ) .
the effects of measurement non - ideality are also investigated .
our findings have important implications for precision local electrical measurements .
following the work of engquist and anderson @xcite , markus bttiker developed a paradigm @xcite of quantum voltage measurement carried out by a probe consisting of a reservoir of non - interacting electrons coupled locally to a system of interest .
the probe exchanges electrons with the system until it reaches local electrical equilibrium with the system : @xmath1 where @xmath2 is the mean electrical current flowing into the probe . once equilibrium is established , the chemical potential @xmath3 of the probe constitutes a measurement of the local electrochemical potential ( and voltage @xmath4 ) within the nonequilibrium quantum system @xcite .
the condition ( [ eq : equil0 ] ) implies the probe has a large electrical input impedance , a necessary condition for a faithful voltage measurement . scanning potentiometers satisfying these conditions @xcite are now a mature technology , and many experiments in mesoscopic electrical transport utilize voltage probes as circuit components @xcite .
although the average electric current into the probe is zero , electrons are constantly being emitted from the system into the probe , and replaced by electrons from the probe reservoir whose quantum mechanical phase is uncorrelated with those emitted by the system . in this way , such a voltage probe serves as an _
inelastic scatterer
_ @xcite .
indeed , much of the theoretical interest in bttiker s model of a voltage probe is as a convenient way to introduce inelastic scattering in a quantum coherent conductor at the expense of introducing one additional electrical terminal .
bttiker s early analysis @xcite was confined to systems at absolute zero temperature , where thermoelectric effects are absent .
later , voltage probes at finite temperature were considered in the limit where the thermal coupling of the probe to the environment is large , so that the probe remains at ambient temperature despite its coupling to the nonequilibrium quantum system @xcite .
this limit is consistent with the original analysis of engquist and anderson @xcite , which did not consider thermoelectric effects .
however , considered as a model of an inelastic scatterer , a voltage probe can not be a steady - state source or sink of heat @xcite .
this suggests that in generalizing the voltage probe concept @xcite to finite temperatures , the probe should be not only in local electrical equilibrium , but also in local thermal equilibrium with the system : @xmath5 where @xmath6 is the heat current flowing into the probe .
condition ( [ eq : equil1 ] ) is required for a probe with a large thermal input impedance .
further support for the additional condition ( [ eq : equil1 ] ) is provided by considering thermoelectric effects in the three - terminal circuit formed by the system with source , drain , and probe .
even if both source and drain electrodes are held at ambient temperature , an electrical bias between source and drain can drive peltier cooling / heating within the system , resulting in hot and cold spots differing significantly from ambient temperature .
if the probe is not allowed to equilibrate thermally with the system under these conditions , a voltage will develop across the system - probe junction due to the seebeck effect .
then the probe voltage can no longer be interpreted as a measurement of the local electrochemical potential in the system .
we thus define an _ ideal voltage measurement _ as one satisfying both conditions ( [ eq : equil0 ] ) and ( [ eq : equil1 ] ) .
a precision voltage measurement thus requires a simultaneous precision temperature measurement .
a significant challenge to achieving such an ideal voltage measurement is posed by thermal coupling of the probe to the environment @xcite , including to the system s lattice @xcite , which may not be in local thermal equilibrium with the nonequilibrium electron system .
furthermore , this coupling may be many times as large as the probe s local thermal coupling to the system s electrons @xcite .
the probe s thermal coupling to anything other than the nonequilibrium electron system of interest leads to a deviation of the probe s voltage from the ideal value associated with the local electrochemical potential of the system , and thus must be considered a non - ideality .
the probe s thermal coupling to the system s lattice can be minimized when it is operated in the tunneling regime @xcite , and continued advances in scanning thermal microscopy ( sthm ) @xcite promise to further reduce the probe s thermal coupling to the environment . |
quasars are among the most luminous sources in the universe and can be therefore detected and investigated at huge distances and early cosmic times .
for this reason , their study can be used to probe a variety of phenomena from cosmology models to gas content of the intergalactic medium ( from intervening absorption systems ) and for an understanding of many aspects of the processes of the formation and evolution of galaxies .
the latter issue gained a lot of interest with the discovery that most of massive galaxies host a dormant supermassive black hole ( smbh , di matteo et al .
2005 ; callegari et al .
2011 , and references therein ) .
this means that accretion onto a smbh , that is the most likely mechanism for the qso phenomenon , is strictly linked to the whole process of formation of galaxies .
the coevolution of the smbh and their host galaxies therefore represents a fundamental issue in the modern extragalactic studies .
while direct dynamical measurements of the mass of the bh can be derived only for the nearest galaxies ( see peterson 2014 for a review ) , for more distant objects ( [email protected] , e.g. koekemoer et al .
2009 ) it is possible to deduce the bh mass only from the relationships , derived from a small number of nearby galaxies , between the bh mass ( m(bh ) ) and the global properties of the galaxies ( e.g. stellar velocity dispersion , luminosity of the spheroidal component ; see e.g. ferrarese & merritt 2000 , gebhardt et al .
2000 ; tremaine et al .
2002 , marconi & hunt 2003 , haring & rix 2004 , gultekin et al .
2009 ) . in the case of galaxies with active nuclei ( as the case of qsos )
it is possible to estimate directly the bh mass from the dynamical properties of the gas that is under the sphere of influence of the central bh .
this method was applied by various authors to derive m(bh ) of several qsos at various redshift ( e.g. mclure & dunlop 2004 , shen et al . 2013 and references therein ) .
the direct measurements of both bh mass and the mass ( luminosity ) of their host galaxies allow to investigate the relation between these two fundamental ingredients and to probe their evolution with the cosmic time . while for the m(bh ) estimation of the qso the spectroscopic observation ( provided if a sufficient signal - to - noise is reached ) is able to gather a measurement of bh mass virtually at any redshift , in the case of the properties of the qso host galaxies the measurement is increasingly more challenging for higher redshift .
this is because the contrast between the bright central nucleus and the starlight from the host galaxy critically depends on the size and shape of the latter compared with that of the point spread function ( psf ) . indeed , for a robust decomposition of the nuclear and host galaxy emission of a qso , it is extremely important to obtain deep images with as narrow as possible psf in order to reduce the emission contribution of the central source .
this can be done either from space based observations , as from the hubble space telescope ( hst , see e.g. disney et al .
1995 , bahcall et al .
1997 , mclure et al . 1999
; mcleod & mcleod 2001 , dunlop et al . 2003 ) or using ground based imaging under excellent seeing conditions ( hutchings et al . 1999 ,
mrquez et al .
2001 , lacy et al .
2002 , kotilainen et al .
this situation becomes particularly critical for high redshift sources because of the faintness of the starlight emission and/or the very small size of the galaxies .
observations from space with hst can indeed provide an excellent narrow psf but because of the faintness of high z host galaxies they are usually limited by the small collecting area of the telescope and the limited performances in the near - ir ( e.g. kukula et al .
2001 , peng et al .
the ground based near - ir imaging of high - z quasars , obtained with large telescopes under excellent seeing conditions , can compete with space based observations because of the significantly larger collecting power and thus allow to characterize the properties of distant qsos ( e.g. falomo et al .
2004 , kotilainen et al .
2009 ) to further explore quasar hosts at high - z the use of adaptive optics , that produces narrow psf , is therefore seen as a natural cure for reducing the light of the central source and to emphasize the more extended emission of its host galaxy and their features in the very close environments . in the last decade
adaptive optics ( ao , see davies & kasper 2012 for a review ) images of qsos were obtained either to investigate with better spatial resolution the immediate environment of low redshift objects ( e.g. mrquez et al .
2001 , 2003 ; guyon et al .
2006 , rosario et al . 2011 ) or to characterize the properties of the host galaxies of high redshift ( z @xmath1 23 ) quasars ( e.g. croom et al .
2004 ; hutchings et al . 1999 , 2004 ;
falomo et al .
2005 , 2008 ; kuhlbrodt et al .
2005 ; wang et al .
2013 ) . in the latter case
the quality of the psf and its characterization are essential ingredients to disentangle the light of the host galaxy from that of the bright central nucleus .
the large majority of the ao imaging systems requires the presence of a bright star close ( @xmath2 10 - 30 arcsec ) to the target in order to derive the signal of the wavefront perturbation and apply the adaptive corrections .
this implies that only the targets that are angularly very close to relatively bright stars could be actually observed with ao systems .
this limit could be overcome by some present ( e.g. keck , wizinowich et al .
2006 ) and future ao instruments using artificial laser guide stars .
another approach to cope with this limitation is to use many natural guide stars ( ngss ) in the field of view of the target as reference for ao correction and more than one deformable mirror conjugated to different altitudes of the perturbed wavefront . with
this multi conjugated adaptive optics ( mcao ) it is possible to obtain a more stable and uniform psf over the observed field of view and to increase significantly the sky coverage even at high galactic latitudes ( moretti et al .
2009 , arcidiacono et al .
2010 , meyer et al .
2011 ) .
in this paper we report on observations of high - z luminous qsos gathered with multi - conjugate adaptive optics demonstrator ( mad ) at the european southern observatory ( eso ) very large telescope ( vlt ) .
other examples of nearby active galaxies observed by mad have been presented by falomo et al .
( 2009 ) and liuzzo et al .
( 2011 , 2013 ) .
the selected qsos are high - luminous sources with v magnitude ranging between 18.6 and 20.1 .
the detection of the host galaxy in those objects appears therefore more difficult with respect to low - luminous sources , because of the presence of very bright central nuclei . due to non optimal observing conditions all
qsos but one ( observed in the conventional ao mode ) were observed in the ground layer ao ( glao ) configuration ( marchetti et al .
2007 ) using ngss with r magnitude in the range between 9.8 and 15.6 .
in particular , in glao mode a set of reference stars are available to provide the control of a deformable mirror conjugated to the ground thus allowing the correction of the optical turbulence in the lower part of the atmosphere volume above the telescope .
although glao aims just to a partial correction of the turbulence , it allows for a much larger compensated fov with respect to the conventional single conjugate ao mode ( scao ) which uses only one reference star in the field . in spite of the modest observing sky conditions , we are able to resolve the host galaxy of two objects and to set upper limits to the host luminosity for the others . finally , future perspectives for the capabilities in this field with european extremely large telescope ( e - elt ) imaging are presented .
ackermann , m. , ajello , m. , albert , a. , et al .
2013 , , 771 , 57 + ackermann , m.,ajello , m. , allafort , a. , et al .
2011 , , 741 , 30 + amorim , a. , lima , j. , alves , j. , et al .
2006 , , 6269 , 626957 + amorim , a. , melo , a. , alves , j. , et al .
2004 , , 5492 , 1699 + arcidiacono , c. , lombini , m. , moretti , a. , et al .
2010 , , 7736 , 77363d + callegari , s. , kazantzidis , s. , mayer , l. , et al .
2011 , , 729 , 85 + croom , s. m. , schade , d. , boyle , b. j. , et al .
2004 , , 606 , 126 + colless , m. , peterson , b. a. , jackson , c. , et al .
2003 , arxiv : astro - ph/0306581 + condon , j. j. , & et al . 1998 , observational cosmology with the new radio surveys , 226 , 37 + davies , r. , ageorges , n. , barl , l. , et al .
2010 , , 7735 , 77352a + davies , r. , & kasper , m. 2012 , , 50 , 305 + decarli , r. , falomo , r. , kotilainen , j. k. , et al .
2012 , advances in astronomy , 2012 , 782528 + di matteo , t. , springel , v. , & hernquist , l. 2005 , , 433 , 604 + disney , m. j. , boyce , p. j. , blades , j. c. , et al .
1995 , , 376 , 150 + dunlop , j. s. , mclure , r. j. , kukula , m. j. , et al .
2003 , , 340 , 1095 + falomo , r. , kotilainen , j. k. , scarpa , r. , & treves , a. 2005 , , 434 , 469 + falomo , r. , pian , e. , treves , a. , et al .
2009 , , 501 , 907 + falomo , r. , kotilainen , j. k. , pagani , c. , scarpa , r. , & treves , a. 2004 , , 604 , 495 + falomo , r. , treves , a. , kotilainen , j. k. , scarpa , r. , & uslenghi , m. 2008 , , 673 , 694 + falomo , r. , fantinel , d. , & uslenghi , m. 2011 , in proc .
spie , vol .
8135 , applications of digital image processing xxxiv , ed .
g. tescher , i d . 813523 + falomo , r. , bettoni , d. , karhunen , k. , kotilainen , j. k. , & uslenghi , m. 2014 , , 440 , 476 + ferrarese , l. , & merritt , d. 2000 , , 539 , l9 + francis , p. j. , hewett , p. c. , foltz , c. b. , et al .
1991 , , 373 , 465 + garofalo , d. 2013 , , 434 , 3196 + gltekin , k. , richstone , d. o. , gebhardt , k. , et al .
2009 , , 698 , 198 + gebhardt , k. , bender , r. , bower , g. , et al .
2000 , , 539 , l13 + guyon , o. , sanders , d. b. , & stockton , a. 2006 , , 166 , 89 + hring , n. , & rix , h .- w .
2004 , , 604 , l89 + hutchings , j. b. , crampton , d. , morris , s. l. , durand , d. , & steinbring , e. 1999 , , 117 , 1109 + hutchings , j. b. 2004 , the interplay among black holes , stars and ism in galactic nuclei , 222 , 523 + koekemoer , a. m. , batcheldor , d. , postman , m. , & somerville , r. 2009 , astro2010 : the astronomy and astrophysics decadal survey , 2010 , 157 + kotilainen , j. k. , falomo , r. , decarli , r. , et al .
2009 , , 703 , 1663 + kotilainen , j. k. , falomo , r. , labita , m. , treves , a. , & uslenghi , m. 2007 , , 660 , 1039 + kovalev , y. y. 2009 , , 707 , l56 + kuhlbrodt , b. , rndahl , e. , wisotzki , l. , & jahnke , k. 2005 , , 439 , 497 + kukula , m. j. , dunlop , j. s. , mclure , r. j. , et al .
2001 , , 326 , 1533 + lacy , m. , gates , e. l. , ridgway , s. e. , et al .
2002 , , 124 , 3023 + liuzzo , e. , falomo , r. , treves , a. , et al .
2011 , , 528 , a34 + liuzzo , e. , falomo , r. , treves , a. , et al .
2013 , , 145 , 73 + lyu , f. , liang , e .- w . ,
liang , y .- f . , et al .
2014 , , 793 , 36 + mrquez , i. , petitjean , p. , thodore , b. , et al .
2001 , , 371 , 97 + mrquez , i. , & petitjean , p. 2003 , revista mexicana de astronomia y astrofisica conference series , 16 , 135 + mannucci , f. , basile , f. , poggianti , b. m. , et al .
2001 , , 326 , 745 + marchetti , e. , brast , r. , delabre , b. , et al .
2005 , comptes rendus physique , 6 , 1118 + marchetti , e. , brast , r. , delabre , b. , et al .
2007 , the messenger , 129 , 8 + marconi , a. , & hunt , l. k. 2003 , , 589 , l21 + mcleod , k. k. , & mcleod , b. a. 2001 , , 546 , 782 + mclure , r. j. , kukula , m. j. , dunlop , j. s. , et al . 1999 , , 308 , 377 +
mclure , r. j. , & dunlop , j. s. 2004 , , 352 , 1390 + moretti , a. , piotto , g. , arcidiacono , c. , et al .
2009 , , 493 , 539 + meyer , e. , krster , m. , arcidiacono , c. , ragazzoni , r. , & rix , h .- w .
2011 , , 532 , a16 + murphy , t. , sadler , e. m. , ekers , r. d. , et al .
2010 , , 402 , 2403 + nakamura , o. , fukugita , m. , yasuda , n. , et al .
2003 , , 125 , 1682 + peng , c. y. , ho , l. c. , impey , c. d. , & rix , h .- w .
2002 , , 124 , 266 + peng , c. y. , impey , c. d. , rix , h .- w . , et al .
2006 , , 649 , 616 + peterson , b. m. 2014 , , 183 , 253 + ragazzoni , r. , & bonaccini , d. 1996 , european southern observatory conference and workshop proceedings , 54 , 17 + ragazzoni , r. , farinato , j. , & marchetti , e. 2000 , , 4007 , 1076 + rosario , d. j. , mcgurk , r. c. , max , c. e. , et al .
2011 , , 739 , 44 + shen , y. , liu , x. , loeb , a. , & tremaine , s. 2013 , , 775 , 49 + tremaine , s. , gebhardt , k. , bender , r. , et al .
2002 , , 574 , 740 + uslenghi m. , and falomo r. , proc .
6th international workshop on data analysis in astronomy , modelling and simulation in science hackensack , nj : world scientific ; 2008:313 , di ges v. , bosco g. l. , maccarone m. c. , editors .
+ wang , y. p. , yamada , t. , tanaka , i. , iye , m. , & ji , t. 2013 , iau symposium , 292 , 195 + wizinowich , p. l. , le mignant , d. , bouchez , a. h. , et al .
2006 , , 118 , 297 | we present near - ir images of five luminous quasars at z@xmath02 and one at z@xmath04 obtained with an experimental adaptive optics instrument at the eso very large telescope .
the observations are part of a program aimed at demonstrating the capabilities of multi - conjugated adaptive optics imaging combined with the use of natural guide stars for high spatial resolution studies on large telescopes .
the observations were mostly obtained under poor seeing conditions but in two cases . in spite of these non optimal conditions ,
the resulting images of point sources have cores of fwhm @xmath0 0.2 arcsec .
we are able to characterize the host galaxy properties for 2 sources and set stringent upper limits to the galaxy luminosity for the others .
we also report on the expected capabilities for investigating the host galaxies of distant quasars with adaptive optics systems coupled with future extremely large telescopes .
detailed simulations show that it will be possible to characterize compact ( 2 - 3 kpc ) quasar host galaxies for qsos at z = 2 with nucleus k - magnitude spanning from 15 to 20 ( corresponding to absolute magnitude -31 to -26 ) and host galaxies that are 4 mag fainter than their nuclei . |
a two - dimensional electron gas ( 2deg ) confined into a semiconductor quantum well and subjected to a perpendicular magnetic field @xmath0 experiences a great diversity of ground states as the filling factor @xmath1 is varied . here
@xmath2 is the number of electrons and @xmath3 is the landau level degeneracy with @xmath4 the 2deg area and @xmath5 the magnetic length . besides the well known ( uniform ) laughlin states responsible for the fractional and integer quantum hall effects@xcite , many nonuniform phases are also possible .
these states have spatial modulations of the electronic and/or spin densities .
they have been the subject of intense theoretical and experimental work using a variety of techniques over the past 35 years .
famous examples are the wigner crystal,@xcite the bubble crystals , the quantum hall stripe staterevuebulles , microwave , yoshioka , goerbig , cotebubble , cotestripes and the skyrme crystal.@xcite experimental evidence for the crystal phases can be obtained from microwave absorption experiments@xcite which detect the gapped phonon mode of the crystal pinned by disorder . in the case of the skyrme crystal which has an exotic spin texture ,
optically pumped nuclear magnetic resonance ( op - nmr)@xcite has been particularly useful in measuring the magnetization of the crystal as a function of the filling factor and zeeman coupling as well as its low - lying collective excitations which involve two gapless ( goldstone ) modes.@xcite in op - nmr , the magnetization is obtained by measuring the knight shift ( @xmath6 ) of the @xmath7ga nuclei around @xmath8 in gaas / algaas multiple quantum wells .
this technique exploits the fermi contact hyperfine interaction that exists between the magnetic moment of the @xmath7ga nuclei in the gaas quantum wells and the spin of the electrons in the confined 2deg .
a nonzero local spin polarization of the electrons modifies the local magnetic field seen by the nuclei and shift their resonance frequency by an amount @xmath9 which is proportional to the electronic spin polarization . in a related technique : resistively detected nuclear magnetic resonancedesrat ( rd - nmr ) , the knight shift of the @xmath10as nuclei in the quantum well is obtained from the change in the longitudinal resistance @xmath11 of the 2deg .
this change , @xmath12 is related to the change in the electronic zeeman energy which is caused by the coupling via the hyperfine interaction of the average nuclear magnetic moment @xmath13 and the electron zeeman energy @xmath14 where @xmath15 is the external magnetic field and @xmath16 is a constant .
this technique has been used to study the liquid - solid phase transition of the 2deg at small filling factor in landau levels @xmath17 and @xmath18@xcite it has also been applied to the skyrme crystal ( or skyrmion liquid ) near @xmath8 where an anomalous spectral line shape is detected.desrat,gervais,gervais2,tracy,barrett the rd - nmr technique has been used recently@xcite to study the spatial modulations of the electronic density ( or , more precisely , of the spin density ) i.e. the topography of the electron solids in the integer and quantum hall regimes .
it is well known ( see for example ref . ) that the density pattern of a wigner crystal changes with the landau level index @xmath19 because the wave function of an electron in a magnetic field depends on @xmath19 .
this change in topography is reflected to some degree in the nmr spectral line shape .
indeed , in experiments , the wigner crystals detected at small filling factors in landau levels @xmath17 and @xmath20 show very different rd - nmr spectra.@xcite in the present work , we want to find out how different the spectral line shape of other crystal states ( bubble , stripe , and skyrme crystals ) are from one another and how well the rd - nmr technique can discriminate between them .
is the rd - nmr technique sufficiently sensitive , for exemple , to distinguish between bubble crystals with @xmath21 or @xmath22 electrons per site or to separate a bubble crystal from a stripe state ? to study this sensitivity
, we compare the rd - nmr spectra of several crystal phases where the quasiparticles that crystallize can be electrons , bubbles , skyrmions or the corresponding anti - quasiparticles .
the calculation of the spectral line shape requires a knowledge of the average local spin polarization @xmath23 along the direction of the applied magnetic field . here
@xmath24 is a vector in the plane of the 2deg and @xmath25 is in the perpendicular ( confining ) direction . to compute @xmath26
, we use an equation of motion method developed previously for the study of the wigner crystal.@xcite in this method , the spin polarization is computed from the hartree - fock equation of motion of the single - particle green s function .
this method has the advantage over the maki - zotos wave function@xcite that it can easily be adapted to study crystal structures with ferromagnetic or antiferromagnetic order,@xcite crystals with more complex spin textures such as meron or skyrmion crystals,@xcite or bubble and stripe crystals.@xcite a limitation of this approach , however , is that the hartree - fock approximation does not take into account the quantum or thermal fluctuations of the crystals nor the disorder . as previously discussed in the nmr litterature , desrat , gervais2,tracy these effects may alter significantly the nmrspectra . however , by introducing a gaussian blurring factor@xcite in the calculation of the crystal polarization , it is possible to model the effect of these fluctuations on the spectral line shape . in this work , we limit ourselves to computing the rd - nmr spectra at zero temperature i.e. in the ideal case of frozen crystals . in this regime ,
the nuclei in the quantum well see an effective magnetic field that depends on their location and so their knight shift varies spatially .
the rd - nmr signal is obtained by summing the knight - shifted signal of all nuclei as discussed in the next section .
our paper is organized in the following way . in secs .
ii and iii , we give a brief summary of the calculation of the rd - nmr spectral line shape for crystal states and of the spin polarization @xmath26 in the hartree - fock approximation . in sec .
iv , we describe the different crystal phases that we include in our study . in sec .
v , we compute the rd - nmr spectra of these crystals and discuss how well the rd - nmr technique allows to discriminate between them .
we conclude in sec . vi .
we consider a 2deg confined in a gaas / algaas quantum well .
the 2deg is in a perpendicular magnetic field @xmath0 that quantizes the kinetic energy into landau levels .
we define @xmath27 as the electronic density in the last partially filled landau level @xmath19
. the filling factor of that level is @xmath28 . for convenience ,
we define the dimensionless density @xmath29 of each spin specie @xmath30 in the partially filled landau level @xmath19 by@xmath31where @xmath32 is the density of electrons with spin @xmath33 .
the averaged spin polarization density in the direction of the applied magnetic field is then@xmath34where@xmath35and @xmath36 is the wave function of the first bound state of the quantum well which is the only occupied state at sufficiently low temperature .
this wave function depends on the exact shape of the quantum well which in turn depends on the potential of the gate used to control the electron density .
the notation @xmath37 stands for the quantum and thermal averages of the spin density . in this paper
, we work at zero temperature and so @xmath37 is the ground state average of the spin density .
following tiemann _
et al_. ( see supplemental material of ref . ) , we write the spectral function for the rd - nmr signal from the @xmath10as nuclei in the quantum well as@xmath38where @xmath39 is a constant and @xmath40 is the width of the quantum well . in eqs .
( [ une ] ) , @xmath41 is the subband wave function at filling factor @xmath42 used to account for the position dependence of the readout sensitivity and @xmath43 is a parameter to be determined experimentally .
the function @xmath44 is the intrinsic line shape for the @xmath10as nuclei in the crystal environment ( in the absence of the 2deg ) .
it is assumed to have the gaussian form @xmath45where @xmath46 is the spectral line shape and @xmath47 the resonance frequency . in eq .
( [ une ] ) , it is implicitly assumed that the subband wave function is negligible outside the quantum well .
this wave function leads to an asymmetrical broadening of the spectral line shape .
the variations of the electronic density in the plane of the well introduce a supplementary inhomogeneous broadening of the line shape . in our numerical calculation , we compute the wave functions @xmath48 and @xmath49 in the well at @xmath50 points @xmath51 and @xmath52 at @xmath53 points @xmath54 in the unit cell of the crystal . upon defining @xmath55(and a similar relation for @xmath56 ) which satisfies @xmath57we
get@xmath58where @xmath59 is now the constant to be found experimentally . since @xmath53 can be large in eq .
( [ trois ] ) , @xmath60 is easier to evaluate if we first compute a function giving the probability of occurrence of @xmath61 .
we find the minimal and maximal values of @xmath61 in a unit cell , divide the interval between these two values into @xmath62 bins of width @xmath63 with middle values @xmath64 and write @xmath65 for the probability of occurrence of @xmath64 .
the signal then becomes@xmath66with@xmath67for a uniform state with @xmath68@xmath69even in this simple case , the line shape is complex since it depends on the form of the confining wave function . in the experimental setup described in ref .
, the magnetic field is kept fixed .
a doping layer provides a finite density of electrons in the well .
this density can be modified by changing the potential of a gate allowing the rd - nmr signal to be studied in some range of filling factors
. the ionized doping layer , the applied electric field and the finite density of electrons in the well all modify its potential profile and consequently the subband wave functions @xmath36 and @xmath70 these functions must then be found by solving the schrdinger - poisson equations . to fit the experimental rd - nmr signal correctly ,
these two wave functions must be computed from the exact potential profile of the quantum well at filling factors @xmath71 and @xmath72 in this work , we are interested in studying how a qualitative change in the crystal structure ( such as the transition from a wigner to a bubble crystal with two electrons per site ) affects the rd - nmr signal . to isolate the effect of a change in the spin polarization , we assume an ideal situation where the quantum well profile is fixed , i.e. unchanged by a variation of the filling factor .
we thus approximate both functions @xmath73 and @xmath41 by the simple form@xmath74which is the first bound state of an infinite quantum well .
our final expression for the rd - nmr spectral line shape is thus @xmath75 for greater clarity in the figures on this paper , we plot @xmath76 as a probability density function @xmath77 instead of as an histogram of values of @xmath64 .
we normalize @xmath78 such that:@xmath79 in our calculations , we discretize the confining wave functions into @xmath80 points , the unit cell of each crystal into @xmath81 points ( @xmath82 ) that are further classified into @xmath83 bins .
the electronic density in the partially filled level is written as@xmath84where @xmath85 is an associated laguerre polynomial and the operator @xmath86the set of vectors @xmath87 are the reciprocal lattice vectors of the crystal considered while the set of ground - state average values @xmath88 can be considered as the order parameters of the crystal .
they are related to the single - particle matsubara green s function by@xmath89where @xmath90and@xmath91with @xmath92 the time - ordering operator and @xmath93 the imaginary time . in eq .
( [ rho ] ) , the operator @xmath94 creates an electron in landau level @xmath19 with guiding - center index @xmath95 in the landau gauge and with spin @xmath33 .
since we consider the filled levels below @xmath19 as inert ( they do not contribute to the spin polarization ) , we hereafter drop the index @xmath19 to simplify the notation the single - particle green s function is computed by using an equation of motion method@xcite which is generalized to include the spin degree of freedom and the possibility of a spin texture in the plane of the quantum well .
the hartree - fock equation of motion for @xmath96 where @xmath97 are the matsubara frequencies , is@xmath98 g_{\alpha , \beta } \left ( \mathbf{g},\omega _ { n}\right ) = \delta _ { \mathbf{g},0}\delta _ { \alpha , \beta } \label{equation } \\ & & + \frac{1}{\hslash } \sum_{\mathbf{g}^{\prime } \neq \mathbf{g}}u^{h}\left ( \mathbf{g - g}^{\prime } \right ) e^{-i\mathbf{g}\times \mathbf{g}^{\prime } \ell ^{2}/2}g_{\alpha , \beta } \left ( \mathbf{g}^{\prime } , \omega
_ { n}\right ) \notag \\ & & -\frac{1}{\hslash } \sum_{\mathbf{g}^{\prime } } \sum_{\gamma } u_{\alpha , \gamma } ^{f}\left ( \mathbf{g - g}^{\prime } \right ) e^{-i\mathbf{g}\times \mathbf{g}^{\prime } \ell ^{2}/2}g_{\gamma , \beta } \left ( \mathbf{g}^{\prime } , \omega _ { n}\right ) , \notag\end{aligned}\]]with the potentials@xmath99where @xmath100 and @xmath101 is the dielectric constant of the host semiconductor .
the non - interacting single - particle energy is given by@xmath102where @xmath103 is the effective @xmath104factor of bulk gaas and @xmath105 is the bohr magneton .
the hartree and fock interactions are defined by@xmath106 ^{2 } , \\
x\left ( \mathbf{q}\right ) & = & \left ( \frac{e^{2}}{\kappa \ell } \right ) \sqrt{2}\int_{0}^{\infty } dxe^{-x^{2}}\lambda \left ( x\right ) \left [ l_{n}^{0}\left ( x^{2}\right ) \right ] ^{2 } \\ & & \times j_{0}\left ( \sqrt{2}xq\ell \right ) , \notag\end{aligned}\]]where @xmath107 is the bessel function of order zero . because of the laguerre polynomial @xmath85 , these interactions depend on the landau level index @xmath108 the function @xmath109 takes into account the finite width of the quantum well .
it is defined by @xmath110and depends on the magnetic field @xmath111 for an ideal 2deg , @xmath112 equation ( [ equation ] ) is self - consistent and can be solved numerically by successive iterations.@xcite a gaussian blurring @xmath113 of the density @xmath114 can be obtained by multiplying the summand in eq .
( density ) by the factor @xmath115 this blurring can be used to mimic the effects of quantum and thermal fluctuations on the crystal .
the parameter @xmath116 is then adjusted , in this ase , so as to fit the experimental spectra.@xcite
we give now a brief description of the different crystal phases for which we want to compute the nmr spectral line shape . in the hartree - fock approximation , even the quantum hall stripe phase has density modulations along the direction of the stripes and so can be described as a crystal although a very anisotropic one ( a face - centered rectangular a large aspect ratio ) .
all these states can by described formally as crystals of some quasiparticle .
the quasiparticle can be an electron , an electron bubble , a skyrmion or the corresponding quasi - antiparticles : hole , hole bubble , antiskyrmion .
we write @xmath117 for the density of these quasiparticles and define @xmath118 $ ] as the filling factor of the partially filled level @xmath19 ( the total filling factor is @xmath119 ) .
the filled levels are considered as inert .
a level with both spin sub - levels filled is unpolarized and so does not contribute to the nmr spectra .
we consider the following phases : * bc1 ( @xmath120 ) . a bubble crystal , as described in ref . , where the quasiparticle at each site is a @xmath121electron bubble with spin up .
the densities @xmath122 and the spin polarization density is @xmath123 * bc2 ( @xmath124 ) . a bubble crystal where the quasiparticle is a @xmath121hole bubble with spin down .
the densities @xmath125 and @xmath126 * bc3 ( @xmath127 ) . a bubble crystal where the quasiparticle is a @xmath128-electron bubble with spin down .
the densities @xmath129 and @xmath126 * bc4 ( @xmath130 ) . a bubble crystal where the quasiparticle is a @xmath121hole bubble with spin up
. the densities @xmath131 and @xmath123 the special case @xmath132 corresponds to the wigner crystal .
the hartree - fock hamiltonian being electron - hole symmetric , @xmath117 is the same function for all four phases .
it follows that bc1 and bc4 have the same spectrum and similarly for bc2 and bc3 .
moreover , the spectrum for bc2 ( or bc4 ) is obtained from the polarizations @xmath64 computed for bc1 ( or bc3 ) by replacing @xmath64 in eq .
( [ signal ] ) by @xmath133 .
the description given above for bc1 to bc4 applies with no change to the quantum hall stripe phases when considered as face - centered rectangular wigner crystal ( @xmath132 ) . in this case , the aspect ratio that minimizes the ground - state energy depends on the landau level index @xmath19 and filling factor @xmath71 . to distinguish the stripe phases from the bubble crystals , we use for the former the abbreviations sc1 to sc4 .
the last phases we consider are the skyrme crystals .
these occur for @xmath71 close to @xmath134 : * sk1 ( @xmath135 ) . a crystal where the quasiparticle at each site is an anti - skyrmion of topological charge @xmath136 the densities can be described by @xmath137 they depend on the ratio of the zeeman coupling to the coulomb energy @xmath138 and on the filling factor@xmath139 the density @xmath140 while @xmath141 * sk2 ( @xmath142 ) .
a crystal where the quasiparticle at each site is a skyrmion of topological charge @xmath143 the densities @xmath144 and so @xmath145 and @xmath146 by electron - hole symmetry , sk1 and sk2 have the same nmr spectrum.note1 in the hartree - fock calculation,@xcite the lowest - energy lattice structure for the skyrme crystal if @xmath71 is not too close to @xmath134 is a square lattice with two skyrmions of opposite phases per unit cell i.e. the square lattice antiferromagnetic ( sla ) phase .
it has lower energy than a triangular lattice where all skyrmions have the same phase ( tlf ) . for @xmath71 close to one , a three - sublattice crystal where skyrmions on different sublattice are rotated by @xmath147 should be the ground state
however , it is difficult to stabilize this phase numerically so that we will only consider the sla and tlf crystals in our calculation .
we remark that , of all the crystal states considered here , sk1 and sk2 are the only ones where @xmath148 can be locally negative .
as we will show , this makes the nmr skyrme crystal s spectrum quite different from that of the other crystal states .
in this section , we compute the rd - nmr spectra of the crystal states described in sec .
because of the electron - hole symmetry , we can restrict ourselves to the bc1,bc2,sc1,sc2 and sk1 phases .
the order parameters for the crystals are calculated for a quantum well of width @xmath149 nm and in a magnetic field @xmath150 t as in the experiments.@xcite we work , however , at zero temperature .
we define the knight shift as the displacement of the peak of the spectrum from the bare resonance frequency @xmath151 at @xmath150 t , the @xmath10as nucleus has @xmath152 mhz with a linewidth @xmath153 khz@xcite .
we use this value of @xmath46 in our calculations . figure 4(b ) of ref .
gives the expected behavior of the knight shift of a fully polarized uniform system for filling factors @xmath154 $ ] in landau level @xmath17 .
we use this figure to get the constant @xmath59 that enters eq .
( [ signal ] ) .
we get @xmath155 . since @xmath156 a positive spin polarization leads to a negative knight shift in eq .
( [ signal ] ) . in landau level @xmath17 with @xmath157 and for @xmath158 with @xmath159(b )
probability density @xmath77 in the wigner crystal phase for @xmath158 , @xmath160 ] figure 1(a ) shows the spectra for the bc1 phase with @xmath132 i.e. a wigner crystal at filling factors @xmath161 in landau level @xmath162 these spectra were discussed previously@xcite but we reconsider them here for completeness .
the spectra are characterized by a single peak and a long tail , absent in the the uniform state , which is due to the large spread in the values of @xmath61 in the crystal as shown by the plot of @xmath78 in fig .
1(b ) . from eq .
( [ signal ] ) , a region of space with a certain @xmath61 produces a local shift in frequency of @xmath163 since @xmath164 this shift is @xmath165 khz for the largest possible value @xmath166 .
the function @xmath77 is dominated by the regions of small values of @xmath61 so that the knight shifts in the spectra are small . as @xmath71 increases , the most probable value of @xmath61 increases ( not shown in the figure ) and the knight shift becomes more negative . at the same time , the maximal intensity of the signal decreases .
the most probable value of @xmath61 at these filling factors is smaller that the spatially averaged value of @xmath61 so that the knight shift of the crystal is less negative than that of the uniform state@xmath139 all spectra start at @xmath167 even though the crystal and the liquid have no region with @xmath168 this is because the subband wave function term @xmath169 in eq .
( [ signal ] ) goes to zero at @xmath170 at @xmath171 there is an important difference between the theoretical and experimental spectra.@xcite this is attributed to the fact that the hartree - fock approximation for @xmath148 neglects both quantum and thermal fluctuations . to get a spectra closer to the experimental result ,
it is necessary to use a finite blurring factor of order @xmath172 at @xmath173@xcite as shown in fig .
1(a ) , this shifts the peaks to a more negative value and cuts the low - frequency tail . in landau level @xmath17 with @xmath174(b )
probability density @xmath77 in the wigner crystal phase for @xmath175 ] figure 2(a ) shows the spectra for the bc2 phase with @xmath132 at filling factors @xmath176 in landau level @xmath162 a comparison with fig .
1(a ) shows that the spectra for bc2 are distinctively different from those of bc1 . in the plot of @xmath77 in fig .
2(b ) , the most probable value for @xmath61 is now close to @xmath166 and so the knight shift is large .
this is also true for the uniform phase . as in bc1 ,
the spread in @xmath61 is important and the signal is intense in a large range of frequency shift@xmath139 in contrast to bc1 , however , the maximal @xmath61 decreases with decreasing @xmath71 and so does the spread of the signal .
the knight shift becomes less negative with decreasing filling factor because the maximum value of @xmath61 decreases with decreasing @xmath71 .
the most probable value of @xmath61 in the bc2 phase is , in contrast to bc1 , bigger than that of the uniform state so that the knight shift in the crystal bc2 state is more negative than in the uniform state .
for : ( a ) the skyrme crystal phases ( sk1 ) at different filling factors @xmath177 and for a zeeman coupling @xmath178 ( unless indicated otherwise ) with @xmath174 ( b ) the triangular ( tlf ) and square ( sla ) skyrme crystals for @xmath179 ; ( c ) knight shift ( measured at the peak position ) and spatially averaged polarization @xmath180 as a function of the filling factor . ] figure 3(a ) shows the spectra of the sla skyrme crystal sk1 at filling factors @xmath176 in landau level @xmath17 for a zeeman coupling @xmath181 and at @xmath182 for @xmath183 for comparison , we include the spectrum of the uniform state at @xmath184 for @xmath185 the polarization @xmath186 is shown as a function of @xmath71 and @xmath187 in fig . 1 of ref . .
( in that calculation , the well width was taken to be zero but taking @xmath149 nm does not change much that result . ) that hartree - fock calculation of @xmath188 gives a fairly accurate description of the magnetization of the skyrme crystal observed in nmr experiments.@xcite as fig .
3(a ) shows , the spectrum for the skyrme crystal is different from that of the wigner crystal [ compare fig .
2(a ) and fig .
3(a ) ] or the uniform state . for the skyrme crystal , there is large frequency range of positive knight shift which is due to the spatial regions of reversed spin ( negative @xmath61 ) in the skyrmion s texture . in the wigner crystal , there is no spin texture and @xmath148 is always positive . for @xmath189
the skyrmion s size and so the number of reversed spins increases meaning that the spatial regions of positive knight shift increase too . the regions of positive shift decrease with increasing zeeman coupling @xmath187 since that coupling decreases the skyrmion s size .
another important difference between the spectrum of the wigner and the skyrme crystals is that the former is independent of @xmath190 we compare the spectra of different skyrme crystal lattices at @xmath191 in fig .
the sla and triangular skyrme crystals in @xmath17 are evaluated at @xmath185 they have a spatially averaged polarization @xmath192 and @xmath193 respectively . the knight shift is larger for the triangular lattice because of the regions of higher polarization but the positive knight shift extends to higher frequencies for the sla structure which has more reversed spins . the skyrme crystal for a perfect 2deg is not expected to be the ground state for @xmath194 indeed , in @xmath195 we were not able to stabilize a skyrme crystal even for a zeeman coupling as small as @xmath196 with a large number of iterations , the solution converges towards a triangular wigner crystal of holes ( bc2 ) with no spin texture [ its spectrum
is shown in fig .
this is consistent with a recent experiment@xcite where no skyrme crystal are detected in @xmath20 around @xmath197 a finite temperature .
the low - temperature phase of the 2deg in this case seems to be a wigner crystal of holes ( type bc2 with @xmath132 ) .
the difference in the line shapes of bc2 crystals for @xmath17 and @xmath20 [ compare fig .
2(a ) with fig .
4(b ) arises because of the difference in the electronic wave functions of the two landau levels as we explain in more detail below .
figure 3(c ) shows the knight shift measured at the position of the peak in the spectrum of fig .
3(a ) and the spatially averaged polarization @xmath61 as a function of the filling factor @xmath198 for @xmath199 the uniform state has @xmath200 as can be seen by comparing the two curves , the relation between the knight shift measured in this way and the average magnetization is not exactly linear ( as is the case for zero - width uniform 2deg ) .
the spectra that we gave in this section apply to an ideal situation where the skyrme crystals are frozen . with thermal fluctuations
, the line shapes may be very different . at finite temperature
, one must know the exact regime where the experiment is carried on i.e. the skyrme s crystal dynamics.barrett,sinova,villares,gervais2 experiments@xcite close to @xmath8 have measured an anomalous line shape for the skyrme crystal that our simple ground - state calculation can not reproduce .
in fact , it has even been suggested that the dispersive line shape found experimentally may be inconsistent with the usual model of rd - nmr where the signal is due solely to a hyperfine - induced increase of the electronic zeeman energy.@xcite ( wigner crystal ) at filling factor @xmath201 in landau levels @xmath202 and for the uniform phase at @xmath203 ( b ) the bc1 , sc1 and bc2 phases in landau level @xmath20 at different filling factors ; ( c ) probability density @xmath77 for the same phases as in ( a ) .
the function @xmath77 for @xmath17 extends to @xmath166 ( not shown on the graph ) . ]
the laguerre polynomial @xmath85 that enters the calculation of the density as well as the hartree and fock potentials causes these functions to change with the landau level index . in @xmath205
the density for a bubble crystal with @xmath132 has a gaussian profile centered at each lattice site . for @xmath204 and @xmath206
the density profile has the form of a ring centered on each lattice site .
a new ring is added each time @xmath128 increases with the exception of the last bubble phase with @xmath207 in landau level @xmath108 in this case , a peak is added at the center of the existing @xmath128 rings ( see fig . 2 of ref . ) .
we compare the spectra of the @xmath132 bc1 phase for @xmath202 at @xmath208 in fig .
the corresponding probability densities @xmath77 are shown in fig .
the most probable density is the minimal density which increases with @xmath19 while its probability decreases with @xmath108 consequently , the knight shift in fig .
4(a ) becomes more negative with @xmath19 but the peak intensity in the spectral line shape decreases with @xmath19 .
the next contribution to these spectra comes from the ring . at fixed @xmath209 the diameter of the ring increases with @xmath19 while @xmath61 in the ring decreases .
this behavior is reflected in the spectra shown in fig .
4(a ) : the frequency and the intensity of the shoulder in the spectrum increases with @xmath19 since , as the ring expands , it covers a greater area . for comparison , we include in fig .
4(a ) the spectrum of the uniform phase at @xmath208 which is independent of @xmath108 the peak in the density of the crystal for @xmath210 produces a long tail in its spectrum that is absent of the spectrum of a @xmath211 crystal in higher landau levels . figure 4(b ) shows the spectra of several phases in landau level @xmath212 the bc1 crystals with @xmath132 ( @xmath208 ) and @xmath213 ( @xmath214 ) , the stripe phase sc1 at @xmath215 and the bc2 crystal with @xmath132 at @xmath216 the stripe phase is not present in @xmath20 in the hfa but appears in dmrg calculation.@xcite on the contrary , the @xmath213 bubble crystal is present in hfa but not in dmrg .
the transition from @xmath132 to @xmath213 is predicted to occur at @xmath217 in the hfa .
the phase diagram obtained in dmrg is richer than that of the hfa because it contains the fractional quantum hall states and other correlated liquid states .
in contrast to the hfa , it predicts some form of stripe state in @xmath17 ( whose structure is however different from the stripe state found in higher landau levels ) and a stripe state in @xmath20 when @xmath218 $ ] whose structure is similar to the stripe state in higher landau levels predicted by the hfa . up to now
, only the @xmath132 bc1 phase has been detected experimentally.@xcite the long tail in the @xmath219 phase , like that of the wigner crystal with @xmath210 is due to the peak in density .
this makes its spectrum different from that of the other phases in @xmath20 .
the spectra of the different phases in @xmath20 are quite different and their detection should allow the unambiguous identification of the corresponding phases .
along @xmath220 for @xmath221 and at different filling factors in @xmath222 for the : ( a ) @xmath132 bubble crystals ; ( b ) @xmath213 bubble crystals ; ( c ) @xmath223 bubble crystals ( bottom @xmath220 axis ) and stripe phase ( top @xmath220 axis ) .
the numbers identify a region that contributes to a peak in the function @xmath78 shown in fig .
6(a)@xmath139 here @xmath224 is the lattice spacing and @xmath225 is the magnetic length .
the dashed curve in ( b ) shows the magnetization trace for @xmath226 at @xmath227 for filling factor @xmath228 the curves are offset by @xmath229 for clarity .
the intensity is in arbitrary units , the same for all curves . ]
we now consider landau level @xmath230 figure 1(a ) of ref
. shows the phase diagram of the 2deg in landau level @xmath222 for @xmath231 the density pattern for each phase ( bc1 and sc1 ) is shown in fig . 2 of that reference .
in the hfa , there is a transition at @xmath232 from @xmath132 to @xmath213 and at @xmath233 a second transition to @xmath223 .
the stripe phase sc1 is reached after @xmath234 these numbers are for a perfect 2deg .
they are slightly modified when @xmath40 is finite as is the case here .
figure 5 shows the polarization @xmath148 along @xmath220 for @xmath221 in the bubble and stripe phases at different values of the filling factors@xmath139 the corresponding functions @xmath77 and line shapes @xmath235 are shown in fig .
6(a ) and fig .
6(b ) respectively . the bubble crystals and stripes of holes bc2 , sc2 follow the same phase diagram than bc1 , sc1 but in reverse order , i.e.
, the crystal of electrons at @xmath236 is replaced by a crystal of holes at filling factor @xmath237 and @xmath61 is replaced by @xmath238 in the calculation of the line shape .
the spectra of the phases bc2 , sc2 are shown in fig .
the density @xmath61 being bigger in the hole phases than in the electron phases , the knight shift is larger in the former than in the latter . as @xmath71 increases , the radius of the ring in the @xmath132 bubble crystal increases .
the transition to the @xmath213 crystal occurs when rings from adjacent sites just touch . at this point
, there is an abrupt transition to the @xmath213 bubble crystal which has two rings .
as @xmath71 increases further , these two rings expands and there is again an abrupt transition to the @xmath223 bubble crystal when the outer rings from adjacent sites just touch .
the @xmath223 bubble crystal has two rings and one central maximum on each site . from fig .
5 , as @xmath71 increases , the average value of the magnetization increases and there are more oscillations in @xmath148 and consequently more structure in @xmath239 in each case , there is one oscillation with a large value of @xmath240 and other smaller ones . in the stripe phase , the oscillations are much less pronounced than in the crystal phases . as regards its nmr spectrum , this phase is close to that of the uniform phase . in fig . 5 , the numbers above the lines for @xmath241 and @xmath242 identify the regions in @xmath148 that are responsible for a peak in the corresponding plot of @xmath77 shown in fig . 6(a ) .
we see in fig .
5(b ) that , in comparison with @xmath206 the @xmath213 bubble crystal has one extra ring which adds the extra peaks numbered @xmath243 and @xmath244 in fig .
6(a ) for @xmath245 the spectral line shape of the bc1 , sc1 crystals changes with increasing @xmath71 and is different from that of the uniform phase .
the signal from the @xmath213 crystal at @xmath242 looks different from that of the @xmath132 crystal at @xmath208 but is not very different from the @xmath132 crystal at @xmath246 even though the functions @xmath77 shown in fig .
6(a ) are quite different , many of the peaks in @xmath77 are too close in @xmath61 to be resolved in the spectra .
for example , the peaks numbered @xmath22 and @xmath244 in fig .
6(a ) ( for @xmath247 ) are spaced by @xmath248 this leads to peaks separated by @xmath249 khz in the line shape .
this separation is smaller than the linewidth @xmath153 khz considered in our calculation .
it follows that most of the interesting features in @xmath77 are lost in the integrated signal @xmath235 ( not considering the extra complication due to confining wave function ) .
the signal from the @xmath223 bc1 crystal , however , is different from that of the @xmath250 crystal and the stripe phase .
it has a characteristic long low frequency tail ( not visible in fig .
6(a ) but it is finite up to @xmath166 and in fig .
6(b ) it extends to @xmath251 khz ) . as we mentioned above
, this phase may not be present in the ground state according to dmrg calculations .
we expect the same conclusions to hold for landau levels @xmath252 since there are more oscillations in bubbles with higher values of @xmath128 and these oscillations are smaller in amplitude [ see the traces for @xmath253 and @xmath254 in fig .
5(b ) ] . except in some range of @xmath71 ( fort @xmath255 for example )
most spectra are sufficiently different from that of the corresponding uniform state to allow a discrimination between the uniform and crystal states .
unfortunately , it does not seem possible to infer the number of electrons in each bubble crystal from its spectrum alone when @xmath256 ( with the exception of @xmath207 ) since the distinguishing features in the density are lost in the integrated signal
. it may also be difficult to differentiate between the @xmath257 crystal and the stripe phase as their spectra are not dramatically different .
moreover , the oscillations are small in the stripe phase and the corresponding nmr spectrum is close to that of the uniform state . one marked difference between the @xmath213 and @xmath132 spectra , however , is that the maximal knight shift ( in absolute value ) is larger in the former than in the latter . indeed , as fig .
6(a ) shows , the maximum value of @xmath148 increases discontinuously with @xmath258 rd - nmr spectra ( full curves ) for the @xmath132 and @xmath213 bubble crystals at @xmath259 and @xmath158 and for the stripe phase at @xmath260 calculated with a linewidth @xmath261 khz .
the dashed curves are the spectra of the corresponding uniform phases . ] at some filling factors in @xmath222 , the difference between the crystal and liquid spectra is small and may be washed out by fluctuations .
we give some examples of that in fig .
6(b ) where the dashed - dotted curve represent the crystal phases calculated with a gaussian blurring factor @xmath262 which is the value necessary in @xmath17 to make the calculated spectrum agrees with the experimental result according to ref . .
we see that , with the exception of @xmath263 the spectra for @xmath242 and @xmath264 are almost identical with those of the corresponding uniform states .
we checked that the value of @xmath116 that brings the crystal spectrum identical with that of the uniform state for the wigner crystal at @xmath208 in @xmath265 are @xmath266 respectively .
the crystal state is thus more fragile in higher landau levels .
we have presented in this paper a study of the rd - nmr spectra of various crystal phases that can theoretically exist as ground states of the 2deg in the quantum hall regime . in landau levels @xmath267 , wigner crystal , bubble crystal , stripe phase and skyrme crystal ( when present ) produce distinct rd - nmr line shapes that should allow their unambiguous experimental identification . in higher landau levels , the rd - nmr spectra are not so distinct with the exception of the @xmath207 bubble crystal . in these levels , crystal and uniform states can be distinguished but it would be difficult to differentiate between the wigner , bubble ( with @xmath268 and @xmath269 ) and stripe phases .
our calculation assumed the ideal case of a frozen solid at zero temperature .
as shown before@xcite , thermal fluctuations will render the crystal spectra closer to that of the uniform state .
this will not help the differentiation of the different phases in @xmath256 since the small difference in the spectra will probably be washed out unless the temperature is very small
. the nmr spectra would be more discriminating with a stronger hyperfine coupling between the electron gas and the nuclei or with a reduced linewidth @xmath46 but these parameters are fixed by the quantum well .
we have used the value @xmath153 khz in all our calculations@xcite but , as a test , we show in fig .
8 the spectra for the @xmath222 stripe phase at @xmath270 and the @xmath271 bubble crystals at @xmath241 and @xmath272 with a smaller linewidth @xmath273 khz .
the distinction between the solid and uniform phases is much clearer [ compare with fig .
6(b ) ] . apart from the important difference in the range of the knight shift for crystal with different values of @xmath128 , the spectra for the @xmath213 crystal and the stripe phase still have a similar shape . in a real experimental situation ,
the filling factor is tuned by changing the potential on a gate in the gaas / algaas quantum well .
this modifies the shape of the confining wave function that enters the calculation of the nmrspectra .
we have considered this wave function as fixed , but our calculation can easily take this effect into account ( by solving the schrdinger - poisson equations ) if an exact description of the quantum well is given . for reviews ,
see _ physics of the electron solid _ , edited by s. t. chui ( international , boston , 1994 ) ; h. fertig and h. shayegan , in _ perspectives in quantum hall effects _ , edited by s. das sarma and a. pinczuk ( wiley , new york , 1997 ) , chaps . 5 and 9 ,
respectively ; yu .
p. monarkha and v. e. syvokon , low temp
. phys . * 38 * , 1067 ( 2012 ) . for a review , see michael m. fogler , _ stripe and bubble phases in quantum hall systems _ , pp . 98 - 138 in high magnetic fields : applications in condensed matter physics and spectroscopy , ed . by c. berthier ,
levy , g. martinez ( springer - verlag , berlin , 2002 ) .
m. o. goerbig , p. lederer , and c.m .
smith , phys .
b * 68 * , 241302(r ) ( 2003 ) ; _ ibid_. phys . rev .
b * 69 * , 115327 ( 2004 ) ; _ ibid_. phys . rev . lett . * 93 * , 216802 ( 2004 ) ; n. thiebault , n. regnault , m. o. goergib , arxiv : 1503.09132 [ cond-mat.mes-hall ] .
r. ct , c. b. doiron , j. bourassa , and h. a. fertig , phys .
b * 68 * , 155327 ( 2003 ) . note that there is a typo in the wave function given for the bubble in eq .
( 42 ) of this reference .
the wave function should read : p. d. ye , l. w. engel , d. c. tsui , r. m. lewis , l. n. pfeiffer , and k. west , phys .
lett . * 89 * , 176802 ( 2002 ) ; yong p. chen , g. sambandamurthy , z. h. wang , r. m. lewis , l. w. engel , d. c. tsui , p. d. ye , l. n. pfeiffer , and k. w. west , nature phys . * 2 * , 452 ( 2006 ) ; y. p. chen , r. m. lewis , l. w. engel , d. c. tsui , p. d. ye , l. n. pfeiffer , and k. w. west , phys .
91 * , 016801 ( 2003 ) ; han zhu , g. sambandamurthy , yong p. chen , p. jiang , l. w. engel , d. c. tsui , l. n. pfeiffer , and k. w. west , phys .
. lett . * 104 * , 226801 ( 2010 ) . n. shibata and d. yoshioka , phys .
lett . * 86 * , 5755 ( 2001 ) ; d. yoshioka and n. shibata , physica e * 12 * , 43 ( 2002 ) ; n. shibata and d. yoshioka , physica e * 22 * , 111 ( 2004 ) ; n. shibata and d. yoshioka , j. phys .
. jpn . * 73 * , 1 ( 2004 ) .
s. e. barrett , r. tycko , l. n. pfeiffer , and k. w. west , phys .
lett . * 72 * , 1368 ( 1994 ) ; s. e. barrett , g. dabbagh , l. n. pfeiffer , k. w. west , and r. tycko , phys .
. lett . * 74 * , 5112 ( 1995 ) ; r. tycko , s. e. barrett , g. dabbagh , l. n. pfeiffer , and k. w. west , science * 268 * , 1460 ( 1995 ) ; n. n. kuzma , p. kandelwal , s. e. barrett , l. n. pfeiffer , k. w. west , science * 281 * , 686 ( 1998 ) .
l. brey , h. a. fertig , r. ct , and a. h. macdonald , phys .
* 75 * , 2562 ( 1995 ) ; r. ct , a. h. macdonald , luis brey , h. a. fertig , s. m. girvin , and h. t. c. stoof , phys . rev . lett . * 78 * , 4825 ( 1997 ) .
it is clear then that our calculation can not account for the tilted plateau that is reported for the knight shift around @xmath8 in p. khandelwal , a. e. dementyev , n. n. kuzma , s. e. barrett , l. n. pfeiffer , and k. w. west , phys .
* 86 * , 5353 ( 2001 ) . | transport experiments on the two - dimensional electron gas ( 2deg ) confined into a semiconductor quantum well and subjected to a quantizing magnetic field have uncovered a rich variety of uniform and nonuniform phases such as the laughlin liquids , the wigner , bubble and skyrme crystals and the quantum hall stripe state .
optically pumped nuclear magnetic resonance ( op - nmr ) has also been extremely useful in studying the magnetization and dynamics of electron solids with exotic spin textures such as the skyrme crystal .
recently , it has been demonstrated that a related technique , resistively - detected nuclear magnetic resonance ( rd - nmr ) , could be a good tool to study the topography of the electron solids in the fractional and integer quantum hall regimes . in this work ,
we compute theoretically the rd - nmr line shapes of various crystal phases of the 2deg and study the relation between their spin density and texture and their nmr spectra .
this allows us to evaluate the ability of the rd - nmr to discriminate between the various types of crystal states . |
during the last decade , ground- and space - based surveys have been very successful in detecting transiting exoplanets .
in addition to their primary science goal , the large photometric data sets acquired by them allow studying millions of stars for variability .
numerous projects thus provide an exceedingly increasing number of detections that are collected by variable star catalogs such as the general catalogue of variable stars ( gcvs ; * ? ? ? * ) or the variable star index ( vsx ) .
such catalogs not only broaden the statistical sample of variable stars , but are also important to gain further knowledge about the different processes that cause stellar variability .
several methods have been proposed to search for periodic signals in astronomical time series ( for a good overview , see , e.g. , * ? ? ?
one of the most widely applied algorithms is the analysis of variance ( aov ) statistic @xcite , which provides an optimal period search in uneven sampled observations .
it has been used very successfully by projects like hat @xcite , wasp ( e.g. , * ? ? ?
* ) , or ogle ( e.g. , * ? ? ?
in addition , the @xmath0 index @xcite is frequently used to quantify variability in general and/or for selecting candidate stars prior to a period search in order to minimize computation time ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . however , both methods the aov period search and stetson s variability index are strongly affected by systematic trends present in ground - based data sets ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
most dominant are diurnal systematics , introducing artificial variability with periods of one day or multiples thereof . such trends generally yield a higher ranking of non - variable stars , thus increasing the false alarm rate . a common approach to account for candidates with systematic variability
is to set limits , e.g. , to exclude detections within certain period ranges .
however , any such manual mechanism is usually not well applicable to other data sets or projects , and the number of missed detections ( false negatives ) is often unknown .
therefore , a more sophisticated treatment of systematic variability in combination with period search and ranking is needed .
the berlin exoplanet search telescope ( best ; * ? ? ? * ) and best ii @xcite are used to perform ground - based support of the corot space mission @xcite . by obtaining high - precision and long - time series photometry of the corot target fields prior to the satellite s observations , planetary candidates
can quickly be checked in the best data archive ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in addition , the obtained light curves can be used to identify variable stars .
the best project has already yielded the detection of several hundreds of new periodic variable stars @xcite .
a first characterization of periodic stellar variability in the corot field lra02 has been published by @xcite . following the detections from this first publication ( denoted as paper i hereafter )
, it was possible to optimize the best ii reduction pipeline .
the large , well - characterized data set lra02 was used as a proxy to analyze and automatically include systematic biases in the variable star candidate selection and period search , yielding a significantly reduced false alarm rate .
the optimized procedure to rank periodic variable stars is described in this paper .
in addition , we present a large catalog extension to the variable star classification in lra02 that was obtained through the reanalysis and compare our results to publicly available corot data .
best ii is located at the observatorio cerro armazones , chile .
it is operated by the _
institut fr planetenforschung _ of the _ deutsches zentrum fr luft- und raumfahrt _ in robotic mode from berlin since summer 2007 .
the system consists of a 25 cm baker - ritchey - chrtien telescope with a focal ratio of @xmath1 , yielding a wide field of view ( fov ) of @xmath2 .
it is equipped with a 4k @xmath3 4k , 16 bit finger lakes imager ccd ( kaf-16801e1 ) with a pixel size of 9 @xmath4 m , and an angular resolution of @xmath5 pixel@xmath6 . best ii observes without any filter to maximize the photon yield the ccd sensitivity peaks at 650 nm and is roughly comparable to the johnson @xmath7 band .
the corot long - run field lra02 was observed by best ii for 41 nights from 2007 november to 2008 february prior to the satellite observations .
as the fov of corot is slightly larger than best ii , we split the field into two subfields ( called lra02a and lra02b , respectively ) and pointed at them alternating .
paper i indicates their corresponding center coordinates and shows the orientation with respect to the corot fov ( figure 1 in paper i ) .
the acquired data set was calibrated and reduced using the best automated photometric data pipeline as outlined in @xcite . the actual reduction procedure for the field lra02
has already been described in detail in paper i and below only those steps relevant to the new results are discussed . in paper
i , kabath et al . reported the detection of 350 periodic variable stars ( of which five were previously known ) .
their selection was based upon the variability index @xmath0 @xcite , calculated for each star by @xmath8 where @xmath9 is indexing individual data points .
@xmath10 is calculated from each pair of subsequent magnitudes @xmath11 and @xmath12 using the corresponding normalized residuals @xmath13 and @xmath14 : @xmath15 where @xmath16 denotes the uncertainty of measurement @xmath9 , @xmath17 the mean magnitude , and @xmath18 the number of measurements for the selected star . the weights @xmath19 in equation ( [ eq : jindex1 ] )
were calculated following the approach of @xcite as @xmath20 where @xmath21 is the time of observation @xmath9 and @xmath22 is the median of all pair time spans @xmath23 . in paper
i , the limit of @xmath24 was applied in order to distinguish variable from constant field stars .
this preselection yielded 1,858 stars in lra02a and 1,868 stars in lra02b , respectively . for each star
, the aov statistic @xmath25 @xcite was calculated for a period range of 0.135 days .
the light curves were folded with the frequency @xmath26max}}}}$ ] corresponding to the maximum aov value , @xmath27max } } } } ) = \max \left ( \theta(\omega ) \right),\ ] ] and then inspected visually .
most folded light curves showed no clear periodic variability or an artificial period of one day or multiples thereof , which can be caused by systematic effects due to the observational cycle .
@xcite finally identified 173 periodic variables in lra02a and 177 in lra02b .
in the best archive , some stars are marked as clear variables with large @xmath0 indices ( up to 10 and higher ) .
figure [ fig : jhist ] shows the count of all stars in subfield lra02b and the corresponding number of variable star detections in paper i as a function of the @xmath0 index .
altogether , the large number of false positives shows that the @xmath0 index alone is not an effective criterion for selecting variable stars .
furthermore , a number of clear detections with low @xmath0 values indicates that several variables must have been missed in paper i due to the cutoff . in order to improve the quality of our preselection process and to maximize the detection yield
, we performed a deeper analysis of our data set lra02 .
field lra02 was chosen for this purpose because it was observed with longest duration within the best project so far .
the performed reanalysis consisted of three main steps .
first , the best period was searched in all light curves from the initial data set without any preselection criterion , i.e. , directly using the aov multiharmonic algorithm for _ all _ stars in both subsets .
each of the resulting 37,361 folded light curves in lra02a and 66,974 in lra02b was then examined visually for periodic stellar variability .
in addition to the variable stars from paper i , this first step already revealed 189 _ additional _ variable stars that were not detected in our first publication . during this step
, we also discovered a bug in our implementation of the @xmath0 index that yielded systematically lower values especially for short periods , which is why many of these detections show periods of less than a day .
second , the combined data set of variable stars from paper i and the additional manual detections were used to optimize the best ii selection process ( see next section [ sec : limitimprove ] ) . with a very good knowledge of this data set
, it was possible to compare different selection methods and to adjust their corresponding parameters .
third , the new search algorithm was applied to the data set with optimized parameters .
the improvements to the pipeline finally lead to an increase of detections by another 135 variable stars that were not found during all previous steps ( see section [ sec : results : newmethod ] ) .
the deep analysis of the best ii data set lra02 gives us the opportunity to study the performance of variability search algorithms in detail .
the aim is to recover all variable stars in the data set automatically and to minimize the number of false alarms ( and manpower ) at the same time . in this section ,
we show the limitations of the @xmath0 index with regard to systematic trends ( section [ sec : jlimit ] ) , describe how to quantify the performance of a variability search ( section [ sec : limitimprove1 ] ) , present the algorithms tested ( sections [ sec : limitimprove2 ] and [ sec : limitimprove3 ] ) , and finally compare the performance of different approaches and parameters ( section [ sec : limitimprove4 ] ) .
after the first step of the reanalysis , the visual inspection , the majority of new detections showed @xmath0 indices below the limit of 0.5 applied before , which is why they were not detected in paper i. index plotted vs. determined period without treatment of systematic effects for all stars in the best ii field lra02a ( top ) and lra02b ( bottom ) .
variable stars identified in paper i are marked with blue dots , whereas variable stars from the manual reanalysis appear red .
the dashed line shows the selection limit of @xmath28 as applied in paper i. [ fig : jper],title="fig : " ] index plotted vs. determined period without treatment of systematic effects for all stars in the best ii field lra02a ( top ) and lra02b ( bottom )
. variable stars identified in paper i are marked with blue dots , whereas variable stars from the manual reanalysis appear red .
the dashed line shows the selection limit of @xmath28 as applied in paper i. [ fig : jper],title="fig : " ] the relation between periods and variability is shown in figure [ fig : jper ] . a clear bulk of stars can be found at about @xmath29 , with the limits being widely period independent and populated by stars showing noise only .
most new variable star detections are found in the region between @xmath30 and the cutoff limit of @xmath28 from paper i. a small number of new variables with @xmath31 were not detected in paper i because their @xmath0 indices have been initially underestimated ( see section [ sec : reanalysis ] ) .
the dominant variation in many light curves is due to diurnal systematics , aliasing , or a combination of both .
figure [ fig : jper ] shows a large accumulation of stars having periods of one day or integral fractions / multiples thereof , often in combination with large @xmath0 indices .
consequently , this leads to a very high number of false alarms when using the stetson index as the only criterion for variability selection . in the example of the data set lra02 , a cutoff limit of @xmath30 would be sufficient to include all variable stars in the selection sample , but
only @xmath32 of all light curves would be sorted out .
the remaining large sample of 31,000 stars is mainly affected by systematic effects and contains only 681 stars with real physical variability ( see section [ sec : results ] ) .
the corresponding false alarm rate of about 98% shows the need for an automated treatment of systematic variability , which is not part of the @xmath0 index .
two fundamental criteria are used to assess the quality of period search algorithms : the significance of the detection itself and the correct determination of the frequency of variability .
first , a quantity @xmath33 is introduced to evaluate the detection efficiency of any given search algorithm .
detection methods are usually based upon a single numerical value @xmath34 ( e.g. , the stetson index , @xmath35 ) that can be used to prioritize a candidate list .
the success of ranking variable stars high in the list is measured with @xmath33 for each tested search algorithm .
it ranges from 0 for the perfect algorithm ( all previously identified variable stars listed first ) to 1 ( listed last ) .
for details on the calculation of @xmath33 see the appendix .
the second criterion is tested by comparing the frequency @xmath26correct}}}}$ ] that was verified manually with the frequency of a tested algorithm .
we consider a tolerance range of 2% around @xmath26correct}}}}$ ] for a correct determination .
also included are 2% deviation around half or twice that value , because the distinction between these is often ambiguous from the light curve itself .
the fraction @xmath36 of correctly identified frequencies can then be used for a quantitative comparison between tested algorithms .
because systematics and their aliases are usually limited to a set of few well - defined frequencies @xmath37sys}}}}\}$ ] , they can be excluded by searching the best frequency @xmath26max}}}}$ ] only on a subset @xmath38sys}}}}\}$ ] ( figure [ fig : sysfreq - sets ] ) .
we tested three different methods to account for systematic frequencies , both independent of each other as well as in combination .
_ master power spectrum .
systematic periodic signals affect many light curves in a data set and can thus be distinguished from real stellar variability by analyzing many power spectra @xmath39 of individual stars @xmath40 statistically .
we use the mean of all @xmath41 spectra to build a _ master _
spectrum @xmath42 in order to identify significant peaks in the master spectrum , a baseline fit is determined . among several investigated functions , a polynomial log - log fit , @xmath43base}}}}(\omega ) ) = \sum_{i=0}^{4}{c_i \cdot ( \ln \omega)^i},\ ] ] with coefficients
@xmath44 maps the baseline best and most reliable for various tested data sets .
the baseline - subtracted spectrum can then be searched for systematics .
figure [ fig : masterspec ] shows how @xmath45base}}}}$ ] peaks clearly at the diurnal frequencies that we aim to identify .
finally , a simple cutoff using the standard deviation @xmath46base}}}}$ ] of the subtracted spectrum @xmath45base}}}}$ ] is applied to filter systematics automatically .
the corresponding set of frequencies having peaks at least @xmath47base}}}}$ ] above the average power spectrum is defined by @xmath48sys}}}}_1(n_1 ) = \ { \omega \ |\ \theta_m(\omega ) > \theta_m{\ensuremath{^{\mbox{\smaller[2]base}}}}(\omega ) + n_1\cdot \sigma_m{\ensuremath{^{\mbox{\smaller[2]base } } } } \},\ ] ] where the parameter @xmath49 can be adjusted to quantify the degree of exclusion .
_ frequencies with empty phases .
_ ground - based observations are strongly affected by periodic gaps in the data , most commonly due to the diurnal cycle .
the incomplete phase coverage leads to aliasing and can often cause false positive detections .
we use a simple model to exclude frequencies with poor phase coverage : first , the folded light curve is split into @xmath50boxes}}}}=100 $ ] intervals of the same length .
the number of empty intervals @xmath50boxes}}}}{\ensuremath{^{\mbox{\smaller[2]empty}}}}(\omega)$ ] is then counted for each sampled frequency . a frequency
is considered systematic by this criterion if the fraction of empty intervals is larger than a relative threshold parameter @xmath51 , i.e. , @xmath52sys}}}}_2(n_2 ) = \ { \omega \ |\ n{\ensuremath{_{\mbox{\smaller[2]boxes}}}}{\ensuremath{^{\mbox{\smaller[2]empty}}}}(\omega ) > n_2 \cdot n{\ensuremath{_{\mbox{\smaller[2]boxes } } } } \}.\ ] ] both criteria [ eq : n1 ] and [ eq : n2 ] are merged to exclude systematic frequencies determined in either way , i.e. , the overall set of non - systematic frequencies ( figure [ fig : sysfreq - sets ] ) is defined by @xmath53sys}}}}_1(n_1 ) \cap \omega{\ensuremath{^{\mbox{\smaller[2]sys}}}}_2(n_2 ) \right)\ ] ] and the search for @xmath26max}}}}$ ] is restricted from the whole frequency range ( equation ( [ eq : wmax ] ) ) to the subset @xmath54 : @xmath55max}}}}^{(1)}).\ ] ] note that the function @xmath50boxes}}}}{\ensuremath{^{\mbox{\smaller[2]empty}}}}(\omega)$ ] is similar but not equivalent to the window function @xmath56 @xcite . limiting the latter
was tested as an alternative criterion ; it shows a very similar ranking performance , but slightly less correctly determined frequencies .
being more simple , the empty phase criterion was chosen for the final test setup .
_ power spectrum scaling .
_ in addition to the exclusion of systematic frequencies , we also investigated a method to include the information about systematics into the aov results directly . instead of searching for the maximum of @xmath57 , an artificial spectrum @xmath58
is created by dividing the aov spectrum @xmath25 of every star by the master power spectrum @xmath59 .
its maximum is found at the frequency @xmath26max}}}}^{(2)}$ ] in analogy to equation ( [ eq : wmax1 ] ) : @xmath60max}}}}^{(2)})\ .\ ] ] in addition to the @xmath0 index , two methods to prioritize variable star candidates were tested .
the first method takes the aov result directly , i.e. , @xmath61max}}}}^{(1)}).\ ] ] in its special case of no excluded systematic frequencies ( @xmath62 , @xmath63 ) , this is a widespread method for prioritizing variable star candidates .
likewise , the maximum of the divided power spectrum @xmath64 could serve as a variability indicator : @xmath65max}}}}^{(2)}).\ ] ] the aov statistic @xmath66 compares the quadratic norm of a model @xmath67 ( with @xmath68 free parameters ) with the residuals that remain after subtraction of the model from @xmath18 observations @xmath69 @xcite .
because it has optimum period detection properties @xcite , we expect @xmath70 to yield the best ranking .
however , as an empirical alternative we also tested the light curve s standard deviation with and without the periodic signal obtained by the aov multiharmonic fit : @xmath71max}}}}^{(1 ) } ) } \hspace{5 mm } \mbox{and } \ \ \ q_2^{(2 ) } = \frac{\sigma}{\sigma'(\omega{\ensuremath{_{\mbox{\smaller[2]max}}}}^{(2)})}\ ] ] it is dependent upon the frequency @xmath26max}}}}$ ] determined in the previous section , which is why different choices of @xmath26max}}}}$ ] lead to different rankings .
the quoted @xmath72 refers to the standard variation _
after _ subtraction of the corresponding fit ( for calculation of coefficients , see * ? ? ?
ccp5mmccp5mmccp5mmc @xmath73 & lra02a & & @xmath74 & @xmath75 & & @xmath76 & @xmath77 & & @xmath78 + & lra02b & & @xmath79 & @xmath80 & & @xmath81 & @xmath82 & & @xmath83 + + @xmath84 & lra02a & & & & + & lra02b & & & & + + the quantities @xmath33 and @xmath36 have been calculated for both frequencies @xmath26max}}}}^{(1)}$ ] and @xmath26max}}}}^{(2)}$ ] and both tested ranking methods @xmath85 . for each run , the parameters @xmath49 and @xmath51 were varied independently on the following values : @xmath86 the results of the comparison are summarized in table [ tab : alg - comparison ] , which shows the best value for @xmath33 and @xmath36 achievable with each tested method .
paper i and the first step of the reanalysis ( section [ sec : reanalysis ] ) determined the stellar variability using aov periodograms with two harmonics , which was also used for this comparison .
furthermore , the number of harmonics was set to @xmath87 in a second test in order to increase the sensitivity on sharp signals that are , e.g. , caused by eclipsing binaries ( for the sensitivity dependence on the number of harmonics , see , e.g. , * ? ? ?
both results show very similar ranking performances ( for the best ranking method @xmath70 , we find @xmath88 in both data sets ) , but the @xmath89 test naturally yields a slightly better ( 12%15% ) frequency match with the initial run that was obtained using the same number of harmonics .
however , the test with seven harmonics revealed a number of additional interesting eclipsing binaries that could not be detected using the smaller number of model parameters ( see sections [ sec : results : newmethod ] and [ sec : results : cat ] ) . therefore , we prefer the latter for our improved variability search ( section [ sec : newselmethod ] ) and focus in the following on the details of the search performance with @xmath87 harmonics . the performance of the ranking differs only slightly between the tested methods .
figure [ fig : xicomparison ] shows how the quantities @xmath70 ( aov ) and @xmath90 ( @xmath91-ratio ) both provide a sorting that lists stars with real variability first . the numerical quantity @xmath33 yields with @xmath92 for @xmath93 a minimum close to the optimal ranking ( @xmath94 ) .
it is an order of magnitude lower than the previously used @xmath0 index ( @xmath95 ) and significantly lower than ranking the aov power without exclusion of systematics ( @xmath96 for @xmath97 with @xmath98 and @xmath99 ) . in particular , this corresponds to a drastically decreased false alarm rate .
figure [ fig : xi : n1n2 ] shows the dependence of the ranking performance on the parameters @xmath49 and @xmath51 for the two best methods @xmath97 and @xmath93 . in both cases ,
the most restrictive exclusion of systematic frequencies yields the best sorting .
thereby , the cut in the master power spectrum ( equation ( [ eq : n1 ] ) ) has a slightly larger impact than the exclusion of empty phases ( equation ( [ eq : n2 ] ) ) .
the minimum of @xmath33 is reached for @xmath100 , but is almost independent of @xmath51 , because the first criterion is more restrictive .
the frequencies found in paper i and by manual reanalysis of the data set are in good agreement with the values of @xmath26max}}}}^{(k)}$ ] ( see table [ tab : alg - comparison ] and figure [ fig : foundfrac ] ) . without master spectrum division ( @xmath101 ) , about two - thirds of the frequencies are recovered .
however , the yield increases to about 80% if the procedure is applied ( @xmath102 ) .
interestingly , the frequency exclusion from the first criterion ( equation ( [ eq : n1 ] ) ) now has the reverse effect
the maximal agreement is reached if it is almost switched off by setting @xmath103 .
smaller values of @xmath49 are too restrictive and can increase the number of wrong periods by up to about 10% .
on the other hand , the exclusion of empty phases has again a small influence , although a value of @xmath104 yields a slight improvement for the majority of tested scenarios ( figure [ fig : foundfrac ] ) .
the remaining small group of variable stars detected with a different period has been analyzed carefully .
the majority of them shows multi - period variation and was identified with a rational multiple ( e.g. , 1/7 , 2/5 ) of the original frequency .
for some stars , the original period had to be revised during the reanalysis ( see also section [ sec : results ] ) .
a small rest shows amplitudes close to the noise level , such that the period could not be determined unambiguously . based upon the results from the comparison
, a new procedure was set up to search for variable stars within the best project . 1 .
the @xmath0 index is used to exclude non - variable stars in order to save computation time .
high values of @xmath0 can originate from either real variability or systematic trends , leading to a very high false alarm rate when being used as the only ranking criterion .
however , low @xmath0 indices give a reliable criterion for non - variability , i.e. , neither physical nor systematic variations . in the studied data set lra02 ,
_ no _ star shows clear variability below @xmath30 .
this limit is therefore used for analyses of best ii data sets , for which it typically excludes 50%75% of all stars .
the aov algorithm is applied with @xmath87 harmonics to the selected subset ( @xmath105 ) in order to obtain power spectra .
the improved selection method is applied to rank all investigated stars . following the results from section [ sec : limitimprove4 ] ,
a master power spectrum is calculated , the number of empty boxes is counted for each test period , and each individual power spectrum is divided by the master power spectrum ( equation ( [ eq : psscale ] ) ) . for the ranking , the frequency @xmath26max}}}}^{(2)}$ ] is determined from the subset of non - systematic frequencies in the divided spectrum @xmath106 by following equations ( [ eq : n1])([eq : n12 ] ) and ( [ eq : wmax2 ] ) with the parameters @xmath107 and @xmath104 .
the corresponding maximum @xmath108max}}}}^{(2)})$ ] ( equation ( [ eq : q12 ] ) ) serves as the quantity @xmath34 for priorization ( see example in figure [ fig : specexamples ] ) .
however , in order to improve the final period @xmath109 , @xmath26max}}}}^{(2)}$ ] is recalculated without exclusion of systematic frequencies from the master power spectrum , i.e. , by setting @xmath110 , @xmath104 and applying equations ( [ eq : n1])([eq : n12 ] ) and ( [ eq : wmax2 ] ) again .
all light curves are folded with their respective final periods @xmath109 and analyzed visually in descending order of @xmath111max}}}}^{(2)})$ ] .
in addition to the 350 variables already published in paper i , 279 stars in lra02 were identified with clear periodic variability ( 114 in lra02a and 165 in lra02b ) .
furthermore , we identified 52 suspected periodic variable stars ( 17 in lra02a and 35 in lra02b ) . for the latter ,
the quality of the light curves is not sufficient to fully exclude systematic errors as sources of variability , the folded light curves are partly incomplete or the point - spread functions of two neighboring stars overlap each other such that the signal can not be well separated .
the number of detections in this work is compared with paper i in table [ tab : oldnew ] .
lccc & lra02a & lra02b & total + paper i & 173 ( 4 ) & 177 ( 1 ) & 350 ( 5 ) + this paper & 114 ( 2 ) & 165 ( 5 ) & 279 ( 7 ) + this paper ( suspected ) & 17 ( 0 ) & 35 ( 0 ) & 52 ( 0 ) + total & 304 ( 6 ) & 377 ( 6 ) & 681 ( 12 ) + * note . *
the number of previously known variables in the field confirmed by best ii is given in brackets ( included in first number ) .
a large number of 189 new variable stars was already identified in the best ii data set lra02 by the first step of the reanalysis , the manual screening ( see section [ sec : reanalysis ] ) .
after the search procedure was tested and optimized using the results from paper i and this additional sample of detections , the most successful sorting method ( see description in section [ sec : limitimprove ] ) was finally applied to search the data set lra02 once more .
in addition to the 350 variable stars published in paper i and the 189 found by the manual reanalysis , this improved search yielded another 135 previously unknown variable stars .
reasons why these went undetected by previous searches are as follows . *
systematic trends or aliases were found instead of the real periodicity . *
the aov algorithm was run with @xmath89 harmonics for paper i and the manual reanalysis , but with @xmath87 for the latest search .
this leads to a higher sensitivity for non - sinusoidal variations , which is particularly important for the detection of algol type eclipsing binaries ( at least 27 additional ea variables with long periods ( @xmath112 days ) can be attributed to the increase of @xmath113 , some of them being very eccentric ) . * a total of 21 variable stars on the edge of the best ii fov with few datapoints were excluded by earlier reductions of the data set .
due to an improved data quality ( figure [ fig : rmsplot ] ) and increased sensitivity for non - sinusoidal events , we could also refine the periods for 17 of the variable stars published in paper i ( see table 4 ) .
the newly discovered periodic variable stars are classified by the shape , amplitude , and period of their folded light curves according to the scheme used by the gcvs .
we identified pulsating stars of the delta scuti type ( dsct ) , rr lyrae type ( rr ) , and cepheid variables ( cep ) .
eclipsing binary star systems were classified as algol type ( ea ) , beta lyrae type ( eb ) , or w ursae majoris type ( ew ) .
stars having sinusoidal - like light curves and showing eclipses are marked as rotating ellipsoidal variables ( ell ) , whereas some light curves exhibit characteristic features of spotted stars ( sp ) . in case the type of variability
could not be determined from the light curve alone , we classified a star as var , meaning that further observations are needed to constrain the proper physical origin of stellar variability .
stars that are variable on time scales longer than the observational baseline are classified as long periodic ( lp ) .
this class can also include non - periodic variables .
the newly identified periodic variable stars of the best ii data set lra02 are listed in table [ tab : vars1 ] .
a set of stars from paper i with revised ephemerides is presented in table [ tab : vars2 ] . due to the reanalysis ,
the internal numbering is not consistent with paper i , which is why ids were given different prefixes ( i.e. , _
lra02a2 _ and _ lra02b2 _ , respectively ) .
corresponding ids of the nearest two micron all sky survey ( 2mass ) object are given if the catalog coordinates do not differ from our astrometry by more than 2 arcsec . in some cases
, two neighboring stars could be spatially resolved , but their photometric apertures overlap so that the actual source of variation is ambiguous .
such objects have been marked with a `` c '' ( crowded ) , whereas suspected variables are indicated with an `` s '' flag .
all instrumental magnitudes given are obtained without any filter so that their absolute value should only be considered as a rough estimation .
magnitude of usno - a2 , yielding an rms deviation with the catalog of about 0.5 mag for all matched stars . ] amplitudes and ephemerides ( @xmath114 in rhjd = hjd-2,454,400 ) are the results of the aov algorithm . however , after visual inspection , several periods have been refined manually ( usually to rational multiples of the initial result ) .
lccccccccccl best i d & flag & 2mass i d & @xmath115 & @xmath116 & @xmath117 ( mag ) & t@xmath118 [ rhjd ] & @xmath119 ( d ) & @xmath120 ( mag ) & @xmath0 index & type & other names + + lra02a2_00759 & & 06473611 - 0352264 & @xmath121 & @xmath122 & 18.11 & 34.638 & @xmath123 & @xmath124 & 0.306 & ea & + lra02a2_01976 & & 06483833 - 0312255 & @xmath125 & @xmath126 & 16.34 & 35.163 & @xmath127 & @xmath128 & 1.37 & ea / sp & + lra02a2_03383 & & 06481122 - 0346301 & @xmath129 & @xmath130 & 17.13 & 35.097 & @xmath131 & @xmath132 & 0.323 & cep & + lra02a2_04701 & & 06472903 - 0432171 & @xmath133 & @xmath134 & 15.25 & 38.772 & @xmath135 & @xmath136 & 0.445 & ell / sp & + lra02a2_05225 & & 06483988 - 0339178 & @xmath137 & @xmath138 & 16.51 & 34.937 & @xmath139 & @xmath132 & 0.812 & ell / sp & + lra02a2_05867 & & 06475987 - 0416581 & @xmath140 & @xmath141 & 17.72 & 34.841 & @xmath142 & @xmath124 & 0.457 & eb & + lra02a2_06108 & & 06475271 - 0424518 & @xmath143 & @xmath144 & 17.00 & 34.601 & @xmath145 & @xmath146 & 0.469 & ea & + lra02a2_06147 & & 06492768 - 0308050 & @xmath147 & @xmath148 & 13.87 & 34.711 & @xmath149 & @xmath150 & 23.6 & eb & + lra02a2_06168 & & 06483769 - 0348472 & @xmath151 & @xmath152 & 16.15 & 35.256 & @xmath153 & @xmath154 & 0.807 & ea & + lra02a2_06363 & @xmath155 & 06482994 - 0356402 & @xmath156 & @xmath157 & 15.80 & 35.944 & @xmath158 & @xmath159 & 0.548 & ell & + lra02a2_06993 & & 06482849 - 0403248 & @xmath160 & @xmath161 & 13.82 & 34.705 & @xmath162 & @xmath163 & 5.75 & ew / ell & corot 300002950 + lra02a2_07010 & & 06481669 - 0413092 & @xmath164 & @xmath165 & 14.56 & 35.923 & @xmath166 & @xmath136 & 0.658 & ea & corot 300002493 + lra02a2_07056 & & 06484660 - 0349153 & @xmath167 & @xmath168 & 17.20 & 34.934 & @xmath169 & @xmath170 & 0.838 & ew & + lra02a2_07090 & & 06493379 - 0311214 & @xmath171 & @xmath172 & 17.24 & 34.925 & @xmath173 & @xmath174 & 0.374 & ew & + lra02a2_07148 & & 06480066 - 0427255 & @xmath175 & @xmath176 & 16.74 & 34.676 & @xmath177 & @xmath178 & 1.62 & ew & + * notes .
* stars have been matched with the closest 2mass object within a radius of 2 arcsec .
coordinates are given for epoch j2000.0 and were derived by an astrometric match of ccd to usno - a2 coordinates .
given magnitudes are instrumental and reflect the ccd sensitivity as observations have been obtained without filter .
overlapping apertures of neighboring stars can lead to contaminated light curves
such cases are marked with a `` @xmath179 '' flag .
suspected variables are marked by `` @xmath155 '' .
amplitudes @xmath120 and ephemerides ( variability period @xmath119 and times of minimum brightness @xmath114 , given in rhjd = hjd-2454400 ) are the results of the aov algorithm .
( this table is available in its entirety in machine - readable and virtual observatory ( vo ) forms in the online journal .
a portion is shown here for guidance regarding its form and content . ) [ cols="<,^,^,^,^,^,^,^,^,^,^,<",options="header " , ] the corresponding phase - folded light curves can be found in figures [ fig : lcs : new ] and [ fig : lcs : revised ] . for stars classified as lp ,
light curves are shown unfolded and no ephemerides are given in tables [ tab : vars1 ] and 4 .
light curves in a machine - readable format as well as finding charts are available upon request .
we searched the vsx and the gcvs for previously known variable stars within a radius of 10 arcsec around each of our new detections .
a total number of seven variables was found in these catalogs .
for comparison , the new best ii results for these are included in table [ tab : vars1 ] and marked with `` k '' as known variables .
the three stars corot 110742676 , nsvs 12579155 , and nsvs12585233 have periods longer than 50 days .
best ii confirms their long - time periodicity , but the phase coverage of their cycles is insufficient to confirm the periods quantitatively .
for the four eclipsing binaries
asas j064835 - 0534.3 , dymon , [ kee2007]1318 , and [ kee2007 ] 1334 , both the classifications and periods are completely confirmed .
the latter two have first been detected by @xcite in the corot ir01 field with best , for which the chosen fov shows a small overlap with the best ii field lra02a .
however , the precision in the periods of these two binaries could be significantly improved because the lra02 data set covers a much larger time period ( 41 compared to 12 nights ) and the photometric quality of best ii is better .
the lra02 data sets of best ii and corot are , except for the fact that they point to the same field , completely independent of each other and thus provide a great opportunity to compare the scientific results of the two surveys .
a step in the scientific analysis of corot data consists of an automatic stellar variability classification @xcite .
this method was also applied to the corot observations of field lra02 , and the results are meanwhile together with the full light curves publicly available through the corot archive . the corot lra02 data set contains 11,448 targets , from which 10,392 ( 91% ) are matching a best ii target within a maximum angular distance of @xmath180 .
because lra02 was not covered completely by best ii ( figure 1 in paper i ) , 454 corot targets are located outside of the best ii fov .
furthermore , the magnitude ranges do not overlap completely , so that 425 bright corot stars are saturated on the best ii ccd .
the remaining 177 corot targets are within the fov and right magnitude range , but have no best ii counterpart due to technical issues such as blooming . in the same way ,
best ii observed a total of 93,943 stars in both pointings that have not been given a corot mask or are located outside the corot fov . from the 681 variable stars presented together in paper
i and this work , 262 variables ( 190 from paper i , 72 from this work ) match a corot target .
the corresponding corot ids are shown for the new detections in tables [ tab : vars1 ] and 4 .
corot observed the lra02 field about 1 year after best ii , and the observing times do not overlap . because this work focuses on the improvement of variable star detection for ground - based telescopes , corot and best ii light curves
have not been combined .
however , we note that such a combination might yield improved ephemerides . best ii light curves of interesting objects will be provided on request . for stars both observed with corot and best ii , it is possible to investigate the performance of best ii more in detail .
the automatic classification @xcite provides information about the amplitude of variation for corot targets . because the satellite has a much higher photometric precision , most of these amplitudes are well below the detection limit of best ii .
however , the knowledge of stellar variability with higher precision can be used to evaluate the detection efficiency of best ii .
of particular interest are two questions : first , how many stars with sufficiently high variation in the corot data set have been detected as variables from best ii data .
second , if these are clearly distinguished by the new detection algorithm from the rest of stars having variabilities below the threshold of best ii .
the full amplitude @xmath181 of variation was derived from the fit coefficients given in the corot classification .
as this contains each a low- and high - frequency entry for every corot target , the amplitude was calculated as the maximum of both .
the value @xmath181 itself contains no information on whether the signal can be detected by best ii or not , which strongly depends on the magnitude of a given star .
therefore , the quantity @xmath182min}}}}(r_b)\ ] ] is used to estimate the variability signal to noise ( s / n ) .
the noise @xmath183min}}}}(r_b)$ ] gives the photometric precision achievable with best ii in the given data set for a star of magnitude @xmath117 .
it was determined by a fit to the @xmath91-magnitude plot of the field ( figure [ fig : rmsplot ] ) .
figure [ fig : detefficiency ] shows the expected s / n for all stars that are contained in both data sets . from 680 stars with corot amplitudes that should be visible in the best ii data ( @xmath184 ) , 162 were detected as variable stars in paper i and this work .
448 stars are expected to have a large @xmath184 , but show no significant variability in the best ii data set ( @xmath185 ) .
70 stars with both expected and measured variability ( @xmath184 , @xmath186 ) were not detected by best ii .
we have checked corot light curves with significant variability in the corot classification , but that are not variable in the best ii data .
many of them show strong instrumental effects ( hot pixels ; * ? ? ? * ) which obviously mislead the automatic classification algorithm . however ,
from the 448 stars in this region , only 88 show a probability larger than 95% to belong to any class so that most can be considered false alarms .
stars that are expected to be variable ( @xmath184 ) and show variation in the best ii light curve ( @xmath186 ) , but were not detected as variable stars after visual inspection have been re - inspected .
such targets have been missed due to the smaller phase coverage of best ii , or because the real noise of individual light curves is underestimated using @xmath183min}}}}$ ] ( e.g. , due to higher order extinction effects for very red stars ) .
the distinction between variable and non - variable stars works very well using the new ranking quantity @xmath111max}}}}^{(2)})$ ] ( see section [ sec : limitimprove ] ) . if the limit @xmath187 is chosen to separate variables from the bulk of non - variable stars , 92% of all matched stars are found below the limit .
only seven variables have @xmath185 , but their light curves and low s / n indicate rather false detections than too low variability values . from all 780 stars with @xmath186 , one third belongs to the set of variable star detections . for comparison ,
figure [ fig : detefficiency ] also shows the @xmath0 index vs. the expected s / n .
the plot shows clearly that the separation between real and artifical variability is much weaker .
the strength of the new ranking is particularly clear in the regime of @xmath188 , i.e. , close to the detection limit of best ii .
the variable star detections of best ii were compared in detail with the automatic corot classification for the 262 matched variables .
the overall agreement between both methods is very good ; details regarding the period and magnitude determination as well as the classifications obtained by best ii and corot are given in the remainder of this section .
figure [ fig : lcs : corot ] shows some instructive examples of variable stars in agreement ( a ) and with differences in the determined periods and/or classifications ( ( b)(f ) ) .
( a ) ( b ) ( c ) ( d ) ( e ) ( f ) the mean magnitudes of matched stars are in reasonably good agreement ( figure [ fig : bestcorot : mag ] ) .
only a few very long - term variables like the example in figure [ fig : lcs : corot](d ) show differences in the order of 1 mag and above because best ii and corot observed during different phases ; the remaining majority differs by only @xmath189 mag .
periods determined by best ii have been compared with the main frequency from the automatic corot characterization . for 72.5% of the matched stars , the periods are equal or integral ( @xmath190 ) multiples of each other to a precision of at least 1% .
the histogram of period ratios in figure [ fig : bestcorot : per ] shows that most detections have been identified in the corot data with half the period compared to best ii .
this is because many periods are doubled during the visual inspection of best ii light curves in order to show full cycles ( e.g. , figures [ fig : lcs : corot](c ) and [ fig : lcs : corot](e ) ) , in particular for w ursae majoris eclipsing binaries ( ew ) .
the classes from visual inspection of best ii light curves match the automatic classification of the corot data set well .
from the 262 variables present in both data sets , 196 stars have best ii variability classes that are consistent with either the short- or long - periodic classification in the corot data set . note that the variability classes used by best ii and corot are slightly different : for example , best ii distinguishes within the corot class ecl between the eclipsing binary types ea , eb , and ew , while the corot scheme includes , e.g. , slowly pulsating b - stars , which are simply identified as var within the best ii study .
all such consistent refinements are considered a classification match .
the 64 stars with a clear disagreement in the variability classifications were checked carefully by reviewing the light curves from both best ii and corot . for 26 stars , the variability classes obtained by best
ii appear more realistic .
most of such cases are lp variable stars ( e.g. , figure [ fig : lcs : corot](d ) ) that have been identified as such in the best ii data set by visual inspection . for these cases ,
even the longer corot baseline does not cover a full cycle , so that the period and classification obtained from the corot pipeline are not conclusive .
furthermore , some eclipsing binaries are clearly misclassified by the automatic corot procedure . figure [ fig : lcs : corot](e ) shows an example of a w ursae majoris type eclipsing binary that was identified as an rr lyrae pulsator
most likely because it was detected with half of its physical period from the corot data set . for three stars similar to the example in figure [ fig : lcs : corot](b ) ,
narrow eclipse events were not detected by the corot analysis .
most of the stars with implausible variability types were classified as be by the automatic classification , which was described by @xcite as a `` trash '' class regarding its wide parameter spread . for 16 stars like the example in figure [ fig : lcs : corot](f ) , the corot classifications are in better agreement with the measurements . for most of these cases ,
this clearly results from the better photometric quality of the satellite data . for 22 cases ,
the photometric data itself are insufficient to choose between the corot and best ii classifications ( e.g. , ew / ell ) .
no automatic classification data were available for the two corot targets 110833621 and 300001413 .
the corot target field lra02 was observed with best ii during 41 nights from 2007 november to 2008 february ( see paper i ) .
we reanalyzed the data set in order to improve the detection method and to maximize the number of detections .
in addition to the 350 periodic variable stars already published earlier ( paper i ) , we present a catalog of 272 new variables and 52 stars with suspected variability . for seven known variables both the classifications and periods are confirmed , and the periods could significantly be improved for two of them .
revised ephemerides are presented for 17 variable stars from paper i. from a manual inspection of all light curves without any preselection criterion , it turned out that most of the new periodic variable stars went undetected in paper i because the applied variability criterion using the stetson @xmath0 index was too restrictive . but
because the @xmath0 index is heavily biased by systematic effects , a smaller cutoff limit leads to very high false alarm rates ( @xmath191 ) and is therefore not a practical alternative .
however , although the @xmath0 index is not capable to distinguish between systematic and stellar variability , it can still be used to exclude non - variable stars from the analysis : no variable star is falsely rejected if light curves with low variability indices @xmath192 are sorted out in both data sets of this study .
this separation can be particularly useful if the full aov process is too time consuming for a whole data set .
the deep characterization of the data set enabled us to compare and train different automatic methods for an improved variability ranking .
in particular , a master power spectrum was calculated as the mean of all individual aov spectra .
this method proved a valuable tool for exclusion of systematic frequencies and hence the ranking of real variability .
the best algorithm found separates variable stars very effectively from the non - variable background population and in parallel recovers their frequencies well .
the new ranking method is particularly superior to the @xmath0 index in regimes where the amplitude of variation becomes comparable to the noise level , i.e. , close to the detection limit of the photometric system .
it shows an almost equal performance for both independent subsets lra02a and lra02b , so that it should be easily applicable to other data sets .
finally , the results from both paper i and this work have been compared with the publicly available corot data set .
best ii obtained light curves for 91% of all corot targets in lra02 . from all 681 variable stars observed with best ii in the field , 262 are matched with corot targets .
stellar amplitudes measured by corot show that best ii detects variability efficiently in its parameter range .
the results of period determination and variability classification are found to be in very good overall agreement , confirming the validity of the measurements , data reduction , and scientific analysis of both projects .
196 stars show consistent classifications .
corot yields a better classification in case of small variability for 16 targets , e.g. , be stars . on the other hand , the final visual screening of folded best ii light curves proves to be an important step in the analysis . for 26 stars
it was found to yield a more realistic classification , in particular for eclipsing binaries and stars with variability on time scales comparable to the observational baseline .
* acknowledgments .
* this work was funded by deutsches zentrum fr luft- und raumfahrt and partly by the nordrhein - westflische akademie der wissenschaften .
we kindly thank julian petrasch for his great commitment of scanning all folded light curves .
furthermore , we appreciate valuable comments on the initial manuscript by the anonymous referee .
our research made use of the 2mass , usno - a2 and gcvs catalogs , the aavso variable star search index and the simbad database , operated at cds , strasbourg , france .
let @xmath193 with @xmath194 be the group of all stars in a data set .
the selection of variable stars
e.g. , by sorting all stars by a ranking quantity @xmath34 can then be considered a permutation @xmath195 that splits @xmath196 into a part @xmath197 containing all @xmath198 variable stars and another part with the rest : @xmath199 a check of the first @xmath198 within @xmath200 would thus reveal all variable stars in the data set .
unfortunately , such an _ optimal _ sorting @xmath201 is usually unknown . in practice
, a given permutation @xmath202 aims at a similar splitting of variable and non - variable stars , but contaminates both groups with false positives .
the number of identified variable stars @xmath203 thus depends on the number of stars @xmath204 that are actually checked : for a comparison between different approaches , it is useful to evaluate the performance of @xmath202 as a whole . for that , one can define the quality parameter @xmath208 the parameter @xmath33 sums missed variable stars for all values of @xmath209 .
the denominator accounts for normalization such that @xmath94 for @xmath201 ( best selection ) and @xmath210 for its counterpart ( worst selection ) .
alonso , r. , guenther , e. w. , almenara , j .-
m . , et al .
2012 , , submitted auvergne , m. , bodin , p. , boisnard , l. , et al .
2009 , , 506 , 411 baglin , a. , auvergne , m. , boisnard , l. , et al .
2006 , in 36th cospar scientific assembly , 3749 , g. , noyes , r. w. , kovcs , g. , et al .
2004 , , 116 , 266 , j. , sarro , l. m. , aerts , c. , et al .
2007 , , 475 , 1159 , j. , sarro , l. m. , lpez , m. , et al .
2009 , , 506 , 519 , h. j. , gillon , m. , shporer , et al .
2009 , , 506 , 343 , t. j. 1975 , , 36 , 137 , j. d. , bakos , g. . , noyes , r. w. , et al .
2011 , , 141 , 166 kabath , p. , eigmller , p. , erikson , a. , et al .
2007 , , 134 , 1560 kabath , p. , eigmller , p. , erikson , a. , et al . 2008 , , 136 , 654 kabath , p. , erikson , a. , rauer , h. , et al .
2009a , , 506 , 569 ( paper i ) kabath , p. , fruth , t. , rauer , h. , et al . 2009b , , 137 , 3911 karoff , c. , rauer , h. , erikson , a. , et al .
2007 , , 134 , 766 , g. , dimitrov , d. , neuhuser , r. , et al .
2011 , , 411 , 1204 pasternacki , t. , csizmadia , sz . , cabrera , j. , et al .
2011 , , 142 , 114 , j. , & burke , c. j. 2006 , , 132 , 1177 rauer , h. , eislffel , j. , erikson , a. , et al .
2004 , , 116 , 38 rauer , h. , erikson , a. , kabath , p. , et al .
2010 , , 139 , 53 samus , n. n. , durlevich , o. v. , et al .
2009 , vizier online data catalog , b / gcvs schwarzenberg - czerny , a. 1996 , , 460 , 107 schwarzenberg - czerny , a. 1999 , , 516 , 315 schwarzenberg - czerny , a. , & kaluzny , j. 1998 , , 300 , 251 , i. , poleski , r. , udalski , a. , et al .
2008 , , 58 , 163 stetson , p.b .
1996 , , 108 , 851 zhang , x .- b . , deng , l .- c . , xin , y. , & zhou , x. 2003 , chin . j. astron .
astrophys . , 3 , 151 | the corot field lra02 has been observed with the berlin exoplanet search telescope ii ( best ii ) during the southern summer 2007/2008 .
a first analysis of stellar variability led to the publication of 345 newly discovered variable stars .
now , a deeper analysis of this data set was used to optimize the variability search procedure .
several methods and parameters have been tested in order to improve the selection process compared to the widely used @xmath0 index for variability ranking .
this paper describes an empirical approach to treat systematic trends in photometric data based upon the analysis of variance statistics that can significantly decrease the rate of false detections .
finally , the process of reanalysis and method improvement has virtually doubled the number of variable stars compared to the first analysis by kabath et al .
a supplementary catalog of 272 previously unknown periodic variables plus 52 stars with suspected variability is presented .
improved ephemerides are given for 19 known variables in the field .
in addition , the best ii results are compared with corot data and its automatic variability classification . |
weakly - interacting electrons in a three - dimensional ( 3d ) periodic potential are well - described by landau - fermi liquid theory @xcite , in which the free electrons of a fermi gas become dressed quasiparticles with renormalised dynamical properties .
conversely , in the one - dimensional ( 1d ) limit a tomonaga - luttinger liquid ( tll ) is formed @xcite , where single - particle excitations are replaced by highly correlated collective excitations .
so far , it has proved difficult to interpolate theoretically between these two regimes , either by strengthening electron - electron ( @xmath3-@xmath3 ) interactions in 3d , or by incorporating weak transverse coupling into 1d models @xcite .
the invariable presence of disorder in real materials places further demands on theory , particularly in the description of ordered electronic ground states .
q1d systems such as nanowire ropes , filamentary networks or single crystals with uniaxial anisotropy therefore represent an opportunity to experimentally probe what theories aspire to model : strongly correlated electrons subject to disorder in a highly anisotropic 3d environment . physical properties of q1d materials may vary considerably with temperature .
tll theory is expected to be valid at elevated temperatures , since electrons can not hop coherently perpendicular to the high - symmetry axis and q1d systems behave as decoupled arrays of 1d filaments .
phase - coherent single - particle hopping can only occur below temperature @xmath4 ( where @xmath5 is the transverse hopping integral ) , at which a dimensional crossover to an anisotropic quasi - three - dimensional ( q3d ) electron liquid is anticipated @xcite .
the properties of such q3d liquids remain largely unknown , especially the role of electronic correlations in determining the ground state . at low temperature ,
a tll is unstable to either density wave ( dw ) or superconducting fluctuations , depending on whether the @xmath3-@xmath3 interaction is repulsive ( due to coulomb forces ) or attractive ( from electron - phonon coupling ) .
following dimensional crossover , the influence of such interactions in the q3d state is unclear . as an example , electrical transport in the tll state of the q1d purple bronze li@xmath6mo@xmath2o@xmath7
is dominated by repulsive @xmath3-@xmath3 interactions @xcite , yet a superconducting transition occurs below 1.9 k. disorder adds further complication to q1d materials due to its tendency to localise electrons at low temperature . for dimensionality @xmath8 , localisation occurs for any non - zero disorder ; in contrast , for @xmath9 a critical disorder is required and a mobility edge separates extended from localised states .
the question of whether a mobility edge can form in q1d materials after crossover to a q3d liquid state is open , as is the microscopic nature of the localised phase .
disorder also renormalises @xmath3-@xmath3 interactions , leading to a dynamic amplification of the coulomb repulsion @xcite and a weaker enhancement of phonon - mediated @xmath3-@xmath3 attraction , i.e. cooper pairing @xcite .
we therefore anticipate that disorder should strongly suppress superconductivity in q1d materials , unless the coulomb interaction is unusually weak or screened . in this work ,
we show that the q1d superconductor na@xmath1mo@xmath2se@xmath2 provides a unique environment in which to study the interplay between dimensionality , electronic correlations and disorder .
although na@xmath1mo@xmath2se@xmath2 is metallic at room temperature , the presence of na vacancy disorder leads to electron localisation and a divergent resistivity @xmath10 at low temperature , prior to a superconducting transition .
in contrast with all other known superconductors , the onset temperature for superconducting fluctuations @xmath11 is positively correlated with the level of disorder .
normal - state electrical transport measurements also display signatures of an attractive @xmath3-@xmath3 interaction , which is consistent with disorder - enhanced superconductivity .
a plausible explanation for these phenomena is an intrinsic screening of the long - range coulomb repulsion in na@xmath1mo@xmath2se@xmath2 , arising from the high polarisability of disordered q1d materials .
the combination of disorder and q1d crystal symmetry constitutes a new recipe for strongly correlated electron liquids with tunable electronic properties .
* crystal and electronic structure of na@xmath1mo@xmath2se@xmath2 * na@xmath1mo@xmath2se@xmath2 belongs to the q1d @xmath12mo@xmath2se@xmath2 family @xcite ( @xmath13 = group ia alkali metals , tl , in ) which crystallise with hexagonal space group p6@xmath14 m .
the structure can be considered as a linear condensation of mo@xmath2se@xmath15 clusters into infinite - length ( mo@xmath2se@xmath2)@xmath16 chains parallel to the hexagonal @xmath17 axis , weakly coupled by @xmath13 atoms ( fig .
[ fig1]a ) .
the q1d nature of these materials is apparent from the needle - like morphology of as - grown crystals ( fig .
[ fig1]b ; see methods for growth details ) .
_ ab initio _ calculations ( supplementary note i ) using density functional theory reveal an electronic structure which is uniquely simple amongst q1d metals . a single spin - degenerate band of predominant mo @xmath18 character crosses the fermi energy @xmath19 at half - filling ( fig . [ fig1]c , supplementary fig . 1 ) ,
creating a 1d fermi surface composed of two sheets lying close to the brillouin zone boundaries at @xmath20 ( where @xmath17 is the @xmath17-axis lattice parameter ) .
the warping of these sheets ( and hence the coupling between ( mo@xmath2se@xmath2)@xmath16 chains ) is controlled by the @xmath13 cation , yielding values for @xmath5 ranging from 230 k ( @xmath13 = tl ) to 30 k ( @xmath13 = rb ) ( supplementary fig . 2 ) .
in addition to tuning the dimensionality , @xmath13 also controls the ground state : @xmath13 = tl , in are superconductors @xcite , while @xmath13 = k , rb become insulating at low temperature @xcite . within the @xmath12mo@xmath2se@xmath2 family ,
@xmath13 = na is attractive for two reasons .
firstly , we calculate an intermediate @xmath21 120 k , suggesting that na@xmath1mo@xmath2se@xmath2 lies at the threshold between superconducting and insulating instabilities . secondly , the combination of the small na cation size and a high growth temperature ( 1750@xmath22c ) result in substantial na vacancy formation during crystal synthesis . since the na atoms are a charge reservoir for the ( mo@xmath2se@xmath2)@xmath16 chains , these vacancies will reduce @xmath19 and lead to an incommensurate band filling . despite the reduction in carrier density , the density of states @xmath23 remains constant for na@xmath24 ( fig .
[ fig1]d , supplementary note i ) .
energy - dispersive x - ray spectrometry on our crystals indicates na contents from 1.7 to 2 , comfortably within this range .
this is confirmed by synchrotron x - ray diffraction ( xrd ) on three randomly - chosen crystals : structural refinements reveal na deficiencies of 11@xmath251% , 11@xmath252% and 13@xmath254% ( i.e. @xmath26 = 0.22 , 0.22 , 0.26 ) , but the ( mo@xmath2se@xmath2)@xmath16 chains remain highly ordered . no deviation from the @xmath12mo@xmath2se@xmath2 structure is observed between 293 k and 20 k , ruling out any lattice distortions such as the peierls transition , which often afflicts q1d metals . to probe the na vacancy distribution , we perform diffuse x - ray scattering experiments on the @xmath26 = 0.26 crystal .
no trace of any huang scattering ( from clustered na vacancies ) or structured diffuse scattering from short - range vacancy ordering is observed ( supplementary fig . 3 , supplementary note ii ) .
na vacancies therefore create an intrinsic , random disorder potential in na@xmath1mo@xmath2se@xmath2 single crystals .
* normal - state electrical transport *
we first examine the electrical transport at high energy for signatures of disorder and one - dimensionality .
the temperature dependence of the resistivity @xmath10 for 6 randomly - selected na@xmath1mo@xmath2se@xmath2 crystals @xmath27-@xmath28 is shown in fig .
@xmath29(300k ) increases by more than one order of magnitude from crystal @xmath27 to @xmath28 ( fig .
[ fig2]b ) : such large differences between crystals can not be attributed to changes in the carrier density due to na stoichiometry variation and must instead arise from disorder . despite the variance in @xmath29(300k ) , the evolution of @xmath10 is qualitatively similar in all crystals . upon cooling ,
@xmath10 exhibits metallic behaviour before passing through a broad minimum at @xmath30 and diverging at lower temperature .
@xmath30 falls from 150 k to @xmath31 70 k as @xmath29(300k ) decreases ( fig .
[ fig2]c ) , suggesting that the divergence in @xmath10 and the disorder level are linked .
upturns or divergence in @xmath10 have been widely reported in q1d materials and variously attributed to localisation @xcite , multiband tll physics @xcite , dw formation @xcite , incipient density fluctuations @xcite and proximity to mott instabilities @xcite .
differentiating between these mechanisms has proved challenging , in part due to the microscopic similarity between localised electrons and randomly - pinned dws in 1d .
we briefly remark that the broad minimum in @xmath10 in na@xmath1mo@xmath2se@xmath2 contrasts strongly with the abrupt jumps in @xmath10 for nesting - driven dw materials such as nbse@xmath32 @xcite , while any mott transition will be suppressed due to the non - stoichiometric na content .
instead , a disordered tll provides a natural explanation for this unusual crossover from metallic to insulating behaviour . at temperatures
@xmath33 , power - law behaviour in @xmath10 is a signature of tll behaviour in a q1d metal .
fitting @xmath34 in the high - temperature metallic regime of our crystals consistently yields @xmath35 ( fig .
[ fig2]a ) . in a clean half - filled tll
, this would correspond to a luttinger parameter @xmath36 1 , i.e. non - interacting electrons . however , disorder renormalises the @xmath3-@xmath3 interactions : for a commensurate chain of spinless fermions , @xmath37 and a critical point separates localised from delocalised ground states at @xmath38 @xcite .
our experimental values for @xmath39 therefore indicate that na@xmath1mo@xmath2se@xmath2 lies close to this critical point .
although the effects of incommensurate band filling on a disordered tll remain unclear , comparison with clean tlls suggests that removing electrons reduces @xmath40 . for @xmath41 , @xmath10 is predicted to be metallic at high temperature , before passing through a minimum at @xmath30 ( which rises with increasing disorder ) and diverging at lower temperature .
these features are consistently reproduced in our data . within the disordered tll paradigm ,
our high - temperature transport data indicate that the @xmath3-@xmath3 interaction is attractive , i.e. @xmath42 .
this implies that electron - phonon coupling dominates over coulomb repulsion and suggests that the coulomb interaction may be intrinsically screened in na@xmath1mo@xmath2se@xmath2 .
a quantitative analysis of the low - temperature divergence in @xmath10 provides further support for the influence of disorder as well as a weak / screened coulomb repulsion .
we have attempted to fit @xmath10 using a wide variety of resistive mechanisms : gap formation ( arrhenius activation ) , repulsive tll power laws , weak and strong localisation ( supplementary fig . 4 , supplementary note iii ) .
among these models , only mott variable range hopping @xcite ( vrh ) consistently provides an accurate description of our data .
vrh describes charge transport by strongly - localised electrons : in a @xmath43-dimensional material @xmath44 $ ] , where @xmath45 is the characteristic vrh temperature ( which rises as the disorder increases ) and @xmath46 .
although mott s original model assumed that hopping occurred via inelastic electron - phonon scattering , vrh has also been predicted to occur via @xmath3-@xmath3 interactions in disordered tlls @xcite .
figure [ fig3]a displays vrh fits for crystals @xmath27-@xmath28 , while fits to @xmath10 in three further crystals which cracked during subsequent measurements are shown in supplementary fig . 5 .
all our crystals yield values for @xmath43 ranging from 1.2 to 1.7 ( supplementary table i ) , in good agreement with the @xmath47 1.5 predicted for arrays of disordered conducting chains @xcite .
coulomb repulsion in disordered materials opens a soft ( quadratic ) gap at @xmath19 , leading to vrh transport with @xmath48 regardless of the actual dimensionality .
we consistently observe @xmath0 , implying that localised states are present at @xmath19 and no gap develops in na@xmath1mo@xmath2se@xmath2 .
a small paramagnetic contribution also emerges in the dc magnetisation below @xmath30 and rises non - linearly with @xmath49 ( supplementary fig . 6 ) .
similar behaviour has previously been attributed to a progressive crossover from pauli to curie paramagnetism due to electron localisation ( supplementary note iv ) .
although @xmath10 exhibits vrh divergence in all crystals prior to peaking at @xmath11 , a dramatic increase in @xmath50 by four orders of magnitude occurs between crystals @xmath51 and @xmath52 .
this is reminiscent of the rapid rise in resistivity upon crossing the mobility edge in disordered 3d materials .
our data are therefore suggestive of a crossover to strong localisation and the existence of a critical disorder or `` q1d mobility edge ''
. such behaviour may also originate from proximity to the @xmath38 critical point .
interestingly , the critical disorder approximately correlates with the experimental condition @xmath53 , where @xmath54 is the estimated single - particle dimensional crossover temperature ( fig .
[ fig2]c ) .
this suggests a possible role for dimensional crossover in establishing the mobility edge .
further evidence for criticality is seen in the frequency dependence of the conductivity @xmath55 within the divergent @xmath10 regime ( fig .
[ fig3]b ) . for crystals with sub - critical disorder
, @xmath55 remains constant at low frequency , as expected for a disordered metal .
in contrast , @xmath55 in samples with super - critical disorder rises with frequency , following a @xmath56 trend .
this is quantitatively compatible with both the mott - berezinskii formula for localised non - interacting electrons in 1d @xcite and the expected behaviour of a disordered chain of interacting fermions @xcite .
the strong variation of @xmath55 even at sub - khz frequencies implies that the localisation length @xmath57 is macroscopic , in contrast with the @xmath58 100 nm expected from mott vrh theory @xcite .
however , it has been predicted that the relevant localisation lengthscale for a weakly - disordered q1d crystal is the larkin ( phase distortion ) length , which may be exponentially large @xcite .
the evolution of the magnetoresistance ( mr ) @xmath59 with temperature also supports a localisation scenario . above @xmath30 , @xmath59 is weakly positive and follows the expected @xmath60 dependence for an open fermi surface ( fig . [
fig3]c ) . at lower temperature ,
the divergence in @xmath10 correlates with a crossover to strongly negative mr within the vrh regime ( fig .
[ fig3]d ) .
the presence of a soft coulomb gap at @xmath19 would lead to a positive mr within the vrh regime @xcite ; in contrast , our observed negative mr in na@xmath1mo@xmath2se@xmath2 corresponds to a delocalisation of gapless electronic states @xcite and provides additional evidence for a screened coulomb interaction .
the mr switches sign again below @xmath11 and becomes positive ( fig .
[ fig3]e ) : as we shall now demonstrate , this is a signature of superconductivity . * superconducting transitions in na@xmath1mo@xmath2se@xmath2 * the presence of a superconducting ground state @xcite in tl@xmath61mo@xmath2se@xmath2 and in@xmath61mo@xmath2se@xmath2 implies that the peak in @xmath10 below 6 k is likely to signify the onset of superconductivity in na@xmath1mo@xmath2se@xmath2 .
upon cooling crystals @xmath62 in a dilution refrigerator , we uncover a 2-step superconducting transition characteristic of strongly anisotropic q1d superconductors @xcite ( fig .
[ fig4]a - c ) .
below @xmath11 , superconducting fluctuations initially develop along individual ( mo@xmath2se@xmath2)@xmath16 chains and @xmath10 is well - described by a 1d phase slip model ( supplementary note v ) .
subsequently , a weak hump in @xmath10 emerges ( fig .
[ fig4]d - f ) at temperatures ranging from @xmath31 0.95 k ( crystal @xmath27 ) to @xmath31 1.7 k ( crystal @xmath51 ) .
this hump signifies the onset of transverse phase coherence due to inter - chain coupling .
cooper pairs can now tunnel between the chains and a meissner effect is expected to develop , but we are unable to observe this since 1.7 k lies below the operational range of our magnetometer .
analysis of the current - voltage characteristics indicates that a phase - coherent superconducting ground state is indeed established at low temperature ( supplementary fig . 7 , supplementary note vi ) .
we estimate an anisotropy @xmath63 6.0 in the coherence length , which is lower than the experimental values for tl@xmath61mo@xmath2se@xmath2 and in@xmath61mo@xmath2se@xmath2 ( 13 and 17 respectively @xcite ) in spite of the smaller @xmath5 in na@xmath1mo@xmath2se@xmath2 ( supplementary fig . 2 ; see methods for magnetic field orientation details ) .
this anisotropy is also far smaller than the measured conductivity ratio at 300k : @xmath64 57 . in comparison
, close agreement is obtained between the anisotropies in @xmath65 and @xmath66 for li@xmath6mo@xmath2o@xmath7 @xcite , where the effects of disorder are believed to be weak @xcite .
the disparate anisotropies in na@xmath1mo@xmath2se@xmath2 arise from a strong suppression of @xmath67 , thus illustrating the essential role of disorder in controlling the low temperature properties of na@xmath1mo@xmath2se@xmath2 . although superconducting fluctuations are observed regardless of the level of disorder in na@xmath1mo@xmath2se@xmath2 , it is important to identify whether phase - coherent long range order develops in crystals @xmath52-@xmath28 which exhibit super - critical disorder . in fig .
[ fig4]g - i we demonstrate that @xmath10 in these samples still follows a 1d phase slip model , albeit with a strongly enhanced contribution from quantum phase slips due to the increased disorder @xcite ( supplementary note v ) .
the fitting parameters for our 1d phase slip analysis are listed in supplementary table ii .
a weak meissner effect also develops in the magnetization below @xmath31 3.5 k in crystals @xmath52 and @xmath68 ( fig .
[ fig4]g , h , j ) , but is rapidly suppressed by a magnetic field . low transverse phase stiffness is common in q1d superconductors : for example , bulk phase coherence in carbon nanotube arrays is quenched by 2 - 3 t , yet pairing persists up to 28 t @xcite .
the superconducting volume fraction corresponding to the magnitude of this meissner effect is also unusually low : @xmath69 0.1% .
magnetic measurements of the superconducting volume fraction in q1d materials invariably yield values below 100% , since the magnetic penetration depth @xmath70 normal to the 1d axis can reach several microns @xcite and diamagnetic flux exclusion is incomplete . for a typical na@xmath1mo@xmath2se@xmath2 crystal of diameter @xmath71 100 @xmath72 m , we estimate that a 0.1% volume fraction would require @xmath73 10 @xmath72 m , which seems excessively large .
conversely , an array of phase - fluctuating 1d superconducting filaments would not generate any meissner effect at all .
we therefore attribute the unusually small meissner signal to inhomogeneity in the superconducting order parameter , which is predicted to emerge in the presence of intense disorder @xcite . in an inhomogeneous superconductor ,
meissner screening is achieved via josephson coupling between isolated superconducting islands @xcite . within a single super - critically disordered na@xmath1mo@xmath2se@xmath2 crystal
, we therefore anticipate the formation of multiple josephson - coupled networks comprising individual superconducting filaments
. the total magnitude of the diamagnetic screening currents flowing percolatively through each network will be much smaller than that in a homogeneous sample due to the smaller @xmath74 ratio , thus diminishing the meissner effect .
* enhancement of superconductivity by disorder * we have established a clear influence of disorder on electrical transport in na@xmath1mo@xmath2se@xmath2 ( figs . [ fig2],[fig3 ] ) and demonstrated that the peak in @xmath10 at @xmath11 corresponds to the onset of superconductivity ( fig . [ fig4 ] ) .
let us now examine the effects of disorder on the superconducting ground state .
figure [ fig5]a illustrates @xmath11 rising monotonically from crystal @xmath27 to @xmath28 . plotting @xmath11 as a function of @xmath29(300k ) ( which is an approximate measure of the static disorder in each crystal ) , we observe a step - like feature between crystals @xmath51 and @xmath52 , i.e. at the critical disorder ( fig .
[ fig5]b ) .
strikingly , the characteristic vrh temperature @xmath45 which we extract from our @xmath10 fits ( fig .
[ fig3]a ) displays an identical dependence on @xmath29(300k ) .
this implies that disorder controls both the superconducting ground state and the insulating tendency in @xmath10 at low temperature . the positive correlation between @xmath11 and @xmath45 ( fig .
[ fig5]c ) confirms that the onset temperature for superconducting fluctuations ( and hence the pairing energy @xmath75 ) is enhanced by localisation in na@xmath1mo@xmath2se@xmath2 .
a concomitant increase in the transverse coherence temperature ( supplementary note vi ) implies that some enhancement in the phase stiffness also occurs . super - critical disorder furthermore enables superconducting fluctuations to survive in high magnetic fields ( fig .
[ fig5]d - g ) . in crystal @xmath51 ( which lies below the q1d mobility edge )
, superconductivity is completely quenched at all temperatures ( i.e. @xmath76 ) by @xmath77 4 t ( fig .
[ fig5]d , f ) . a giant negative mr reappears for @xmath78 t ( fig .
[ fig3]e ) , confirming that superconductivity originates from pairing between localised electrons .
in contrast , the peak at @xmath50 in the highly - disordered crystal @xmath28 is strikingly resistant to magnetic fields ( fig .
[ fig5]e , g ) : at @xmath79 = 4.6 k , our observed @xmath80 14 t which exceeds the weak - coupling pauli pair - breaking limit @xmath81 3 t by a factor @xmath82 4 ( see supplementary note vii for a derivation of @xmath83 ) .
a similar resilience is evident from the positive mr in crystal @xmath52 , which persists up to at least 14 t at 1.8 k ( fig . [ fig3]e ) .
triplet pairing is unlikely to occur in na@xmath1mo@xmath2se@xmath2 ( since scattering would rapidly suppress a nodal order parameter ) and orbital limiting is also suppressed ( since vortices can not form across phase - incoherent filaments ) .
our data therefore suggest that disorder lifts @xmath84 , creating anomalously strong correlations which raise the pairing energy @xmath75
@xcite above the weak - coupling @xmath85 . a direct spectroscopic technique would be required to determine the absolute enhancement of @xmath75 , since spin - orbit scattering from the heavy mo ions will also contribute to raising @xmath84 .
the emergence of a superconducting ground state in na@xmath1mo@xmath2se@xmath2 places further constraints on the origin of the normal - state divergence in @xmath10 .
our electronic structure calculations indicate that the q1d fermi surface of na@xmath1mo@xmath2se@xmath2 is almost perfectly nested : any incipient electronic dw would therefore gap the entire fermi surface , creating clear signatures of a gap in @xmath10 and leaving no electrons at @xmath19 to form a superconducting condensate .
in contrast , our vrh fits and mr data do not support the formation of a dw gap , and a superconducting transition occurs at low temperature .
electrons must therefore remain at @xmath19 for all @xmath86 , indicating that @xmath10 diverges due to disorder - induced localisation rather than any other insulating instability .
it has been known since the 1950s that an @xmath87-wave superconducting order parameter is resilient to disorder @xcite , provided that the localisation length @xmath57 remains larger than the coherence length ( i.e. the cooper pair radius ) .
however , experiments have invariably shown superconductivity to be destroyed by disorder , due to enhanced coulomb pair - breaking @xcite , phase fluctuations @xcite or emergent spatial inhomogeneity @xcite . in particular
, increasing disorder in li@xmath6mo@xmath2o@xmath7 ( one of the few q1d superconductors extensively studied in the literature ) monotonically suppresses superconductivity @xcite .
therefore , the key question arising from our work is why the onset temperature for superconductivity rises with disorder in na@xmath1mo@xmath2se@xmath2 , in contrast to all other known materials ?
disorder acts to enhance the matrix element for @xmath3-@xmath3 interactions .
this may be explained qualitatively by considering that all conduction electron wavefunctions experience the same disorder - induced potential , developing inhomogeneous multifractal probability densities @xcite and hence becoming spatially correlated .
such enhanced correlations have been predicted to increase the cooper pairing energy @xcite : in the absence of pair - breaking by long - ranged coulomb interactions , this will lead to a rise in the superconducting transition temperature @xcite .
a proposal to observe this effect in superconducting heterostructures with built - in coulomb screening @xcite ( by depositing superconducting thin films on substrates with high dielectric constants ) has not yet been experimentally realised . however , our vrh dimensionality @xmath0 ( fig .
[ fig3]a ) and negative mr ( fig . [ fig3]d , e ) both point towards a weak or screened coulomb repulsion , while the power - laws and broad minima in @xmath10 at high temperature ( fig . [ fig2]a ) indicate a luttinger parameter @xmath42 .
these results all imply that @xmath3-@xmath3 interactions in na@xmath1mo@xmath2se@xmath2 are attractive .
( for comparison , @xmath88 in li@xmath6mo@xmath2o@xmath7 and the @xmath3-@xmath3 interaction is repulsive @xcite . )
phonon - mediated coupling - the cooper channel - therefore appears to dominate over the coulomb repulsion in na@xmath1mo@xmath2se@xmath2 , suggesting that the usual disorder - induced coulomb pair - breaking may be avoided . below the q1d mobility edge ,
our rise in @xmath11 is quantitatively compatible with a weak multifractal scenario ( supplementary fig .
8 , supplementary note viii ) , providing a possible explanation for the enhancement of superconductivity which merits further theoretical attention .
the fact that no experimental examples of q1d materials with attractive @xmath3-@xmath3 interactions have yet been reported poses the question why na@xmath1mo@xmath2se@xmath2 should be different .
although strong electron - phonon coupling is known to play an important role in the physics of molybdenum cluster compounds @xcite , we propose that the disordered q1d nature of na@xmath1mo@xmath2se@xmath2 may instead play the dominant role , by suppressing the coulomb repulsion . in the presence of disorder
, a q1d material can be regarded as a parallel array of `` interrupted strands '' @xcite , i.e. a bundle of finite - length nanowires .
the electric polarisability of metallic nanoparticles is strongly enhanced relative to bulk materials @xcite , although this effect is usually cancelled out by self - depolarisation .
the geometric depolarisation factor vanishes for q1d symmetry , leading to giant dielectric constants @xmath89 which rise as the filament length increases @xcite .
this effect was recently observed in au nanowires @xcite , with @xmath89 reaching @xmath90 . in na@xmath1mo@xmath2se@xmath2
, we therefore anticipate that the long - range coulomb repulsion in an individual ( mo@xmath2se@xmath2)@xmath91 filament ( @xmath92 ) will be efficiently screened by neighbouring filaments @xcite .
this intrinsic screening provides a natural explanation for attractive @xmath3-@xmath3 interactions and suppresses coulomb pair - breaking in the superconducting phase .
it has been suggested that impurities can increase the temperature at which transverse phase coherence is established in q1d superconductors @xcite .
this effect can not be responsible for our observed rise in @xmath11 , which corresponds to the onset of 1d superconducting fluctuations on individual ( mo@xmath2se@xmath2)@xmath91 filaments .
we also point out that the finite - size effects which influence critical temperatures in granular @xcite or nanomaterials @xcite are not relevant in na@xmath1mo@xmath2se@xmath2 : quantum confinement is absent in homogeneously - disordered crystalline superconductors and hence no peaks form in @xmath23 .
these mechanisms are discussed in detail in supplementary note ix . in summary , we have presented experimental evidence for the enhancement of superconductivity by disorder in na@xmath1mo@xmath2se@xmath2 .
the combination of q1d crystal symmetry ( and the associated dimensional crossover ) , disorder and incommensurate band filling in this material poses a challenge to existing 1d
/ q1d theoretical models .
although the normal - state electrical resistivity of na@xmath1mo@xmath2se@xmath2 is compatible with theories for disordered 1d systems with attractive electron - electron interactions , we establish several unusual low - temperature transport properties which deserve future attention .
these include a resistivity which diverges following a q1d vrh law for all levels of disorder , the existence of a critical disorder or q1d mobility edge where @xmath53 , and a strongly frequency - dependent conductivity @xmath93 in crystals with super - critical disorder . at temperature @xmath11 ,
1d superconducting fluctuations develop , and a phase - coherent ground state is established via coupling between 1d filaments at lower temperature .
as the disorder rises , @xmath11 increases : in our most - disordered crystals , the survival of superconducting fluctuations in magnetic fields at least four times larger than the pauli limit suggests that the pairing energy may be unusually large .
we conclude that deliberately introducing disorder into q1d crystals represents a new path towards engineering correlated electron materials , in remarkable contrast with the conventional blend of strong coulomb repulsion and a high density of states . beyond enhancing superconductivity , the ability to simultaneously modulate band filling , disorder and dimensionality
promises a high level of control over emergent order , including density waves and magnetic phases . more generally , na@xmath1mo@xmath2se@xmath2 and other similar interrupted strand materials may be ideal environments in which to study the evolution of many - body electron localisation beyond the non - interacting anderson limit
* * crystal growth and initial characterisation**a series of na@xmath1mo@xmath2se@xmath2 crystals was grown using a solid - state synthesis procedure .
the precursor materials were mose@xmath61 , inse , mo and nacl , all in powder form . before use
, the mo powder was reduced under h@xmath61 gas flowing at 1000@xmath22c for ten hours , in order to eliminate any trace of oxygen .
the mose@xmath61 was prepared by reacting se with h@xmath61-reduced mo in a ratio 2:1 inside a purged , evacuated and flame - baked silica tube ( with a residual pressure of @xmath31 10@xmath94 mbar argon ) , which was then heated to @xmath31 700@xmath22c for two days .
inse was synthesised from elemental in and se in an evacuated sealed silica tube at 800@xmath22c for 1 day .
powder samples of na@xmath1mo@xmath2se@xmath2 were prepared in two steps .
firstly , in@xmath61mo@xmath2se@xmath2 was synthesised from a stoichiometric mixture of inse , mose@xmath61 and mo , heated to 1000@xmath22c in an evacuated sealed silica tube for 36 hours . secondly ,
an ion exchange reaction of in@xmath61mo@xmath2se@xmath2 with nacl was performed at 800@xmath22c , using a 10% nacl excess to ensure total exchange as described in reference @xcite .
all starting reagents were found to be monophase on the basis of their powder x - ray diffraction patterns , acquired using a d8 bruker advance diffractometer equipped with a lynxeye detector ( cuk@xmath95 radiation ) .
furthermore , in order to avoid any contamination by oxygen and moisture , the starting reagents were kept and handled in a purified argon - filled glovebox . to synthesise single crystals , a na@xmath1mo@xmath2se@xmath2 powder sample ( of mass @xmath31 5 g ) was cold - pressed and loaded into a molybdenum crucible , which had previously been outgassed at 1500@xmath22c for 15 minutes under a dynamic vacuum of @xmath31 10@xmath96 mbar .
the mo crucible was subsequently sealed under a low argon pressure using an arc - welding system .
the na@xmath1mo@xmath2se@xmath2 powder charge was heated at a rate of 300@xmath22c / hour up to 1750@xmath22c , held at this temperature for 3 hours , then cooled at 100@xmath22c / hour down to 1000@xmath22c and finally cooled naturally to room temperature within the furnace .
crystals obtained using this procedure have a needle - like shape with length up to 4 mm and a hexagonal cross section with typical diameter @xmath97 150 @xmath72 m .
initial semi - quantitative microanalyses using a jeol jsm 6400 scanning electron microscope equipped with an oxford inca energy - dispersive - type x - ray ( edx ) spectrometer indicated that the na contents ranged between 1.7 and 2 , i.e. up to 15% deficiency . the na deficiency results from the high temperatures used during the crystal growth process coupled with the small size of the na ion : it can not be accurately controlled within the conditions necessary for crystal growth .
since in@xmath61mo@xmath2se@xmath2 is known to be superconducting below 2.85 k @xcite , it is important to consider the possibility of in contamination in our samples .
the na / in ion exchange technique used during synthesis is known to be highly efficient @xcite and in@xmath61mo@xmath2se@xmath2 decomposes above 1300@xmath22c , well below our crystal growth temperature ( 1750@xmath22c ) .
this precludes the presence of any superconducting in@xmath61mo@xmath2se@xmath2 ( or in - rich ( in , na)@xmath61mo@xmath2se@xmath2 ) filaments in our crystals .
diffuse x - ray scattering measurements accordingly reveal none of the huang scattering or disk - like bragg reflections which would be produced by such filaments .
furthermore , edx spectrometry is unable to detect any in content in our crystals , while inductively - coupled plasma mass spectrometry indicates a typical in residual of less than 0.01% , i.e. @xmath69 0.0002 in atoms per unit cell .
the electronic properties of na@xmath1mo@xmath2se@xmath2 crystals will remain unaffected by such a tiny in residual in solid solution . * * electrical transport measurements**before all measurements , the as - grown crystal surfaces were briefly cleaned with dilute hydrochloric acid ( to remove any residue from the mo crucible and hence minimise the contact resistance ) , followed by distilled water , acetone and ethanol .
four au contact pads were sputtered onto the upper surface and sides of each crystal using an al foil mask ; 50 @xmath72 m au wires were then glued to these pads using silver - loaded epoxy cured at 70@xmath22c ( epotek e4110 ) .
especial care was taken to thoroughly coat each end of the crystal with epoxy , to ensure that the measurement current passed through the entire crystal .
all contacts were verified to be ohmic at room temperature before and after each series of transport measurements , and at @xmath98 k after cooling .
typical contact resistances were of the order of 2 @xmath99 at 300 k. the transverse conductivity @xmath100 was estimated at room temperature using a four - probe technique , with contacts on opposite hexagonal faces of a single crystal .
the temperature dependence of the transverse resistivity @xmath101 has never been accurately measured in @xmath12mo@xmath2se@xmath2 due to the exceptionally large anisotropies , small crystal diameters and high fragility , even in the least anisotropic tl@xmath61mo@xmath2se@xmath2 which forms the largest crystals @xcite .
low - frequency four - wire ac conductivity measurements were performed in two separate cryogen - free systems : a variable temperature cryostat and a dilution refrigerator , both of which may be used in conjunction with a superconducting vector magnet .
the ac conductivity was measured using a keithley 6100 current source , a stanford srs850 lock - in amplifier with input impedance 10 m@xmath99 and ( for low resistances , i.e. weakly - disordered samples ) a stanford sr550 preamplifier with input impedance 100 m@xmath99 .
data from several crystals were cross - checked using a quantum design physical property measurement system ( ppms ) with the standard inbuilt ac transport hardware : both methods generate identical , reproducible data . with the exception of the frequency - dependence studies in fig .
[ fig3]b , all the transport data which we present in our manuscript are acquired with an ac excitation frequency of 1 hz , i.e. we are measuring in the dc limit . at 1 hz
, the phase angle remained zero at all temperatures in all crystals .
therefore , no extrinsic capacitance effects are present in our data .
the typical resistance of a weakly disordered crystal lies in the 1 - 10 @xmath99 range . in contrast , the absolute resistances of crystals @xmath102 at @xmath11 are 41.9 k@xmath99 , 33.7k@xmath99 and 27.6k@xmath99 respectively : the crystal diameter increases from @xmath102 , thus explaining the rise in resistivity despite a fall in resistance .
these values remain much smaller than our lock - in amplifier input impedance , ruling out any current leakage in highly disordered crystals .
our measurement current @xmath103 10 @xmath72a leads to a maximum power dissipation @xmath104 @xmath72w .
this is negligible compared with the @xmath31 2 mw cooling power at 2k on our cryostat cold finger and we may hence rule out any sample heating effects in our data .
we acquire transverse magnetotransport data ( figs .
[ fig3]c - e , [ fig5]d - g ) with the magnetic field perpendicular to both the @xmath17 axis and the crystal faces , i.e. at 30@xmath22 to the hexagonal @xmath105 axis .
q1d bechgaard salts and blue / purple bronzes exhibit monoclinic crystal symmetry , and hence strong anisotropies along all three crystallographic axes .
in contrast , @xmath12mo@xmath2se@xmath2 crystallise in a hexagonal lattice : any azimuthal ( @xmath106 ) anisotropy in na@xmath1mo@xmath2se@xmath2 will therefore reflect this hexagonal symmetry . in tl@xmath61mo@xmath2se@xmath2 , this anisotropy has been variously reported to be small or entirely absent : it is at least an order of magnitude lower than the polar anisotropy at low temperature @xcite .
our conclusions regarding the reduced low temperature anisotropy in na@xmath1mo@xmath2se@xmath2 are therefore robust . in common with
most highly one - dimensional materials , na@xmath1mo@xmath2se@xmath2 crystals are extremely fragile , with a tendency to split into a forest of tangled fibres if mishandled .
the crystals therefore exhibit a finite experimental lifetime , with thermal cycling from 2 k to room temperature presenting a particular risk to their structural integrity : this explains why we were unable to obtain complete data - sets in crystals @xmath27-@xmath28 ( the magnetoresistance @xmath59 at high temperature in crystal @xmath51 and @xmath107 in crystal @xmath28 are missing , for example ) .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 .
_ _ * * , ( ) . .
_ _ * * , ( ) . .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) .
_ _ ( , ) .
_ et al . _
_ _ * * , ( ) . , & .
_ _ * * , ( ) . .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) . .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) .
_ et al . _ .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , , , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & _ _ ( , , ) . & .
_ _ * * , ( ) . & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
, , & _ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
& _ _ ( , ) . , & . _
_ * * , ( ) . , & _ _ * * , ( ) . , , &
_ _ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . , , &
_ _ * * , ( ) . , , &
_ _ * * , ( ) . .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . & _ _ * * , ( ) . , , , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) .
et al . _ . _
_ * * , ( ) .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) .
the authors declare that the data supporting the findings of this study are available within the article and its supplementary information files .
we thank alexei bosak ( beamline id28 , esrf grenoble ) for assistance with data collection and processing , and igor burmistrov , vladimir kravtsov , tomi ohtsuki and vincent sacksteder iv for stimulating discussions .
the swiss - norwegian beamlines ( esrf grenoble ) are acknowledged for beam time allocation .
this work was supported by the national research foundation , singapore , through grant nrf - crp4 - 2008 - 04 .
app and cp conceived the project ; ds , pg and mp grew the crystals ; dc performed the xrd measurements with mh and app ; da carried out the transport experiments ; app and da analysed the data ; lb contributed the electronic structure calculations ; app , da and cp wrote the paper with input from all the authors ; cp supervised the entire project .
the authors declare no competing financial interests .
* quasi - one - dimensional crystal and electronic structures in na@xmath1mo@xmath2se@xmath2a * hexagonal crystal structure of na@xmath1mo@xmath2se@xmath2 , viewed perpendicular and parallel to the @xmath17 axis . from synchrotron x - ray diffraction experiments , we measure the @xmath105 and @xmath17 axis lattice parameters to be 8.65 and 4.49 respectively at 293 k ( supplementary note ii ) .
* b * electron micrograph of a typical na@xmath1mo@xmath2se@xmath2 crystal .
the scale bar corresponds to 300 @xmath72 m .
* c * calculated energy - momentum dispersion of the conduction band within the hexagonal brillouin zone , highlighting the large bandwidth and minimal dispersion perpendicular to the chain axis . *
d * electronic density of states @xmath108 around the fermi level in na@xmath61mo@xmath2se@xmath2 . ]
[ hp ] * power - laws and minima in the normal - state resistivity @xmath10a * @xmath10 for crystals @xmath27-@xmath28 , together with power - law fits @xmath109 ( black lines , fitting range @xmath110 k ) .
@xmath30 corresponds to the minimum in @xmath10 for @xmath86 .
* b * @xmath10 plotted on a semilogarithmic scale for crystals @xmath27 and @xmath28 : @xmath111 as @xmath112 .
* c * evolution of @xmath30 with @xmath113 , which is a measure of the disorder in each crystal .
the horizontal shading indicates the estimated @xcite single - particle dimensional crossover temperature @xmath114 104 k , obtained using @xmath115 , where @xmath116 is the conduction bandwidth ( supplementary note i ) , @xmath117 and @xmath38 .
no anomaly is visible in @xmath10 at @xmath54 , suggesting either that @xmath54 may be further renormalised due to competing charge instabilities @xcite , or that signatures of tomonaga - luttinger liquid behaviour may persist even for @xmath118 @xcite .
, title="fig : " ] [ hp ] * resistive and magnetic superconducting transitions in na@xmath1mo@xmath2se@xmath2a - c * electrical resistivity @xmath119 for crystals @xmath62 .
coloured points represent experimental data ; black lines are fits to a 1d model incorporating thermal and quantum phase slips ( supplementary note v ) . * d - f * zoom views of @xmath10 in crystals @xmath62 , plotted on a semi - logarithmic scale .
the low temperature limit of our 1d phase slip fits is signalled by a hump in @xmath10 , highlighted by the transition from solid to dashed black fit lines : this corresponds to the onset of transverse phase coherence . in quasi - one - dimensional ( q1d ) superconductors , such humps form due to finite size or current effects during dimensional crossover @xcite . *
g - i * @xmath119 for the highly - disordered crystals @xmath102 .
coloured points represent experimental data ; black lines are fits to the same 1d phase slip model as in * a - c * , which accurately reproduces the broad superconducting transitions due to an increased quantum phase slip contribution ( supplementary note v ) .
inhomogeneity and spatial fluctuations of the order parameter are expected to blur the characteristic hump in @xmath10 at dimensional crossover , thus explaining its absence from our data as the disorder rises . in * g , h
* we also plot zero - field - cooled / field - cooled ( zfc / fc ) thermal hysteresis loops displaying the meissner effect in the magnetic susceptibility @xmath107 ; * j * shows a zoom view of the susceptibility in crystal @xmath68 .
data were acquired with the magnetic field parallel to the crystal @xmath17 axis and a paramagnetic background has been subtracted .
the small diamagnetic susceptibilities @xmath120 1 are due to emergent pairing inhomogeneity creating isolated superconducting islands @xcite ; @xmath121 is further decreased by the large magnetic penetration depth perpendicular to the @xmath17 axis in q1d crystals.,title="fig : " ] | a powerful approach to analysing quantum systems with dimensionality @xmath0 involves adding a weak coupling to an array of one - dimensional ( 1d ) chains .
the resultant quasi - one - dimensional ( q1d ) systems can exhibit long - range order at low temperature , but are heavily influenced by interactions and disorder due to their large anisotropies .
real q1d materials are therefore ideal candidates not only to provoke , test and refine theories of strongly correlated matter , but also to search for unusual emergent electronic phases . here
we report the unprecedented enhancement of a superconducting instability by disorder in single crystals of na@xmath1mo@xmath2se@xmath2 , a q1d superconductor comprising mose chains weakly coupled by na atoms .
we argue that disorder - enhanced coulomb pair - breaking ( which usually destroys superconductivity ) may be averted due to a screened long - range coulomb repulsion intrinsic to disordered q1d materials .
our results illustrate the capability of disorder to tune and induce new correlated electron physics in low - dimensional materials . |
the precise measurement and control of magnetic fields and field fluctuations is of crucial importance in many fundamental physics experiments .
the suppression of systematic uncertainties in experiments searching for permanent electric dipole moments ( edms ) in atoms and neutrons is one prominent example .
new generations of edm experiments with ultracold neutrons ( ucns ) aim at a higher statistical precision by the use of higher ucn flux and larger storage volumes . in such experiments magnetic field drifts and fluctuations can produce false edm signatures thereby putting stronger constraints on the magnetic field control .
although magnetometers based on superconducting quantum interference devices ( squids ) are the most sensitive magnetometers available to date they are of limited interest for monitoring magnetic fields in large volumes . moreover , squids do not measure absolute field values .
two distinct magnetometric techniques were used in past edm experiments . in the ill experiment@xcite , which has produced the presently lowest upper bound on the neutron edm , a vapor of @xmath2hg atoms filled into the ultracold neutron storage chamber ( 20 liter ) served as `` cohabitating '' magnetometer .
the pnpi experiment @xcite , on the other side , used a set of two self - oscillating cesium vapor magnetometers placed above and below the storage chamber for monitoring the field in the chamber .
both techniques have pros and contras .
co - magnetometers yield only a volume - averaged field value , which yields no information on field gradients and their fluctuations .
external magnetometers , on the other hand , do not measure the field in the volume of interest directly , but allow - if used in sufficient number - to access field distributions , thereby permitting the active control of specific multipole moments of the field .
borisov et al .
have proposed a large volume ( external ) magnetometer based on nuclear spin precession in @xmath3he @xcite .
that device uses a double pulse ramsey resonance technique , which besides its lack of spatial resolution also suffers from a lack of temporal resolution .
the pnpi experiment used two conventional state - of - the - art discharge lamp pumped self - oscillating cesium vapor magnetometers ( opm)@xcite . such types of magnetometers - developed since the 1950 s - have a shot - noise limited performance and large bandwidths .
the high spatial and temporal resolution of optically pumped alkali magnetometers thus make such devices interesting alternatives for the continuous monitoring of fields , gradients , and fluctuations thereof .
the use of alkali opms for the field control in larger volumes calls for a substantially larger number of sensor heads , which suffers from the fact that a single discharge lamp can only drive a limited number of sensors .
the steady development in the past decades of narrow - band semiconductor diode lasers makes such light sources attractive alternatives to discharge lamps .
owing to the high spectral density of its radiation a single diode laser of moderate power ( a few mw ) can be used to drive dozens of magnetometer heads . having a multichannel external magnetometer approach for a planned neutron edm experiment in mind we have performed a comparative study of state - of - the - art discharge lamp pumped magnetometers ( lpopm ) and laser pumped magnetometers ( lsopm ) using similar room temperature sensor cells ( 7 cm and 6 cm diameter respectively ) and identical electronics .
we discuss the principle of operation and details of their practical realization .
the devices were operated in two distinct modes , viz . , the self - oscillating mode and the phase - stabilized mode .
details of the development and performance of the lsopm have been published earlier @xcite .
we have determined the intrinsic sensitivities of the magnetometers and present measurements of the fluctuations of a @xmath1 field recorded simultaneously by the lsopm and the lpopm in a multi - layer magnetic shield .
the lpopm reached its ultimate shot - noise limited performance while the lsopm showed a superior intrinsic sensitivity , although its performance still lies 50% above its fundamental shot - noise limit .
an opm measures the larmor precession frequency @xmath4 of a vapor sample of spin polarized atoms in an external magnetic field @xmath5 .
in small magnetic fields @xmath6 and the field measurement reduces to a frequency measurement . in eq .
[ eq : larmorfrequency ] the subscript of @xmath7 refers to the total angular momentum of the precessing atomic state . although opms based on nuclear spin polarization were demonstrated in the past @xcite we restrict the present discussion to alkali vapors in which the precessing levels are one or both of the hyperfine ground states with total angular momentum @xmath8 , where @xmath9 is the nuclear spin .
magnetic resonance is used to measure the precession frequency by inducing resonant spin flips by a weak magnetic field @xmath10 oriented at right angles with respect to @xmath5 and oscillating at the frequency @xmath11 .
although for the magnetometer discussed here @xmath11 lies in the audio range of frequencies the index `` rf '' ( radio - frequency ) is used to comply with common notation . optical pumping with a resonant circularly
polarized light beam creates spin polarization in the medium ( room temperature alkali atom vapor contained in a glass cell ) and hence an associated net bulk magnetization .
it has been realized for many years that the pumping process is most efficient for @xmath12 resonance light driving the transition @xmath13 , although magnetometers can also be realized using @xmath14 ( @xmath15 ) light . in general the optically pumped medium
becomes transparent with respect to the pumping light , except for the spectrally resolved closed @xmath16 transition @xcite , in which case the absorption of the pumped medium increases .
the fact that the optical properties of the medium depend on its spin polarization is used to detect the magnetic resonance transition by monitoring either the power or the polarization of the transmitted or scattered light beam .
the technique is known as optically detected magnetic resonance ( odmr ) . in the present study
we have used a particular realization of the odmr technique , the so - called @xmath17-method , in which @xmath5 is oriented at 45@xmath18 with respect to the direction of propagation ( @xmath19 ) of the circularly polarized light beam .
the particular feature of that technique is that the transmitted light intensity is modulated at the frequency @xmath11 of the oscillating field , when @xmath11 is tuned close to @xmath4 .
the amplitude of the modulation depends as @xmath20 with @xmath21 .
the amplitude and phase of the modulation depend on @xmath11 as a classical lorentz oscillator with a resonance frequency @xmath4 . on resonance
the phase shift between the oscillating @xmath10-field and the transmitted light modulation is @xmath22 and for a small detuning @xmath23 the phase shift varies linearly with @xmath24 .
the width of the resonance(s ) are determined by the transverse relaxation rate of the spin polarization , which is limited by several effects . in atom - atom collisions
only the sum of the angular momenta of the collision partners is preserved but spin - exchange processes can change the individual polarizations . the rate of spin exchange depolarization is proportional to the collision rate , i.e. , to the vapor density , the spin - exchange cross - section , and the relative velocities of the collision partners .
the dominant depolarization mechanism is due to collisions of the atoms with the cell walls and the depolarization rate depends on the adsorption time of the atoms on the walls and on the wall collision rate .
this process can be suppressed by either preventing the atoms from reaching the walls through the addition of an inert buffer gas or by reducing the sticking time on the walls by a suitable coating of the wall surfaces by paraffin or silanes . however , these coatings may act as a sink for alkali atoms thereby significantly lower the atomic density @xcite . a stable vapor pressure is established by having the vapor in thermal equilibrium with a droplet of alkali metal contained in a sidearm . depolarizing collisions with the bulk metal
are suppressed by connecting the cell proper to the sidearm via a small aperture .
the intrinsic linewidth resulting from the combined action of the mentioned depolarization effects depends on temperature , quality of the wall coating and cell geometry .
besides those intrinsic broadening mechanisms the interaction with the optical field and with the oscillating magnetic field further broaden the magnetic resonance line .
these processes are known as optical and r.f .
power broadening respectively .
the ground state of alkali atoms with a nuclear spin @xmath9 splits into two hyperfine levels with total angular momenta @xmath25 with @xmath26 zeeman sublevels labeled by the magnetic quantum number @xmath27 , respectively .
the general evolution of the hyperfine levels in a magnetic field is described by the breit - rabi - formula @xcite . in low magnetic fields ( zeeman interaction @xmath28 hyperfine interaction )
the energy of the state @xmath29 is shifted by @xmath30 . here , @xmath31 is the bohr magneton , and the g - factors @xmath32 are given by @xmath33 where @xmath34 is the electronic g - factor , defined via @xmath35 and @xmath36 is the nuclear g - factor , defined via @xmath37 .
the magnetic resonance process consists in driving transitions between adjacent sublevels with a resonance frequency @xmath4 given by @xmath38 which is equivalent to eq .
[ eq : larmorfrequency ] with @xmath39 for @xmath40cs
one has @xmath41 in second order in the field @xmath5 the levels acquire an additional energy shift depending on @xmath42 and @xmath43 ( quadratic zeeman effect ) which shifts the @xmath44 where @xmath45 is the ground state hyperfine splitting .
the quadratic zeeman effect thus splits the magnetic resonance into a series of equidistant lines separated by @xmath46 for cesium ( @xmath47 ) , one has @xmath48 in the 2@xmath49 t field used here @xmath50hz , which is much smaller than the resonance linewidth . the low field approximation ( eq .
[ eq : larmorfrequency ] ) is therefore valid for the present work .
the light source is the discharge lamp or the diode laser as explained in the text .
l : lens , p : polarizer and d@xmath51 interference filter ( in case of lamp pumping ) , @xmath52 : quarter - wave plate , pd : photodiode . ] : phase shifter , vco : voltage - controlled oscillator , pid : feedback controller . ] in this work we compare the performance of lamp pumped and laser pumped magnetometers .
both devices have a common basic design consisting of the light source , the sensor head , the detector and feedback electronics ( figs .
[ fig : mag_genprinc ] , [ fig : ps_so_electronics ] ) .
the sensor head contains the sensor proper , a spherical glass cell ( 60 mm diameter for the lpopm , produced in the group of one of the authors , a.s.p . , and 70 mm diameter for the lsopm , purchased from magtech ltd . , st .
petersburg , russia ) coated with paraffin in which cesium vapor is in thermal equilibrium with a droplet of metallic cesium at room temperature .
the cell is mounted in a ( 200 mm long , 110 mm diameter ) cylindrical housing .
the pumping light is carried from the light source to the sensor cell by a multimode fiber ( 800@xmath53 diameter ) in the lsopm and by a fibre bundle ( 6 mm diameter ) in the lpopm .
the light transmitted through the cell is carried back to a detector ( photodiode ) by an identical fiber in the lsopm and by an 8 mm diameter fiber bundle in the lpopm .
the lengths of the fibers are 8 m and 5 m length respectively for the laser and the lamp pumped device .
the sensor head contains also polarization optics ( linear polarizer and quarter - wave plate ) for making the light circularly polarized prior to entering the cell as well as lenses for collimating the incoming light and focussing the outgoing light into the return fiber ( bundle ) .
particular care was taken to use only non - magnetic components in the sensor head .
the coils producing the oscillating field consist of two 70 mm diameter loops with 12 turns of copper wire each , separated by 52 mm .
when two sensors are operated in close proximity , the cross - talk of the respective r.f .
fields is avoided by sliding a 1 mm thick al cylinder over the heads .
the optical , electronic and mechanical components of the lsopm were produced at the university of fribourg , while the lpopm was realized at the ioffe institute . .
c : magnetic field coil.,title="fig : " ] + .the fribourg magnetic shield .
@xmath54 : inner diameter , @xmath55 : inner length , @xmath56 : layer thickness .
note that all layers are closed by endcaps except of layer 3 . [ cols="^,^,^,^,^",options="header " , ] the characterization of the magnetometers described in sec .
[ sec : intrsens ] was performed in fribourg .
two magnetometers were placed inside of a multi - layer cylindrical magnetic shield as shown in fig .
[ fig : fribourgshield ] and tab .
[ tab : fribourgshield ] .
the magnetic field of 2@xmath49 t was produced by a 50 cm long , 15 cm diameter solenoid driven by an ultra - low noise current supply .
for a direct comparison of the noise performance with an accuracy below 0.1pt the level of magnetic field fluctuations at the experimental site have to be kept below that level , a performance , which is hard to realize . magnetic field variations ( in a 1hz bandwidth ) in unshielded environments are of the order of several @xmath57 or more . a shielding factor exceeding 10000
is thus needed to suppress fluctuations at that level . in both types of magnetometers
the d@xmath51 transition @xmath58 of cs at a wavelength of 894 nm is used for optical pumping .
the lpopm is driven by an electrodeless discharge lamp , in which a power stabilized high frequency generator ( @xmath59 100mhz ) produces a discharge in a 12 mm diameter glass bulb containing cesium vapor and xenon as buffer gas .
the pumping light is collimated and filtered by a d@xmath51 interference filter centered at 894.5 nm with a fwhm=11.5 nm .
because of the high temperature of the discharge plasma the spectrum of the emitted d@xmath51 radiation is considerably broader than the doppler width of the room temperature absorption line in the sensor cell .
all four hyperfine components of the d@xmath51 line are excited simultaneously as indicated in fig .
[ fig : csd1](a ) . as the same light is used for detecting the ground state spin precession the lpopm detects magnetic resonance in both the f=4 and the f=3 hyperfine ground states . because of the differing g - factors of the two states ( eq . [ eq : gplusminusnumerical ] ) the corresponding magnetic resonance lines are split by 22hz in the 2@xmath60 field , which is larger than the width of the magnetic resonance lines ( @xmath592.55hz ) under optimized conditions .
the f=3 component is much weaker than the f=4 component , so that the former plays a minor role for magnetometry .
the arrows indicate the transitions driven by the discharge lamp ( a ) and the laser ( b ) . ]
the lsopm is pumped by a tunable extended cavity laser in littman configuration ( sacher lasertechnik , model tec500 ) .
the output power of more than 10mw exceeds the power required for magnetometry by more than three orders of magnitude .
therefore a single laser can be used to drive dozens of magnetometers in experiments calling for the simultaneous monitoring of the magnetic field in different locations .
the laser frequency is actively stabilized to the doppler broadened @xmath61 hyperfine transition using a dichroic atomic vapor laser lock ( davll ) @xcite in an auxiliary ( vacuum ) cesium cell .
this lock is very stable and allows a reliable operation of the magnetometer over measurement periods of several weeks . in both magnetometers the states @xmath62 are dark states and @xmath63 are also dark states . ] which can not absorb circularly polarized light . at the same time the repopulation of absorbing ( bright ) states by the magnetic resonance r.f .
transitions from these levels is used for the optical detection of the magnetic resonance .
the lsopm drives the @xmath64 hyperfine state , which do not interact with the narrow band light .
this process , called hyperfine pumping , degrades the efficiency of the optical pumping process and reduces the overall spin polarization which can be achieved with laser radiation . a straightforward way to reduce that loss
is to empty the @xmath65 sublevels using a repumping laser tuned to a transition emanating from the f=3 hyperfine ground state .
an experimental study of that process is underway .
in the lpopm all four hyperfine components are pumped simultaneously . as a result the loss due to hyperfine pumping
is excluded and a larger spin polarization is obtained in the f=4 state .
the drawback of the large spectral width of the beam from the lamp is that an appreciable part of its spectrum lies outside of the room temperature absorption spectrum of the sensor cell .
the corresponding photons carry no spectroscopic information , but produce excess shot noise in the detected photocurrent . in that respect narrow - band laser light
leads to a better detection efficiency in the lsopm .
both types of magnetometers were operated in two different modes . the self - oscillating mode ( som ) ( fig . [
fig : ps_so_electronics ] a ) uses the fact that at resonance the driving r.f . field and the modulated photocurrent are dephased by 90@xmath18 . for that reason
the sinusoidal part of the photocurrent can be used , with an appropriate amplification and phase shift to drive the r.f .
coils in a feedback loop .
in such a configuration the system will auto - oscillate at the larmor frequency .
an amplitude gain control ensures that the amplitude of the coil current is kept constant independently of phase and frequency , so that any r.f .
power dependent systematic effects are suppressed . the phase - stabilized mode ( psm ) ( fig .
[ fig : ps_so_electronics ] b ) also uses the characteristic phase dependence between the applied oscillating field and the modulation of the detected photocurrent for locking the frequency of an external oscillator to the larmor frequency .
the phase , the in - phase component , and the quadrature component are detected simultaneously by a lock - in amplifier ( stanford research systems sr830 ) .
both the in - phase signal ( dispersive lorentzian ) and the 90@xmath18 phase shifted phase signal ( arctan - dependence ) show a linear zero - crossing near zero detuning ( @xmath66 ) .
either of the two signals can thus be used as discriminant in a feedback loop , which stabilizes the phase to 90@xmath18 .
it can be shown that from a statistical point of view both signals yield an equivalent magnetic field sensitivity .
the phase signal is less sensitive to light power fluctuations , which may be advantageous to suppress systematic effects related to power fluctuations @xcite .
however , the used commercial digital lock - in amplifier had only a moderate update rate of 400hz of its phase output , so that the much faster in - phase signal was used in the feedback loop . in principle
the psm can be understood as a variant of the som , in which the phase - detector , the vco , and the feedback controller form a tracking filter . in both modes of operation changes of the magnetic field lead to instantaneous changes of the larmor frequency and thus to instantaneous changes of the transmitted modulation frequency .
the time needed for the radio - frequency to adjust to a new value after a sudden field change depends on the filters and other delays in the feedback loop .
if a very fast response is not required , as in our applications , the bandwidth can be decreased by appropriate filters . in the som
it is the preamplifier of the photodetector which limits the bandwidth to 10khz , whereas in the psm the feedback loop filters provide a bandwidth up to 1khz @xcite . from a practical point of view
the registered bandwidth is limited by the data acquisition system . because of the frequency dependence of the phase - shifter the som is , in general , optimized only for a given larmor frequency , which reduces the dynamic range of the som device .
the phase - stabilized magnetometer keeps the phase - shift at @xmath22 independently of the larmor frequency . in practice
the dynamic range is limited by the frequency range of the voltage - controlled oscillator ( vco ) used to generate the oscillating field .
the long - term stability of both feedback schemes is limited by temperature dependent phase drifts .
the al shield slipped over each magnetometer for avoiding cross - talk effectively acts as a low - pass filter for external magnetic field fluctuations ( skin effect ) which reduces the response bandwidth of all devices to about 250hz ( fig .
[ fig : bwdeponalshield ] ) .
the sensitivity of the magnetometer is defined as the noise equivalent magnetic flux density ( nem ) , which is the flux density change @xmath67 equivalent to the total noise of the detector signal . in a perfectly stable external magnetic field
the smallest detectable field changes are limited by the intrinsic magnetometer noise @xmath68 . for a measurement bandwidth @xmath69 the intrinsic resolution @xmath68 depends on the magnetic resonance linewidth @xmath70 ( hwhm ) and on the signal - to - noise ratio @xmath71 of the magnetometer signal according to @xmath72 for a feedback operated magnetometer using an optically thin medium ( @xmath73 ) the light power ( expressed in terms of photocurrent ) detected after the sensor cell is given by @xmath74 where @xmath75 is the incident power , @xmath76 the resonant optical absorption coefficient and @xmath77 the sample length .
the absorption coefficient can be written as @xmath78 where @xmath79 is the mean absorption coefficient and @xmath80 is the modulation depth which depends - among others - on the degree of spin polarization and the amplitude of the oscillating field .
we can then write eq .
[ eq : transmittedcurrent ] as @xmath81 where @xmath82 and @xmath83 .
the contrast @xmath84 is the ratio of the modulation amplitude and the average photocurrent in the approximation @xmath85 .
the signal @xmath86 is given by the rms value of the oscillating part of the magnetometer signal @xmath87 the fundamental limit of the magnetometric sensitivity is obtained for a shot - noise limited signal , with a noise level @xmath88 of @xmath89 .
the shot - noise limited sensitivity then reads @xmath90 the magnetometer signal of interest is contained in the amplitude and phase of the sinusoidal modulation of the light power after the sensor cell .
the power spectrum of the spectral power density of an ideal magnetometer in a perfectly stable magnetic field thus consists of a delta function centered at the larmor frequency , superposed on a flat background of shot - noise fluctuations . in practice
the peak is broadened by the resolution ( 1hz ) of the fft analyzer ( stanford research systems , model sr760 ) used for its recording .
the relevant noise contributions which define the s / n ratio are fluctuations of the photocurrent at the larmor frequency , i.e. , the value of the background below the larmor peak . in practice that ideal spectrum is modified by various imperfections which degrade the magnetometer performance . in the following we address contributions from light power fluctuations and magnetic field fluctuations .
the light power @xmath75 has a continuous ( technical ) noise spectrum , which lies above the shot - noise level , in particular at frequencies below 100hz , as shown for the laser in fig .
[ fig : lowfreqlasernoise ] .
the individual peaks are even and odd harmonics of the 50 hz line frequency .
power fluctuations contribute to the photocurrent noise at @xmath4 by two distinct processes .
first , there is a direct contribution via the noise component of @xmath91 in eq .
[ eq : ipc ] at the fourier frequency @xmath4 . at @xmath92= 7khz
this noise level is close to the shot - noise level .
the second contribution is due to the second term in ( [ eq : ipc ] ) .
each fourier component ( at @xmath93 ) of the power fluctuations is multiplied by @xmath94 , and this mixing produces sidebands at @xmath95 in the power density spectrum . in this way the continuous low frequency part of the technical noise around @xmath96 ( fig .
[ fig : lowfreqlasernoise ] ) produces a symmetric background under the larmor peak . although the power noise around @xmath97 is 18 times ( fig .
[ fig : lowfreqlasernoise ] ) larger than the shot noise around @xmath98 , it is suppressed - according to eq .
[ eq : ipc ] - by a factor @xmath84 , which has a value of approximately 0.05 in the lsopm . as a consequence
the contribution of the modulation term in eq .
[ eq : ipc ] to the photocurrent noise is less than the contribution from @xmath91 .
we have verified that this is indeed fulfilled in a carefully calibrated auxiliary experiment .
w and the corresponding shot - noise level is shown as solid line .
the noise at 7khz , which lies 50% above the shot noise level , is also shown.,title="fig : " ] + in presence of uncorrelated magnetic field fluctuations , @xmath99 , the highest resolution with which a magnetic field change can be detected is given by @xmath100 .
the fluctuations @xmath99 of the external magnetic field can be parameterized by the equivalent noise @xmath101 that they produce on the signal , and @xmath67 can be expressed in a form similar to eq .
[ eq : nemint ] by @xmath102 where @xmath103 .
fourier components of the field fluctuations at frequency @xmath93 will mix with the magnetometer oscillation frequency @xmath4 in eq .
[ eq : ipc ] .
monochromatic field fluctuations , such as the 50hz line frequency and harmonics thereof produce symmetric sidebands , while low frequency magnetic field fluctuations produce a continuous background underlying the larmor peak .
the intrinsic linewidth of the magnetic resonance transition was measured by extrapolating the experimental linewidth to zero light power and zero r.f .
we found a hwhm of 1.63hz for the cell in the lpopm and 2.35hz for the cell in the lsopm .
as all cells were manufactured by the same person the difference of the intrinsic linewidths is probably due to the slightly larger aperture between the cell and the sidearm in the case of the lsopm . after optimizing the signal - to - noise ratio with respect to light and r.f .
power the ( power - broadened ) magnetic resonance linewidth ( hwhm ) is @xmath104 = 2.5hz for the lpopm and 5.0hz for the lsopm .
are 29000 ( a ) and 98000 ( b ) respectively .
the photocurrents are 4.3@xmath49a for the lpopm and 5@xmath49a for lsopm .
the corresponding shot - noise levels are represented by the horizontal lines .
the dashed areas indicate the pedestal discussed in the text.,title="fig : " ] + and the corresponding shot - noise levels are represented by horizontal lines.,title="fig : " ] + figures [ fig : fftlamplaserpsm ] and [ fig : fftlamplasersom ] show typical power density spectra - recorded in the phase - stabilized and self - oscillating mode of operation - of the r.m.s .
voltage fluctuations at the output of the current - to - voltage preamplifier of the photodiode current .
the spectra contain all structures discussed above . the pedestal ( indicated as dashed area in fig .
[ fig : fftlamplaserpsm ] ) underlying the larmor peak contains contributions from fluctuations of the light power and of the magnetic field . the white noise floor in the far wings of the central structure represents the intrinsic opm noise @xmath105 . its numerical value for use in eq .
[ eq : nemint ] was measured 70hz above the carrier . at that frequency
this noise represents the noise component of @xmath91 of the power noise in eq .
[ eq : ipc ] .
we have verified that the noise component of the modulated contribution in eq .
[ eq : ipc ] is 8 times less than the noise contribution of @xmath91 under present experimental conditions .
the pedestal under the larmor peak is thus most probably due to low - frequency field fluctuations . after optimizing the magnetometric sensitivity with respect to light power and r.f .
power the lpopm yields a s / n ratio of @xmath106 , while the lsopm reaches @xmath107 in a bandwidth @xmath108 of 1hz .
it can be seen from the figures that the signal - to - noise ratio does not depend on the mode of operation ( som or psm ) . for the lpopm the shot - noise level
is nearly reached while in the lsopm @xmath109 . according to eq .
[ eq : nemint ] the measured s / n ratios and linewidths under optimized conditions result a nem @xmath68 of @xmath110 for the lpopm and of @xmath111 for the lsopm in a bandwidth of 1hz .
the lsopm is thus 1.7 times more sensitive than the lpopm , although its performance is not yet shot - noise limited . of the photocurrent produced by laser light ( black dots ) and by light from the discharge lamp ( open cicles ) at 7khz .
the measurement bandwidth was 1hz .
the dashed lines indicate the calculated photodiode shot noise ( @xmath112 ) and the fitted technical laser noise ( @xmath113 ) .
b ) ratio of @xmath114 to the calculated shot noise @xmath115 .
the technical noise appears to be proportional to @xmath116 in this representation .
the full lines in both plots represent the calculated sum of all noise contributions according to eq .
[ eq : overalllasernoise ] .
the arrows refer to the ( optimized ) photocurrent for both devices . ] in order to get a better understanding of the excess power noise in the lsopm we measured the dependence of the photocurrent noise on the light power for the laser in comparison to that of the lamp . in those measurements
only the noise component of @xmath91 ( eq . [ eq : ipc ] ) at 7khz was recorded as it represents the dominant noise contribution of @xmath117 . for each of the measurements
the cesium cell was removed and the light beam was detected directly by the photodiode using a transimpedance amplifier for the photocurrent .
the noise @xmath114 measured in this way gives a lower limit of the intrinsic magnetometer performance .
the results for the laser and lamp source are shown in fig .
[ fig : lasernoise ] . the noise can be written as @xmath118 where @xmath115 is the shot - noise of the photocurrent , @xmath119 the intrinsic detector ( photodiode and amplifier ) noise , and @xmath120 technical noise of the light source , which is proportional to the photocurrent @xmath91 .
@xmath119 was measured with the light beams blocked for all ( discrete ) amplification stages of the transimpedance amplifier . with the highest amplification ( @xmath121a / v ) used for @xmath122 we measured @xmath123 .
this corresponds to the shot noise of a current @xmath91 of 7na , so that above 7na the intrinsic detector noise can be neglected .
this dark current is responsible for the deviation of the measured noise from the ( dotted ) shot - noise line ( fig .
[ fig : lasernoise ] ) at low currents . for larger photocurrents technical noise @xmath120 , which is proportional to the light power , dominates over the shot noise .
we found @xmath124 , with @xmath125 and @xmath126 for the laser and the lamp source respectively . with the laser source
the shot noise @xmath115 becomes equal to the technical noise @xmath120 for a photocurrent @xmath91 of @xmath127 , which thus yields a noise level @xmath128 at that photocurrent .
the laser power for optimized magnetometer parameters corresponds to @xmath129 , thereby explaining the excess noise of the lsopm in figs .
[ fig : fftlamplaserpsm ] and [ fig : fftlamplasersom ] . in the case of lamp
pumping the technical noise becomes important for @xmath130 , so that the magnetometer is shot - noise limited for the ( optimized ) photocurrent of 4.3@xmath49a .
if one succeeds in eliminating the excess noise of the laser power , e.g. , by an active power stabilization of the lsopm one can achieve an intrinsic shot - noise limited sensitivity of @xmath131 , thereby outperforming the lpopm by a factor of @xmath132 .
this is compatible with earlier results @xcite obtained from a comparative study of lamp and laser pumped magnetometers using @xmath133k , in which the sensitivity of the lsopm version was found to be 2.3 times higher than of the corresponding lpopm device .
it is also interesting to compare those results with the present results on an absolute scale .
the sensitivity of the @xmath133k - lsopm was found to be 1.8ft/@xmath134 .
this superior performance compared to the @xmath40cs magnetometer discussed here is mainly due to the two times larger diameter of the potassium sensor cell , which , combined with the appreciably smaller spin exchange cross section of potassium , led to an operating linewidth of 1hz compared to 5hz with the present cs - lsopm
. furthermore the g - factor of @xmath133k is twice as large than that of @xmath40cs .
these two factors explain the superior performance of the k - magnetometer ( @xmath135 ) compared to the cs magnetometer ( @xmath111 ) .
+ . individual readings of the magnetometers ( top ) and of their difference frequency ( bottom).,title="fig : " ] + .
open circles : lpopm measurement , black dots : lsopm measurement , black squares : field gradient measured over 21 cm . for times below 200 seconds the lsopm data and the lpopm data points overlap and can not be distinguished .
the lines are drawn to guide the eye .
in the two lower graphs the single points at @xmath1360.5s indicate the nem of the lpopm ( open circle ) and of the lsopm ( black dot ) , and their extrapolation to other integration times assuming white noise.,title="fig : " ] + we performed a direct comparison of the performance of the lpopm and the lsopm in simultaneous measurements of magnetic field fluctuations inside the multilayer shield described above ( fig .
[ fig : fribourgshield ] ) .
both devices were mounted coaxially in the shield and the centers of their sensor cells were separated by 21 cm .
the oscillatory signal of each magnetometer was filtered by a resonance amplifier , centered near 7khz with a fwhm of 500hz , and analyzed by a frequency counter ( stanford research systems , model sr620 ) with a gate time of 0.1s .
an example of such a recording over a continuous interval of 6 hours with the lpopm operated in psm and the lsopm operated in the som is shown in fig .
[ fig : tsls2lp4_6h ] .
one sees that both devices oscillate a different average frequencies , which can be explained by the presence of a magnetic field gradient of 66pt / cm which drifts by 0.4% over the 6 hour interval .
the drift is probably due to a thermal drift of the shield s magnetization .
similar gradients were measured after interchanging the positions of the two opms .
figure [ fig : tsls2lp4_5min ] shows a 5 minutes time slice of the data in fig . [ fig : tsls2lp4_6h ] .
there are highly correlated irregular field jumps of approximately 3.6pt in both traces .
these fluctuations correspond to relative fluctuations of the solenoid current at a level of 10@xmath137 and are suspected to be caused by the current source .
the allan standard deviation @xcite of the data is a convenient way for characterizing the field drifts on various time scales .
[ fig : asdls2lp4 ] shows the allan plot of the data from fig .
[ fig : tsls2lp4_6h ] as a function of integration time both in absolute and in relative units .
both magnetometers show the same field stability and the data points are indistinguishable for small integration times .
while the short term stability is governed by white noise , the bump between 1 and 200s is due to the irregular field jumps .
the long term stability is determined by long term field drifts of the imperfectly shielded external field and thermal drifts of the solenoid support structure . at the encountered level of field fluctuations
possible false effects due to the light shift are negligible @xcite .
+ we have further investigated the performance of the magnetometers in an active magnetic field stabilization system using a phase - locked loop . for that purpose the phase of the lpopm oscillation ( psm ) relative to the phase of a reference oscillator was measured by a lock - in amplifier and used as error signal driving a correction coil . while field fluctuations common to both sensors are strongly suppressed by this method , gradient drifts and fluctuations are not compensated and thus detected by the free - running magnetometer .
the performance of the stabilization scheme is shown in fig .
[ fig : tsasdlp4stab ] , where the top traces represent time series of the field readings ( 1s gate time of the frequency counter ) of the free - running lsopm and the lpopm used for the feedback . the lower part of fig .
[ fig : tsasdlp4stab ] shows the allan standard deviation of those data .
the jumps in the magnetic field measured by the lsopm ( fig .
[ fig : tsls2lp4_5min ] ) are completely suppressed by the feedback loop and only white noise in the short term stability and contributions from gradient drifts in the long term stability remain .
the field fluctuations reach a minimum of 25 ft for @xmath136100s , which corresponds to a relative field stability of 1.3@xmath138 . due to the fact that the field is stabilized to the lpopm
one would expect a significant lowering of the lpopm s allan standard deviation for short integration times .
the white noise behavior ( slope @xmath139 in the allan plot ) of the lpopm trace is entirely due to the resolution of the frequency counter , which is limited by trigger time jitter due to amplitude noise of the measured sine wave .
we also realized a setup , in which the roles of the lsopm and the lpopm were reversed . the observed performance was identical with the one described .
opms based on cs vapor are well suited for operation in magnetic fields smaller than @xmath140 t ( such as typical fields used in neutron edm experiments ) . in such fields asymmetries of the magnetic resonance line due to the quadratic zeeman effect will not affect their long - term stability ( accuracy ) . in fields between 10@xmath49 t and 200@xmath49 t the line asymmetry due to the quadratic zeeman effect depends on the light power and
power fluctuations induce fluctuations of that asymmetry and thus of the resonance frequency . in magnetic fields above 300@xmath49 t
the quadratic zeeman shift exceeds the linewidth and the magnetic resonance spectrum consists of 8 resolved lines @xcite .
this offers a further possibility for the use of cesium magnetometers .
the performance of the lsopm in that field range is expected not to be worse than the one measured in a 2@xmath49 t field .
this expectation is based on the fact that the ( performance determining ) laser power noise does not increase with frequency , so that a shot - noise limited sensitivity can be reached .
a magnetometer based on the quadratic zeeman effect in potassium has been demonstrated earlier @xcite .
in contrast to buffer gas cells , which require the whole cell volume to be illuminated in order to achieve a maximum sensitivity , paraffin - coated cells can be pumped with a laser beam of much smaller diameter .
this allows one to adapt the spatial dimensions of the paraffin - coated sensors to specific experimental requirements .
one can think , e.g. , of using very large cells of several liters for measuring volume averaged fields . compared to other
proposed large cell schemes@xcite , the use of a cs - opm offers the further advantage of a high temporal resolution .
recently , a novel type of optically pumped magnetometer with a sub - ft ( gradiometric ) sensitivity was demonstrated @xcite . besides its use of very high buffer gas pressures and its operation at a temperature of 190@xmath18c , specific features of that magnetometer are its limited operation range near zero field and its reduced bandwidth of 20hz .
we thank g. bison for fruitful discussions and careful reading of this manuscript . we acknowledge financial support from schweizerischer nationalfonds , intas , and paul - scherrer - institute ( psi ) .
one of the authors ( a.s.p . ) acknowledges financial support from the university of fribourg . | we have performed a comparison of laser ( lsopm ) and lamp ( lpopm ) pumped cesium vapor magnetometers . although the lsopm operated 50% above its shot - noise limit we found an intrinsic sensitivity of 15ft/@xmath0 and 25ft/@xmath0 for the lsopm and lpopm respectively .
two modes of operation , viz . , the phase - stabilized and the self - oscillating mode were investigated and found to yield a similar performance .
we have compared the performance of the lsopm and the lpopm directly by simultaneous measurements of field fluctuations of a @xmath1 magnetic field inside a multilayer magnetic shield and have used one of the magnetometers for an active field stabilization . in the stabilized mode we found a gradient instability of 25 ft within an integration time of 100s , which represents an upper limit of the long - term stability of the magnetometers . |
a sensor node is a small size , low - power devices with limited computation and communication capabilities .
a wireless sensor network ( wsn ) is a set of sensor nodes and each sensor of the network can measure certain physical phenomena like temperature , pressure , intensity of light or vibrations around it .
wireless networks of such sensor nodes have many potential applications , such as surveillance , environment monitoring and biological detection @xcite .
the objective of such network is to process some high - level sensing tasks and send the data to the application @xcite . since the sensor nodes are battery operated , so minimizing energy consumption is a critical issue for designing topology of wireless sensor networks to increase its lifetime .
a message packet transmitted by a sensor device to another is often received by many nodes in the vicinity of the receiver node .
this causes collision of signals and increased interference in its neighboring nodes , which leads to message packet loss .
due to packet loss the sender node needs to retransmit the message packets .
therefore , large interference of networks may result delays for delivering data packets .
this would further enhance the energy consumption of nodes in network .
thus interference reduction of nodes is a crucial issue for minimizing ( i ) delays for delivering data packets and ( ii ) energy consumption of a wireless sensor network .
@xmath13 denotes the euclidean distance between two sensor nodes @xmath14 and @xmath15 .
a range assignment is @xmath16 .
a communication graph @xmath17 is a directed graph , where @xmath0 is the set of sensor nodes and @xmath18 .
@xmath19 is said to be strongly connected if there is a directed path between each pair of vertices @xmath20 in @xmath19 .
@xmath19 is said to be an arborescence rooted at @xmath4 if there is a directed path from @xmath15 to all other vertices @xmath21 in @xmath19 .
different models have been proposed to minimize the interference in sensor networks @xcite . in this paper , we focus on the following two widely accepted models : * sender interference model ( sim ) : the interference of a node @xmath21 is the cardinality of the set of nodes to whom it can send messages directly .
the set of nodes creating interference with @xmath14 with respect to the range assignment @xmath22 is denoted by @xmath23 and defined by @xmath24 , where @xmath25 is the range of the node @xmath14 .
therefore the interference value of node @xmath14 is equal to @xmath26 with respect to range assignment @xmath22 . *
receiver interference model ( rim ) : the interference of a node @xmath21 is the cardinality of the set of nodes from which it can receive messages directly .
the set of nodes creating interference with @xmath14 with respect to the range assignment @xmath22 is denoted by @xmath27 and defined by @xmath28 , where @xmath29 is the range of the node @xmath15 .
therefore the interference value of node @xmath14 is equal to @xmath30 with respect to range assignment @xmath22 . in the interference minimization problems ,
the objective is to minimize the following objective functions : * * minmax si ( mmsi ) : * given a set @xmath0 of @xmath1 sensors , find a range assignment function @xmath22 such that the communication graph @xmath19 satisfy the connectivity predicate @xmath5 and maximum sender interference of the sensors in the network is minimum i.e. , @xmath31 * * min total si ( mtsi ) : * given a set @xmath0 of @xmath1 sensors , find a range assignment function @xmath22 such that the communication graph @xmath19 satisfy the connectivity predicate @xmath5 and total sender interference of the entire sensor network is minimum i.e. , @xmath32 * * minmax ri ( mmri ) : * given a set @xmath0 of @xmath1 sensors , find a range assignment function @xmath22 such that the communication graph @xmath19 satisfy the connectivity predicate @xmath5 and maximum receiver interference of the sensors in the network is minimum i.e. , @xmath33 * * min total ri ( mtri ) : * given a set @xmath0 of @xmath1 sensors , find a range assignment function @xmath22 such that the communication graph @xmath19 satisfy the connectivity predicate @xmath5 and total receiver interference of the entire sensor network is minimum i.e. , @xmath34 in the communication graph corresponding to a range assignment @xmath22 , the number of out - directed edges and in - directed edges are same , which leads to the following result : [ theorem-1 ] for any range assignment @xmath22 , * mtsi = mtri*. in this paper , we propose algorithm for _ minimizing maximum interference _ and _ minimizing total interference _ of a given wireless sensor network . for minimizing maximum interference we present optimum solution with running time @xmath6 for connectivity predicate @xmath5 like strong connectivity , broadcast , @xmath8-edge(vertex ) connectivity , spanner . here
@xmath9 is the time complexity for checking the connectivity predicate @xmath5 .
the running time of the previous best known solution was @xmath10 @xcite . for the minimizing total interference
we propose optimum algorithm for the connectivity predicate broadcast .
the running time of the propose algorithm is @xmath11 .
for the same problem , the previous best known result was @xmath12-factor approximation algorithm @xcite .
we also propose a heuristic for minimizing total interference in the case of strongly connected predicate and compare our result with the best result available in the literature .
experimental results demonstrate that our heuristic outperform existing result .
we organize remaining part of this paper as follows : in section [ relatedwork ] , we discuss existing results in the literature .
the algorithm for the optimum solution of minimizing maximum interference for different connectivity predicate appears in section [ minmax ] . in section [ mintotal ] , we present optimum algorithm to minimize total interference for the connectivity predicate broadcast .
a heuristic for minimizing total interference for the strongly connected predicate appears in section [ heuristic ] .
finally , we conclude the paper in section [ conclusion ] .
tan et al . @xcite studied minimization of the average interference and the maximum interference for the highway model , where all the nodes are arbitrarily distributed on a line . for the minimum average interference problem they proposed an exact algorithm , which runs in @xmath35 time , where @xmath1 is the number of nodes and @xmath36 is the maximum node degree in the communication graph for the equal range assigned to each nodes equal to the maximum consecutive distance between two nodes . for the minimization of maximum interference problem , they proposed @xmath37 time algorithm , where @xmath38 .
lou et al .
@xcite improves the time complexity to @xmath39 for the minimization of average interference problem .
rickenbach et al .
@xcite proved that minimum value of maximum interference is bounded by @xmath40 and presented an @xmath41{\delta})$]-factor approximation algorithm , where @xmath36 is the maximum node degree in the communication graph for some equal range @xmath42 assigned to all the nodes . for 2d networks ,
buchin @xcite considered receiver interference model and proved that minimizing the maximum interference is np - hard whereas bil@xmath43 and proietti @xcite considered sender interference model and proposed a polynomial time algorithm for minimizing the maximum interference .
their algorithm works for many connectivity predicate like simple connectivity , strong connectivity , broadcast , @xmath8-edge(vertex ) connectivity , spanner , and so on .
they also proved that any polynomial time @xmath44-approximation algorithm for minimum total range assignment problem with connectivity predicate @xmath5 can be used for designing a polynomial time @xmath44-approximation algorithm for minimum total interference problem for @xmath5 .
panda and shetty @xcite considered 2d networks with sender centric model and proposed an optimal solution for minimizing the maximum interference and a 2-factor approximation algorithm for average interference .
moscibroda and wattenhofer @xcite proposed @xmath45-factor greedy algorithm for minimizing average interference .
in this section we consider sender centric interference model . given a set @xmath46 of @xmath1 nodes distributed on a 2d plane , the objective is to find a range assignment @xmath16 such that the corresponding communication graph @xmath19 contains connectivity predicate @xmath5 like strong connectivity , broadcast , @xmath8-edge(vertex ) connectivity , spanner etc . bil@xmath43 and proietti considered the same problem and proposed @xmath10 time algorithm for optimum solution , where @xmath47 is the time required to check predicate @xmath5 for a given communication graph @xcite . here
we propose an algorithm to solve the above problem optimally .
the running time of our algorithm is @xmath6 , which leads to a big improvement over @xcite . in the network ,
the number of nodes is @xmath1 , which means maximum possible interference of a node is @xmath48 .
the main idea of our algorithm is very simple : first we start range assignment to each of the node in such a way that the interference of each node is @xmath8 ( @xmath49 ) .
next we test whether the communication graph contains the connectivity predicate @xmath5 or not . if the answer is _ yes _
, then we try for lower values of @xmath8 .
otherwise we try for higher values of @xmath8 .
the pseudo code for the minimizing maximum sender interference ( mmsi ) algorithm described in algorithm [ alg-1 ] . in the algorithm
we use a matrix @xmath50 of size @xmath51 and @xmath52 , where @xmath21 such that if @xmath53 , then @xmath54 for all @xmath55 and @xmath56 .
elements of the @xmath57-th row are @xmath58 , @xmath59 in ascending order from left to right . a set @xmath60 of @xmath1 nodes and a connectivity predicate @xmath5 .
a range assignment @xmath22 such that communication graph @xmath19 contains connectivity predicate @xmath5 and an interference value ( mmsi ) .
construct the matrix @xmath50 as described above .
@xmath61 , @xmath62 @xmath63 construct the communication graph @xmath19 corresponding to @xmath22 .
test whether @xmath19 contains connectivity predicate @xmath5 or not .
@xmath64 / * maximum interference is at most @xmath65 * / @xmath66 / * minimum interference is greater than @xmath65 * / @xmath67 return@xmath68 [ alg-1 ] [ theorem-2 ] algorithm [ alg-1 ] computes minimum of maximum sender interference ( mmsi ) optimally and its worst case running time is @xmath6 .
let @xmath22 ( resp .
@xmath69 ) be the range assignment to the nodes in @xmath0 when the sender interference of each node is @xmath70 ( resp .
@xmath71 ) .
the correctness of the algorithm [ alg-1 ] follows from the fact ( i ) the algorithm [ alg-1 ] stops when @xmath72 such that @xmath19 does not contain connectivity predicate @xmath5 , whereas @xmath73 contains connectivity predicate @xmath5 and ( ii ) if @xmath74 , then @xmath75 .
the construction time of matrix @xmath50 ( line number 3 of the algorithm [ alg-1 ] ) takes @xmath76 time in worst case .
construction of the communication graph @xmath19 ( line number 9 of the algorithm [ alg-1 ] ) and testing connectivity predicate ( line number 10 of the algorithm [ alg-1 ] ) take @xmath9 time .
again , each execution of * while * loop in line number 5 reduce the value @xmath77 by half of its previous value .
therefore , algorithm [ alg-1 ] call this * while * loop @xmath45 time .
thus , the time complexity results of the theorem follows .
given a set @xmath60 of @xmath1 nodes and a source node @xmath3 distributed on a 2d plane , the objective is to find a range assignment @xmath16 such that the corresponding communication graph @xmath19 contains an arborescence rooted at @xmath7 ( connectivity predicate is broadcast ) and the total interference ( sender / receiver ) is minimum .
bil@xmath43 and proietti considered the same problem and proposed @xmath78-factor approximation algorithm @xcite . here
we propose a very simple optimum algorithm .
the running time of the algorithm is linear .
the pseudo code for the minimum total sender interference ( mtsi ) algorithm described in algorithm [ alg-2 ] .
though the propose algorithm is trivial , we are proposing it for completeness of the literature .
[ lemma-1 ] for any range assignment @xmath22 such that the communication graph @xmath17 contains arborescence rooted at any node @xmath14 in the network of @xmath1 nodes , the minimum total sender interference of the networks is at least @xmath48 .
since the communication graph @xmath19 contains an arborescence rooted at @xmath14 , each node @xmath79 has an incoming edge .
therefore , the total receiver interference is at least @xmath48 .
thus , the lemma follows from theorem [ theorem-1 ] . a set @xmath60 of @xmath1 nodes and a source node @xmath3 .
a range assignment @xmath22 such that communication graph @xmath19 contains an arborescence rooted at @xmath7 and total interference ( mtsi ) .
@xmath80 find the farthest node @xmath21 from @xmath7 .
@xmath81 return@xmath82 / * @xmath83 * / [ alg-2 ] [ theorem-3 ] algorithm [ alg-2 ] produces optimum total interference in @xmath11 time .
the correctness of the algorithm follows from ( i ) the fact that total interference produce by algorithm [ alg-2 ] is @xmath48 and ( ii ) lemma [ lemma-1 ] .
time complexity follows from the * for * loop ( line number 3 ) and line number 6 .
in this section , we propose a heuristic to minimize total interference for strongly connected predicate i.e given a set @xmath60 of @xmath1 nodes distributed on a 2d plane , we design a heuristic for range assignment @xmath22 such that the communication graph @xmath19 is strongly connected . experimental results presented in the subsection [ expres ]
demonstrate that our heuristic perform very well compare to existing result in the literature . a set @xmath60 of @xmath1 nodes .
a range assignment @xmath22 such that communication graph @xmath19 is strongly connected and total interference of the network .
@xmath84 / * initial range assignment * / @xmath85 / * initial interference assignment * / @xmath86 / * initial total interference * / @xmath87 and @xmath88 choose @xmath89 and @xmath90 such that @xmath91 for all @xmath92 and @xmath93 .
@xmath94 / * total interference update * / @xmath95 and @xmath96 / * new range assignments * / @xmath97 and @xmath98 + / * new interference assignments * / @xmath99 and @xmath100 return(@xmath101 ) [ alg-3 ] [ theorem-4 ] the running time of the algorithm [ alg-3 ] is polynomial in input size . the input size of the algorithm [ alg-3 ] is @xmath1 . in each execution of *
while * loop ( line number 9 of the algorithm [ alg-3 ] ) , the size of the set @xmath102 is decreasing by 1 . choosing the vertices @xmath103 ( line number 10 of the algorithm [ alg-3 ] ) needs @xmath104 time .
thus , the time complexity result of the theorem follows .
the model consists of @xmath1 sensor nodes randomly distributed in a @xmath105 square grid . for different values of @xmath1
, we execute our heuristic 100 times for different input instances .
we have taken average value of the total interference of the nodes in the network . in figure
[ figure-1 ] , we have shown average interference of our heuristic and compare it with the algorithm improved smit @xcite . in table [ table-1 ] , we have shown the comparison of the total interference between our heuristic and the algorithm proposed in @xcite .
these experimental results demonstrate that our heuristic outperform existing algorithm .
[ table-1 ] .simulation results for total interference [ cols="^,^,^ " , ]
in this paper we considered 2d networks i.e. , the wireless nodes are distributed on a 2d plane .
all the results presented in this paper are applicable in 3d networks also . here
we considered interference ( sender interference model ) minimization problem in wireless networks .
we proposed algorithm for _ minimizing maximum interference _ and _ minimizing total interference _ of a given networks . for minimizing maximum interference we presented optimum solution with running time @xmath6 for connectivity predicate @xmath5 like strong connectivity , broadcast , @xmath8-edge(vertex ) connectivity , spanner , where @xmath1 is the number of nodes in the network and @xmath9 is the time complexity for checking the connectivity predicate @xmath5
the running time of the previous best known solution was @xmath10 @xcite .
therefore , our solution is a significant improvement over the best known solution with respect to time complexity . for the minimizing total interference we proposed optimum algorithm for connectivity predicate broadcast .
the running time of the proposed algorithm is @xmath11 .
for the same problem , the previous best known result was @xmath12-factor approximation algorithm @xcite .
therefore , our solution is a significant improvement over the existing solution in the literature .
we also proposed a heuristic for minimizing total interference in the case of strongly connected predicate and compare our result with the best result available in the literature .
experimental results demonstrate that our heuristic outperform existing result .
m. burkhat , p. von rickenbach , r. wattenhofer and a. zollinger , _ does topology control reduce interference _ , proc . of the 5th international symposium on mobile ad - hoc networking and computing ,
9 - 19 , 2004 .
t. lou , h. tan , y. wang , f. c. m. lau , _ minimizing average interference through topology control _ , proc .
of the 7th international symposium on algorithms for sensor systems , wireless ad hoc networks and autonomous mobile entities , lncs - 7111 , pp .
115 - 129 , 2012 .
a. mainwaring , j. polastre , r. szewczyk , d. culler and j. anderson , _ wireless sensor networks for habitat monitoring _ , proc . of the 1st international workshop on wireless sensor networks and applications , pp .
88 - 97 , 2002 .
p. von rickenbach , s. schimid , r. wattenhofer and a. zollinger , _ a robust interference model for wireless adhoc networks _ , proc . of the 5th international workshop on algorithms for wireless , mobile , ad hoc and sensor networks , 2005 .
h. tan , t. lou , f. c. m. lau , y. wang and s. chen , _ minimizing interference for the highway model in wireless ad - hoc and sensor networks _ , proc . of the 37th conference on current trends in theory and practice of computer science , lncs - 6543 , pp .
520 - 532 , 2011 . | given a set @xmath0 of @xmath1 sensor node distributed on a @xmath2-dimensional plane and a source node @xmath3 , the _ interference problem _ deals with assigning transmission range to each @xmath4 such that the members in @xmath0 maintain connectivity predicate @xmath5 , and the maximum / total interference is minimum .
we propose algorithm for both _ minimizing maximum interference _ and _ minimizing total interference _ of the networks . for minimizing maximum interference we present optimum solution with running time @xmath6 for connectivity predicate @xmath5 like strong connectivity , broadcast ( @xmath7 is the source ) , @xmath8-edge(vertex ) connectivity , spanner , where @xmath9 is the time complexity for checking the connectivity predicate @xmath5 .
the running time of the previous best known solution was @xmath10 @xcite . for the minimizing total interference
we propose optimum algorithm for the connectivity predicate broadcast .
the running time of the propose algorithm is @xmath11 .
for the same problem , the previous best known result was @xmath12-factor approximation algorithm @xcite .
we also propose a heuristic for minimizing total interference in the case of strongly connected predicate and compare our result with the best result available in the literature .
experimental results demonstrate that our heuristic outperform existing result .
* keywords : * wireless sensor networks , interference . |
various stochastic partial differential equations ( spdes ) have emerged over the last two decades in different areas of mathematical finance .
a classical example is the heath - jarrow - morton interest rate model @xcite of the form @xmath0 where @xmath1 is the forward rate of tenor @xmath2 at time @xmath3 and @xmath4 its instantaneous volatility . in @xcite ,
a similar equation has been proposed more recently to model electricity forwards .
most of the spdes studied share with ( [ hjm ] ) the property that the derivatives of the solution only appear in the drift term ; in the case of ( [ hjm ] ) the volatility of the brownian driver does not depend on the solution @xmath5 at all .
numerical methods for hyperbolic spdes of the type ( [ hjm ] ) have been studied , for example , in @xcite .
this article , in contrast , considers the parabolic spde @xmath6 where @xmath7 is a standard brownian motion , and @xmath8 and @xmath9 are real - valued parameters .
it is clear that the behaviour of this equation is fundamentally different from those with additive or multiplicative noise .
the significance of ( [ spde ] ) for the following applications is that it describes the limiting density of a large system of exchangeable particles .
specifically , if we consider the system of sdes @xmath10 for @xmath11 , with @xmath12 and @xmath13 , where @xmath14 are assumed i.i.d . with finite second moment , the empirical measure @xmath15 has a limit @xmath16 for @xmath17 ,
whose density @xmath18 satisfies ( [ spde ] ) in a weak sense . for a derivation of this result in the more general context of quasi - linear pdes
see @xcite .
while the motivation in @xcite is to use a large particle system ( [ sdesys ] ) to approximate the solution to the spde ( [ spde ] ) , our view point is to use ( [ spde ] ) as an approximate model for a large particle system , and we will argue later the ( computational ) advantages of this approach in situations when the number of particles is large . as a first possible application , one may consider @xmath19 as the log price processes of a basket of equities , which have idiosyncratic components @xmath20 and share a common driver @xmath7 ( the `` market '' ) . if the size of the basket is large enough , the solution to the spde can be used to find the values of basket derivatives . in this paper , we study an application of a similar model to basket credit derivatives .
we mention in passing that equations of the form ( [ spde ] ) arise also in stochastic filtering .
to be precise , ( [ spde ] ) is the zakai equation for the distribution of a signal @xmath21 given observation of @xmath7 , see e.g. @xcite .
it is interesting to note that the solution to the spde ( [ spde ] ) without boundary conditions can be written as the solution of the pde @xmath22 shifted by the current value of the brownian driver , @xmath23 in particular , if @xmath24 , then @xmath25 the intuitive interpretation of this result is that the independent brownian motions @xmath20 have averaged into a deterministic diffusion in the infinite particle limit , whereas the common factor @xmath7 , which moves all processes in parallel , shifts the whole profile ( and also adds to the diffusion , via the it term ) . in @xcite , the analysis of the large particle system
is extended to cases with absorption at the boundary ( @xmath26 ) , @xmath27 it is shown that there is still a limit measure @xmath16 , which may now be decomposed as @xmath28 where @xmath29 is the proportion of absorbed particles ( the `` loss function '' ) , and the density @xmath18 of @xmath30 satisfies ( [ spde ] ) in @xmath31 with absorbing boundary condition @xmath32 @xcite consider applications to basket credit derivatives . for the market pricing examples ,
they consider a simplified model , where defaults are monitored only at a discrete set of dates . between these times , the default barrier is inactive and ( [ spde ] ) is solved on the real line by using ( [ pde ] ) and ( [ shift ] ) . for the initial - boundary value problem ( [ spde ] ) , ( [ bc ] ) ,
however , such a semi - analytic solution strategy is no longer possible and an efficient numerical method is needed .
moreover , there is a loss of regularity at the boundary in this case , such that @xmath33 but not @xmath34 , as is documented in @xcite .
recent papers on the numerical solution of spdes deal with cases relevant to ours , yet structurally crucially different .
a comprehensive analysis of finite difference and finite element discretisations of the stochastic heat equation with multiplicative white noise and non - linear driving term is given in @xcite and @xcite , respectively .
@xcite shows a lax equivalence theorem for the sde @xmath35 in a hilbert space , driven by a process @xmath7 from a class including brownian motion , where @xmath36 is a suitable ( e.g. elliptic differential ) operator and @xmath37 a lipschitz function . in this paper
, we propose a milstein finite difference discretisation for ( [ spde ] ) and analyse its stability and convergence in the mean - square sense by fourier analysis .
a main consideration of this paper is the computational complexity of the proposed methods , and we will demonstrate that a multilevel approach achieves a cost for the spde simulation no larger than that of direct monte carlo sampling from a known univariate distribution , @xmath38 for r.m.s .
accuracy @xmath39 , and is in that sense optimal .
multilevel monte carlo path simulation , first introduced in @xcite , is an efficient technique for computing expected values of path - dependent payoffs arising from the solution of sdes .
it is based on a multilevel decomposition of brownian paths , similar to a brownian bridge construction .
the complexity gain can be explained by the observation that the variance of high - level corrections involving a large number of timesteps is typically small , and consequently only a relatively small number monte carlo samples is required to estimate these contributions to an acceptable accuracy .
overall , for sdes , if a r.m.s .
accuracy of @xmath40 is required , the standard monte carlo method requires @xmath41 operations , whereas the multilevel method based on the milstein discretisation @xcite requires @xmath38 operations .
the first extension of the multilevel approach to spdes was for parabolic pdes with a multiplicative noise term @xcite .
there have also been recent extensions to elliptic pdes with random coefficients @xcite .
our approach , for a rather different parabolic spde , is similar to the previous work on sdes and spdes in that the solution is decomposed into a hierarchy with increasing resolution in both time and space
. provided the variance of the multilevel corrections decreases at a sufficiently high rate as one moves to higher levels of refinement , the number of fine grid monte carlo simulations which is required is greatly reduced .
indeed , the total cost is only @xmath38 to achieve a r.m.s
. accuracy of @xmath40 compared to an @xmath42 cost for the standard approach which combines a finite difference discretisation of the spatial derivative terms and a milstein discretisation of the stochastic integrals .
the rest of the paper is structured as follows .
section [ sec : finite - difference ] outlines the finite difference scheme used , and analyses its accuracy and stability in the standard monte carlo approach .
section [ sec : multilevel ] presents the modification to multilevel path simulation of functionals of the solution .
numerical experiments for a cdo tranche pricing application are given in section [ sec : results ] , providing empirical support for the postulated properties of the scheme and demonstrating the computational gains achieved .
section [ sec : conclusions ] discusses the benefits over standard monte carlo simulation of particle systems and outlines extensions .
integrating ( [ spde ] ) over the time interval @xmath43 $ ] gives @xmath44 making the approximation @xmath45 for @xmath46 in the first integral and @xmath47 in the second , and noting the standard it calculus result that @xmath48 where @xmath49 with @xmath50 , @xcite , leads to the milstein semi - discrete approximation @xmath51 using a spatial grid with uniform spacing @xmath52 , standard central difference approximations to the spatial derivatives @xcite then give the finite difference equation @xmath53 in which @xmath54 is an approximation to @xmath55 .
the spatial domain is truncated by introducing an upper boundary at @xmath56 and using the boundary condition @xmath57 .
since the initial distribution will be assumed localised , both the localisation error for a given path @xmath7 , and the expected error of functionals of the solution , can be made as small as needed by a large enough choice @xmath58 .
the system of sdes can then be written in matrix - vector form @xmath59 where @xmath60 is the vector with elements @xmath61 and @xmath62 and @xmath63 are the matrices corresponding to first and second central differences as explicitly given in appendix [ sec : matanal ] .
[ remark - mol ] an alternative discretisation arises if the spatial derivatives are discretised first , and the milstein scheme is applied to the resulting system of sdes .
the practical implication is that the it term then contains a second finite difference with twice the step size , @xmath64 instead of @xmath63 , resulting in pentadiagonal discretisation matrices instead of tridiagonal ones , specifically @xmath65 the accuracy and stability of the two schemes are similar , hence we will not go into details . to approximate the initial condition on the grid , we apply the initial measure @xmath66 to a basis of ` hat functions ' @xmath67 , where @xmath68 giving @xmath69 for the corresponding pde ( @xmath70 ) , this is known to result in @xmath71 convergence provided
@xmath72 satisfies a certain stability limit , even when the initial data are not smooth @xcite .
we will see that an extension of this analysis holds for spdes .
finite difference fourier stability analysis ignores the boundary conditions and considers fourier modes of the form @xmath73 which satisfy ( [ discrete ] ) provided @xmath74 where @xmath75 b(\theta ) & = & -\ , \frac{i \sqrt{\rho\ , k}}{h}\ , \sin \theta , \\[0.05 in ] c(\theta ) & = & -\ , \frac{2\ , \rho\ , k}{h^2 } \sin^2 { { \textstyle \frac{\theta}{2 } } } .\end{aligned}\ ] ] following the approach of mean - square stability analysis from @xcite , we obtain @xmath76 & = & { \mathbb{e}}\left [ ( a + b\ , z_n + c\ , z_n^2)(a^ * + b^ * z_n + c^ * z_n^2)\ |g_n|^2 \right ] \\[0.1 in ] & = & \left ( \ , |a\!+\!c|^2 + |b|^2 + 2 |c|^2 \,\right)\ { \mathbb{e}}\left[\ , |g_n|^2 \right],\end{aligned}\ ] ] where @xmath77 denotes the complex conjugate of @xmath78 . mean - square stability therefore requires @xmath79 and enforcing this for all @xmath80 leads to the two conditions summarised in the following theorem .
the milstein finite difference scheme ( [ discrete - sys ] ) is stable in the mean - square sense provided @xmath81 the analysis in appendix [ sec : matanal ] combines mean - square and matrix stability analysis to prove that in the limit @xmath82 the condition @xmath83 is also a sufficient condition for mean - square stability of the initial - boundary value problem with the boundary conditions at @xmath84 and @xmath85 .
fourier analysis can also be used to examine the accuracy of the finite difference approximation in the absence of boundary conditions .
considering a fourier mode of the form @xmath86 the pde ( [ pde ] ) reveals that @xmath87 and therefore @xmath88 where @xmath89 .
fourier analysis of the discretisation gives @xmath90 where @xmath91 , @xmath92 , @xmath93 are as defined before .
writing @xmath94 we obtain , after performing lengthy expansions using matlab s symbolic toolbox , @xmath95 where , for all @xmath96 , @xmath97 \le c(\kappa , p ) \left(k^2 + k h^2\right)^p.\ ] ] when summing over @xmath98 timesteps , @xmath99 and @xmath100 are both @xmath101 since they have zero expectation and @xmath102 variance .
hence , it follows that @xmath103\right)^{1/2}= o(k , h^2),\ ] ] so the rms error in the fourier mode over the full simulation interval is @xmath104 . following the method of analysis in @xcite
, it can be deduced from this that the rms @xmath105 and @xmath106 errors for @xmath60 are both @xmath107 .
this is consistent with the usual @xmath107 accuracy of the forward - time central space discretisation of a parabolic pde @xcite , and the @xmath108 strong accuracy of the milstein discretisation of an sde , see e.g.@xcite .
we now test the accuracy and stability of the scheme numerically .
the chosen parameters for ( [ spde ] ) are @xmath109 , @xmath110 , @xmath111 , @xmath112 , @xmath113 with @xmath114 .
these values are typical for the applications later on .
the upper boundary for the computation is @xmath115 , and chosen to ensure that the use of zero dirichlet data has negligible influence on the solution .
we approximate the initial value problem without boundary conditions on @xmath116 $ ] ( note @xmath117 is roughly in the centre ) , and also the initial - boundary value problem with zero dirichlet conditions on @xmath118 $ ] .
figure [ fig : solns ] plots several solutions to the latter problem at @xmath119 , each corresponding to a different brownian path @xmath120 . for different driving brownian motions @xmath121 .
] it can be seen that for those realisations for which @xmath120 has been largely positive , the boundary at @xmath122 has had negligible influence and so the solution is approximately equal to the displaced normal distribution given by ( [ eq : gaussian ] ) .
for the unbounded case , we approximate the mean - square @xmath105-error by @xmath123 \\ & \approx & \frac{1}{m } \sum_{m=1}^m \sum_{j=0}^j ( v_j^n(\omega_m)-v(n k , j h;\omega_m))^2 \ , h , \label{errex}\end{aligned}\ ] ] where @xmath124 is the number of grid intervals , @xmath125 the number of timesteps , and the expectation is taken over brownian paths @xmath126 , of which @xmath127 are @xmath7 samples . to study the convergence , we introduce decreasing grid sizes @xmath128 and timesteps @xmath129 , motivated by the second order convergence in @xmath52 and first order convergence in @xmath72 as well as the stability limit for the explicit scheme , and denote @xmath130 the mean - square @xmath105-error at level @xmath131 . for the initial - boundary value problem ,
no analytical solution is known , but we can compute error indicators via the difference between a fine grid solution @xmath132 with mesh parameters @xmath72 and @xmath52 , and a coarse solution @xmath133 with mesh parameters @xmath134 and @xmath135 , @xmath136 \\ & \approx & \frac{1}{m } \sum_{m=1}^m \sum_{j=0}^{j/2 } ( f_{2j}^n(\omega_m)-c_j^{n/4}(\omega_m))^2 \ , h , \label{errest}\end{aligned}\ ] ] and define @xmath137 . on the coarsest level , @xmath138 , @xmath139 , such that @xmath140 does not coincide with a grid point .
[ fig : unbddandbddplots ] shows both the computed values of @xmath141 and @xmath142 for the unbounded case , and @xmath142 for the bounded case .
difference between true and numerical solution , @xmath143 as in ( [ errex ] ) , and coarse and fine grid solutions , @xmath144 as in ( [ errest ] ) ; for the bounded case , difference between coarse and fine grid solutions @xmath144 . ]
the results confirm the theoretical @xmath107 convergence in the unbounded case , and support the conjecture that the convergence order is unchanged in the bounded case .
we now formalise this conjecture about the error due to the spde discretisation .
denote @xmath145 the solution to the initial boundary value problem at time @xmath146 for a given brownian path and @xmath147 its numerical approximation with grid size @xmath52 and timestep @xmath148 .
[ conj ] the error in the solution at time @xmath146 satisfies the strong error estimate in the @xmath105 norm @xmath149 @xmath150 } = o(h^{2}).\ ] ] a lipschitz payoff function @xmath151 has a similar strong error bound @xmath152 } = o(h^{2}).\ ] ] we know conjecture [ conj ] to be true for the initial value problem on @xmath153 , and it conjectures that the introduction of the boundary condition at @xmath122 does not affect the ( weak and ) strong error , which is supported by the numerical results .
the conjecture assumes @xmath154 ; in practice , a finite value for @xmath155 will introduce an additional truncation error which will decay exponentially in @xmath155 . numerical tests with different parameters , not reproduced here ,
indicate that the error in the bounded case increases for values of @xmath140 close to 0 , in which case more paths contribute solutions with large higher derivatives close to the zero boundary , even if the asymptotic convergence order is still @xmath107 .
we analyse this loss of regularity further in appendix [ subsec : regularity ] .
we now consider estimating the expectation of a scalar functional of the spde solution , an expected `` payoff '' @xmath156 in a computational finance context . defining @xmath157 to be the payoff arising from a single spde approximation , the standard monte carlo approach is to average the payoff from @xmath125 independent spde simulations , each one using an independent vector of @xmath158 normal variables @xmath159 .
thus the estimator for the expected value is @xmath160 the mean square error for this estimator can be expressed as the sum of two contributions , one due to the variance of the estimator and the other due to the error in its expectation , @xmath161 \right)^2 \right ] = n^{-1 } { \mathbb{v}}[{\widehat{p } } ] + \left ( { \mathbb{e}}[{\widehat{p } } ] \!-\ ! { \mathbb{e}}[p ] \right)^2.\ ] ] to achieve a r.m.s .
error of @xmath40 requires that both of these terms are @xmath162 .
this in turn requires that @xmath163 , @xmath164 and @xmath165 , based on the conjecture that the weak error @xmath166 \!-\ !
{ \mathbb{e}}[p]$ ] is @xmath107 .
since the computational cost is proportional to @xmath167 this implies an overall cost which is @xmath42 .
the aim of the multilevel monte carlo simulation is to reduce this complexity to @xmath38 .
consider monte carlo simulations with different levels of refinement , @xmath168 , with @xmath169 being the coarsest level , ( i.e. the largest values for @xmath72 and @xmath52 ) and level @xmath170 being the finest level corresponding to that used by the standard monte carlo method .
let @xmath171 denote an approximation to payoff @xmath156 using a numerical discretisation with parameters @xmath172 and @xmath173 .
because of the linearity of the expectation operator , it is clearly true that @xmath174 = { \mathbb{e}}[{\widehat{p}}_0 ] + \sum_{l=1}^l { \mathbb{e}}[{\widehat{p}}_l \!-\ ! { \widehat{p}}_{l-1 } ] .
\label{eq : identity}\ ] ] this expresses the expectation on the finest level as being equal to the expectation on the coarsest level plus a sum of corrections which give the difference in expectation between simulations using different numbers of timesteps . the multilevel idea is to independently estimate each of the expectations on the right - hand side in a way which minimises the overall variance for a given computational cost .
let @xmath175 be an estimator for @xmath176 $ ] using @xmath177 samples , and let @xmath178 for @xmath179 be an estimator for @xmath180 $ ] using @xmath181 samples .
each estimator is an average of @xmath181 independent samples , which for @xmath179 is @xmath182 the key point here is that the quantity @xmath183 comes from two discrete approximations using the same brownian path .
the variance of this simple estimator is @xmath184 = n_l^{-1 } v_l $ ] where @xmath185 is the variance of a single sample . combining this with independent estimators for each of the other levels ,
the variance of the combined estimator @xmath186 is @xmath187 the corresponding computational cost is @xmath188 where @xmath189 represents the cost of a single sample on level @xmath131 . treating the @xmath181 as continuous variables , the variance is minimised for a fixed computational cost by choosing @xmath181 to be proportional to @xmath190 , with the constant of proportionality chosen so that the overall variance is @xmath38 .
the total cost on level @xmath131 is proportional to @xmath191 .
if the variance @xmath185 decays more rapidly with level than the cost @xmath189 increases , the dominant cost is on level @xmath192 .
the number of samples on that level will be @xmath38 and the cost savings compared to standard monte carlo will be approximately @xmath193 , reflecting the different costs of samples on level @xmath192 compared to level @xmath170 .
on the other hand , if the variance @xmath185 decays more slowly than the cost @xmath189 increases , the dominant cost will be on the finest level @xmath170 , and the cost savings compared to standard monte carlo will be approximately @xmath194 , reflecting the difference between the variance of the finest grid correction compared to the variance of the standard monte carlo estimator , which is similar to @xmath195 . this outline analysis is made more precise in the following theorem : [ ml - theorem ] let @xmath156 denote a functional of the solution of an spde for a given brownian path @xmath120 , and let @xmath171 denote the corresponding level @xmath131 numerical approximation . if there exist independent estimators @xmath178 based on @xmath181 monte carlo samples , and positive constants @xmath196 such that @xmath197 and 1
. @xmath198 \right| \leq c_1\ , 2^{-\alpha\ , l } $ ] 2 .
@xmath199 = \left\ { \begin{array}{ll } { \mathbb{e}}[{\widehat{p}}_0 ] , & l=0 \\[0.1 in ] { \mathbb{e}}[{\widehat{p}}_l \!-\ ! { \widehat{p}}_{l-1 } ] , & l>0 \end{array}\right .
[ ml - iii ] @xmath184 \leq c_2\ , n_l^{-1 } 2^{-\beta\ , l } $ ] 4 .
[ ml - iv ] @xmath200 where @xmath189 is the computational complexity of @xmath178 then there exists a positive constant @xmath201 such that for any @xmath202 there are values @xmath170 and @xmath181 for which the multilevel estimator @xmath203 has a mean - square - error with bound @xmath204\right)^2\right ] < { \varepsilon}^2\ ] ] with a computational complexity @xmath205 with bound @xmath206 c_4\ , { \varepsilon}^{-2 } ( \log { \varepsilon})^2 , & \beta=\gamma , \\[0.1 in ] c_4\ , { \varepsilon}^{-2-(\gamma\!-\!\beta)/\alpha } , & 0<\beta<\gamma .
\end{array}\right.\ ] ] the proof is a slight generalisation of the proof in @xcite . in our application , we choose @xmath207 and @xmath208 so that the ratio @xmath209 is held fixed to satisfy the finite difference stability condition .
the computational cost increases by factor 8 in moving from level @xmath131 to @xmath210 , so @xmath211 .
given that the payoff is a lipschitz function of the loss approximations at various dates , conjecture [ conj ] implies that the weak error is also @xmath71 and so @xmath212 .
also , due to the triangle inequality @xmath213 } & \leq & \sqrt{{\mathbb{v } } [ { \widehat{p}}_l \!-\ ! p ] } + \sqrt{{\mathbb{v } } [ { \widehat{p}}_{l-1 } \!-\ ! p ] } \\ & \leq & \sqrt{{\mathbb{e } } [ ( { \widehat{p}}_l \!-\ !
p)^2 ] } + \sqrt{{\mathbb{e } } [ ( { \widehat{p}}_{l-1 } \!-\ !
p)^2 ] } , \end{aligned}\ ] ] conjecture [ conj ] and its corollary give @xmath214 .
consequently , the computational cost to achieve a r.m.s .
error of @xmath40 is @xmath38 .
in this section we study the numerical performance of the algorithms presented earlier on the example of cdo tranche pricing in a large basket limit .
basket credit derivatives provide protection against the default of a certain segment ( ` tranche ' ) of a basket of firms .
a typical example is that of a _
collateralized debt obligation _ where the protection buyer receives a notional amount , minus some recovery proportion @xmath215 , if firms in a specified tranche of the basket default , and in return pays a regular spread until the default event occurs .
the arbitrage - free spread depends crucially on the ( risk - neutral , when hedged with defaultable bonds ) probability of joint defaults . for a tranche with attachment point @xmath216 and detachment point @xmath217 , the outstanding tranche
notional @xmath218 where the loss variable @xmath219 is the proportion of losses in the basket up to time @xmath3 , determines the spread and default payments related to that tranche .
the risk - neutral value of the tranche spread can be derived as @xmath220 } { \delta \sum_{i=1}^{n } { \ , { \rm e}}^{-r t_i } \mathbb{e}^{\mathbb{q}}[p(l_{t_{i}})]},\ ] ] see , e.g. , @xcite . here
, @xmath221 is the maximum number of spread payments , @xmath146 the expiry , @xmath222 the payment dates for @xmath223 , @xmath224 the interval between payments .
spreads are quoted as an annual payment , as a ratio of the notional , but assumed to be paid quarterly . for an extensive survey of credit derivatives and pricing models
we refer the reader to @xcite , and note only that they typically fall in one of two classes : so - called reduced - form models , which model default times of firms directly as ( dependent ) random variables ; structural models , which capture the evolution of the firms values , and model default events as the first passage of a lower default barrier .
we will consider the latter here . in a structural model in the spirit of the classical works of @xcite and @xcite ,
as extended to multiple firms e.g. in the work of @xcite , a company @xmath225 s firm value , @xmath226 , is assumed to follow a model of the type @xmath227 where @xmath228 is a correlation parameter , @xmath5 is the risk - free interest rate , @xmath229 the volatility , and @xmath7 , @xmath20 are standard brownian motions . here
, the individual firms are correlated through a common ` market ' factor @xmath7 , but independent conditionally on @xmath7 .
we make in the following the assumption that the firms are _ exchangeable _ in the sense that their dynamics is governed by the same set of parameters , specifically @xmath230 , @xmath231 .
their initial values @xmath232 are not necessarily identical which allows for different default probabilities for individual firms , consistent with their cds spreads . in this framework ,
the default time @xmath233 of the @xmath225-th firm is modelled as the first hitting time of a default barrier @xmath234 , for simplicity constant , and the _ distance - to - default _ @xmath235 evolves according to ( [ sdesys ] ) , where @xmath236 , @xmath237 .
@xmath238 as in ( [ defaulttime ] ) is precisely the default time . in the majority of applications ,
monte carlo simulation of the firm value processes is used for the estimation of tranche spreads , see e.g. @xcite or @xcite .
this is largely enforced by the size of @xmath125 , for instance in the case of index tranches @xmath239 .
this , however , puts the model precisely in the realm of the large basket approximation ( [ spde ] ) .
see @xcite for a numerical study of this large basket approximation .
the loss functional is thereby approximated by @xmath240 in terms of the numerical spde solution @xmath241 at time @xmath222 , which feeds into the estimator for the outstanding tranche notional ( [ tranche - notional ] ) and subsequently the tranche spreads ( [ tranche - spreads ] ) .
all following results are for a representative set of parameter values , taken from a calibration performed in @xcite to 2007 market data , @xmath109 , @xmath110 , @xmath111 , @xmath112 .
while the initial distribution used in @xcite is somewhat spread out to match individual cds spreads of obligors , most of the mass is around @xmath114 and for simplicity we centre all firms there for the numerical tests , @xmath113 .
the upper boundary is chosen as @xmath115 .
we consider a maturity @xmath119 and tranches @xmath242 = [ 0,0.03]$ ] , @xmath243 $ ] , @xmath244 $ ] , @xmath245 $ ] , @xmath246 $ ] , @xmath247 $ ] .
figure [ fig : spde_results ] shows the multilevel results for the expected protection payment from the first tranche , @xmath248,\ ] ] expressed as a fraction of the initial tranche notional . we pick mesh sizes @xmath249 and @xmath250 for @xmath251 and @xmath252 , @xmath139 .
this is motivated by the second and first order consistency in @xmath52 and @xmath72 respectively , and within the stability limit ( [ expl - stab - lim ] ) .
the top - left plot shows the convergence of the variance @xmath253 $ ] as well as the variance of the standard single level estimate , while the top - right plot shows the convergence of the expectation @xmath180 $ ] .
the bottom - left plot shows results for different multilevel calculations , each with a different user - specified accuracy requirement . for each value of the accuracy @xmath40 ,
the multilevel algorithm determines ( by formula ( 12 ) from @xcite ) the numbers of levels of refinement which are needed to ensure that the contribution to the mean square error ( mse ) due to the weak error on the finest grid is less than @xmath254 , and it determines ( by formula ( 10 ) from @xcite ) the optimal number of samples on each level so that the combined variance of the multilevel estimate is also less than @xmath254 , and hence the mse is less than @xmath255 . treating a single finite difference calculation on the coarsest level as having a unit cost ,
the bottom - right plot compares the total cost of the standard and multilevel algorithms . since the objective is to achieve a computational cost which is approximately proportional to @xmath256 , it is the cost @xmath205 multiplied by @xmath255 which is plotted versus @xmath40 .
the results confirm that @xmath257 does not vary much as @xmath258 for the multilevel method , in line with the prediction that @xmath259 , whereas it grows significantly for the standard method since @xmath260 , where @xmath170 is the index of the finest level .
hence , there is a big jump in the cost of the standard method each time it is necessary to switch to a finer level to ensure the weak error is less than @xmath261 , whereas the jump is minimal for the multilevel method . in practice
, default is not monitored continuously but only at a discrete set of times , and for pricing often the simplifying assumption is made that default is only determined at the spread payment dates @xmath262 .
this can be incorporated as follows . in the time intervals @xmath263
, we solve ( [ spde ] ) on a sufficiently large domain ( e.g. @xmath264 $ ] ) , and apply the following interface conditions at @xmath262 : @xmath265 to maintain quadratic grid convergence in spite of the discontinuity introduced by ( [ discont ] ) , we choose the mesh such that a grid point coincides with the 0 boundary ( e.g. by setting @xmath266 in the above example ) , and set the numerical solution after default monitoring to 0 for grid coordinates below 0 and to @xmath267 its previous value at 0 , see e.g. @xcite .
it is seen in fig .
[ fig : spde_results_discr ] that convergence is very similar to the previous case .
here we discuss the computational complexity of the multilevel solution of the spde compared with the alternative use of the multilevel method for solving the sdes which arise from directly simulating a large number of sdes .
we have already explained that to achieve a r.m.s .
accuracy of @xmath40 requires @xmath42 work when solving the spde by the standard monte carlo method , but the cost is @xmath38 using the multilevel method , provided conjecture [ conj ] is correct .
consider now the alternative of using a finite number of particles ( firms ) , @xmath7 , to estimate the tranche loss in the limit of an infinite number of particles ( firms ) . in this case
, empirical results suggest that there is an additional @xmath268 error ( see also @xcite ) , and the proof of this convergence order is the subject of current research . taking this to be the case , the optimal choice of @xmath7 to minimise the computational complexity to achieve an r.m.s .
error of @xmath40 is @xmath269 . using the standard monte carlo method ,
the optimal timestep is @xmath270 , and the optimal number of paths is @xmath38 , so the overall cost is @xmath271 .
using the multilevel method for the sdes reduces the cost per company to @xmath38 , so the total cost is @xmath41 .
this complexity information is summarised in table [ table : comparison ] .
there is also a practical implementation aspect to note .
the computational cost per grid point in the finite difference approximation of the spde is minimal , requiring just three floating point multiply - add operations if equation ( [ discrete ] ) is re - cast as @xmath272 with the coefficients @xmath273 computed once per timestep , for all @xmath274 . if we let @xmath205 be the cost of generating all of the gaussian random numbers @xmath275 for a single spde simulation , then the cost of the rest of the finite difference calculation with 20 points in @xmath2 ( as used on the coarsest level of our multilevel calculations ) is probably similar , giving a total cost of @xmath276 for each spde . on the other hand
, each sde needs its own gaussian random numbers for the idiosyncratic risk , and so the cost of simulating _ each _ sde is approximately @xmath205 , roughly half of the cost of the spdes on the coarsest level of approximation , giving a total cost of @xmath277 . ' '' '' method / model & sde & spde + ' '' '' standard mc & @xmath271 & @xmath42 + ' '' '' multilevel mc & @xmath41 & @xmath38 + we have shown that stochastic finite differences combined with a multilevel simulation approach achieve optimal complexity for the computation of expected payoffs of an spde model . in the case of an absorbing boundary , the complexity estimate is a conjecture in so far it relies on the convergence order of the finite difference scheme , which does not follow from the fourier analysis of the unbounded case .
the matrix stability analysis in appendix [ sec : matanal ] could form part of a rigorous analysis if a lax equivalence theorem could be proved . in the case of multiplicative white noise
this is shown in @xcite .
one difficulty in the present case of an spde with stochastic drift is the loss of regularity towards the boundary , which may be accounted for by weighted sobolev norms of the solution , but even then it is not clear that convergence of the functionals of interest follows .
there are several possible extensions of the present basic model as discussed in @xcite , ranging from stochastic volatility and jump - diffusion to contagion models .
the methods developed in this paper should be of use there also , building for example on multilevel versions of jump - adapted discretisations for jump - diffusion sdes @xcite .
bain , a. & crisan , d. _ fundamentals of stochastic filtering _ , springer , 2009 .
barth , a. , schwab , c. , & zollinger , n. multi - level monte carlo finite element method for elliptic pdes with stochastic coefficients , _ num .
_ , 119(1):123 - 161 , 2010 .
benth , f.e . & koekebakker , s. stochastic modeling of financial electricity contracts , _ energy economics _ , 30(3):11161157 , 2008 .
black , f. and cox , j. valuing corporate securities : some effects of bond indenture provision , _ j. finance _
, 31:351367 , 1976 .
bruti - liberati , n. & platen , e. _ numerical solution of stochastic differential equations with jumps in finance _ ,
springer , 2010 .
buckwar , e. & sickenberger , t. a comparative linear mean - square stability analysis of maruyama- and milstein - type methods , _ math . comp .
_ , 81:11101127 , 2011 .
bujok , k. & reisinger , c. numerical valuation of basket credit derivatives in structural jump - diffusion models , _ j. comp .
_ , to appear
. bush , n. , hambly , b. , haworth , h. , jin , l. , & reisinger , c. stochastic evolution equations in portfolio credit modelling , _ siam fin .
_ , 2(1):627664 , 2011 .
carmona , r. , fouque , j .- p . ,
& vestal , d , interacting particle systems for the computation of rare credit portfolio losses , _ fin .
_ , 13(4):613633 , 2009 .
carter , r. & giles , m.b .
sharp error estimates for discretisations of the 1d convection / diffusion equation with dirac initial data , _
i m a j. num .
_ , 27(2):406425 , 2007 .
cliffe , k.a .
, giles , m.b . , scheichl , r. , & teckentrup , a.l .
multilevel monte carlo methods and applications to elliptic pdes with random coefficients , _ computing and visualization in science _
, 14(1):315 , 2011 .
fouque , j. , wignall b. , & zhou , x. modeling correlated defaults : first passage model under stochastic volatility .
_ j. comp .
_ , 11(3 ) : 4378 , 2008 .
giles , m.b .
multi - level monte carlo path simulation , _ operations research _
, 56(3):981986 , 2008 .
giles , m.b . improved multilevel monte carlo convergence using the milstein scheme , pp .
343358 in _ monte carlo and quasi - monte carlo methods 2006 _ , editors keller , a. , heinrich , s. , & niederreiter , h. , springer - verlag , 2008 .
glasserman , p. _ monte carlo methods in financial engineering _ , springer , 2004 .
graubner , s. multi - level monte carlo methoden fr stochastische partielle differentialgleichungen .
diplomarbeit , tu darmstadt , 2008 .
gyngy , i. lattice approximations for stochastic quasi - linear parabolic partial differential equations driven by space - time white noise ii , _ potential anal .
_ , 11:137 , 1999 .
gyngy , i. & nualart , d. implicit schemes for stochastic quasi - linear parabolic partial differential equations driven by space - time white noise , _ potential anal .
_ , 7:725757 , 1997 .
heath , d. , jarrow , r. , & morton , a. bond pricing and the term structure of interest rates : a new methodology for contingent claims valuation , _ econometrica _ , 60(1):77105 , 1992 .
higham , d.j .
mean - square and asymptotic stability of the stochastic theta method , _ siam j. num .
_ , 38(3):753769 , 2000 .
hull , j. , predescu , m. , & white , a. the valuation of correlation - dependent credit derivatives using a structural model , _
j. credit risk _
, 6(3):99132 , 2005 . kloeden , p.e . &
platen , e. _ numerical solution of stochastic differential equations _ , springer , 1992 .
krylov , n.v .
a @xmath278-theory of the dirichlet problem for spdes in general smooth domains , _ probab .
theory relat . fields _ , 98:389421 , 1994 .
kurtz , t.g .
& xiong , j. particle representations for a class of nonlinear spdes , _ stoch .
_ , 83:103126 , 1999 .
lang , a. a lax equivalence theorem for stochastic differential equations , _ j. comp .
_ , 234(12):33873396 , 2010 .
merton , r. , on the pricing of corporate debt : the risk structure of interest rates , _ j. fin .
_ , 29:449470 , 1974 .
pooley , d.m . ,
vetzal , k.r . , & forsyth , p.a .
remedies for non - smooth payoffs in option pricing , _ j. comp . fin .
_ , 6:2540 , 2003 .
richtmyer , r.d .
& morton , k.w .
_ difference methods for initial - value problems _ , wiley - interscience , 1967 .
roth , c. difference methods for stochastic partial differential equations , _ z. angew
_ , 82(1112):821830 , 2002 .
roth , c. a combination of finite difference and wong - zakai methods for hyperbolic stochastic partial differential equations , _ stoch .
_ , 24(1):221240 , 2006 .
saito , y. & mitsui , t. stability analysis of numerical schemes for stochastic differential equations , _
siam j. num .
_ , 33(6):22542267 , 1996 .
schnbucher , p.j .
_ credit derivatives pricing models _ , wiley , 2003 .
walsh , j.b .
finite element methods for parabolic stochastic pdes , _ potential anal .
_ , 23:143 , 2005 .
xia , y. & giles , m.b .
multilevel path simulation for jump - diffusion sdes , in _
monte carlo and quasi - monte carlo methods 2010 _ , editors wozniakowski , h. & plaskota , l. , springer - verlag , 2012 .
zhou , c. , an analysis of default correlations and multiple defaults , _ the review of financial studies _
, 14:555576 , 2001 .
if @xmath60 is the vector with elements @xmath61 then the finite difference equation can be expressed as @xmath279 a & = & i - \frac{\mu\ , k}{2h}\ , d_1 + \frac{(1\!-\!\rho)\ , k}{2h^2}\ , d_2 , \\[0.05 in ] b & = & -\ , \frac{\sqrt{\rho\ , k}}{2h}\ , d_1 , \\[0.05 in ] c & = & \frac{\rho\ , k}{2h^2}\ , d_2,\end{aligned}\ ] ] where @xmath280 is the identity matrix and @xmath62 and @xmath63 are the matrices corresponding to central first and second differences , which for @xmath281 are @xmath282 from the recurrence relation we get @xmath283 & = & { \mathbb{e}}\left [ v_n^t ( a^t + b^t\ , z_n + c^t\ , z_n^2)(a + b\ , z_n + c\ , z_n^2)\ v_n \right ] \\[0.1 in ] & = & { \mathbb{e}}\left [ v_n^t \left ( ( a\!+\!c)^t ( a\!+\!c ) + b^t b + 2\ ,
c^t c \,\right ) v_n \right].\end{aligned}\ ] ] noting that @xmath62 is anti - symmetric and @xmath63 is symmetric , and that @xmath284 where @xmath285 corresponds to a central second difference with twice the usual span , @xmath286 ( with the end values of @xmath287 being chosen to correspond to @xmath288 and @xmath289 ) , and @xmath290 and @xmath291 are each entirely zero apart from one corner element , @xmath292 then after some lengthy algebra we get @xmath293 } \\ & = & { \mathbb{e}}\left [ v_n^t m v_n \right ] - \left ( e_1 + e_2 \right ) { \mathbb{e}}[(v_1^n)^2 ] - \left ( e_1 - e_2 \right ) { \mathbb{e}}[(v_{j-1}^n)^2],\end{aligned}\ ] ] where @xmath294 and @xmath295 it can be verified that the @xmath296 eigenvector of @xmath7 has elements @xmath297 for @xmath298 , and the associated eigenvalue is @xmath299 where @xmath300 are the same functions as defined in the mean - square fourier analysis . in addition , in the limit @xmath301 , @xmath302 , and therefore in this limit the fourier stability condition @xmath303 is also a sufficient condition for mean - square matrix stability .
figure [ fig : conv ] shows the convergence behaviour as the computational grid is refined .
level 0 has @xmath304 ; @xmath52 is reduced by factor 2 and @xmath72 by factor @xmath305 in moving to finer levels .
the top - left plot shows the convergence of @xmath306 $ ] , with notation as in sections [ sec : multilevel ] and [ subsec : pricing ] ; the top - right plot shows the convergence of @xmath307 $ ] .
the bottom two plots show the behaviour of @xmath308 , which is estimated on each grid level using the standard second difference @xmath309 the left plot indicates that the mean of this quantity is well behaved , but the right plot indicates a singular behaviour of its variance , with the value increasing rapidly with increased grid resolution . | in this article , we propose a milstein finite difference scheme for a stochastic partial differential equation ( spde ) describing a large particle system .
we show , by means of fourier analysis , that the discretisation on an unbounded domain is convergent of first order in the timestep and second order in the spatial grid size , and that the discretisation is stable with respect to boundary data .
numerical experiments clearly indicate that the same convergence order also holds for boundary - value problems .
multilevel path simulation , previously used for sdes , is shown to give substantial complexity gains compared to a standard discretisation of the spde or direct simulation of the particle system .
we derive complexity bounds and illustrate the results by an application to basket credit derivatives . |
numi is a tertiary beam from the main injector @xcite .
protons will be extracted via single turn extraction ( @xmath2sec pulse , cycle time @xmath3 sec . ) from the main injector and focused downward by 58 mrad , where they strike a 0.94 m graphite target to produce the secondary hadron ( @xmath4 , @xmath5 ) beam .
it is anticipated to achieve @xmath6 protons / pulse .
two toroidal magnetic `` horns '' sign- and momentum - select the secondary beam and are moveable so as to achieve different neutrino energy spectra@xcite .
after being focussed forward , the hadrons enter a 675 m long , 1 m radius evacuated decay volume .
soft pions tend to decay in the upstream end of the decay pipe , so must decay at wide angles to send neutrinos to the minos detectors , while stiff pions tend to travel further down the decay pipe .
minos @xcite is a 2-detector neutrino experiment .
the near detector measures the neutrino energy spectrum and rate produced at fermilab . a second , `` far '' , detector is located 735 km away in the soudan mine in minnesota .
the far detector looks for differences from the near spectrum that could be explained by new physics such as neutrino oscillations .
figure [ lespectra ] shows the expected spectra in the two detectors in the absence of new physics .
it is desirable to have the spectra at the two detectors as similar as possible .
differences due to beamline acceptances , etc .
, must be characterized by the `` far - over - near ratio '' ( see figure [ farnear ] ) , which is the factor by which the near spectrum would be multiplied to predict the far spectrum in the absence of new physics .
systematic errors on this factor limit the ultimate reach of oscillation searches . ) minos detectors with and without the hadronic hose for the numi low - energy beam ( ph2le).,width=264 ] the neutrinos in the peaks of the spectra of figure [ lespectra ] come from pions focused by the horns , whereas the neutrinos in the high energy tail come from poorly focused pions ( pions that travel through the necks of the horns ) .
the high energy tail is nonetheless crucial to the experiment because for low @xmath7 signals it provides us it provides a control sample to demonstrate a region without oscillation effects .
it is proposed to add an additional focusing system for the secondary pion beam , the hadronic hose. it consists of a 2.38 mm diameter aluminum wire at the center of the numi decay tunnel and carries a 1000 a peak current pulse .
the current provides a toroidal magnetic field which focuses positive particles back toward the center of the numi decay pipe . a typical meson executes 3 - 4 orbits around the wire over the length of the decay pipe
the idea is similar to a proposal by van der meer , but here implemented for a neutrino beam.@xcite the hadronic hose contributes two essential features : ( 1 ) @xmath4 s and @xmath5 s that otherwise diverge out to the decay pipe walls and interact before they decay are given a restoring force back to the decay pipe center .
thus , @xmath4s/@xmath5 s travel farther and have a greater chance to decay , so the neutrino flux is increased ( see table [ fluxes ] ) ; ( 2 ) the pion orbits effectively randomize the decay angle between the pion direction and the neutrino that hits the minos near and far detectors . because of the kinematic correlation @xmath8 between neutrino energy and decay angle , this randomization is important to make the neutrino energy spectra in the near and far minos detectors more similar ( see figure [ lespectra ] )
differences in the far - over - near ratio due to different predictions@xcite of pion cross sections off the target are reduced ( see figure [ farnear ] )
. .charged current event yields in the far minos detector in the 3 different beam configurations , with and without hadronic hose .
the peak refers to the regions @xmath96 , 9 , and 30 gev , for low , medium , and high energy beams , while overall refers to @xmath10 gev .
the units are events per kiloton of far detector mass per year of running . [
cols="<,^,^",options="header " , ] besides minimizing the experiment s sensitivity to variations in pion productions cross sections , the hadronic hose loosens accuracy criteria for other beamline components .
the eccentricity of the inner conductor of horn 1 , required to be @xmath110.08 mm at its neck position in order to minimize fringe fields which affect high energy pions , is relaxed to @xmath110.12 mm . the spatial alignment of horn 1 transverse to the beamline is relaxed from @xmath12 mm to @xmath13 mm .
relative current variations between the two horns can now be as large as @xmath14% ( c.f .
@xmath13% ) .
finally , the alignment of the straightness of the numi decay pipe , previously required to be within 9 mm all along its 675 m length , is now relaxed considerably .
the hose guides the beam along even long arc - like excursions in the beamline geometry .
the hadronic hose is built up out of 72 9 m sections of wire , each of which is an independent circuit .
the electrical connection of each section to the transmission line passes through an electrical feedthrough in the decay pipe wall consisting of an aluminum pin and ceramic insulating jacket .
the wire is suspended at the center of the decay pipe by 0.1
diameter invar guide wires attached to the decay pipe at @xmath15 m intervals along the decay pipe .
the guide wires are insulated from the decay pipe wall by ceramic pins .
the ceramics will be exposed to @xmath16 rad / year in numi .
tensioning springs at the end of the wire maintain a 2 lb .
tension on the wire .
a 2 m hose segment at the most upstream end of the decay pipe is left unpulsed and absorbs some of the unreacted proton beam .
each hose segment is connected in parallel to a transmission line that runs down the side passageway of the decay tunnel , through a 3:1 transformer .
the transmission line is energized with a 5000 v pulse 620 @xmath17sec in duration .
the upstream end of a hose segment is center - tapped to the same transformer as the downstream end of the preceding hose segment .
all the hose segments are effectively in series , receiving the same currents to within 1 a according to simulations , where the difference is due to the 80@xmath18 c temperature variation of the hose segments inside the decay tunnel .
if a hose segment should fail and sever during operation , all the current passes through the transmission line .
the 620 @xmath17sec rms , 1000 a peak , pulse causes a 250 v voltage drop across each of the 9 m segments , dominated by the inductive voltage drop ( @xmath19 for the wire in the 1 m vacuum decay pipe ) .
the resistance of the 9 m long , 2.4 mm diameter al1350 wire is @xmath20 m@xmath21 , giving a resistive drop of 83 v. this current pulse deposits @xmath22 j into the wire . at 250
v , the voltage on the wire is well below the breakdown point measured for gas pressures of 0.1 - 1.0 torr expected in numi ( see figure [ heat ] ) .
pion interactions in the wire material motivate the choice of low @xmath23 wire ( see figure [ material ] ) , and also small ( 2.4 mm ) diameter wire to avoid reductions in neutrino flux .
aluminum alloy 1350 has a conductivity approximately 80% of pure copper .
pure aluminum experiences plastic flow ( `` creep '' ) at elevated temperature , and we have measured several alloys creep rates at various temperatures and tensions on the wires over a period of 400 days by placing them in a high temperature vessel ( see figure [ heat ] ) . in the numi beam it is expected the wire will operate at 120 - 150 @xmath18c , at which the creep rate is low enough to allow expansion into the 20 cm gaps between wire segments over 10 years of operation without causing voltage breakdown between segments .
the hose wire is hit by the protons that did nt react in the numi target . approximately @xmath24 protons per beam pulse enter the decay pipe with a @xmath25 cm spot size after multiple scattering in the target , and this cloud is almost perfectly focused into the hose wire . ) , so acquire a component of their momentum in the azimuthal direction looking down the beamline which causes their orbits in the hose field circle the wire . ]
the energy deposition in the hose wire , as calculated using the mars@xcite beamline monte carlo , agrees qualitatively with the energy deposition expected from protons focused back into the hose further downstream in the decay tunnel and from the proton beam directly hitting the end of the first hose wire segment ( see figure [ deposition ] ) .
the peak of this energy deposition is 18 m downstream in the decay pipe , and is 4.5 w / m .
the hose wire dissipates heat largely through conduction through the residual gas ( @xmath26torr in the decay pipe ) or through blackbody radiation .
the hose wire is coated with a 17 @xmath17 m thick aluminum oxide layer10% . ] to improve its emissivity from 0.1 for bare aluminum to @xmath27 .
the emissivity has been measured in a 25 foot long , 6 diameter vacuum chamber with blackened interior walls and a hose wire running down its center at dc current to provide a known power in .
the temperature was inferred from the elongation of the wire viewed through a window and the coefficient of thermal expansion measured for the wire .
we have investigated the potential impact of a new focusing system , the hadronic hose , for conventional neutrino beams . such a system is being considered for the numi beam at fermilab , and may be of benefit to future conventional super beams because of its increase in neutrino flux and ability to control systematic uncertainties due to particle production in the target or imperfections in the rest of the neutrino beam .
j. hylen _ et al .
_ `` conceptual design for the technical components of the neutrino beam for the main injector ( numi ) , '' fermilab - tm-2018 , sept . , 1997 .
v. garkusha , f. novoskoltsev , and v. zarucheisky , `` the ph2 m two horn focussing system for the numi project , '' fermilab note numi - b-471 ( 1999 ) . the minos collaboration , `` the minos detectors technical design report , '' fermilab numi - l-337 , oct .
1998 , s. wojicki , spokesman .
s. van der meer , _ the beam guide _ , cern preprint number cern-62 - 16 , april 17 , 1962 .
mokhov , `` the mars monte carlo '' , fermilab fn-628 ( 1995 ) ; o.e .
of the third and fourth workshops on simulating accelerator radiation environments ( sare3 and sare4 ) , fermilab - conf-98/043(1998 ) and fermilab - conf-98/379(1998 ) .
m. bonesini , a. marchionni , f. pietropaolo , and t. tabarelli de fatis , `` on particle production for high energy neutrino beams , '' eur .
j. * c20 * , 13 - 27 ( 2001 ) .
malensek , `` empirical formula for thick target particle production , '' fermilab fn-341 , oct .
1981 . , cern program library , w5013 ( 1994 ) . j. hylen _ et al .
_ , `` proposal to include hadronic hose in the numi beamline , '' fermilab numi - b-542 , oct . 1999
. j. hylen _ et al .
_ , `` technical design report for the hadronic hose '' , numi - b-610 ( 2000 ) . | the neutrinos at the main injector ( numi ) beam supplies an intense @xmath0 beam to the main injector neutrino oscillation search ( minos ) .
the @xmath0 s are derived from a secondary @xmath1 beam that is allowed to decay within a 675 m decay tunnel .
we are developing a continuous toroidal magnetic focusing system , called the hadronic hose , to better steer this secondary beam .
the hose will both increase the net neutrino flux reaching the minos detectors and reduce systematic differences in the neutrino energy spectra at the two detectors due to solid angle acceptances .
.3 cm |
lipids are polymer molecules composed of hydrophobic acyl chains attached to a hydrophilic polar head group . in the presence of a polar solvent ( e.g. , water ) lipids can spontaneously self - assemble to form a bilayer membrane ( fig .
[ lipid ] ) .
the fluid ( l@xmath6 ) phase of lipid bilayers behaves as a two dimensional ( 2d ) fluid and the lateral 2d self - diffusion coefficient of individual lipids within a leaflet of the bilayer has been determined both experimentally by a variety of methods @xcite and through computer simulations @xcite .
the experiments suggests that there are at least two relevant length / time scales associated with the lateral diffusion of lipids in a bilayer .
experiments designed to probe motion on picosecond ( ps ) time scales measure a diffusion coefficient @xmath7 that can be one to two orders of magnitude larger than the diffusion coefficient @xmath8 measured in experiments which probe motion on tens or hundreds of nanosecond ( ns ) time scales .
there have been several models used to explain the difference in the diffusion coefficients @xmath7 and @xmath8 .
vaz and almeida @xcite suggested a free - volume jump - diffusion model where the lipid moves in discrete steps when a void forms next to the lipid .
after a jump , the lipid can either return to its original position or another lipid can jump into the empty space left behind . in this model ,
the short time diffusion coefficient @xmath7 is associated with the `` rattling '' motion of the lipid inside the `` cage '' created by the surrounding lipids .
the jump diffusion mechanism was investigated by falck _ _ et al.__@xcite in a simulation of diyristoyl - phosphatidylcholine ( dmpc ) bilayers , and they did not observe enough jump events to provide evidence for the jump diffusion model . to explain their md simulation results of a dmpc bilayer , wohlert and edholm @xcite proposed a model where the motion of the lipid in the plane of the bilayer can be regarded as diffusion ( with a coefficient @xmath7 ) confined in a circular `` cage '' whose center undergoes free diffusion ( with a coefficient @xmath8 ) .
the two diffusion coefficients obtained by fitting their simulated data to the analytical solution of this model were in reasonable agreement with the experimental values .
none of the above diffusion models explicitly take into account the microscopic polymeric structure of lipids , but effects of this structure has been observed in simulations . for example , doxastakis _ et al . _
@xcite observed in a md simulation of 1,2-dipalmitoyl - sn - glycero-3-phosphocholine ( dppc ) bilayers that the atoms at the ends of the lipid tails fluctuate more than those in the head group , but the overall diffusion of the lipid is limited by the diffusion of the head group .
this observation led them to examine a diffusion in a sphere model where the size of the sphere depends on the position of the chain .
however , the authors describe this model as `` not particularly accurate '' , and the model was extended to include a distribution of spheres at each tail position .
the dynamics of the atoms in lipid molecules are more complex than those in ordinary liquids .
in a simple liquid @xcite atoms move ballistically at short times .
thus the mean square displacement ( msd ) @xmath9 ^ 2 \rangle \approx \langle ( v t)^2 \rangle \sim t^2 $ ] , which is followed by a crossover to fickian diffusion , characterized by @xmath9 ^ 2
\rangle \sim t$ ] for long times . in dense fluids
a caging effect , where the atoms are trapped by their neighbors , is observed between the ballistic and diffusive regimes , leading to a plateau in @xmath9 ^ 2 \rangle$ ] . in a lipid bilayer ,
the motion of lipid atoms is further complicated by the polymeric structure , characterized by connectivity and flexibility , of the lipids .
a more realistic description of the diffusion of lipids should be based on theoretical models designed for polymers .
two such theories are the rouse model @xcite and the mode - coupling theory @xcite . in the rouse model over - damped brownian motion
is assumed for the individual monomers with a harmonic potential connecting adjacent monomers .
the motion is subdiffusive with @xmath10 ^ 2 \rangle \sim t^{1/2}$ ] then there is a crossover to diffusive motion @xmath9 ^ 2 \rangle \sim t^{1}$ ] at later times .
the mode - coupling theory ( mct ) , generally used to study the dynamics of glass - forming liquids , has recently been extended for flexible macromolecules and polymers @xcite , though it has not been applied to study lipid bilayers to our knowledge .
the mct is a method used to approximately calculate space - time correlation functions ( e.g. , intermediate scattering function ) by solving a set of integro - differential equations obtained from first principles .
an important feature of the extended polymer version of the mode coupling theory is the prediction of a sub - diffusive region between the short time ballistic and the long time fickian diffusion regimes . specifically , the theory predicts @xmath9 ^ 2 \rangle \sim t^{\beta}$ ] , with @xmath11 typically between 0.5 and 0.65 . besides other parameters ,
the exponent @xmath11 depends on the length of the polymer chain and approaches 0.5 , the rouse limit , for an infinite chain of identical monomers@xcite . since the only input to the theory is the static structure , the subdiffusion predicted by the extended mct is due to the polymeric structure and not to the specifics of the interactions .
thus , one expects that the dynamics of the atoms in a lipid bilayer should exhibit a pronounced sub - diffusive regime before it crosses over to normal fickian diffusion .
we argue that the extended sub - diffusive region in the msd is the cause of the difference between the experimentally measured diffusion coefficients @xmath7 and @xmath8 . in this paper , we report an extensive computational investigation of the complex dynamics of both individual lipid atoms and entire lipid molecules in the fluid phase of a hydrated dmpc bilayer by employing a microscopic description of the system .
our study is based on a @xmath12s long all atom molecular dynamics ( md ) simulation of the dmpc bilayer . by calculating the time evolution of the mean square displacement of lipid atoms in the plane of the bilayer for a time interval ranging from @xmath13s to @xmath14s ,
we identify the short time ( @xmath15 ps ) ballistic region , the long time ( @xmath16 ns ) fickian diffusion region , and in between an extended _ subdiffusive _ region characterized by a power law @xmath17 ^ 2 \rangle \sim t^{\beta}$ ] , with @xmath3 .
we find that @xmath11 depends on the atom type and , most importantly , on its position within the lipid molecule .
the dynamics is different for atoms in the lipid tail versus atoms in the head group .
we also examine the implications of the heterogeneous subdiffusion of the individual atoms on the diffusion of the lipid molecule as a whole , and on the interpretation of neutron scattering experiments .
this paper is organized as follows . in sec .
[ sim ] we describe our md study of a solvated dmpc lipid bilayer . in sec .
[ msd ] we provide a detailed numerical and theoretical study of the mean square displacement of lipid atoms and lipid molecules . in sec .
[ sf ] we examine the implications of this study for the incoherent intermediate scattering function , and the dynamic structure factor that is measured in neutron scattering experiments .
lastly , in sec .
[ sec : conc ] we present a discussion and our conclusions . ,
title="fig:",width=307 ] , title="fig:",width=307 ]
to investigate the dynamics of selected lipid atoms and of entire lipid molecules , we performed a @xmath0s long all atom md simulation of a fully solvated dmpc lipid bilayer ( fig .
[ lipid ] ) .
the pre - equilibrated structure of the bilayer was obtained from mikko karttunen s web site ( www.apmaths.uwo.ca/@xmath18mkarttu/downloads.shtml ) @xcite and contained 128 lipid molecules .
the system was solvated by adding two 11 thick layers of water to each side of the membrane using the _ solvate _ plugin in vmd @xcite .
the final system contained a total of 2577 tip3 @xcite water molecules .
the md simulation was performed with namd-2.6 @xcite using the charmm27 @xcite force field .
the equations of motion were integrated with a multiple time step algorithm with time steps of 1 fs for bonding interaction , 2 fs for non - bonding interactions , and 4 fs for long - range electrostatic interactions .
the non - bonded interactions were cutoff at 12 with a smooth switching function starting at 10 .
long - range electrostatic interactions were computed by employing the smooth particle mesh ewald method @xcite with a grid spacing of 1 .
the system was brought to @xmath19 k and normal pressure ( @xmath20 atm ) through several stages of equilibration .
first , the system was subjected to @xmath21 energy minimization steps by harmonically restraining the phosphorous ( p ) atoms of the lipid headgroups along the normal direction to the surface of the membrane .
next , the system was gradually heated to the desired temperature @xmath19 k during a 0.5 ns period . after removing all the restraints ,
the system was equilibrated through a 75 ns long npt simulation . at the end of this process
the area per lipid ( apl ) was 56.2 @xmath22 . to increase the apl to the desired value of @xmath23 @xmath22 , we performed nvt simulations and gradually increased the size of the system in the @xmath24-plane of the membrane .
the final system size was @xmath25 @xmath26 , with an apl of @xmath27 @xmath22 .
the equilibration process was concluded with an additional 10 ns nvt simulation .
a production run of @xmath0s was performed in the nvt ensemble .
we employed a langevin thermostat with a coupling constant of 0.05 ps@xmath28 .
the coordinates of all the atoms were stored every 2 fs for the first 100 ps , then saved every 20 fs for the next 10 ns , then every 100 fs for the next 90 ns . this allowed us to study the short , intermediate , and long time dynamics of the individual atoms and lipids . a snapshot of the lipid bilayer system is shown in fig .
[ lipid]a .
the md simulations were carried out on 40 cpus of a dual core 2.8ghz intel xeon em64 t cluster with a performance of around 0.2 days / ns .
in this section we examine the time dependence of the lateral mean square displacement ( msd ) of selected atoms within a lipid , and of entire lipids within the bilayer .
we compare the results with the general predictions of the extended mct for polymers @xcite , and we propose a memory function method capable of describing the time dependence of msd in lipid bilayers .
we begin by outlining some of the qualitative predictions of the extended mct for polymers . to derive the extended mode coupling equations ,
one first applies the mori - zwanzig projection operator formalism in the so called site - site formalism .
this leads to a set of coupled integro - differential equations for monomer density fluctuations .
the mct is a set of approximations of the memory kernel that permits the corresponding integro - differential equations to be numerically solved with the only input being the static structure .
the monomer and the polymer center of mass msd can be computed from the solution of the mode - coupling equations from the small wave - vector , @xmath29 , limit ( see ref . for details ) . for the msd @xmath30 ^ 2 \rangle = \langle [ \vec{r}\,^a(t ) -
\vec{r}\,^a(0)]^2 \rangle$ ] for monomer @xmath31 , the mct predicts an initial ballistic region where @xmath30 ^ 2 \rangle \sim t^2 $ ] .
after the ballistic regime , the growth of the mean square displacement is suppressed due to a local caging effect caused by the trapping of the polymer by its neighbors . at larger times
there is a crossover to fickian diffusion with @xmath30 ^ 2 \rangle = 2d d_l t$ ] , where @xmath32 is the dimension of the system and @xmath33 is the long time self diffusion coefficient for the monomer . in this crossover region
there is a sub - diffusive region where @xmath30 ^ 2 \rangle \sim t^\beta$ ] with @xmath3 .
the value of @xmath11 is smaller for longer chains and has a limiting value of 0.5 for an infinitely long chain .
note that the sub - diffusive region is absent for simple liquids , and is a _ polymer
_ specific feature . to test the mct predictions for the lipids
, we examined the motion of three representative atoms in the lipid ( see fig .
[ lipid]b ) , namely the phosphorous ( p ) , the carbon c15 ( ch ) atoms in the lipid head group , and the carbon atom c214 ( ct ) located at the end of one of the lipid tails .
because the surroundings of the head group atoms is different from those in the lipid tails , we expect that the msd of the p and ch atoms will show distinctive features from that of the ct atoms . shown in fig .
[ pmsd ] is the time evolution of the lateral msd of the p atom ( squares ) @xmath34 ^ 2 \rangle = \langle |r^p(t)-r^p(0)|^2 \rangle$ ] on a log - log scale , which follows the qualitative predictions of the mct .
specifically , an initial ballistic region is followed by a narrow caging region that gradually crossover to linear diffusion through an extended sub - diffusive region characterized by @xmath35 ^ 2
\rangle \sim t^{\beta}$ ] , with @xmath36 . .
inset : @xmath37 ^ 2 \rangle /\partial \ln(t)$ ] .
the dashed horizontal lines correspond to the ballistic ( @xmath38 ) , sub - diffusive ( @xmath39 ) and fickian diffusion ( @xmath40 ) values.,width=307 ] also plotted in fig .
[ pmsd ] is the msd for the ch ( triangle ) and ct ( circles ) atoms .
the ballistic , the sub - diffusive , and the diffusive time scales are identifiable for all the atoms .
the msd in the ballistic regime is the same for ch and ct since they have the same mass and are at the same temperature , but their mean square displacements differ starting in the caging region .
the msd for ct is larger than that of ch , which indicates that the size of the cage is larger for atoms at the end of the tail than atoms close to the head group .
this difference is a direct consequence of the structure of the lipid bilayer , and has been observed in previous simulations @xcite .
for all the atoms the crossover to fickian diffusion begins around 10 ns and @xmath30 ^ 2 \rangle\approx 100 $ ] @xmath22 , which approximately corresponds to the nearest neighbor distance of the lipids ( @xmath41 ) . thus the linear diffusion can be observed only after the lipid has moved around one lipid diameter . to determine the power - law exponent @xmath42 for atom @xmath31 in the sub - diffusive region
, we calculated @xmath43 ^ 2 \rangle ] } { \partial \ln(t ) } = \frac{t}{\langle[\delta{r}^a(t)]^2\rangle}\frac{\mbox{d}\langle[\delta{r}^a(t)]^2\rangle}{\mbox{d}t } , \ ] ] and averaged the result from @xmath44 ns to @xmath45 ns , the region where @xmath46 is approximately constant .
the result for @xmath46 for the center of mass ( cm ) of the lipids is shown in the inset to fig .
the horizontal lines in the inset correspond to the ballistic ( @xmath47 ) , sub - diffusive ( @xmath48 ) and fickian diffusion ( @xmath49 ) regimes .
the value of @xmath42 depends on the atom type and ranged from @xmath50 for the cm to @xmath51 for ct , the carbon atom at the end of the tail .
the dynamics in the ps time scale strongly depends on the position of the atom within the lipid .
the atoms at the end of the lipid tails have more freedom to move , and hence a larger cage . however , these atoms are connected to those within the head group , which ultimately determine their diffusion on long time scales . for @xmath16 ns ,
@xmath52 for each of the individual atoms ( @xmath53 p , ch , ct ) and the center of mass , thus the lipids appear to be undergoing fickian diffusion .
we calculated the diffusion coefficient @xmath54 ^ 2 \rangle/4t$ ] by fitting @xmath55 ^ 2 \rangle$ ] for @xmath56 ns to a straight line .
the diffusion coefficients ( @xmath33-msd ) are shown in table [ dtable ] , and their values range between @xmath57 cm@xmath22s@xmath28 and @xmath59 cm@xmath22s@xmath28 .
these values are within statistical error and are close to the diffusion coefficient of the cm
. however , these diffusion coefficients are almost a factor of two larger than the value of @xmath60 cm@xmath22s@xmath28 obtained from frap measurement @xcite of dmpc at 34@xmath61c .
this difference may be due to a combination of several factors , e.g. , imperfection of the empirical charmm27 force field used in the md simulations , discrepancy between the apl of the experimental and the simulated systems , finite size of the dmpc bilayer system and insufficient sampling .
however , the fact that we have obtained similar long time diffusion coefficient for all lipid atoms indicates that our md trajectory is sufficiently long to capture the linear diffusion regime .
.[dtable]the diffusion coefficient calculated from a linear fit to the the mean square displacement and the memory function approach described by eqs . [ msd - proj ] and [ memory - function ] for the atoms shown in fig .
[ lipid ] and the center of mass ( cm ) .
the experimental value obtain from frap experiments is 0.69@xmath62 cm@xmath22 s@xmath28 @xcite . [
cols="^,^,^",options="header " , ]
in spite of extensive literature dealing with the experimental and computational study of the dynamics of lipid bilayers , the determination of the lateral self - diffusion coefficient @xmath33 of lipid molecules in the membrane remains a controversial problem .
the results for @xmath33 appear to depend significantly on the time and length scales probed by the employed measurement or computational method .
this observation prompted researchers to propose different models that have been predominantly influenced by models of diffusion in dense fluids .
however , lipid molecules are polymers , characterized by flexibility and connectivity , and their lateral diffusion in the leaflets of lipid bilayers is qualitatively different from the diffusion of molecules in dense fluids .
this difference is reflected in the time dependence of the msd .
we observe a broad sub - diffusive region , not present in simple fluids , between the short time ballistic region and the long time fickian diffusion region .
here we developed a phenomenological memory function approach , which can be used to describe the behavior of the msd of lipid atoms and molecules over the whole time interval spanning from the ballistic regime ( t @xmath63 ) to the fickian diffusion regime ( t @xmath64 ) . by fitting the mean square displacement , we were able to identify three clearly separated time scales that correspond to three different dynamic regimes in the lipid system .
overall , our memory function fits to the msd matched remarkably well @xmath1 ^ 2\rangle$ ] from the md simulation .
however , some features in the fairly broad crossover region between the ballistic and sub - diffusive regimes are missed by our approximation .
our memory function approach allows us to calculate the lateral self - diffusion coefficient @xmath33 of lipid atoms and molecules through eq . , and shows that , while the dynamics of the different atoms are very different at ps time scales , the motion at @xmath65 is best described by simple diffusion .
we also investigated the consequence of the sub - diffusive behavior of the msd by examining the first two terms in the cumulant expansion of the sisf .
the first term is referred to as the gaussian approximation and is exact if the probability distribution @xmath66 of the displacements is gaussian . by calculating the first non - gaussian correction , characterized by the parameter @xmath67 ( eq . [ nongauss ] ) , we found that @xmath66 is nearly gaussian outside the sub - diffusive region . while for the p atoms and the center of mass of individual lipid molecules @xmath66 remained nearly gaussian even in the sub - diffusive region , @xmath66 for the carbon and hydrogen atoms showed significant deviation from gaussian in this region .
interestingly , a gaussian distribution for the center of mass , whose width did not increase as the square root of time , for lipid displacements has been observed in previous simulations @xcite .
when @xmath66 is gaussian , the sisf ( e.g. , measured in inelastic neutron scattering ( ins ) experiments of protonated lipid membrane samples ) can be expressed in terms of the msd , eq . .
the sub - diffusive dynamics strongly influences the scattering from the lipids whenever the probing scattering vector , @xmath29 , and frequency , @xmath68 , correspond to lengths and times in the sub - diffusive regime of the lipid atoms and molecules . in this case
the sisf can be fairly well fitted with a stretched exponential ( kww ) function , @xmath69^{\beta(q)}]$ ] .
note , however , that for a given @xmath29 the stretched exponential will not be able to reproduce the short ( ballistic ) and long ( fickian diffusion ) time regimes .
we examined the @xmath29 dependence of @xmath70 and @xmath71 . for the phosphorus atoms we find that @xmath72 for @xmath73 , which agrees with the exponent determined from the msd .
we emphasize that in most ins experiments on lipid bilayers primarily one explores the sub - diffusive region and not the long time fickian diffusion region , thus making the use of eq .
therefore , it should come as no surprise that the values of the lipid lateral self - diffusion coefficient extracted from ins experiments using eq .
have a strong @xmath29 dependence and generally overestimate the real @xmath33 obtained by other ( e.g. , frap ) experiments .
furthermore , it would be interesting to investigate the temperature dependence of the analysis presented in this paper . in some liquids
, it has been observed that activated events become more pronounced as the temperature is lowered .
it would be reasonable to expect a similar behavior in lipid membranes , thus the diffusion could change character at lower temperatures and in the gel phase .
it is also unknown how the picture presented in this work would change in the gel phase of the lipid membrane .
since the tails of the lipids are aligned , do the lipids behave as if they are more rigid ? this may change the sub - diffusive region dramatically .
the addition of cholesterol or proteins to the lipids may also change the nature of the sub - diffusive region and dramatically change the dynamics of the lipids .
we gratefully acknowledge the computer time that was generously provided by the university of missouri bioinformatics consortium . e.f . acknowledges partial support of nsf grant no .
che 0517709 . | we use a long , all - atom molecular dynamics ( md ) simulation combined with theoretical modeling to investigate the dynamics of selected lipid atoms and lipid molecules in a hydrated diyristoyl - phosphatidylcholine ( dmpc ) lipid bilayer . from the analysis of a @xmath0s md trajectory we find that the time evolution of the mean square displacement , @xmath1 ^ 2\rangle$ ] , of lipid atoms and molecules exhibits three well separated dynamical regions : ( i ) ballistic , with @xmath1 ^ 2\rangle \sim t^2 $ ] for @xmath2 ; ( ii ) subdiffusive , with @xmath1 ^ 2 \rangle \sim t^{\beta}$ ] with @xmath3 , for @xmath4 ; and ( iii ) fickian diffusion , with @xmath1 ^ 2\rangle \sim t$ ] for @xmath5 .
we propose a memory function approach for calculating @xmath1 ^ 2\rangle$ ] over the entire time range extending from the ballistic to the fickian diffusion regimes .
the results are in very good agreement with the ones from the md simulations .
we also examine the implications of the presence of the subdiffusive dynamics of lipids on the self - intermediate scattering function and the incoherent dynamics structure factor measured in neutron scattering experiments . |
the disordered network structure of glasses provides a highly complex environment for the transport of mobile ions .
a basic consensus has developed that the ionic motion consists of jumps between distinct minima of the energy landscape @xcite . because of the strong coulomb interaction between the ions
the effective energy landscape will perpetually change with the motion of the ions .
the dispersion in the frequency dependence of ionic conductivity can be interpreted @xcite as a consequence of correlated backward jumps forced by the interaction with other ions .
the complexity of the ion dynamics in glasses is highlighted by the mixed
alkali effect @xcite : in glasses containing two different alkali species , a deep minimum in the mobility of the ions is observed .
even small concentrations of a second alkali species effectively slow down the total ion dynamics .
several important structural properties of mixed alkali ( ma ) glasses are already known .
exafs measurements @xcite have established that the cations retain the same specific coordination environments as in single alkali ( sa ) glasses .
this was also supported by reverse monte carlo calculations @xcite .
furthermore , nmr measurements @xcite have shown that the different alkali cations are intimately mixed on the microscopic scale and do not belong to phase segregated parts .
specific information about the dynamics of cations in disordered ion conductors can be obtained via molecular dynamics ( md ) simulations . for sa glasses
it has been shown that regions of high ionic mobility form channels @xcite that arise statistically @xcite .
proof for some structure at the scale of these channels can also be found in measurements of the dynamic structure factor @xcite .
one of the most important md results for ma glasses is the observation that ions only rarely jump into sites vacated by the other species @xcite and that the site characteristics of both species are very different @xcite .
several models @xcite , inspired by the observations of site selectivity , feature distinct sets of adapted sites for each alkali species .
taking the site selectivity as a starting point one may ask whether the dynamics in the ma system of composition @xmath0 can be viewed as the _ independent _ superposition of two sa systems @xmath1 and @xmath2 , respectively .
if this were true the dramatic decrease of the dc conductivity in the ma system would be related to the well - known observation that the alkali dynamics in sa systems become much slower for lower alkali concentration and one might just speak of a dilution effect .
experimental results tell that reality is more complex . for example , in recent measurements of dc conductivities and tracer diffusivities in na rb
borate glasses the mobility of sodium and rubidium is compared to that in pure na borates and rb borates with identical alkali content @xcite . despite a significant decrease of the conductivity for the ternary glasses the conductivity is still significantly larger than the sum of those in the respective sa glasses .
furthermore , the experimental data reveal asymmetric behavior with respect to the two alkali species . a dramatic difference is visible when comparing the limit where the smaller ion ( here : na ) is the minority species as compared to the opposite limit .
it turns out that the small ion as minority species is still relatively mobile and its diffusion constant , as determined by tracer diffusion experiments , only weakly depends on concentration .
in contrast , in the other limit the large ion becomes immobile .
similar observations have been already made in earlier experiments ; see e.g. @xcite . a simple reason for the apparent increase of mobility in mixed alkali systems as compared to the sum of the individual glasses may be given by the different network structures . due to the higher
alkali content in the mixed alkali system the network structure is more discontinuous , giving rise to higher network mobility and thus , possibly , to higher alkali mobility . beyond this effect it is discussed whether the adaption of the individual sites may change with time .
the rearrangement of free volume has been put forward as a possible mechanism for the readaption , based on measurements of the pressure dependence of ionic conductivity @xcite : a site entered by a too large ion has to expand , while a site entered by a smaller ion can shrink .
if both happens at the same time in close vicinity , an isochoric redistribution of site volume via relaxations of the matrix could facilitate the readaption of both sites .
if this mechanism of _ matrix mediated coupling _ were indeed present , an important correlation between the unlike cations would be a present in the ma system @xcite .
actually , the presence of the release of mismatches and thus the readaption of sites in the glassy phase has been already postulated in the dynamic structure model @xcite .
others suppose that the adaption of a site is definitively fixed during the glass transition @xcite . in a very different approach
it is assumed that a distribution of sites is open in principle to all ions @xcite . in this class of models , the competition for favorable sites
is governed by the different radii of the cation species .
the goal of our work is to elucidate the mechanism of ion dynamics in mixed
alkali systems via md simulations on a microscopic level .
the analysis is based on our method @xcite to identify individual ionic sites from the md trajectories ; see also @xcite .
this procedure was also successfully applied to simulations of phosphate glasses @xcite and rationalized in more general terms @xcite . in this investigation ,
the method is for the first time applied to a ma system . from comparison of the dynamics in sa systems and ma systems
, three major aspects will be treated .
first , the site selectivity is quantified in detail .
second , the trajectories of the alkali ions are characterized and compared with the typical behavior in sa systems .
third , dynamic correlations between like and unlike cations are analyzed .
we performed molecular dynamics simulations of alkali silicate glasses @xmath3 with @xmath4 .
we use a modified version of the software moldy @xcite , using buckingham
type pair potentials developed by habasaki _
@xcite , that are well tested for comparable systems @xcite .
the simulations were done in the nvt
ensemble , with systems of 1215 particles , i.e. 270 alkali ions .
the size of the simulation boxes was set according to experimental densities @xcite , resulting in @xmath5 . in @xcite
the experimental densities for the two sa systems could be reproduced from npt - simulations at identical pressure .
in contrast , we obtain differing pressures in our nvt runs with these experimental densities . at @xmath6k ,
where equilibrium runs are easily possible , the pressure of the two limiting systems for our simulations are @xmath7 gpa and @xmath8 gpa . yet
the pressure of the ma system is roughly the average of the two limiting sa systems , with a deviation of @xmath9 gpa .
this relation does not change at different temperatures .
the additive behavior supports the idea that the observed slowdown in the ma system ( see below ) is not generated by our choice of densities .
the systems were prepared at high temperatures and cooled down in steps . here
we present data from runs at @xmath10 , which is well below the glass transition .
the alkali subsystem was equilibrated at this temperature for @xmath11 .
subsequently trajectories were generated over @xmath11 for the sa glasses , and over @xmath12 for the ma system .
the determination of cation sites based on cation trajectories is described in detail in @xcite .
mean square displacements of alkali ions in single and mixed alkali systems .
inset : time dependence of @xmath13 in the mixed alkali system .
the broken line at @xmath14 indicates that for this work only information from the second part of our data is taken if not mentioned otherwise .
, width=325 ] simulations below the computer glass transition temperature have to be carefully conducted due to expected aging effects of the network .
furthermore one may expect that due to this effect also the lithium dynamics will somewhat change with time .
after an initial simulation period of @xmath11 we have calculated the mean square displacement during time intervals @xmath15 $ ] where different starting times @xmath16 during our production run of 160 @xmath17 have been chosen ; see inset of fig .
[ fig_diff ] . during the first @xmath18
a significant slowdown is present , which is less visible for the second @xmath19 .
thus one finds a dynamic signature of network aging effects . of course
, subsequent aging effects of the network will still be present , but probably on longer time scales .
the subsequent analysis is mainly performed for the second half of our production run . in any event , all general conclusions drawn in this work
could already be obtained from analyzing just the first half of the production run .
figure [ fig_diff ] shows the mean square displacements of the alkali cations in both the sa and ma glasses . in the long time
limit the ionic transport becomes diffusive , corresponding to a slope of one in the double logarithmic plot . in the sa glasses
this is reached around @xmath20 .
the diffusivities at @xmath21 in the sa systems are @xmath22 and @xmath23 .
the curves for lithium and potassium in the ma system do not attain a slope of one during the time interval shown .
this observation is in agreement with earlier findings in sa glasses , where we noted that the diffusive regime is reached only for larger values of the mean square displacement in systems with lower alkali content @xcite . under the assumption of time - temperature superposition and the ability to reach the linear regime at higher temperatures we may estimate the diffusion constants as @xmath24 and @xmath25 .
the slowdown of the diffusivities as compared to the single alkali systems then amounts to a factor of @xmath26 for lithium and @xmath27 for potassium . by mapping the trajectories of the cations onto the sites obtained from our analysis ,
jumps between the sites can be identified . as in our previous work , we registered a jump if an ion leaves a site and moves into a different one .
the duration of an ion s residence in a site , denoted @xmath28 , is defined by the time between its jump into the site and the subsequent jump out of it .
when an ion leaves a site and returns without reaching a different site , the residence is not interrupted . in total
we find 288 sites .
thus , there are ca
. 8% more sites than ions .
this small excess number is close to the result for the sa system @xcite .
distribution of @xmath29 , the fraction of time a site is occupied by lithium cations .
this is a measure of specificity ( see text ) .
the shaded areas correspond to adapted sites , characterized by specificities @xmath30 .
the bold black lines give the cumulative sum over the number of mixed sites , shifted in vertical direction for convenience .
, width=325 ] > from the residence data it is also possible to calculate the average occupation @xmath31 for each site @xmath32 . in the ma system
we distinguish the contributions @xmath33of the different cations @xmath34 .
the specificity of a site can be quantified by @xmath35 and @xmath36 , the probabilities that an occupying ion is lithium or potassium , respectively .
one can define the specificity @xmath37=max(@xmath38 ) as the larger of both @xmath39 .
the distribution of @xmath29 is shown in fig .
[ fig_siteada ] , for the interval @xmath40 and also for @xmath41 for comparison .
the qualitative result is identical . at both times most sites
have an @xmath42 , signalling a high specificity for one predominant ion .
for the subsequent analysis these sites will be termed _ adapted _ sites or , alternatively , lithium or potassium sites . those with lower specificity shall be called _
mixed _ sites .
the regions corresponding to adapted sites are shaded in fig .
[ fig_siteada ] . also shown
are the cumulative sums over the number of mixed sites , shifted to a convenient vertical position .
they clearly show that the number of mixed sites has decreased at the later time .
especially mixed sites with a @xmath43 already close to 0.9 have vanished , with a corresponding increase in the number of adapted sites . at @xmath40 , 94% ( 270 ) of all sites have a @xmath42 , compared to 82% ( 231 ) at @xmath44 .
25% of all sites have been exclusively populated by lithium cations , and 19% have only been visited by potassium cations between @xmath40 .
the values for @xmath41 are slightly lower again , being 23% and 17% , respectively .
the differences between the time slices illustrate the effects of aging on our data .
the observed dependence of site specificity on simulation time rationalizes the observed time dependence of alkali mobility in the ma system . for later times
the site specificity is more pronounced and thus the ma effect , expressed by a slowing down of the alkali dynamics , increases .
distributions of duration of jumps @xmath45 and of @xmath46 , the time a site stays unoccupied .
top : single alkali systems ; bottom : mixed alkali system . from the horizontal line with value 0.5
the median can be read off .
, width=325 ] distributions of residence times @xmath28 .
top : single alkali systems ; bottom : mixed alkali system .
, width=325 ] in previous work on a sa system we have shown that the time scale it takes for an ion to hop from one equilibrium position to a nearby position is very short ( less than 1 ps for lithium ) @xcite . this time scale is denoted @xmath45 .
furthermore , it turned out that on average a site was filled very soon after it was left by an ion .
this is quantified by @xmath46 .
the low values reflect the fact that the number of free sites is very small .
we have determined the cumulative distribution @xmath47 of both quantities for both the sa and the ma system .
thus , @xmath47 expresses the probability , that a value @xmath48 is found .
the results are shown in fig .
[ fig_times_short ] . for both systems
it turns out that both times belong to the ps - time scale .
of course , the short transition times @xmath45 directly imply that ionic transitions are exclusively between adjacent sites .
there is no long - range ionic motion through the network before entering a new site .
interestingly , in the sa system the potassium dynamics between two sites is nearly a factor of 10 slower than the lithium dynamics .
this difference alone can not be explained only by the mass ratio of potassium and lithium . actually , it becomes smaller in the ma system .
it may come as a surprise that the time @xmath46 during which a site is vacant is not much longer than the transition time @xmath45 between adjacent sites .
this reflects strongly cooperative dynamics , as stressed in previous simulation work @xcite .
interestingly , the distributions are significantly broader for the ma system .
this is a first hint that the dynamics is more heterogeneous in the ma system . in the next step
we calculate for every site its residence time , averaged over all ionic visits of this site . in the ma case
we distinguish whether the visit is in a matched site or a mismatched site .
thus , @xmath49 corresponds to the waiting time of a lithium ion in a potassium site ( characterized by @xmath50 ) . in fig .
[ fig_times ] we show the corresponding cumulative distributions @xmath47 . comparing the data from the sa systems shown in the top part with those of the ma system ( @xmath51 and @xmath52 , respectively ) it turns out that the distribution is significantly broader in the latter case , in particular for long times .
this means that the dynamics is more heterogeneous in the ma system .
actually , from previous work we know that the width of the distribution of residence times in the sa case does not depend on concentration @xcite .
thus , the increasing relevance of dynamic heterogeneities in the ma case is our first observation , for which the dynamics displays qualitatively new features in ma systems as compared to sa systems .
interestingly , the median values of the distributions of residence times increase by a factor of only ca .
three in the ma system .
thus the increase in residence times is not sufficient to explain the total extent of the mae ( factor 10 - 14 , see above ) .
> from fig .
[ fig_diff ] one can see that for the sa system the crossover to diffusive dynamics occurs on time scales somewhat longer than 100 @xmath53 . here
we define crossover such that the extrapolation from the long - time diffusive behavior deviates by a factor of 2 from the actual mean square displacement .
this time scale is close to the median of the waiting times .
thus one may conclude that for the sa system the dynamics of ions from the slower half of the sites is not important for understanding the long - range transport .
in contrast , for the ma case this crossover time can be estimated to be significantly beyond 10 @xmath17 .
this implies that nearly the whole range of residence times is relevant for diffusive transport .
so far we have analyzed the behavior of matched events , i.e. adapted sites hosting the appropriate ion . analyzing the waiting times for mismatched events we find , in particular for potassium , a strong reduction of the waiting times .
this asymmetry in the behavior of lithium and potassium is also mirrored by the lower number of exclusive potassium sites mentioned in the discussion of fig .
[ fig_siteada ] .
qualitatively , this may be related to the fact that it is difficult for a potassium ion to enter a smaller lithium site .
a more quantitative version of this argument will be presented below .
maximum residence times in sites visited by lithium and potassium ions during the simulation .
for each site , the value for potassium is plotted versus the value for lithium . adapted sites and mixed sites are distinguished by their relative occupation.,width=268 ] although residences for mismatched events are on average significantly shorter than the residences found in matching events , the ranges covered by the distributions in fig .
[ fig_times ] are overlapping .
the duration of mismatching events can thus reach time scales normally present in well adapted sites .
this possibility is explored in fig .
[ fig_maxtaukli ] .
for each site that has been visited by both lithium and potassium ions , the maximum residence time of a potassium ion is plotted against the maximum residence time of a lithium ion in the same site .
values for lithium sites , defined as above by @xmath54 , are noted as vertical lines , potassium sites as horizontal lines .
mixed sites are shown as circles .
the few mixed sites allow long maximum residence times for both species of ions , with values of at least 100@xmath53 , but typically reaching several @xmath17 for at least one species .
all the adapted sites show residence times in this range for ions of the favored kind .
the maximum residence time of a mismatched ion can be very short .
but surprisingly , the possible residence times of mismatching events too cover the whole range of values up to comparable maximum time scales for both species , despite the high specificity of the respective sites .
even some strongly adapted sites seem thus to have the potential to accommodate ions of both species for a long time . the mean distance from a site to the closest oxygen during potassium residences ,
i.e. @xmath55 , is plotted versus the value found during lithium residences in the same site , i.e. @xmath56 . the bars give the typical ranges of values for sites visited by only one species .
, width=268 ] as a further step it is instructive to analyze the structural aspects of site specificity .
the oxygen coordination environment of lithium and potassium sites is very different .
one finds 4.5 oxygens up to @xmath57 for a lithium site and 10 at @xmath58 for a potassium site .
the distances correspond to the first minima of the distribution function . a different way to characterize the differences is the observation that around a lithium ion , on average , three coordinating oxygen atoms are closer than 2.3 @xmath59 , while on average all coordinating oxygen atoms are further apart around a potassium ion .
we determined for each site the mean distances @xmath60 between the center of the site and the i - th closest oxygen atom whenever this site was either occupied by a lithium or a potassium ion .
in particular @xmath61 is an extremely sensitive quantity to the local structure of a site . in fig .
[ fig_rox1kli ] @xmath55 is plotted vs. @xmath56 for all sites that have been visited by ions of both species .
again , data for lithium sites , for potassium sites and for mixed sites are distinguished by different symbols .
the ranges of typical values for pure sites which are exclusively occupied by lithium or potassium , respectively , are given by the bars parallel to the respective axis .
the bars cover more than 90 % of the values . in tab .
[ median ] we have for all combinations between ions and sites compiled the intervals for @xmath61 which apply to 68% of all sites .
.[median ] the intervals for the distance of the center of a site to the nearest oxygen where 68% of all cases are included .
only adapted sites are taken into account .
[ cols="^,^,^,^,^ " , ] several interesting conclusions can be drawn about the distance to the next oxygen atom from fig .
[ fig_rox1kli ] and the corresponding data of tab .
[ median ] .
first we discuss the properties of adapted sites .
( 1 ) except for a single site one has @xmath62 .
this is a natural consequence of the different radii of the two ion species .
( 2 ) the median of @xmath56 and even the total distribution is very similar when comparing pure sites with adapted sites , which at least once have hosted a potassium site .
thus the occasional occupation by potassium ions does not change the local structure of a lithium site .
( 3 ) the reverse is not true .
pure potassium sites have on average larger values of @xmath55 than adapted sites which at least once have hosted an lithium ion .
thus lithium ions can only visit those potassium sites with not too large values of @xmath55 .
( 4 ) lithium ions visiting a potassium site significantly attract the nearest oxygen .
there is , however , only a partial adaption since the median of @xmath56 is still 0.20 larger than the median for lithium sites .
( 5 ) in contrast , lithium sites only weakly adapt if entered by potassium ion .
this is plausible from a chemical point of view because it is more difficult to compress the oxygen coordination shell rather than expanding it .
what can we learn from the properties of the mixed sites ? a priori the mixed sites may have two different origins .
first , they may reflect sites which during the first part of the 40ns were either lithium or potassium sites and in the second part changed their identity .
this would give rise to a mixed average occupation .
second , they are never potassium or lithium sites but just have a structure which allows both ions to enter this site . whereas the results in fig .
[ fig_maxtaukli ] do not allow us to distinguish between both scenarios the data in fig .
[ fig_rox1kli ] clearly demonstrates that the second scenario is valid .
if during some time interval the mixed sites would have acted as potassium sites the values of @xmath55 should be in the range of the values for potassium sites .
clearly , this is not the case .
thus a relevant number of site transitions from lithium to potassium type or vice versa can not be seen .
mean distances @xmath63 and @xmath64 from mixed and adapted sites to the fourth oxygen neighbor .
, width=268 ] the omitted corresponding plots for the second and third neighbors show gradually weaker displacements .
the distance to the fourth oxygen neighbor , given in fig .
[ fig_rox4kli ] , finally does not show any effect of a change in the occupying species .
the reaction of the network around a site to the nature of a visiting cation is thus very limited in range .
> from these observations we would particularly like to stress that lithium and potassium sites only differ in the behavior of the first three oxygen atoms and that the properties of lithium sites do not change significantly , even if they are visited by potassium ions .
this directly rationalizes the behavior , reported above , that the residence time of potassium in lithium sites is very short .
probability of direct backjumps a @xmath65 b @xmath65 c plotted against the average residence time for site @xmath66,width=325 ] .
comparision of @xmath67 out of matched and mismatched sites in the mixed alkali system , for both lithium and potassium ions .
the squares give the median of the distribution of one - site values , the triangles delimit the range including 34% of all sites both below and above the median .
, width=325 ] the local cation neighborhood of a cation determines , of course , the properties of nearest neighbor ( nn ) jumps .
an important observable is the coordination number in the nearest neighbor ( nn ) shell , based on the partial pair correlation functions .
basically one counts the number of neighboring ions until the first minimum of the respective partial pair correlation functions . for li - li pairs
we obtain 3.6 for k - k pairs 4.1 and for li - k pairs 3.0 .
this clearly shows that , on the one hand , there is a slight tendency for clustering of like ions and , on the other hand , potassium ions have more neighbors than lithium ions .
the latter point is a direct consequence of the fact that potassium has a larger ionic radius than lithium and thus has a larger nn sphere .
it directly translates into the observation that also the number of nn sites is larger for potassium ( 8.1 ) than for lithium ( 7.1 ) . of direct relevance for the dynamics
are the number of nn sites which are accessible by a hopping process .
it turns out that only ca .
60% and 40% of all nn sites were used as destinations for jumps by lithium and potassium ions , respectively . for the slowest sites ,
not all suitable neighbors may have been explored during the analyzed interval , especially in the case of potassium . but even considering this possibility , a significant fraction of nn sites must remain where the saddles are too large for a hop to occur . in the next step
one may analyze the nature of those neighbors actually accessed during the simulation .
it turns out that 70% of them are matching for the jumping ion .
this value is significantly larger than the overall fraction of matching sites ( ca .
47% for both species ; see above ) .
matching sites are thus clearly favored as jump destinations . beyond the residence times the long - time dynamics
is strongly influenced by the presence of correlated forward backward jumps . to quantify this effect , we consider sequences of two jumps a @xmath65 b @xmath65 c performed by an ion .
we define @xmath67 as , the probability that an ion in a site @xmath66 jumps back into the same site @xmath68 from where it reached @xmath66 . in fig .
[ fig_pback ] , @xmath67 is plotted against the average residence time of an ion in site @xmath66 , @xmath69 . in all systems ,
high probabilities are found for backjumps through sites @xmath66 with short residence times , with values approaching @xmath70 .
@xmath67 falls off for longer residences , down to @xmath71 and @xmath72 respectively for the lithium and potassium sa systems .
these values are still high compared to the statistical probability of @xmath73 corresponding to the average number of seven available neighbors in both cases . in the ma system , the curves level off to even higher values of @xmath67 for long times , @xmath74 and @xmath75 for lithium and potassium . generally , the probability of backjumps for any given residence time is higher than in the sa systems .
the difference between the sa and the ma systems directly translates into the longer range of subdiffusive behavior ( see above ) and thus to a further decrease of the diffusion constant for the ma systems .
this means that beyond the ( minor ) effect of longer waiting times it is to a large extent the increase of forward
backward correlations which gives rise to the slowdown in the ma system .
for a closer understanding of the backjump characteristics we have studied @xmath67 for different sites . from fig .
[ fig_pback ] it is evident that short residence times are generally related to high backjump probabilities .
this expectation is directly checked in fig . [ fig_pranges ] .
therein the observed ranges for @xmath67 out of lithium and potassium sites are opposed , for both species of ions . for potassium ,
the expected strong rise in @xmath67 is found if @xmath66 is a mismatched site .
but the backjump probability of lithium ions is only slightly affected by the nature of the site @xmath66 .
this underlines our conclusions about the hopping properties of potassium ions . whenever a potassium ion enters a mismatched site it can hardly enter because it is too difficult to create the necessary volume and immediately jumps back to the original site .
relative number of jumps from the neighboring sites a time @xmath76 from a jump out of a given central site . in brackets the ion jumping at the center , in front the species jumping in response .
the distribution is the average over those for any given central site . , width=325 ] schematic of the type of correlated jump processes , considered in our analysis where four distinct sites as start and end points are required .
the central site is shaded.,width=151 ] cumulative amount of additional jumps correlated with a jump at @xmath77 , for the mixed alkali system .
the importance of correlated jumps between four distinct sites is compared to the total extent of correlations including sequential movement as shown in fig .
[ fig_correl ] .
, width=325 ] so far we have analyzed the dynamic properties on a _ single - particle _ level . as discussed in the introduction an influence of _ multi - particle _ correlations
is discussed in literature as a possible mechanism to enable dynamic processes in ma systems .
for example habasaki demonstrated an additional slowdown of the lithium dynamics , if the potassium ions in a ma system were completely fixed @xcite .
the inferred correlations between jumps of different ions should be directly observable as an altered probability for a second jump shortly before or after one jump has happened .
for each site , we have determined the number of jumps originating from one of its nn sites in dependence on the time difference @xmath76 to a jump from the central site .
jumps by the same ion were ignored .
the resulting distribution of jump probabilities was normalized and averaged over all sites .
the value of one corresponds to the number of jumps from the neighbors happening without the influence of a jump from the central site , i.e. the behavior for long @xmath78 . to gain an estimate of the scale of fluctuations in the results ,
this analysis was done separately for four subsequent pieces of the simulation of @xmath79 duration each .
the distributions shown in fig .
[ fig_correl ] are the average over these four partial distributions .
the error bars correspond to the standard deviation of the peak heights .
the peaks show an increase in the number of nearby jumps by a factor of @xmath80 and @xmath81 for the lithium and potassium sa systems , respectively . for the correlations among like ions in the ma system , a factor of @xmath82
is found for both lithium and potassium .
correlations among jumps by unlike ions are weaker , but also present .
the values are @xmath83 for the lithium neighbors of a potassium ion , and also for potassium ions surrounding a lithium . in all cases , the maximum probability for correlated jumps
is found at the time of the first jump , @xmath84 , and the peak is symmetrical .
careful inspection reveals that the peak consists of a narrow peak with half width close to 1 ps and a broad peak with with half width close to 5 ps .
the latter peak will be discussed below .
the most simple scenario compatible with these findings is that a second ion jumps from a neighbor site into the site just vacated by the jump of the first ion .
indeed these correlations are reflected by the sharp peaks . but more complex correlations are also conceivable , which can generally be separated from the former kind by the additional condition that the two jumps use four distinct sites as start and end points ; see fig .
[ scheme ] for a sketch .
the remaining correlations are responsible for the broad peaks in fig .
[ fig_correl ] . for a quantitative analysis
the maximum height is not suitable .
rather we show the integrals @xmath85 over the peaks ; see fig .
[ fig_intcorr ] . on the one hand
, we integrate over the total peaks and , on the other hand , only over the remaining correlations if we use the additional four - site condition introduced above .
the total extent of four site correlations given by the thicker lines is significant in all cases , amounting to ca .
50% of the total .
to characterize the observed correlations , we determined the relative directions of the central jump and the correlated ones , given by the cosine of the enclosed angle .
the directions of jumps using a common site are strongly correlated in all cases , as it is geometrically required .
the jumps involved in four site correlations are instead preferentially antiparallel in the sa systems . in the ma system ,
no correlation is discernible for the four site correlations .
interestingly , the curve labelled li(k ) , i.e. the probability of a lithium jump to follow a potassium jump , displays an asymmetry since it mainly increases for @xmath86 .
accordingly , the curve k(li)shows the opposite asymmetry .
both observations imply that it is more likely that a jump of a potassium ion may trigger a jump of a lithium ion than vice versa .
this observation is consistent our previous results that potassium sites are better suited for lithium ions than vice versa . a more detailed analysis
, however , is necessary to further clarify the origin of this asymmetry . in this analysis of correlations ,
all sites are taken into account with equal weight . when the sites are weighted according to the number of jumps taking place , the general effects remain , but their strength is greatly reduced .
one can thus conclude that dynamic correlation effects are particularly important in the vicinity of slow sites .
the results favor an explanation of the mixed alkali effect based on ionic sites that are specific for the different alkali species .
the strongly bimodal distribution of occupation probabilities @xmath39 shows that @xmath70 of the ions reside in sites with a strong preference for one kind of ion .
we have shown before that the location of the ionic sites is stable on the time scale of ionic transport @xcite .
it appears thus justified to treat the preference of a site for one species of ions as fixed over the course of the simulations .
possible effects of rare readaptions would still be incorporated via the treatment of `` mixed '' sites . trying to quantify the mixed
alkali effect in terms of properties of single sites or ions , the most direct contribution is observed in the distribution of residence times .
the typical duration of residences in the predominant specific sites in the ma system is increased compared to the sa glasses . but
this effect can account for only a small part of the observed mixed alkali effect .
the increased probability for correlated forward
backward jumps will also add to the slowdown .
as already in the sa systems the effect is strongest for sites with low residence times . among them
will be mismatching jump events , which generally have shorter @xmath28 than matching ones .
but only for potassium ions this tendency translates into a further increase of @xmath67 .
the backjump probability of lithium is hardly affected by the adaption of the @xmath66 site .
the maximum residence times and the oxygen coordination distances show that even strongly adapted sites can temporarily accommodate mismatching ions .
lithium again is favored , because it is apparently easier for the coordinating oxygens to approach than it is to recede from a site .
but generally , both types of ions also visit mismatching sites , suggesting also some degree of interaction between the different species .
the data on two ion correlations shows in fact that the dynamics of the different species in the mixed glass are not independent .
the strongest correlations were observed for ions taking up sites just vacated by like ions , but foreign sites are also reached .
potassium ions will jump back with increased probability , but lithium ions are as likely to move on into a third place as from a matching site . for them , the correlations with the different ions should thus in fact facilitate transport . as mentioned above the main difference between lithium and potassium
is the ability of lithium to enter a potassium site without an immediate backjump , helped by the local oxygen environment .
the reverse is strongly suppressed .
this asymmetry is consistent with the larger mismatch evident in the residence times for potassium . for the present case of identical lithium and potassium concentration
the transport is mainly via the matched sites for both species . in the limit
where one species is very dilute the relevance of transport via mixed or mismatched sites may become important and may be relevant in particular for the smaller ion
. it may be interesting to compare the `` first - principle '' observations of this simulation work with the assumptions , used in the most recent version of the dynamic structure model @xcite for the ma effect in solid glasses .
most ingredients like the presence of two different types of sites or the asymmetry between the smaller and the larger cation fully agree with our observations .
there are , however , two aspects which are somewhat more speculative and where the present simulations yield new insight .
do sites change their character on time scales much shorter than the @xmath87 relaxation time ?
do so - called c@xmath88 sites exist ? they are viewed as sites which after being vacated in the melt are somewhat smaller than the sites , actually visited by ions . from the simulations we have seen that on time scales up to the diffusive regime sites
keep their identity so that the notion of a dynamic energy landscape is not supported by the simulation data .
furthermore there is no signature of possible c@xmath88 sites
. they would be relevant if on the diffusion time scale transformations from c@xmath88 sites to regular sites and vice versa would occur .
this would show up in a large number of sites , visited by the ions during the simulation run .
since the total number of sites , however , is just 8% larger than the number of ions there is basically no place for a large number of c@xmath88 sites being transformed to regular sites and vice versa . in summary , a ma slowdown can be observed in the residence times as well as from increased likelihood of backjumps .
independent jump paths are favored .
but the interception is not complete , as even mismatching sites are also accessible .
although actual cooperativity seems limited to the slowest ions , an enhancement of cation mobility is likely to follow .
together with the fast network dynamics in ma systems this rationalizes the enhanced dynamics in ma systems as compared to the predicted dynamics by assuming a dilution effect .
furthermore a significant asymmetry in the behavior of the smaller and the larger cation is observed .
it will be interesting to see , whether an even closer analysis of the ma system , e.g. in terms of site and saddle energies , will be compatible with very recent explicit models of the dynamics in ma systems @xcite . furthermore a closer relation to ion - exchange experiments
might be illuminating .
h. l. acknowledges funding by the german bmbf through the fonds der chemischen industrie .
this work was also supported by the dfg ( sfb 458 ) .
furthermore we would like to thank a. bunde , j. horbach , m.d .
ingram , p. maass , and h. mehrer for stimulating discussions about this work . | the mixed alkali effect on the cation dynamics in silicate glasses is analyzed via molecular dynamics simulations .
observations suggest a description of the dynamics in terms of stable sites mostly specific to one ionic species . as main contributions to the mixed
alkali slowdown longer residence times and an increased probability of correlated backjumps are identified .
the slowdown is related to the limited accessibility of foreign sites .
the mismatch experienced in a foreign site is stronger and more retarding for the larger ions , the smaller ions can be temporarily accommodated . also correlations between unlike as well as like cations are demonstrated that support cooperative behavior . |
+ the combinatorial editing problem is , in general , the problem of finding the smallest number of element - changes such that the resulting combinatorial object satisfies a certain fixed property . the simplest class of objects for which the editing problem was considered is a set of sequences . in fact ,
the first detailed algorithmic study of editing was motivated by bioinformatics , where sequences over finite alphabets are considered and editing corresponds to changes of the elements in the sequence depicting the mutations in biomolecules . when the desired property consists of a single sequence , studying editing corresponds to investigating the hamming distance between sequences .
the notion of graph editing was introduced by the authors and kzdy @xcite and independently by alon and stav @xcite .
the question considered was : `` how many edges does one need to add or delete in a given graph , such that the result belongs to a given class of graphs ? ''
the authors showed in @xcite , that the answer to this question for hereditary classes could be expressed in terms of the so - called binary chromatic number ( also called the colouring number ) of the family .
alon and stav @xcite showed that the largest distance from a hereditary property is achieved , asymptotically , by an erds - rnyi random graph . in this paper ,
the generalized theory is developed for editing of edge - colored complete graphs and digraphs .
the main result for edge - colored graphs , theorem [ thm : multicol : easybounds ] , is in terms of two parameters : the so - called weak and strong @xmath0-ary chromatic numbers .
the main result for directed graphs , theorem [ thm : digraphs : easybounds ] is in terms of two parameters : the weak and strong directed chromatic numbers . in each case , the results come from more general theorems , theorems [ thm : multicol : basics ] and [ thm : digraphs : basics ] respectively , which deal with generalizing the graph notion of types for the above combinatorial objects .
the analysis is based on using a version of szemerdi s regularity lemma , which we state as theorem [ thm : multicol : genreglem ] ( see @xcite for a proof of theorem [ thm : multicol : genreglem ] ) , and applying it to an erds - rnyi - type random edge - colored graph or random digraph , respectively .
general bounds on the edit distance function are given , as well as some editing algorithms and computing methods , all of which result from theorems [ thm : multicol : basics ] and [ thm : digraphs : basics ] .
the paper is structured as follows .
section [ sec : multicol ] deals with the case of multicolorings of the edges of complete graphs .
section [ sec : digraphs ] deals with the case of directed graphs . in each of these sections
we provide definitions , editing algorithms , examples as well as some general theory on the edit distance function .
most proofs are presented in subsection [ sec : multicol : proofs ] and in subsection [ sec : digraphs : proofs ] .
+ [ sec : multicol ] + an equipartition of a finite set is a partition in which each pair of partite sets differ in size by at most one . for a complete graph on vertex set @xmath1 , and a finite set @xmath2
, we shall say that a @xmath2-coloring , or more specifically , a @xmath2-edge - coloring of this graph is a pair @xmath3 , where @xmath4 . since it is sufficient to let @xmath5 for some integer @xmath0 , we will refer to an @xmath6-edge coloring of a complete graph as simply an @xmath0-graph . for any @xmath0-graph @xmath7 ,
disjoint vertex sets @xmath8 and @xmath9 and color @xmath10 , @xmath11 , the expression @xmath12 denotes the set of edges colored @xmath10 with both endpoints in @xmath13 $ ] and @xmath14 denotes the set of edges colored @xmath10 with one endpoint in @xmath8 and the other in @xmath9 .
the density vector of @xmath8 is an @xmath0-dimensional vector @xmath15 , where @xmath16 for @xmath17 .
the density vector of the pair @xmath18 is an @xmath0-dimensional vector @xmath15 , where @xmath19 for @xmath17 .
note that for such density vectors , @xmath20 . in this
setting , a graph property is merely a set of @xmath0-graphs for some positive integer @xmath21 . if @xmath3 and @xmath22 are @xmath0-graphs on @xmath23 labeled vertices , then @xmath24 is the proportion of edges on which the colors differ , i.e. , the number of edges on which the colors in @xmath7 and @xmath25 differ , divided by @xmath26 .
we may call this the normalized edit distance between @xmath7 and @xmath25 . for any property @xmath27 , a coloring @xmath7 , an integer @xmath23 , we define @xmath28 , @xmath29 , and @xmath30 as follows : @xmath31 note that @xmath32 $ ] .
the last parameter @xmath30 is the limit of the largest proportion of the edges necessary to be changed in a coloring of a complete graph bring it to a property @xmath27 ; we show the existence of this limit later . a hereditary property of @xmath0-graphs ( or , simply , hereditary property , where the context is understood ) is a set of @xmath0-graphs that is closed under vertex - deletion and isomorphisms .
let an @xmath0-graph @xmath25 be an induced coloring of an @xmath0-graph @xmath7 if @xmath25 can be obtained from @xmath7 by vertex - deletion . for an @xmath0-graph , @xmath33 , the family @xmath34 consists of all @xmath0-graphs that have no ( induced ) copies of @xmath33 . for every hereditary property , @xmath27 , there is a family , @xmath35 , of @xmath0-graphs such that @xmath36 .
if @xmath37 is a family of @xmath0-graphs , then we use @xmath38 to denote @xmath39 .
+ + for a hereditary property @xmath40 of @xmath0-graphs , a weakly - good tuple @xmath41 is an @xmath0-tuple of non - negative integers such that for some @xmath42 , the vertex set @xmath43 can be partitioned into sets @xmath44 such that , for each @xmath45 with @xmath46 , the partition can be further refined @xmath47 such that each @xmath48 * does not * induce an edge of color @xmath49 .
the weak clique spectrum of @xmath27 is the set of all tuples @xmath50 that are not weakly - good .
the weak @xmath0-ary chromatic number of @xmath27 , @xmath51 , is the maximum @xmath52 such that for some non - negative integers @xmath53 with @xmath54 , the tuple @xmath50 is in the weak clique spectrum of @xmath27 .
+ for a hereditary property @xmath27 , a strongly - good tuple @xmath41 is an @xmath0-tuple of non - negative integers such that for some @xmath42 , the vertex set @xmath43 can be partitioned into sets @xmath44 such that , for each @xmath45 with @xmath46 , the partition can be further refined @xmath47 such that each @xmath48 has * all * edges of color @xmath49 . the strong clique spectrum of @xmath27 is the set of all tuples @xmath50 that are not strongly - good .
the strong @xmath0-ary chromatic number of @xmath27 , @xmath55 , is the maximum @xmath52 such that for some non - negative integers @xmath53 with @xmath54 , the tuple @xmath50 is in the weak clique spectrum of @xmath27 .
if @xmath56 , then we denote @xmath57 and @xmath58 .
+ + * the weak [ strong ] clique spectrum is a downset in the partially ordered set of @xmath0-tuples ordered coordinatewise .
that is , if @xmath50 is in the weak [ strong ] clique spectrum and @xmath59 has the property that @xmath60 for @xmath61 , then @xmath59 is also in that weak [ strong ] clique spectrum . * informally , we can partition @xmath43 into @xmath62 pieces in which the absent color in each piece is arbitrary , but there is some specification of absent colors for which a @xmath63 piece partition is not possible . * similarly , we can partition @xmath43 into @xmath64 pieces in which the required color in each piece is arbitrary , but there is some specification of required colors for which a @xmath65 piece partition is not possible . * for any @xmath21 and any hereditary property of @xmath0-graphs , @xmath27 , @xmath66 . * in the case of @xmath67 , notions of strong and weak colorings are identical . further , if @xmath56 , @xmath68 corresponds exactly to the binary chromatic number of @xmath33 , introduced in @xcite .
this is also called the `` colouring number '' in related literature such as bollobs and thomason @xcite .
+ + * ( 1 ) * let @xmath69 and @xmath27 be a family of @xmath70-colored complete graphs not containing a triangle @xmath71 with colors @xmath72 on its edges and not containing a triangle @xmath73 with colors @xmath74 on its edges .
so , @xmath75 .
first , the weak @xmath76-ary chromatic number : since @xmath69 , and @xmath35 contains a triangle , any @xmath76-tuple @xmath77 with @xmath78 must be weakly - good . indeed , each of @xmath71 and @xmath73 can be vertex - partitioned into three parts such that each part is a single vertex , thus not inducing edges of any colors .
thus , it is sufficient to consider the tuples with @xmath79 .
the tuple @xmath80 is weakly - good since we can partition the vertex set of @xmath73 in one part not containing edges of color @xmath81 .
similarly , @xmath82 is good . by monotonicity ,
all tuples @xmath83 with @xmath84 or @xmath85 are weakly - good .
however @xmath86 is not weakly - good because both @xmath71 and @xmath73 contain edges of color @xmath87 .
but @xmath88 is weakly - good because @xmath71 can be vertex - partitioned in two parts not containing edges of color @xmath87 .
thus , the weak clique spectrum of @xmath27 is @xmath89 . for the weak @xmath76-ary chromatic number , @xmath90 .
second , the strong @xmath76-ary chromatic number .
similar to the above , if @xmath91 , then either @xmath71 or @xmath73 can be partitioned into two parts such that one part is an edge of a specified color and the other part is a vertex . if @xmath92 , then neither @xmath71 nor @xmath73 can be partitioned into a single monochromatic clique .
thus , the the strong clique spectrum of @xmath27 is @xmath93 and for the strong @xmath76-ary chromatic number , @xmath94 also .
+ * ( 2 ) * let @xmath69 and @xmath27 be a family of @xmath70-colored complete graphs not containing a triangle @xmath71 with colors @xmath72 on its edges .
so , @xmath95 . if we follow the previous example , it is easy to see that @xmath96 is weekly good for all @xmath97 .
moreover , @xmath98 is weakly good as long as @xmath99 .
hence , the weak clique spectrum of @xmath27 is @xmath100 and @xmath101 . for the strong clique spectrum , it is easy to see that @xmath102 is in that spectrum , but if @xmath78
, then @xmath77 is strongly - good .
thus , @xmath103 .
+ * ( 3 ) * let @xmath67 , which we can consider to be the graph case .
as we have observed , we may disregard the notions of `` weak '' and `` strong '' in our terminology .
let @xmath33 be a @xmath104 colored with edges colored with colors @xmath81 and @xmath87 such that each color class is a @xmath105-cycle .
let @xmath27 be a family of colorings not containing @xmath33 , i.e. , @xmath106 .
we need only consider @xmath87-tuples ( i.e. , pairs ) @xmath107 with @xmath108 .
it is relatively easy to see that @xmath109 , @xmath110 , @xmath111 and @xmath112 are good .
the pairs @xmath113 , @xmath114 and @xmath115 are not good since @xmath33 has no monochromatic clique on more than 2 vertices , but has a total of 5 vertices . by monotonicity
, @xmath116 and @xmath117 are also not good .
thus the clique spectrum of @xmath27 is @xmath118 , and @xmath119 .
+ + [ sec : multicol : simplealg ] let @xmath27 be a hereditary property of @xmath0-graphs , such that @xmath120 .
further , let @xmath121 and @xmath50 be in the weak clique spectrum with @xmath122 .
partition @xmath1 into @xmath0 sets @xmath44 and further refine the partition such that @xmath47 , for @xmath61 and then recolor the edges in each @xmath123 by recoloring the edges of color @xmath49 with some other arbitrary color .
this new coloring does not contain any @xmath124 , otherwise the tuple @xmath50 would be good for some @xmath33 .
if the sizes of the @xmath123-s differ by at most one ; i.e. , @xmath125 , then the number of changes provided by this algorithm is at most @xmath126 .
thus , @xmath127 + + in @xcite , the authors and kzdy provide a general bound for @xmath128 in the 2-color case .
[ thm : akm ] for any hereditary property of graphs , @xmath27 , with binary chromatic number @xmath129 , @xmath130 furthermore , if @xmath56 such that @xmath33 is self - complementary , then @xmath131 . here
, we show a similar result in the general case .
[ thm : multicol : easybounds ] let @xmath27 be a hereditary property of @xmath6-edge - colorings of complete graphs .
let @xmath132 and @xmath133 be the weak and strong ( respectively ) @xmath0-ary chromatic numbers of @xmath27 .
then , @xmath134 furthermore , if @xmath56 such that all color classes of @xmath33 are isomorphic , then @xmath135 .
we prove theorem [ thm : multicol : easybounds ] in section [ sec : multicol : proofs ] .
the upper bound is found in the simple editing algorithm , but to get the lower bound , we need a more general theory .
this is theorem [ thm : multicol : basics ] which is stated in section [ sec : multicol : algorithm ] .
we also prove the result for symmetric colorings in corollary [ cor : multicol : symm ] .
theorem [ thm : multicol : basics ] gives the basic results that deal with computing the edit distance for given hereditary properties .
to state these results , we need to provide some preliminary material . + + for an @xmath0-graph , @xmath3 , and some color @xmath11 , let @xmath136 denote the graph on vertex set @xmath1 corresponding to the edges with color @xmath10 in @xmath137 . for a positive integer @xmath0 , recall that a density vector @xmath15 ( we also refer to it as a probability vector ) is a nonnegative real vector with the property that @xmath138 . for any density vector @xmath15 , and integer @xmath23 , we denote for @xmath139 and the size of @xmath140 is what remains . since we fix @xmath141 for @xmath17 and let @xmath23 approach infinity , this will make no appreciable difference . ]
@xmath142 in theorem [ thm : multicol : basics ] , we show that the following limit exits , which we call the edit distance function : @xmath143 having the edit distance function at our disposal , we may also define @xmath144 , where the maximum is taken over all density vectors . in section [ sec : multicol : algorithm - analysis ] , we define two functions which are described in terms of types of colorings , which allow us to compute the edit distance function . in section [ sec : multicol : algorithm ] , we shall provide an algorithm to do such computation .
we define a notion which was called a colored regularity graph ( crg ) by alon and stav @xcite , but earlier called a type by bollobs and thomason @xcite .
we adopt latter terminology .
an @xmath0-type ( or just , type , where the context is understood ) , @xmath145 , is a pair @xmath146 , where @xmath147 is a finite set of vertices and @xmath148 , such that @xmath149 and @xmath150 , for all @xmath151 .
informally , we will view an @xmath0-type as a complete graph with a coloring of both vertices and edges using nonempty subsets of @xmath6 , where the whole set is a forbidden color on the vertices .
the sub-@xmath0-type of @xmath145 induced by @xmath152 is the @xmath0-type achieved by deleting the vertices @xmath153 from @xmath145 .
we say that an @xmath0-graph @xmath154 embeds in type @xmath155 if there is a map @xmath156 such that @xmath157 implies @xmath158 . in other words , there is a mapping @xmath159 that brings each edge of color @xmath160 to a vertex or an edge containing @xmath160 in its color set . if @xmath33 embeds in type @xmath145
, we write @xmath161 , otherwise we write @xmath162 . for every hereditary property @xmath27 ,
we let @xmath163 be the set of all @xmath0-types such that none of @xmath35 embeds in that type , i.e. , @xmath164 we say that an @xmath0-graph @xmath165 has type @xmath155 if @xmath25 embeds into @xmath145 with mapping @xmath166 and @xmath159 is surjective .
fact [ fact : multicol : embed ] generalizes the ideas underlying the simple editing algorithm in section [ sec : multicol : simplealg ] .
[ fact : multicol : embed ] if @xmath145 is an @xmath0-type , @xmath25 is of type @xmath145 and @xmath33 does not embed into @xmath145 , then @xmath167 .
+ + [ sec : multicol : algorithm ] let @xmath168 and @xmath169 be density vectors ; i.e. , their entries are nonnegative and sum to 1 .
they play different roles , however .
the vector @xmath170 will represent a vector of densities , @xmath141 .
that is , the graph @xmath7 has @xmath171 edges of color @xmath10 .
the vector @xmath172 will represent a vector of weights , @xmath173 , assigned to the vertices of an @xmath0-type with vertices @xmath174 , respectively .
let @xmath3 be an @xmath0-graph with edges having densities according to the vector @xmath15 , and @xmath27 be a hereditary property .
in order to find an upper bound on @xmath28 , it is sufficient to change @xmath7 to an @xmath0-graph , @xmath25 , such that , for all @xmath124 , @xmath33 does not embed into the new coloring . in particular , if the resulting coloring has type @xmath175 , then @xmath25 is in @xmath27 .
[ algor : multicol : recolor ] fix a @xmath176 and bring @xmath7 to a coloring of type @xmath145 by edge - recoloring .
let @xmath177 .
partition the vertices of @xmath7 randomly into sets @xmath178 such that the probability of a vertex to be in a part @xmath8 is @xmath179 .
consider an edge @xmath180 of @xmath7 , let @xmath181 , @xmath182 , for @xmath183 . if @xmath184 , recolor @xmath180 with a color from @xmath185 .
this gives the new @xmath0-graph @xmath25 which , according to fact [ fact : multicol : embed ] , does not admit an embedding of any @xmath124 , thus @xmath186 .
note that this generalizes the simple algorithm in section [ sec : multicol : simplealg ] . in that algorithm ,
the type had restricted colorings only on the vertices ( possibly of different colors ) but each edge receives the color @xmath187 . consider algorithm [ algor : multicol : recolor ] applied with type @xmath145 .
let @xmath7 be a graph such that the number of edges of color @xmath10 are @xmath188 for @xmath17 .
the expected number of changes is @xmath189 & = & \binom{n}{2}-\sum_{x , y\in v,~x\neq y } \pr(\{x , y\}\mbox { is not changed } ) \\ & = & \binom{n}{2}-\sum_{x , y\in v,~x\neq y}\sum _
{ 1\leq i , j \leq k}\pr(x\in v_i , y\in v_j ) { \bf 1}_{c(\{x , y\})\in\phi(u_i , u_j ) } \\ & = & \binom{n}{2}-\sum _ { 1\leq i , j\leq k}w_iw_j \sum_{x , y\in v , ~x\neq y } { \bf 1}_{c(\{x , y\})\in\phi(u_i , u_j ) } \\ & = & \binom{n}{2}-\sum _ { 1\leq i , j\leq
k}w_iw_j \sum_{\rho\in\phi(u_i , u_j ) } p_\rho\binom{n}{2 } \\\end{aligned}\ ] ] let @xmath190 be a @xmath191 matrix such that the @xmath192-th entry , @xmath193 , is @xmath194 .
thus , if @xmath195 , then @xmath196={\bf w}^t{{\bf m}}_k({{\bf p}}){\bf w}\binom{n}{2 } .\ ] ] finally , we define two functions in terms of the matrix @xmath190 : @xmath197 the @xmath198 and @xmath199 functions can be interpreted as follows : if the vertices of an @xmath0-graph , @xmath7 , are assigned randomly to parts corresponding to the vertices of @xmath145 , then @xmath200 and @xmath201 represent the expectation of the proportion of times that the color of an edge does not map the set of colors in a corresponding vertex or an edge of @xmath145 .
the function @xmath202 is obtained from the uniform distribution , and @xmath203 is obtained using the optimal distribution @xmath204 of the proportion of sizes of parts .
although the @xmath199 function provides a better bound for @xmath205 , the linearity of the @xmath198 function helps prove results from @xmath205 .
+ + theorem [ thm : multicol : basics ] summarizes some facts about the edit distance function that generalize easily from results in both and @xcite or @xcite .
the proof is in section [ sec : multicol : basics ] .
fix a density vector @xmath15 .
formally , the random @xmath0-graph of density @xmath170 , or random @xmath0-graph where the context is clear , is denoted @xmath206 .
it is a random variable that is an @xmath6-coloring of the edges of a labeled @xmath207 in which each edge , @xmath208 , is colored independently such that @xmath208 receives color @xmath10 with probability @xmath188 .
[ thm : multicol : basics ] let @xmath27 be a hereditary property of @xmath0-graphs . fix an @xmath0-dimensional density vector @xmath170 .
then the limit @xmath209 exists .
moreover , 1 . @xmath210 ; [ it : multicol : inf ] 2 . for a fixed @xmath211 , then with probability approaching 1 as @xmath212 , @xmath213 3 . @xmath214 $ ] ; 4 . @xmath215 is continuous over the domain of @xmath0-dimensional density vectors and is concave down ; being concave down means for every pair of density vectors @xmath216 and every real number @xmath217 $ ] , @xmath218 is a density vector and @xmath219.][it : multicol : concon ] 5 .
@xmath205 achieves its maximum , @xmath128 , at some density vector @xmath220 ( in fact , denote the set of all such vectors @xmath221 ) and so , @xmath222 \mbox { ; and}\ ] ] 6 .
both @xmath223 and @xmath30 exist and @xmath223 is a convex and closed set in @xmath224^{r-1}$ ] .
note that @xmath221 typically consists of a single vector , but we abuse notation by denoting the set of such vectors as @xmath221 when the vector at which the maximum is obtained is not unique .
[ cor : multicol : symm ] let @xmath27 be a symmetric hereditary property ; that is , one that has the property such that if the @xmath0-tuple @xmath50 is in the weak clique spectrum of @xmath27 , then for any permutation @xmath225 of @xmath6 , the @xmath0-tuple @xmath226 is also in the weak clique spectrum . then , @xmath227 in particular , if @xmath56 such that all color classes of @xmath33 are isomorphic , then @xmath135 .
corollary [ cor : multicol : symm ] consider an arbitrary density vector @xmath15 and without loss of generality assume that @xmath228 . choose a permutation of the @xmath229-s such that @xmath230 .
let @xmath155 be a @xmath0-type on @xmath231 vertices such that @xmath232 if @xmath233 and there are exactly @xmath234 vertices @xmath235 such that @xmath236 .
the off - diagonal entries of @xmath190 are zero and so it is easy to see that @xmath237 .
we can use a correlation inequality such as fkg @xcite to see that @xmath238 to finish the proof observe that , in the case of @xmath56 , @xmath239 .
+ theorem [ thm : multicol : triangles ] gives some basic results on examples of hereditary properties of @xmath0-graphs defined by triangles .
the proof is in section [ sec : multicol : triangles ] .
[ thm : multicol : triangles ] let @xmath69 and consider hereditary properties of @xmath0-graphs . 1 . if @xmath37 is a family of that consists of a single monochromatic triangle , then @xmath240.[it : multicol : triangles:1 ] 2 . if @xmath37 is a family that consists of a single triangle with two edges colored @xmath81 and one edge colored @xmath87 , then @xmath240.[it : multicol : triangles:2 ] 3 . if @xmath37 is a family that consists of two monochromatic triangles of different colors , then @xmath240.[it : multicol : triangles:3 ] 4 . if @xmath37 is a family that consists of all six bi - chromatic triangles , then @xmath241.[it : multicol : triangles:4 ] 5 . if @xmath37 is a family that consists of a single rainbow triangle , then @xmath242.[it : multicol : triangles:5 ] + + [ sec : multicol : proofs ] the upper bound for this theorem is proven by the simple editing algorithm from section [ sec : multicol : simplealg ] . for the lower bound
, we apply part ( [ it : multicol : inf ] ) of theorem [ thm : multicol : basics ] , which states that @xmath243 .
consider an arbitrary @xmath244 , an @xmath0-type on @xmath245 vertices .
let @xmath246 be an auxiliary graph with vertex set @xmath147 such that @xmath235 and @xmath247 are adjacent in @xmath246 if and only if @xmath248 .
we observe that @xmath246 has no clique on @xmath249 vertices , otherwise for some @xmath124 , @xmath250 .
using turn s theorem , the number of edges of @xmath246 is at most @xmath251 .
let @xmath252 .
consider the matrix @xmath253 and observe that every entry is either zero or is a positive integer multiple of @xmath254 .
the zero entries correspond exactly to pairs with @xmath255 value equal to @xmath6 .
thus , this matrix @xmath253 has at least @xmath256 entries with value at least @xmath254 .
therefore , @xmath257 is at least @xmath258 .
since @xmath145 was arbitrary , this gives a lower bound for @xmath205 .
+ + [ sec : multicol : basics ] let @xmath259 and let @xmath260 . + * a : * upper bound on @xmath205 .
+ let @xmath7 be an @xmath0-graph with the density of its @xmath49-th color class be @xmath141 for @xmath17 .
let @xmath175 .
apply the editing algorithm in section [ sec : multicol : algorithm ] to @xmath7 using @xmath145 .
the analysis of the algorithm in section [ sec : multicol : algorithm - analysis ] gives that the expected number of changes is @xmath261 and so @xmath262 .
+ * b : * equality of @xmath198 and @xmath199 .
+ by the definition of @xmath201 , it is easy to see that @xmath263 for every density vector @xmath170 .
therefore , @xmath264 . for the other direction
, we will use @xmath145 and its optimal weight vector @xmath265 , where @xmath179 corresponds to @xmath266 in order to construct a sequence of crgs , @xmath267 such that @xmath268 .
first , choose @xmath269 large enough to ensure that @xmath270 for @xmath271 .
then , for each vertex @xmath272 , create @xmath273 copies of @xmath274 in the following sense : let @xmath275 and @xmath276 be copies of @xmath274 and @xmath277 , respectively , where @xmath278
. let @xmath255 be the coloring function of @xmath145 and @xmath279 be the coloring function of @xmath280 .
if @xmath233 , then @xmath281 . if @xmath282 and @xmath283 , then @xmath284 . finally , @xmath285 .
the @xmath192-th block is a @xmath286 matrix and each entry of the @xmath192-th block is the same as the @xmath192-th entry of @xmath253 .
if we denote the @xmath192-th entry of @xmath253 by @xmath287 , then @xmath288 taking @xmath289 , we see that @xmath290 .
consequently , for any @xmath291 , @xmath292 take the infimum over all @xmath291 , and we have that @xmath293 .
+ * c : * lower bound on @xmath205 using the random graph . + we apply theorem [ thm : multicol : reglemapp ] , which is given in @xcite .
theorem [ thm : multicol : reglemapp ] is a corollary of theorem [ thm : multicol : genreglem ] , a relatively straightforward generalization to @xmath0-graphs and digraphs of a theorem by alon , fischer , krivelevich and m. szegedy @xcite , which is suitable for induced graphs . in an @xmath0-graph ,
the density vector of a pair of disjoint sets of vertices @xmath18 is simply @xmath294 .
so we can state the general version of the regularity lemma .
for all definitions of regularity , see @xcite .
[ thm : multicol : genreglem ] fix @xmath21 . for every @xmath295 and function @xmath296 with @xmath297
, there exist @xmath298 and @xmath299 with the following property : + if @xmath7 is a graph [ @xmath0-graph , digraph ] with @xmath300 vertices then there exist an equipartition @xmath301 of @xmath7 and an induced subgraph [ induced @xmath0-subgraph , induced subdigraph ] @xmath25 of @xmath7 , with an equipartition @xmath302 of the vertices of @xmath25 that satisfy : * @xmath303 .
* @xmath304 for all @xmath305 , and @xmath306 . * in the equipartition @xmath307 , * all * pairs are @xmath308-regular . *
all but at most @xmath309 of the pairs @xmath310 are such that @xmath311 .
we use theorem [ thm : multicol : genreglem ] in order to prove theorem [ thm : multicol : reglemapp ] , which is the result that we need .
[ thm : multicol : reglemapp ] let @xmath25 be an @xmath0-graph in hereditary property @xmath36 and @xmath312 be a density vector .
then , there exists an @xmath0-type @xmath291 such that @xmath313 for all @xmath124 and with probability going to @xmath81 as @xmath212 , @xmath314 .
the proof of theorem [ thm : multicol : reglemapp ] from theorem [ thm : multicol : genreglem ] is straightforward and the details are given in @xcite .
we begin with @xmath7 distributed according to @xmath206 and typical in the sense that any szemerdi partition will have every pair @xmath315-regular .
let @xmath25 be the graph of smallest distance from @xmath7 and apply theorem [ thm : multicol : genreglem ] .
the resulting partition @xmath307 describes a type @xmath145 which must be in @xmath163 .
furthermore , the number of changes required to ensure that @xmath25 has partition @xmath316 is very close to @xmath200 because almost every pair in @xmath316 has the same density as in @xmath307 . using part a , we see that for any @xmath211 , with probability approaching 1 as @xmath212 , @xmath317 + we can now combine a , b and c. take the limit of ( [ eq : randomlimit ] ) as @xmath212 , and we obtain that for all @xmath211 , @xmath318 .
hence , @xmath319 .
moreover , we can replace the second term with @xmath320 $ ] because that random variable is bounded ( in @xmath224 $ ] ) and so ( [ eq : randomlimit ] ) occurring with high probability implies that the random variable is concentrated around its mean , which approaches @xmath205 .
this verifies parts ( 1 ) , ( 2 ) and ( 3 ) of the theorem .
+ * d : * continuity of @xmath198 .
+ because the set of @xmath0-types is countable , we can linearly order @xmath163 to be @xmath321 . for every density vector @xmath170 ,
set @xmath322 .
we want to show that each function @xmath323 is lipschitz with coefficient 1 with respect to the @xmath324 metric .
let @xmath325 and @xmath326 be density vectors and define @xmath0-types @xmath327 on @xmath328 vertices , respectively , such that @xmath329 and @xmath330 . then , using the matrix definition of @xmath198 and the definition of @xmath323 as a minimum of linear functions , @xmath331 since each of the entries in matrices @xmath332 and @xmath333 is between zero and one , and the number of entries in these matrices is @xmath334 and @xmath335 , respectively , it is the case that @xmath336 since @xmath337 is lipschitz , definition 7.22 from rudin @xcite says that the sequence of functions is equicontinuous .
the sequence is also pointwise bounded above by @xmath338 and below by @xmath339 . by theorem 7.25(b ) from @xcite the sequence @xmath337 has a uniformly convergent subsequence .
since @xmath337 is an equicontinuous , each member is itself continuous .
theorem 7.12 from @xcite gives that the aforementioned uniformly convergent subsequence has a continuous limit .
the monotonicity of @xmath337 gives that the limit of any subsequence is the same as the pointwise limit of the sequence itself , namely @xmath340 .
+ * e : * concavity .
+ let @xmath341 and @xmath342 be density vectors and @xmath217 $ ] be a real number .
observe that @xmath218 is still a density vector and , hence , in the domain .
furthermore , @xmath343 this gives concavity .
+ using d and e , we obtain part ( 4 ) directly and the fact that @xmath344 achieves its maximum follows from continuity ( and compactness ) and theorem 4.16 from @xcite .
let @xmath345 be the set of density vectors @xmath170 such that @xmath346 .
the set @xmath345 must be convex set , because if @xmath347 , then by continuity and concavity , the line segment that connects @xmath341 and @xmath342 must consist of vectors in @xmath345 .
the set @xmath345 must be closed because a corollary to theorem 4.8 from @xcite says that , under a continuous mapping , the inverse image of a closed set is closed . since @xmath205 is a continuous function and @xmath345 is the inverse image of the closed set , @xmath348 , then @xmath345 is closed .
this verifies parts ( 4 ) , ( 5 ) and ( 6 ) of the theorem and concludes the proof.@xmath349 + + [ sec : multicol : triangles ] * ( [ it : multicol : triangles:1 ] ) * in order to destroy all copies of a monochromatic @xmath81-colored triangle in an arbitrary coloring of @xmath207 , it is sufficient to split the vertex set into two parts and recolor all edges within these parts in color @xmath87 .
this requires at most @xmath350 changes . to see the lower bound ,
consider @xmath207 with all edges colored @xmath81 . after all editing is done to ensure that color class @xmath81 has no triangles ,
color class @xmath81 is triangle - free , having at most @xmath351 edges .
thus , at least @xmath352 edges must have been changed .
+ * ( [ it : multicol : triangles:2 ] ) * in order to destroy all such triangles , it suffices to equipartition the vertex set into two parts and recolor all edges within these parts to color @xmath76 .
this requires at most @xmath350 changes . to see the lower bound ,
consider @xmath207 on vertex set with equipartition @xmath353 .
let all edges between @xmath354 and @xmath355 be colored @xmath81 and let all edges within parts @xmath8 , @xmath356 be colored @xmath87 .
we may assume that the only editing operations are recoloring an edge of color @xmath81 into color @xmath76 and recoloring an edge of color @xmath87 into color @xmath76 because this editing will never create a forbidden triangle .
let @xmath137 be such a recoloring not containing triangles with two edges of color @xmath81 and one edge of color @xmath87 .
let @xmath7 be an auxiliary graph corresponding to edges of color @xmath76 in this coloring .
the complement of @xmath7 can not have any triangles with vertices in both @xmath354 and @xmath355 .
it is easy to prove by induction on @xmath23 that a graph with satisfying such a condition could have at most @xmath350 edges .
therefore @xmath7 has at least @xmath350 edges , and this corresponds to the number of changes made .
+ * ( [ it : multicol : triangles:3 ] ) * assume that @xmath37 consists of a triangle with all edges colored @xmath81 and of a triangle with all edges colored @xmath87 . in order to destroy both of these triangles in an any coloring , as in the previous case
, it is sufficient to equipartition the vertex set into two parts and recolor all edges within these parts in color @xmath76 .
this requires at most @xmath350 changes . as to the lower bound , fix @xmath357 and consider a @xmath76-type , @xmath291 , on @xmath245 vertices .
each of the vertices must have color @xmath76 . by turn s theorem
, at least @xmath358 edges can not have color @xmath81 and at least @xmath359 edges can not have color @xmath87 .
hence , if we consider the off - diagonal entries of @xmath190 , the sum is at least @xmath360 .
so , for any such @xmath145 , @xmath361\geq\frac{1}{2 } .\ ] ] as a result , @xmath362 .
+ * ( [ it : multicol : triangles:4 ] ) * it is suffices to recolor edges of colors @xmath81 or @xmath87 into color @xmath76 . as a result
, all forbidden colored triangles will be destroyed via at most @xmath363 changes .
in fact , for fixed @xmath364 , at most @xmath365 changes suffice . to see the lower bound , consider a @xmath76-type @xmath291 on @xmath245 vertices .
the vertices must be monochromatic and , in addition , the edges incident to a vertex must share the color of that vertex .
otherwise , there would be a bichromatic triangle @xmath33 with @xmath250 .
this implies , however , that @xmath145 must be entirely monochromatic .
hence , @xmath366 .
note the this determines not only @xmath30 , but the entire function @xmath367 .
+ * ( [ it : multicol : triangles:5 ] ) * observe that in order to destroy all rainbow triangles using colors @xmath81 , @xmath87 and @xmath76 , it is sufficient to edit the smallest of these color classes , thus performing at most a @xmath368 proportion of changes . for the lower bound , simply observe that no edge in any @xmath291 can be trichromatic . otherwise , that edge , together with any vertex to which it is incident admits a mapping of a rainbow triangle .
hence , each entry of @xmath369 is at least @xmath368 and so @xmath370 .
hence @xmath371 and @xmath372.@xmath349 +
+ [ sec : digraphs ] a simple directed graph or digraph is defined to be a pair @xmath373 where @xmath1 is a labeled vertex set , @xmath374 and @xmath375 denotes the set @xmath376 .
we will also view this as a coloring ; that is , a digraph is a pair @xmath377 where @xmath378 is a function which has the property that , for distinct @xmath379 , let @xmath384 .
here we interpret the color @xmath385 to mean that neither @xmath386 nor @xmath387 are in @xmath388 , the color @xmath389 to mean that both @xmath386 and @xmath387 are in @xmath388 and the color @xmath382 to mean that @xmath390 and @xmath391 .
[ defn : digraphs : basic ] for any digraph @xmath7 on fixed vertex set @xmath392 , disjoint vertex sets @xmath8 and @xmath9 and color @xmath10 , @xmath393 , the expression @xmath12 denotes the set of pairs @xmath394 with @xmath395 , @xmath396 and @xmath397 . the expression @xmath398 denotes the set of pairs @xmath399 with @xmath400 , @xmath401 and @xmath402 . hence , @xmath403 .
as it happens , we will be able to assume , as in the proof of theorem [ thm : multicol : basics ] , that our graphs are random
. we will also be able to assume that , among the pairs that have directed edges , a @xmath404 is as likely as @xmath405 .
hence , we will postpone the definition of a density vector for directed graphs . 1 .
@xmath408 is the most general case . [
it : palette : a ] 2 .
@xmath409 is the case of simple digraphs such that every pair of vertices has at least one arc between them .
[ it : palette : noempty ] 3 .
@xmath410 is the case of oriented graphs ; that is , no pair of vertices has two arcs between them .
[ it : palette : nounarrow ] 4 .
@xmath411 is the case of simple , undirected graphs .
[ it : palette : undir ] 5 .
@xmath412 is the case of tournaments .
[ it : palette : tourn ] if @xmath407 is a fixed palette and @xmath3 and @xmath22 are digraphs with colors in @xmath407 , then @xmath414 is the proportion of edges on which the colors differ ; i.e. , the number of edges on which the colors differ , divided by @xmath26 . determines the color of the pair @xmath387 . ]
we may call this the normalized edit distance between @xmath7 and @xmath25 . for any property @xmath27 , a simple digraph @xmath7 with all edge - colors in palette @xmath407 , an integer @xmath23
, we define @xmath28 , @xmath29 , and @xmath30 similarly to the multicolor case .
a hereditary property of digraphs with respect to palette @xmath407 ( or , simply , hereditary property , where the context is understood ) is a set of digraphs with all edge - colors in @xmath407 that is closed under vertex - deletion and isomorphisms .
let a digraph @xmath25 be an induced digraph of @xmath7 if @xmath25 can be obtained from @xmath7 by vertex - deletion .
for a fixed palette , @xmath407 and a digraph , @xmath33 , the family @xmath34 ( the palette will be understood ) consists of all digraphs with edge - colors in @xmath407 that have no ( induced ) copies of @xmath33 .
for every palette @xmath407 and every hereditary property @xmath27 with respect to @xmath407 , there is a family , @xmath35 , of digraphs such that @xmath36 .
if @xmath37 is a family of digraphs , then we use @xmath38 to denote @xmath39 .
+ for a hereditary property @xmath36 and a palette @xmath407 , a weakly - good triple @xmath415 is a triple of non - negative integers such that for some @xmath42 , the vertex set @xmath43 can be partitioned into sets @xmath416 such that , for each @xmath417 with @xmath46 , the partition can be further refined @xmath47 and * each @xmath418 does not induce a nonedge ( i.e. , does not induce an edge of color @xmath413 ) , * each @xmath419 ensures that the directed edges induced by @xmath419 form an acyclic digraph , and * each @xmath420 does not induce a bidirectional edge ( i.e. , does not induce an edge of color @xmath421 ) . the weak clique spectrum of @xmath27 with respect to a palette @xmath407 is the set of all triples @xmath415 that are not weakly - good and such that @xmath422 if @xmath423 , @xmath424 if @xmath425 , @xmath426 if @xmath427 .
the weak directed chromatic number , @xmath428 , of @xmath27 with respect to a palette @xmath407 is the maximum @xmath429 such that for some non - negative integers @xmath430 , with @xmath431 , the triple @xmath415 is in the weak clique spectrum of @xmath27 .
we merely use @xmath432 for the weak directed chromatic number if the palette is understood .
+ for a hereditary property @xmath36 and a palette @xmath407 , a strongly - good triple @xmath415 is a triple of non - negative integers such that for some @xmath42 , the vertex set @xmath43 can be partitioned into sets @xmath416 such that , for each @xmath417 with @xmath46 , the partition can be further refined @xmath47 and * each @xmath418 induces only nonedges ( i.e. , all edges are of color @xmath413 ) , * each @xmath419 ensures that the directed edges induced by @xmath419 induce a transitive tournament , and * each @xmath420 induces only bidirectional edges ( i.e. , all edges are of color @xmath421 ) .
the strong clique spectrum of @xmath27 with respect to a palette @xmath407 is the set of all triples @xmath415 that are not strongly - good and such that @xmath422 if @xmath423 , @xmath424 if @xmath425 , @xmath426 if @xmath427 .
the strong directed chromatic number , @xmath433 , of @xmath27 with respect to a palette @xmath407 is the maximum @xmath429 such that for some non - negative integers @xmath430 , with @xmath431 , the triple @xmath415 is in the clique spectrum of @xmath27 .
we merely use @xmath434 for the strong directed chromatic number if the palette is understood . + *
the weak [ strong ] clique spectrum with respect to a given palette is again a downset in the partially ordered set of @xmath0-tuples ordered coordinatewise .
that is , if @xmath415 is in the weak [ strong ] clique spectrum and @xmath435 has the property that @xmath60 for @xmath436 , then @xmath435 is also in that weak [ strong ] clique spectrum . * for any palette @xmath407 and any hereditary property of digraphs , @xmath27 , @xmath437 .
* if the palette @xmath438 , the weak and strong directed chromatic numbers are equal and so in those cases , we can use @xmath439 . *
if the palette is @xmath411 , then @xmath440 is both the binary chromatic number of hereditary property @xmath27 . * if @xmath56 and the palette is @xmath412 , the case of tournaments , then @xmath441 is the fewest number of transitive subtournaments into which @xmath43 can be partitioned .
+ + [ sec : digraphs : simplealg ] let @xmath407 be a palette and let @xmath27 be a hereditary property of digraphs such that @xmath120 and each edge of each @xmath124 has a color in @xmath407 .
further , let @xmath442 and @xmath415 be in the weak clique spectrum and @xmath443 . recall
that if a color is not in the palette , then its corresponding @xmath229 value must be set to zero .
partition @xmath1 into @xmath76 sets , @xmath416 and further refine the partition such that @xmath47 , for @xmath436 and then recolor the edges induced by each @xmath123 as follows : * if @xmath444 , then recolor the edges colored @xmath413 into some other arbitrary color in the palette . *
if @xmath445 , then recolor the edges @xmath404 and @xmath405 so that there are no directed cycles among those directed edges . *
if @xmath446 , then recolor the edges colored @xmath421 into some other arbitrary color in the palette
. this new coloring does not contain any @xmath124 , otherwise the triple @xmath415 would be weakly good for some @xmath33 . as in the multicolor case
, if the partition into sets @xmath123 is an equipartition , then @xmath447 + [ thm : digraphs : easybounds ] let @xmath407 be a palette and @xmath27 be a hereditary property of digraphs .
let @xmath448 and @xmath449 be the weak and strong directed chromatic numbers , respectively , of @xmath27 .
recall that if @xmath438 , then @xmath439 . then , 1 .
@xmath450 , if @xmath451 .
@xmath452 , if @xmath453 .
@xmath452 , if @xmath454 .
@xmath455 , if @xmath456 .
[ it : digraphs : easybounds : undir ] 5 .
@xmath457 , if @xmath458 .
[ it : digraphs : easybounds : tourn ] we prove theorem [ thm : digraphs : easybounds ] in section [ sec : digraphs : proofs ] . as in the multicolor case ,
the upper bound is a consequence of the simple editing algorithm .
the lower bound comes from theorem [ thm : digraphs : basics ] , stated below , which is the digraph version of theorem [ thm : multicol : basics ] and deals with computing the edit distance for given hereditary properties of digraphs . in order to do so
, we need to investigate the so - called edit distance function , which computes the edit distance of a digraph such that nonedges , directed edges and undirected edges having a specified density .
+ * @xmath460 is the set of all unordered pairs @xmath461 such that @xmath385 , * @xmath462 is the set of all ordered pairs @xmath386 such that @xmath463 , * @xmath464 is the set of all ordered pairs @xmath386 such that @xmath382 , * @xmath465 is the set of all unordered pairs @xmath461 such that @xmath389 , given a palette , @xmath407 , a directed density vector @xmath466 with respect to @xmath407 ( or , simply , density vector or probability vector where the context is understood ) is a nonnegative real vector with the property that @xmath467 .
furthermore , 1 . if @xmath453 , then @xmath468 .
2 . if @xmath454 , then @xmath469 and @xmath470 .
3 . if @xmath456 , then @xmath471 and @xmath472 .
this is the @xmath0-graph case where @xmath67 or simply the case of undirected graphs .
see @xcite and @xcite .
4 . if @xmath458 , then @xmath469 and @xmath473 , so @xmath474 .
for any density vector @xmath466 , and an integer @xmath23 , we denote , @xmath475 , @xmath476 and the size of @xmath477 is what remains .
since we fix @xmath478 and @xmath479 and let @xmath23 approach infinity , this will make no appreciable difference . ]
latexmath:[\[{{\rm dist}}_n((p , q),{\mathcal{h}})=\max\left\{{{\rm dist}}(g,{\mathcal{h } } ) : \begin{array}{l } |v(g)|=n , |e_{{-}}(g)|=p{\textstyle\binom{n}{2 } } , |e_{\rightarrow}(g)|=q{\textstyle\binom{n}{2 } } , \\
observe that there are , in fact , four densities here ; two are equal and all sum to one .
thus , we only need two parameters .
we choose parameter names as above because the case of @xmath471 gives the classical case of undirected graphs , as we see below .
later in the paper , we show that the following limit exits , which we call the edit distance function : @xmath481 in section [ sec : digraphs : algorithm - analysis ] , we define two functions which are described in terms of dir - types , which allow us to compute the edit distance function . later in the paper , we shall provide algorithms for such computing
. for a palette @xmath407 , a @xmath407-dir - type ( or dir - type or type , where the context and the palette are understood ) , @xmath145 , is a pair @xmath146 , where @xmath147 is a finite set of vertices and @xmath483 , such that * if @xmath492 , then @xmath493 , * if @xmath494 and @xmath495 , then @xmath496 has no pair with color @xmath160 , * if @xmath497 , then @xmath496 has no directed edge , and * if @xmath498 , then @xmath496 has no directed cycle .
in other words , there is a mapping @xmath159 that brings each edge of color @xmath160 to a vertex or an edge containing @xmath160 in its color set , except that if a vertex contains exactly one of @xmath499 then the pre - image of that vertex can be ordered transitively with respect to the oriented edges . if @xmath33 embeds in type @xmath145 , we write @xmath161 , otherwise we write @xmath162 . for every hereditary property @xmath27 ,
we let @xmath163 be the set of all dir - types such that none of @xmath35 embeds in that type , i.e. , @xmath500 + [ sec : digraphs : algorithm ] let @xmath501 be a density vector and let @xmath502 be a density vector .
this latter vector will represent a vector of densities .
the number of ordered pairs @xmath503 with color `` @xmath421 '' will be @xmath504 and the number of ordered pairs with color `` @xmath413 '' will be @xmath505 .
the number of ordered pairs with color `` @xmath404 '' is @xmath506 and the number of _ ordered _ pairs with color `` @xmath405 '' is @xmath507 .
consequently , @xmath508 .
let @xmath509 be a palette , @xmath27 be a hereditary property and @xmath3 be a digraph in @xmath407 such that the density vector is @xmath502 .
in order to find an upper bound on @xmath28 , it is sufficient to change @xmath7 to a digraph such that , for all @xmath124 , @xmath33 does not embed into the new coloring . in particular , if the resulting coloring has type @xmath175 , then this coloring is in @xmath27 .
[ algor : digraphs : recolor ] fix such a @xmath176 and try to bring @xmath7 to a coloring of type @xmath145 by edge - recoloring .
let @xmath177 .
partition the vertices of @xmath7 randomly into sets @xmath178 such that the probability of a vertex to be in a part @xmath8 is @xmath179 . with an ordering of the vertices of @xmath7 and vertices @xmath510 , consider an edge @xmath503 of @xmath7 , let @xmath181 , @xmath182 , for @xmath511 .
if @xmath233 and @xmath512 , recolor @xmath503 with a color from @xmath513 . next , consider the edges in @xmath8 .
if @xmath514 contains exactly one of @xmath499 , then consider a random order of the vertices of @xmath8 , call it @xmath515 .
let @xmath510 and both in @xmath8 .
if @xmath516 , then recolor @xmath503 if and only if @xmath517 . if @xmath518 , then recolor @xmath503 if and only if @xmath519
note that this forces @xmath8 to have no directed cycles .
if @xmath520 for some @xmath485 , then recolor any edge with color @xmath521 to a color in @xmath514 . this concludes the algorithm .
let us first consider a pair @xmath503 .
if @xmath522 , then the probability that the color of @xmath503 is unchanged is @xmath523 if @xmath524 , then the probability that the color of @xmath503 is unchanged is @xmath525 it does nt matter whether we consider the pair @xmath526 or @xmath527 in the last term because@xmath528 is invariant whether @xmath529 or @xmath530 . now , the expected number of changes is @xmath189 & = & \binom{n}{2}-\sum_{x , y\in v,~x <
y } \pr((x , y)\mbox { is not changed } ) \\ & = & \binom{n}{2}-\sum_{1\leq i , j\leq k}w_iw_jp_{{\bigcirc}}\binom{n}{2}{\bf 1}_{{\bigcirc}\in\phi(u_i , u_j)}-\sum_{1\leq i , j\leq k}w_iw_jp_{{-}}\binom{n}{2}{\bf 1}_{{-}\in\phi(u_i , u_j ) } \\ & & -\sum_{1\leq i , j\leq k}w_iw_j\frac{p_{\leftarrow}+p_{\rightarrow}}{2}\binom{n}{2}|\{\leftarrow,\rightarrow\}\cap\phi(u_i , u_j)| \end{aligned}\ ] ] let @xmath531 , @xmath532 and so @xmath533 . for @xmath534 , and @xmath535 ,
the matrix @xmath536 is such that the @xmath537 entry is @xmath81 if @xmath538 and zero otherwise .
the matrix @xmath539 is a @xmath540-matrix with the property that @xmath541 with @xmath542 denoting the @xmath191 all - ones matrix , then we define @xmath543 consequently , if @xmath195 , then @xmath544={\bf w}^t{{\bf m}}_k({{\bf p}}){\bf w}\binom{n}{2}$ ] . in the directed case ,
each ordered pair can receive one of 4 directions , but the density vectors only have two entries rather than three .
this is because the above computation shows that an upper bound on editing any digraph is determined not by the pair @xmath547 but only by @xmath548 .
it is straightforward , by the same arguments as in the proof of theorem [ thm : multicol : basics ] , to see that the lower bound for the maximum edit distance is asymptotically achieved by a random graph in which the probability of a forward arc is equal to the probability of a backward arc
. + theorem [ thm : digraphs : basics ] is a parallel to theorem [ thm : multicol : basics ] and summarizes some facts about the edit distance function . recall that , depending on the palette , there may be further restrictions on the density vector other than the necessary @xmath467 . the dimension , @xmath0 , of the palette , @xmath407 , is the number of members of @xmath549 that @xmath407 has .
1 . @xmath210 ; [ it : digraphs : inf ] 2 .
fix @xmath211 , then with probability approaching 1 as @xmath212 , @xmath213 3 .
@xmath214 $ ] ; 4 .
@xmath215 is continuous over the domain of density vectors with respect to @xmath407 and is concave down;[it : digraphs : concon ] 5 .
@xmath205 achieves its maximum , @xmath128 , at some density vector @xmath220 ( in fact , denote the set of all such vectors @xmath221 ) and so , @xmath222 \mbox { ; and}\ ] ] 6 .
both @xmath223 and @xmath30 exist and @xmath223 is a convex and closed set in @xmath224^{r-1}$ ]
. + the case of tournaments is relatively straightforward . because in tournaments , there are no edges labeled @xmath413 or @xmath421 , there is only one density vector , @xmath552 .
this means that we only need to consider tournaments that are random , that each arc is forward independently with probability @xmath553 .
this leads to a rather simple expression for the edit distance : note that in the case of tournaments , the directed chromatic number of tournament @xmath33 , @xmath555 is the smallest number of transitive subtournaments into which @xmath33 can be partitioned .
we prove theorem [ thm : digraphs : tournaments ] in section [ sec : digraphs : tournaments ] .
+ 1 . if @xmath37 is a family that consists of a single directed triangle , then , regardless of the palette , @xmath240.[it : digraphs : triangles:1 ] 2 .
if @xmath37 is a family that consists of a single transitive triangle and @xmath556 , the palette of tournaments , then @xmath38 is a trivial hereditary property.[it : digraphs : triangles:2 ] 3 . if @xmath37 is a family of that consists of a single transitive triangle , then , if @xmath407 is any palette other than @xmath557 , then @xmath240.[it : digraphs : triangles:3 ] 4 .
if @xmath37 is a family that consists of both a transitive and a directed triangle , and @xmath407 is any palette other than @xmath557 , then @xmath240.[it : digraphs : triangles:4 ] let @xmath558 . for the lower bound
, we apply part [ it : digraphs : inf ] of theorem [ thm : digraphs : basics ] , which states that @xmath243 .
consider an arbitrary @xmath559 , a @xmath407-dir - type on @xmath245 vertices .
let @xmath246 be a graph with vertex set @xmath1 such that @xmath560 and @xmath561 are adjacent in @xmath246 if and only if @xmath562 .
we observe that @xmath246 has no clique on @xmath563 vertices , otherwise for some @xmath124 , @xmath250 .
using turn s theorem , the number of edges of @xmath246 is at most @xmath564 .
let @xmath252 .
consider the matrix @xmath253 and observe that every entry is either zero or is a positive integer multiple of @xmath254 .
the zero entries correspond exactly to pairs with @xmath255 value equal to @xmath407 .
thus , this matrix @xmath253 has at least @xmath565 entries with value at least @xmath254 .
therefore , @xmath257 is at least @xmath566 .
since @xmath145 was arbitrary , this gives a lower bound for @xmath205 .
+ the proof of most of this theorem is identical to that of theorem [ thm : multicol : basics ] , which is found in section [ sec : multicol : basics ] .
the only significant wrinkle is the upper bound .
that is , if @xmath7 is a digraph with @xmath567 edges with color @xmath421 and @xmath568 edges with color @xmath413 , then , with @xmath550 , @xmath569 + [ sec : digraphs : tournaments ] in this case , @xmath552 . let @xmath27 be a hereditary property of tournaments and @xmath570 . in any type @xmath145 on @xmath245 vertices ,
the vertices have color `` @xmath405 '' and the edges either have one direction or both . by the definition of the directed chromatic number , @xmath250 if @xmath145 has a clique of order @xmath571 such that every edge of @xmath145 has color set @xmath499 .
similar to the argument in section [ sec : digraphs : easyboundproof ] , we can use turn s theorem to find a lower bound for @xmath200 .
the bilinear form @xmath572 counts @xmath573 , where @xmath574 is the set of _ ordered _ pairs with color @xmath10 . since @xmath575
, turn s theorem gives that @xmath576 .
consequently , @xmath577=\frac{1}{2({\chi^{{\rm dir}}_{}}-1 ) } .\ ] ] this concludes the proof of theorem [ thm : digraphs : tournaments ] .
+ * ( [ it : digraphs : triangles:1 ] ) * as to the upper bound , linearly order the vertices so that the number of backward edges ( i.e. , pairs @xmath399 such that @xmath529 and @xmath578 ) is minimized . a greedy ordering results in at most half of such edges being present .
reorient such edges so that they become forward edges , hence @xmath579 .
note that this corresponds to a @xmath145 that consists of a single vertex which has color @xmath405 .
observe that there can be no entry @xmath586 because that means , for the corresponding vertex @xmath274 , @xmath587 and a directed triangle maps to such a vertex .
suppose there is some entry @xmath588 .
this implies that there are a pair of vertices , @xmath274 and @xmath277 such that @xmath589 .
we observe that @xmath590 , otherwise the directed triangle would map to these two vertices of @xmath145 .
consequently , @xmath591 .
moreover , for every @xmath592 , we have @xmath593 .
if not , then without loss of generality , we have a triangle @xmath594 in @xmath145 such that two edges contain @xmath499 and the third contains one of @xmath499 .
it is easy to see that a directed triangle maps to three such vertices .
but then , @xmath595 since each sum on the left hand side must be equal to @xmath201 , it must be that @xmath596 . finally , if there is no zero entry in the @xmath49-th row of @xmath253 , then @xmath597 .
thus , in all cases , @xmath596 .
+ * ( [ it : digraphs : triangles:2 ] ) * here we make the easily verified observation that any tournament with at least @xmath598 vertices has a transitive subtournament of size @xmath76 .
so , the hereditary property consists of no tournaments of size @xmath598 or more .
+ * ( [ it : digraphs : triangles:3 ] ) * as to the upper bound , equipartition the vertex set arbitrarily and recolor an each edge inside either part to have a color other than one in @xmath499 .
hence , @xmath579 .
note that this corresponds to a @xmath145 that consists of two vertices colored with some nonempty subset of @xmath599 and an edge colored @xmath407 .
for the lower bound , simply let @xmath7 be a transitive tournament .
after editing @xmath7 to make @xmath25 , there can be no triangles from @xmath7 that remain and so mantel s theorem gives that @xmath600 * ( [ it : digraphs : triangles:4 ] ) * here we can use the trivial fact that if @xmath27 is the hereditary property that forbids both directed and transitive triangles and @xmath601 is the larger hereditary property in ( [ it : digraphs : triangles:3 ] ) which forbids only the transitive triangle , then @xmath602 .
but the example above of a type @xmath145 that consists of two vertices with none of @xmath499 in its color set is in @xmath163 in this case as well
. hence , @xmath603 and so @xmath604 .
+ 99 n. alon , e. fischer , m. krivelevich and m. szegedy , efficient testing of large graphs , _ combinatorica _ * 20*(4 ) ( 2000 ) , no .
4 , 451476 .
n. alon and j. spencer , _ the probabilistic method_. third edition . with an appendix on the life and work of paul erds .
wiley - interscience series in discrete mathematics and optimization . _ john wiley & sons , inc . ,
hoboken , nj _ , 2008 .
n. alon and a. stav , what is the furthest graph from a hereditary property ?
_ random structures algorithms _ * 33 * ( 2008 ) , no .
1 , pp . 87104 . m. axenovich , a. kzdy and r. martin , on the editing distance of graphs , _ j. graph theory _ * 58 * ( 2008 ) , no .
2 , 123138 . m. axenovich and r. martin , a version of szemerdi s regularity lemma for multicolored graphs and directed graphs that is suitable for induced graphs , manuscript .
j. balogh and r. martin , edit distance and its computation , _ electron .
j. combin .
_ * 16*(1 ) ( 2009 ) , research paper 109 , 16pp . ( electronic ) .
brown , on graphs that do not contain a thomsen graph , _ canad .
_ , 1966 , * 9 * , 281 - 285
fortuin , p.w .
kasteleyn and j. ginibre , `` correlation inequalities on some partially ordered sets , '' _ communications in mathematical physics _ , 1971 , * 22 * , 89103 .
d. e. knuth , `` permutations , matrices , and generalized young tableaux , '' _ pacific journal of mathematics _ , *
34*(3 ) , ( 1970 ) , 709727 .
j. komls and m. simonovits , szemerdi s regularity lemma and its applications in graph theory , _
combinatorics , paul erds is eighty , vol . 2 ( keszthely , 1993 ) _ , budapest , 1996 , pp .
e. marchant , ph.d .
thesis , university of cambridge , ( 2010 ) .
e. marchant and a. thomason , extremal graphs and multigraphs with two weighted colours , in _ fete oif combinatorics and computer science _
katona , a. schrijver and t. sznyi ) , _ bolyai soc .
stud . _ * 20 * , springer , berlin , in press .
r. martin , edit distance and localization , preprint .
r. martin and t. mckay , on the edit distance from @xmath605-free graphs i : the case of @xmath606 , preprint . w. rudin , _ principles of mathematical analysis_. third edition .
international series in pure and applied mathematics . _ mcgraw - hill book co. , new york - auckland - dsseldorf _ , 1976 .
b. bollobs and a. thomason , hereditary and monotone properties of graphs , in _ the mathematics of paul erds ii _ ( r.l .
graham and j. nesetril , eds . ) _ algorithms and combinatorics _ * 14 * , springer - verlag ( 1997 ) , 7078 .
b. bollobs and a. thomason , the structure of hereditary properties and colourings of random graphs , _ combinatorica _
* 20 * ( 2000 ) , 173202 .
sidorenko , boundedness of optimal matrices in extremal multigraph and digraph problems , _ combinatorica _ * 13 * ( 1993 ) , 109120 .
r. martin , edit distance and localization , submitted , arxiv:1007.1897v3 . r. martin and t. mckay , on the edit distance from @xmath605-free graphs , submitted , arxiv:1012.0800 .
r. martin , on the computation of edit distance functions , submitted , arxiv:1012.1237 . | the editing of a combinatorial object is the alteration of some of its elements such that the resulting object satisfies a certain fixed property .
the edit problem for graphs , when the edges are added or deleted , was first studied independently by the authors and kzdy @xcite and by alon and stav @xcite . in this paper ,
a generalization of graph editing is considered for multicolorings of the complete graph as well as for directed graphs .
specifically , the number of edge - recolorings sufficient to be performed on any edge - colored complete graph to satisfy a given hereditary property is investigated .
the theory for computing the edit distance is extended using random structures and so - called types or colored homomorphisms of graphs . |
the best known examples of topological magnetic solitons are magnetic bubbles or skyrmions observed in abundance in ferromagnetic films with easy - axis anisotropy @xcite .
the experimental situation is less clear in the case of ferromagnets with easy - plane anisotropy .
the relevant topological structures are theoretically predicted to be half skyrmions or vortices , with a logarithmically divergent energy that may inhibit production of an isolated vortex on an infinite film .
thus the early studies of magnetic vortex dynamics had been mostly theoretical @xcite drawing on various analogies with related work on ferromagnetic bubbles @xcite , with vortex dynamics in classical fluids and superfluids @xcite , as well as with the dynamics of interacting electric charges in a uniform magnetic field .
the situation has changed dramatically in recent years .
it has been realized that a disc - shaped magnetic element , with a diameter of a few hundred of nanometers , provides an excellent geometry for the realization of a magnetic vortex configuration . in particular , the exchange energy is finite on a finite element while the magnetostatic field vanishes everywhere except at the vortex core . as a result ,
the vortex is actually the lowest energy magnetic state in a disc - shaped element . in other words ,
interest in the vortex stems from the fact that it is a nontrivial magnetic state which can , nevertheless , be spontaneously created in magnetic elements @xcite .
it is then natural to ask whether nontrivial magnetic states other than the single vortex may play an important role in the dynamics of magnetic elements @xcite .
an answer to this question comes from a somewhat unlikely direction .
recent experiments have shown a peculiar dynamical behavior of vortices and magnetic domain walls when these are probed by external magnetic fields .
vortices may switch their polarity under the influence of a very weak external magnetic field of the order of a few mt @xcite . the same switching phenomenon was observed by passing an a.c .
electrical current through a magnetic disc @xcite .
since the polarity of the vortex contributes to its topological structure , the switching process clearly implies a discontinuous ( topologically forbidden ) change of the magnetic configuration .
this is certainly a surprise especially because the external field is rather weak .
the key to this phenomenon is the appearance of vortex pairs which are spontaneously created in the vicinity of existing vortices @xcite .
the creation of topological excitations ( vortex pairs ) by alternating external fields had been anticipated by an early study based on collective coordinates @xcite .
in this paper we study vortex - antivortex pairs ( va pairs ) which are nontrivial magnetic states that play an important role in the dynamics of magnetic elements .
however , unlike a single vortex , a va pair is a localized object whose energy remains finite even on an infinite film .
it is then reasonable to expect that the essential features of the dynamics of va pairs can be understood in the infinite - film approximation which is adopted in the following .
a brief summary of the relevant dynamical equations and related topological structures is given in section [ sec : model ] . in section [ sec : kelvin ] we study a va pair in which the vortex and the antivortex carry the same polarity .
such a pair is shown to undergo translational kelvin motion analogous to that observed in fluid dynamics @xcite . in section [ sec : rotate ] we study a va pair in which the vortex and the antivortex carry opposite polarities . such a pair is shown to behave as a rotating vortex dipole @xcite because its topological structure is substantially different than that of the pair in kelvin motion .
the preceding results are combined in section [ sec : switching ] to demonstrate that a rotating vortex dipole may be annihilated by a quasi - continuous process in spite of its nontrivial topological structure . in particular
, no energy barrier has to be overcome in contrast to the usual expectations for topological solitons .
this opens the possibility for switching mechanisms between topologically distinct states in ferromagnets .
the possibility to change the topological structure leads to a dramatic change in the magnetization dynamics as a va pair is created or annihilated .
such a pair annihilation process lies in the heart of the counter - intuitive vortex polarity switching event that was observed in magnetic elements @xcite .
some concluding remarks are summarized in section [ sec : conclusion ] .
finally , an appendix is devoted to a brief description of the dynamics of interacting electric charges in a uniform magnetic field , which also exhibits most of the peculiar features of the dynamics of va pairs .
a ferromagnet is characterized by the magnetization @xmath6 measured in units of the constant saturation magnetization @xmath7 .
hence @xmath8 is a vector field of unit length , @xmath9 , but is otherwise a nontrivial function of position and time @xmath10 that satisfies the rationalized landau - lifshitz ( ll ) equation @xmath11 here distances are measured in units of the exchange length @xmath12 , where @xmath13 is the exchange constant , and the unit of time is @xmath14 where @xmath15 is the gyromagnetic ratio .
typical values are @xmath16 and @xmath17 which set the scales for the phenomena described by eq .
( [ eq : lle ] ) . to complete the description of the ll equation we note that we consider ferromagnetic materials with uniaxial anisotropy .
then @xmath18 in eq .
( [ eq : lle ] ) is a unit vector along the symmetry axis and the dimensionless parameter @xmath19 , where @xmath20 is an anisotropy constant , measures the strength of anisotropy . in particular
, @xmath21 is taken to be positive throughout this paper , a choice that corresponds to easy - plane ferromagnets .
an important omission in eq .
( [ eq : lle ] ) is the demagnetizing field produced by the magnetization itself @xcite .
however , in the limit of a very thin film , the effect of the demagnetizing field is thought to amount to a simple additive renormalization of the anisotropy constant @xcite .
also note that we may perform the rescalings @xmath22 and @xmath23 which further renormalize the units of space and time discussed earlier and lead to a completely rationalized ll equation where we may set @xmath24 without loss of generality . with this understanding ,
all calculations presented in this paper are based on a two - dimensional ( 2d ) restriction of eq .
( [ eq : lle ] ) , i.e. , @xmath25 and @xmath26 , while @xmath21 is set equal to unity without further notice . the effective field @xmath27 in eq .
( [ eq : lle ] ) may be derived from a variational argument : @xmath28\ , dx dy}\ ] ] where @xmath29 is the conserved energy functional .
a standard hamiltonian form is obtained by resolving the constraint @xmath30 through , say , the spherical parameterization @xmath31 the ll equation is then written as @xmath32 where @xmath33 is the canonical momentum conjugate to the azimuthal angle @xmath34 . taking into account the specific form of the energy in eq .
( [ eq : energy ] ) or @xmath35\ , dx dy\ ] ] the hamilton equations ( [ eq : thetaphi ] ) yield @xmath36\,\cos\theta\,\sin\theta\ , , \nonumber \\ \sin\theta\ , \frac{{\partial}\theta}{{\partial}t } & = & - { \bm{\nabla}}\cdot ( \sin^2\theta\ , { \bm{\nabla}}\phi).\end{aligned}\ ] ] another useful parameterization is obtained through the stereographic variable @xmath37 : @xmath38 in terms of which the ll equation reads @xmath39 where @xmath40 is the complex conjugate of @xmath37 .
equations ( [ eq : lle]),([eq : llethetaphi ] ) , and ( [ eq : lleomega ] ) are three equivalent versions of the ll equation and may be used at convenience depending on the specific calculation considered .
a key quantity for describing both topological and dynamical properties of the 2d ll equation is the local topological vorticity @xmath41 defined from @xcite : @xmath42 where the usual summation convention is invoked for the repeated indices @xmath43 and @xmath44 , which take over two distinct values corresponding to the two spatial coordinates @xmath45 and @xmath46 , and @xmath47 is the 2d antisymmetric tensor . in particular , one may consider the total topological vorticity @xmath48 and the pontryagin index or skyrmion number @xmath49 defined from @xmath50 a naive partial integration using eq .
( [ eq : vorticity1 ] ) yields @xmath51 for all magnetic configurations for which such an integration is permissible .
however , nonvanishing values for @xmath48 and @xmath49 are possible and are topologically quantized . specifically , for field configurations that approach a constant ( uniform ) magnetization at spatial infinity , the skyrmion number @xmath49 is quantized according to @xmath52 .
half integer values are also possible in the case of field configurations with more complicated structure at infinity such as half skyrmions or vortices ( see below ) .
the local topological vorticity @xmath15 is also important for an unambiguous definition of conservation laws in the ll equation .
hence the linear momentum ( impulse ) @xmath53 is defined from @xmath54 while the angular momentum ( impulse ) is given by @xmath55 where @xmath56 . since detailed discussions of these conservation laws
have already appeared in the literature @xcite we simply note here that analogous conservation laws were defined as moments of ordinary vorticity in fluid dynamics @xcite .
we first search for static ( time independent ) solutions of the ll equation which may be obtained by omitting time derivatives in eq .
( [ eq : llethetaphi ] ) and further introducing the axially symmetric ansatz @xmath57 and @xmath58 , where @xmath59 and @xmath60 are the usual cylindrical coordinates @xmath61 , @xmath62 will be referred to as the vortex number , and @xmath63 is an arbitrary constant phase reflecting the azimuthal invariance .
the resulting ordinary differential equation for the amplitude @xmath64 is solved numerically with standard boundary condition @xmath65 and the result is shown in fig .
[ fig : vortex ] .
the corresponding magnetization is then given by @xmath66 where @xmath67 will be called the polarity .
the total energy is accordingly reduced to @xmath68\ , ( 2\pi\rho d\rho).\ ] ] where the centrifugal ( second ) term is logarithmically divergent for the assumed boundary condition at spatial infinity . however , the anisotropy energy given by the last term is finite and is actually predicted to be @xmath69 by a careful derivation of a suitable virial relation @xcite . finally , the total vorticity @xmath48 and the skyrmion number @xmath49 are calculated from eq .
( [ eq : skyrmionnumber ] ) to be @xmath70 where the vortex number @xmath62 and the polarity @xmath67 may be taken in any combination .
thus , we must consider four possibilities ; namely , a vortex that comes in two varieties ( @xmath71 ) and thus @xmath72 , and an antivortex which also comes in two varieties ( @xmath73 ) and thus @xmath74 .
in all cases the calculated static solution is a topological soliton with half integer skyrmion number , in contrast to ordinary skyrmions ( such as magnetic bubbles ) which carry integer @xmath49 .
our main aim in the continuation of this paper is to search for nontrivial solutions that may combine a vortex and an antivortex ( va pair ) in a way that the total energy is finite .
we leave aside for the moment the ll equation and construct model va pairs in terms of the basic single - vortex configuration described above .
this is easily accomplished by invoking the stereographic variable @xmath37 of eq .
( [ eq : omega ] ) to write a single ( @xmath75 ) vortex located at the origin of coordinates as @xmath76 where @xmath64 is the vortex profile taken from fig .
[ fig : vortex ] or simply the model profile defined from @xmath77 and @xmath78 .
we now produce two replicas of the basic vortex to describe a pair of vortices by the product ansatz @xmath79 where @xmath80 is configuration ( [ eq : omega - thetaphi ] ) applied for @xmath81 and the origin displaced to , say , @xmath82 while @xmath83 is a @xmath84 vortex located around @xmath85 .
the skyrmion number of this configuration is given by @xmath86 and its energy is finite only if we restrict attention to vortex - antivortex ( va ) pairs @xmath87 . for definiteness
we choose @xmath88 and @xmath89 and thus the skyrmion number @xmath90 depends on the polarities @xmath91 and @xmath92 .
now a va pair with equal polarities @xmath93 is topologically trivial @xmath94 .
in contrast , a va pair with opposite polarities is topologically equivalent to a skyrmion ( @xmath95 , for @xmath96 and @xmath97 ) or an antiskyrmion ( @xmath98 , for @xmath99 and @xmath100 ) . in all three cases configuration ( [ eq :
productomega ] ) carries finite energy because its overall phase cancels out at spatial infinity where the magnetization approaches the uniform configuration @xmath101 modulo an overall azimuthal rotation which depends on the choice of individual phases @xmath63 in the ansatz ( [ eq : productomega ] ) .
this explains , in particular , why va pairs are characterized by an integer skyrmion number ( @xmath102 ) .
needless to say , the va pairs constructed above are not solutions of the ll equation .
yet an interesting picture arises when configuration ( [ eq : productomega ] ) is used as an initial condition in the complete equation ( [ eq : lleomega ] ) .
a topologically trivial ( @xmath103 ) va pair undergoes kelvin motion in which the vortex and the antivortex initially located at a relative distance @xmath104 along the x axis move in parallel along the y axis with nearly constant velocity .
in contrast , a topologically nontrivial ( @xmath105 ) va pair undergoes rotational motion around a fixed guiding center at nearly constant angular velocity . in both cases the main trajectories
are decorated by larmor - type oscillations @xcite which are tamed when the relative distance between the vortex and the antivortex is large .
the effect of the polarity of vortex pairs has been studied experimentally in patterned ferromagnetic ellipses @xcite .
two vortices were created an ellipse and it was found that their dynamics depended on their relative polarity , as is indicated by eq .
( [ eq : subtracts ] ) . in the following two sections we examine the two cases in turn .
in particular , we aim at constructing true steady - state solitary waves that describe va pairs in pure kelvin motion for @xmath103 and pure rotational motion for @xmath105 .
as the title of this section suggests , va pairs in kelvin motion were originally studied in the context of ordinary fluid dynamics @xcite . a further analogy exists with the 2d motion of an electron - positron pair interacting via the coulomb potential and placed in a uniform magnetic field perpendicular to the plane .
if the electron and the positron are initially at rest their guiding centers will move along two parallel straight lines while the actual trajectories will display the familiar larmor oscillations .
however , when the electron and the positron are given a common initial velocity such that the coulomb force is exactly balanced by the magnetic force , both the guiding centers and the actual positions of the charges will move steadily along parallel lines , even though the two sets of trajectories do not coincide .
the resulting special configuration may be thought of as a peculiar electron - positron bound state in steady translational motion ( see our appendix ) .
it is thus reasonable to expect that a solitary wave exists in a 2d easy - plane ferromagnet which describes a va pair that proceeds rigidly ( without larmor oscillations ) in a direction perpendicular to the line connecting the vortex and the antivortex , probably because the mutual force is exactly balanced by a topological `` magnus force '' .
the actual existence of such a solitary wave was established in ref .
@xcite whose main result is briefly reviewed in the remainder of this section . as it turns out , when the relative distance is large , the sought after solitary wave resembles in its gross features the model va pair of eq .
( [ eq : productomega ] ) applied for , say , @xmath106 and @xmath107 ( thus @xmath103 ) .
we may then invoke this model to motivate some important asymptotic results valid for large @xmath104 .
for instance , the local topological vorticity @xmath15 is then peaked around the two points @xmath108 and @xmath109 with weights @xmath110 and @xmath111 respectively , corresponding to the total vorticities @xmath48 of the individual vortex and antivortex .
then the impulse defined from eq .
( [ eq : linmomentum ] ) yields @xmath112 , thanks to reflexion symmetry , while @xmath113 . one may also invoke the derrick - like scaling relation applied to the extended energy functional @xmath114 : @xmath115 where @xmath116 is the total anisotropy energy of the va pair . for large @xmath104
this energy approaches the sum of the anisotropy energies of the individual vortex and antivortex , each given by @xmath117 in view of the virial relation ( [ eq : virial1 ] ) .
hence , @xmath118 and @xmath119 . to summarize ,
@xmath120 one may further consider the familiar group - velocity relation @xmath121 in which we insert the estimate @xmath122 to obtain an elementary differential equation for @xmath29 whose integral is @xmath123 where @xmath124 is an integration constant that can not be fixed by the present leading - order argument .
nevertheless , eq . ( [ eq : kelvinpairenergy ] ) provides the essence of the energy - momentum dispersion at large relative distance @xmath104 or small velocity @xmath125 and hence large momentum @xmath126 . to obtain an accurate numerical solution it is convenient to work with the stereographic variable @xmath37 of eq .
( [ eq : omega ] ) . then the ll eq .
( [ eq : lleomega ] ) restricted to a solitary wave in rigid motion with constant velocity @xmath127 along , say , the y axis reads @xmath128 and is supplemented by the boundary condition @xmath129 at spatial infinity where the magnetization approaches a uniform configuration . once a solution @xmath130 of eq .
( [ eq : lleomega1 ] ) is obtained for a specific value of the velocity @xmath127 , the sought after solitary wave is given by @xmath131 .
equation ( [ eq : lleomega1 ] ) was solved numerically via a newton - raphson iterative algorithm for velocities in the range @xmath132 .
note that @xmath133 is the familiar magnon velocity ( in rationalized units ) and provides an upper bound for the existence of a solitary wave in rigid motion . on the contrary
, there is no lower bound for the velocity the restriction @xmath134 was dictated only by numerical expedience .
now , the calculated solitary wave is illustrated in fig .
[ fig : kelvin02 ] for the relatively low velocity @xmath135 and does indeed describe a va pair ( with @xmath5 ) at a relative distance @xmath136 , as anticipated by the discussion of model va pairs in section [ sec : model ] and the heuristic asymptotic analysis earlier in this section .
but we are now in a position to carry out an accurate calculation practically throughout the allowed range @xmath137 , to discover that there exists a characteristic velocity @xmath138 above which the vortex - antivortex character is lost ( @xmath139 ) and the solitary wave becomes a lump with no apparent topological features , as illustrated in fig .
[ fig : kelvin095 ] for @xmath140 .
the existence of a characteristic velocity @xmath141 becomes apparent when we calculate energy @xmath29 and impulse @xmath142 as functions of @xmath127 , as shown in fig .
[ fig : kelvin_energy ] . note that both @xmath29 and @xmath142 develop a minimum at a common velocity @xmath143 . as a result , the energy vs impulse dispersion shown in fig .
[ fig : kelvin_energy ] develops a cusp at a point @xmath144 that corresponds to the values of @xmath29 and @xmath142 at @xmath145 .
thus the calculated family of solitary waves consists of two branches .
branch i consists of va pairs that propagate with velocities in the range @xmath146 . the corresponding branch in the dispersion of fig .
[ fig : kelvin_energy ] approaches the asymptotic dispersion ( [ eq : kelvinpairenergy ] ) for large @xmath142 ( or @xmath147 ) .
indeed , an excellent fit of the data is obtained by eq .
( [ eq : kelvinpairenergy ] ) if we choose the subleading constant according to @xmath148 .
branch ii consists of lumps with no apparent topological features which propagate with velocities in the range @xmath149 . the corresponding branch in the dispersion of fig . [ fig : kelvin_energy ] is accurately described in the asymptotic ( @xmath150 ) region , where @xmath142 again becomes large , by @xmath151 with @xmath152 .
actually this asymptotic dispersion can be derived by showing that in the limit @xmath153 the 2d landau - lifshitz equation reduces to what is a modified kadomtsev - petviashvili ( kp ) equation @xcite .
a cusp in the energy vs impulse dispersion occurred previously in a calculation of vortex rings in a model superfluid by jones and roberts @xcite .
the same authors together with putterman later argued that the solitary waves that correspond to branch ii are actually unstable @xcite .
interestingly , a dispersion with a cusp appears also in the much simpler problem of electron - positron motion in a uniform magnetic field , where the motion that corresponds to branch ii is also unstable ( see our appendix ) .
hence , while a stability analysis has not yet been carried out for the solitary waves described in this section , it is reasonable to expect that the lumps of branch ii may be unstable .
but there is every reason to believe that the kelvin motion of the va pairs of branch i is indeed stable .
the possibility of topologically nontrivial ( @xmath105 ) va pairs in steady rotational motion was recently examined by one of us @xcite .
again , one may invoke the model va pair of section [ sec : model ] to understand some important features of the rotational motion at large distance @xmath104 .
for definiteness we consider a va pair defined by eq .
( [ eq : productomega ] ) with @xmath106 and @xmath154 , thus @xmath95 .
the local topological vorticity @xmath15 is then peaked around the positions of the vortex and the antivortex , now with weight equal to @xmath111 in both cases .
then , for large @xmath104 , the angular momentum defined by eq .
( [ eq : angmomentum ] ) is estimated to be @xmath155 .
one may also consider the derrick - like scaling relation applied to the extended energy functional @xmath156 : @xmath157 where @xmath116 is the total anisotropy energy of a va pair which approaches asymptotically @xmath158 , and hence @xmath159 , because the anisotropy energy of a single vortex is equal to @xmath117 according to eq .
( [ eq : virial1 ] ) . to summarize , @xmath160 one may further employ the familiar relation @xmath161 in which we insert the estimate @xmath162 to obtain an elementary differential equation for @xmath29 whose integral is @xmath163 where @xmath164 is an integration constant that can not be fixed by the present leading - order argument .
nevertheless , eq .
( [ eq : rotatepairenergy ] ) provides the essence of the energy vs angular momentum dispersion for large relative distance @xmath104 or small angular frequency @xmath165 and hence large angular momentum @xmath166 .
while the preceding heuristic asymptotic analysis is very useful for understanding some basic aspects of a rotating va pair , it does not give us any clue concerning the fate of the pair at small vortex - antivortex separation .
in principle , such a question could be settled by solving numerically the analog of eq .
( [ eq : lleomega1 ] ) for a solitary wave rotating at constant angular frequency @xmath167 : @xmath168 a numerical solution could again be attempted by an iterative newton - raphson algorithm . actually , in this case , we found it more convenient to employ a relaxation algorithm to derive approximate numerical solutions as stationary points of the extended energy functional @xmath156 @xcite . the calculated configuration for @xmath169 is shown in fig .
[ fig : rotate700 ] and does indeed correspond to a topologically nontrivial ( @xmath170 ) rotating va pair consisting of a vortex with negative polarity @xmath171 and an antivortex with positive polarity @xmath172 , as anticipated by the general discussion of section [ sec : model ] .
in particular , for this relatively small value of @xmath167 , the calculated distance @xmath104 in the rotating pair is relatively large @xmath173 , so is the angular momentum @xmath174 ; in rough agreement with the asymptotic estimates of eq .
( [ eq : estimates2 ] ) .
the calculation was repeated for a larger value of angular velocity @xmath175 to find that both the relative distance @xmath176 and the angular momentum @xmath177 are reduced to smaller values .
but the general structure of the solution shown in fig . [ fig : rotate200 ] for @xmath178 remains basically the same as that for @xmath169 , except that the overall size of the va pair is reduced .
for yet larger values of @xmath167 the angular momentum @xmath179 tends to vanish .
this trend is apparent in fig .
[ fig : rotate_energy ] which illustrates the dependence of @xmath29 and angular momentum @xmath179 as functions of angular velocity @xmath167 , as well as the @xmath29 vs @xmath179 dispersion . for large values of @xmath179 the above dispersion exhibits logarithmic dependence , as predicted by the asymptotic result of eq .
( [ eq : rotatepairenergy ] ) .
but the most important feature of the calculated dispersion is that energy approaches the finite value @xmath180 as @xmath181 , while the corresponding rotating va pair becomes vanishingly small .
this result is of central importance for our main argument and will be analyzed in some detail in the remainder of this section .
the best way to describe a vanishing va pair , in the limit @xmath181 , is to invoke yet another model configuration through the stereographic variable @xmath182 where the constant @xmath104 is taken to be real for simplicity .
now , configuration ( [ eq : vaskyrmion ] ) reaches a finite value @xmath183 at spatial infinity and the corresponding magnetization is uniform : @xmath184 which is an appropriate boundary value ( modulo a constant azimuthal rotation ) for easy - plane anisotropy discussed here .
because of ( [ eq : bc ] ) the skyrmion number is expected to be an integer and is actually computed to be @xmath170 by a direct application of eq . ( [ eq : skyrmionnumber ] ) .
furthermore , the spin configuration derived from ( [ eq : vaskyrmion ] ) is an exact solution of the ll equation , if we neglect anisotropy , with exchange energy @xmath185 for any @xmath104 , as discussed long time ago by belavin and polyakov @xcite .
we now return to the main line of argument by noting that configuration ( [ eq : vaskyrmion ] ) may also be thought of as a topologically nontrivial ( @xmath170 ) model va pair consisting of a @xmath171 vortex and a @xmath172 antivortex located at a distance @xmath104 apart , in close analogy with the model va pair constructed in section [ sec : model ] . although ( [ eq : vaskyrmion ] ) is not an exact solution in the presence of anisotropy , it provides a good model for the behavior of a vanishing va pair in the limit @xmath181 .
anisotropy sets a distance scale @xmath186 ( in rationalized units ) beyond which a physically acceptable configuration must reach the uniform value ( [ eq : bc ] ) . in a sense
, this property is shared by configuration ( [ eq : vaskyrmion ] ) because it becomes uniform almost everywhere when @xmath187 while it retains its topological structure as long as @xmath104 remains finite .
the strict limit @xmath188 is not uniform .
if taken naively , all topological structure appears to be lost and both energy @xmath29 and skyrmion number @xmath49 appear to vanish .
however , if integrals are performed before taking the @xmath188 limit , @xmath29 approaches @xmath189 without encountering an energy barrier while @xmath170 for all @xmath104 .
clearly the limit @xmath188 creates a singular point which hides all topological structure .
the main point is the claim that a similar situation arises in the calculated rotating va pair in the limit @xmath181 , as discussed further in section [ sec : switching ] .
this section is completed with a comment concerning the manner in which the @xmath181 limit is reached .
our numerical data as well as virial relation ( [ eq : derrick2 ] ) are consistent with a linear dispersion @xmath190 in the limit @xmath181 , which implies a finite value of the angular velocity @xmath191 in the limit of a vanishing va pair , and thus an upper limit @xmath192 .
applied for long time intervals , our relaxation algorithm revealed some tendency for instability of a rotating va pair , probably due to radiation effects analogous to those expected for a pair of rotating electric charges discussed in the appendix . however , the basic features of rotating va pairs studied in the preceding section persist over sufficiently long time intervals and are thus relevant for practical applications .
in fact , in a realistic ferromagnet , some dissipation is always present and can be modeled by introducing gilbert damping in the ll equation through the replacement @xmath193 in eq .
( [ eq : lle ] ) or eq .
( [ eq : lleomega ] ) , respectively , where @xmath43 is a dissipation constant . the dynamics of a topologically nontrivial ( @xmath105 ) va pair
may thus be summarized as follows . the vortex and the antivortex
rotate around each other , while the pair shrinks due to dissipation . the energy of the pair follows approximately the curve of the right panel of fig .
[ fig : rotate_energy ] as its size ( and its angular momentum ) decreases . at vanishing size
a singular point of the type discussed in the preceding section would be created and the total energy would reach the finite value @xmath180 ( in rationalized units ) . however , the discreteness of the lattice actually interrupts the process when the size of the pair becomes comparable to the lattice spacing .
the va pair disappears ( i.e. , the skyrmion number changes abruptly from @xmath105 to @xmath103 ) and a burst of energy equal to @xmath189 is released into the system , probably in the form of spin waves . in physical units , the amount of energy released is given by @xmath194 where @xmath195 is the film thickness and @xmath13 the exchange constant . for typical values @xmath196 nm and @xmath197 j /
m we obtain the estimate @xmath198 j which is apparently in rough agreement with numerical simulations @xcite .
the scenario described above explains how a topologically forbidden @xmath199 transition can take place in a real ferromagnet but does not by itself account for the experimentally observed vortex core switching .
the complete scenario involves two distinct steps .
first , application of a short burst of an alternating magnetic field creates a va pair in the vicinity of a preexisting single vortex .
second , a three - body collision takes place during which a suitable pair of vortices is annihilated through a @xmath4 transition of the type described above , and the final product is a single vortex with polarity opposite to that of the original vortex @xcite .
we also note that a system with three vortices ( two vortices and an antivortex which form a cross - tie wall ) was observed in a rectangular platelet in @xcite .
their spectrum of eigenmodes was studied experimantally .
here we do not address the question of how a va pair is actually created .
rather we concentrate on step 2 of the process and explain in some detail how vortex core switching may occur in a three - body collision .
specifically , let us assume that a single @xmath200=c vortex is initially at rest at some specified point which is taken to be the origin of the coordinate system .
let us further assume that a topologically trivial @xmath1 va pair consisting of a @xmath201=a vortex and a @xmath202=b antivortex is somehow created in the neighborhood of the original vortex . once created the ab pair will undergo kelvin motion of the type described in section [ sec : kelvin ] and eventually collide with the single vortex c. the process was simulated by a numerical solution of the corresponding initial - value problem in the ll equation .
figure [ fig : switch_v01 ] provides an illustration with three characteristic snapshots in the case of a relatively slow kelvin pair initially moving along the y axis with velocity @xmath203 for which the vortex and the antivortex are separated by a distance @xmath204 . as the pair approaches , the original c=@xmath200 vortex teams up with the b=@xmath202 partner of the ab pair to form a new , topologically nontrivial @xmath205 va pair in quasi - rotational motion .
in fact , b rotates almost a full circle around c before rejoining its original partner a. the new ab pair is again a topologically trivial @xmath1 va pair in kelvin motion that moves away from the target vortex , having suffered a total scattering angle that is greater than @xmath117 from its original direction .
the scattering is inelastic in the sense that the outgoing ab pair moves out with greater velocity @xmath206 . and
, most remarkably , the target vortex c moves away from the origin and comes to rest at a new location in the fourth quadrant of the xy plane .
this rather unusual behavior is explained by the unusual nature of the conservation laws ( 11 ) and ( 12 ) which allow for a transmutation between position and impulse in the case of topologically nontrivial systems , such as the three - vortex system considered here ( a more detailed discussion will be given in a future publication ) .
the preceding numerical experiment was repeated for a kelvin pair with relatively large velocity @xmath207 for which the vortex and the antivortex are tightly bound at a relative distance @xmath208 @xcite .
the process is again illustrated by three characteristic snapshots in figure [ fig : switch_v05 ] .
while the initial stages of the process are similar to those encountered in the case of slow kelvin motion ( figure [ fig : switch_v01 ] ) a substantial departure occurs when the pair now approaches the target vortex . in particular , as soon as antivortex b=@xmath202 begins to rotate around the target vortex c=@xmath200 they collide and undergo a spectacular @xmath209 transition ( annihilation ) leaving behind the a=@xmath201 vortex which may be thought of as the target vortex c=@xmath200 with polarity flipped from @xmath210 to 1 ( vortex core switching ) and a burst of spin waves propagating away from the scattering region . a detailed numerical investigation of the three - vortex process for kelvin waves with velocities in the allowed range @xmath211 suggests the existence of the three characteristic regions separated by two critical velocities @xmath212 and @xmath213 ( such that @xmath214 , with @xmath215 being the critical velocity discussed in section [ sec : kelvin ] ) . for @xmath216 , the kelvin pair undergoes nearly elastic scattering of the type depicted in figure [ fig : switch_v01 ] .
for @xmath217 , the process leads to a topologically forbidden @xmath209 transition of the type illustrated in figure [ fig : switch_v05 ] .
there is also some evidence that fast kelvin waves with velocities in the narrow range @xmath218 undergo a nearly elastic scattering without inversion of the polarity of the target vortex .
the va pairs analyzed in this paper are special examples of solitary waves whose dynamics is closely related to their topological structure .
for example , the va pairs studied in section [ sec : kelvin ] can undergo free translational motion because their topological charge vanishes ( @xmath5 ) .
in contrast , a va pair with nonvanishing @xmath49 performs rotational motion around a fixed guiding center and is thus spontaneously pinned within the ferromagnetic medium , as discussed in section [ sec : rotate ] .
such a peculiar dynamical behavior would have been surprising had it not occurred previously in the case of interacting electric charges in the presence of a magnetic field .
it is then interesting to ascertain conditions under which a certain field theory would exhibit a similar link between topology and dynamics .
a simple criterion was introduced in ref .
@xcite and is briefly summarized as follows .
we restrict attention to 2d hamiltonian systems described in terms of @xmath219 pairs of canonically conjugate variables @xmath220 with @xmath221 .
then one may define the local vorticity @xmath222 which is a simple generalization of the first step of eq .
( [ eq : vorticity1 ] ) , the remaining two steps being special to the specific example considered in the present paper ( ferromagnets ) . now , in general ,
the total vorticity @xmath223 is expected to vanish by a trivial partial integration using eq .
( [ eq : vorticity2 ] ) . on the other hand ,
a nonzero @xmath48 would signal a special topological structure of the field theory under consideration and may lead to peculiar dynamics .
the theory analyzed in the present paper is an example ( with @xmath224 ) which yields a nonzero @xmath48 that may be identified with the pontryagin index @xmath225 .
an example with @xmath226 is provided by a 2d antiferromagnet where @xmath227 except when an external field is present which may lead to @xmath228 and an interesting link between topology and dynamics @xcite .
implicit in the preceding general argument is the fact that a topological charge @xmath49 is conserved . also taking into account that @xmath49 is quantized
, one would expect that a topological ( @xmath229 ) soliton can not be annihilated in a continuous manner .
nevertheless , a quasi - continuous process was described in ref .
@xcite and in the present paper according to which a rotating va pair with @xmath95 may be reduced to a singular point and thereby be eliminated by lattice discreteness without encountering an energy barrier .
a mechanism for changing the topological number of a magnetic configuration makes it possible to obtain controlled switching between topologically distinct ( and thus robust ) magnetic states .
this was achieved in the experiments of refs .
the dynamics underlying both experiments involves a three - vortex process @xcite initiated by the production of a topologically trivial va pair in the vicinity of a preexisting vortex by an alternating magnetic field .
it has been predicted that the same phenomenon would occur if one uses a rotating external field @xcite .
the resulting three - vortex system carries nonzero topological charge and is thus by itself a rotating object spontaneously pinned in the magnet . also due to dissipation
, a quasi - continuous process takes place that changes the topological number by one unit , leaving behind a burst of energy in the form of spin waves and a single vortex with polarity opposite to that of the original vortex .
although our strictly 2d treatment provides a detailed scenario for the three - vortex process that leads to polarity switching , it does not account for the initial production of a topologically trivial va pair .
this is partly due to our approximation of a thin film with infinite extent .
implicit in this approximation is the assumption that the demagnetizing field amounts to a simple ( additive ) renormalization of easy - plane anisotropy . while this assumption appears to be firmly established for static magnetic states @xcite we do not know of a corresponding mathematical derivation for dynamical processes of the type discussed here .
it is also important to visualize how the formation of a singular point discussed in the present strictly 2d context appears within a realistic magnetic element of finite extent .
numerical simulations @xcite show that , in an element of finite thickness , a singular point is first created at one of the surfaces of the element .
the va pair then vanishes by formation and subsequent annihilation of a singular point at successive levels away from the surface . at the stage
when a singular point has been formed and annihilated , say , near the top surface , while the va pair is still present in the bulk of the element , a bloch point ( bp ) is created in the element . this is a somewhat simplified realization of the bp studied in ref .
needless to say , the bp created near the top surface is eventually annihilated when the va pair exits the system through the lower surface .
it is important to emphasize the unusual fact that during creation and annihilation of the bp the system does not have to overcome an energy barrier , unlike the case discussed in @xcite , essentially for the same reasons explained for the strictly 2d va pairs studied in the main text .
is grateful for hospitality at the max - planck institute for the physics of complex systems ( dresden ) where this work was completed .
most of the distinct features of the dynamics of va pairs occur also in the dynamics of electric charges in the presence of a uniform magnetic field @xmath230 .
although we shall mainly be interested in 2d motion in a plane perpendicular to @xmath230 , it is convenient to keep for the moment 3d notation and write the equations of motion for two interacting charges @xmath231 and @xmath232 : @xmath233 where @xmath234 is the mutual force derived from a potential energy @xmath235 and @xmath236 denotes derivative with respect to the argument @xmath237 .
the conserved energy functional is then given by @xmath238 which does not depend explicitly on the magnetic field , while the conserved linear momentum ( impulse ) is now given by @xmath239 and differs from the usual mechanical definition by an important field dependent term .
this ( second ) term actually indicates a rather profound influence of the magnetic field on the dynamics of electric charges .
for instance , note that a shift of the origin of coordinates by a constant vector @xmath240 , thus @xmath241 and @xmath242 , induces a nontrivial change on the impulse @xmath243 of eq .
( [ eq : a4 ] ) given by @xmath244 this unusual behavior is analogous to that of the impulse defined by eq .
( [ eq : linmomentum ] ) in the case of field configurations with nonvanishing total topological vorticity @xmath48 or skyrmion number @xmath49 . here
we may abstract from eq .
( [ eq : a5 ] ) the analog of the topological vorticity in the present problem : @xmath245 an electron - positron pair ( @xmath246 ) may undergo kelvin motion , while two like charges ( @xmath228 ) perform rotational motion around a fixed guiding center , as determined by the explicit solutions constructed and briefly analyzed in the remainder of this appendix .
consider first the case of an electron - positron pair ( @xmath247 ) for which a special 2d solution of eqs .
( [ eq : a1 ] ) is given by @xmath248 where the electron and the positron are located at a constant relative distance @xmath104 along the x axis and move in formation along the y axis with constant velocity @xmath249 in close analogy with the kelvin motion of the va pair discussed in section [ sec : kelvin ] .
we further calculate energy @xmath29 from eq .
( [ eq : a3 ] ) and impulse @xmath250 from eq .
( [ eq : a4 ] ) to find @xmath251 ^ 2 + v(d ) \nonumber \\
p & = & 2 m v + ebd = 2 m \frac{v'(d)}{eb } + ebd.\end{aligned}\ ] ] hence all relevant quantities are given in parametric form as functions of relative distance @xmath104 once the potential energy @xmath252 has been specified .
but there are some generic properties of this solution that are practically independent of the choice of @xmath253
. we may take derivatives with respect to @xmath104 of both sides of eq .
( [ eq : a8 ] ) to find @xmath254 and @xmath255 . an immediate consequence of these relations is the group - velocity relation @xmath256 .
we also note that both @xmath29 and @xmath142 may acquire an extremum at a common value of @xmath104 ( or @xmath127 ) determined from @xmath257 this is actually the reason for the appearance of a cusp in the @xmath29 vs @xmath142 dispersion analogous to that encountered in the kelvin motion of va pairs , which now appears to be a generic feature of a wide class of physical systems . for an explicit demonstration
we make the special choice of potential energy @xmath258 in order to model the behavior of va pairs at large relative distance @xcite . for a graphical illustration
we also make the special choice of constants @xmath259 and @xmath260 , as suggested by eq .
( [ eq : a6 ] ) , to write @xmath261 where we note that both @xmath29 and @xmath142 acquire a minimum at a distance @xmath262 or velocity @xmath263 .
the dependence of @xmath29 and @xmath142 on the velocity @xmath127 as well as the @xmath29 vs @xmath142 dispersion are shown in fig .
[ fig : pn_energy ] and are found to be closely analogous to the results of fig .
[ fig : kelvin_energy ] pertaining to kelvin motion of va pairs . in particular
, for a widely separated pair ( branch i ) we find from eq .
( [ eq : a11 ] ) that @xmath264 , and @xmath265 with @xmath266 , in close analogy with the asymptotic results of eqs . ( [ eq : estimates1])([eq : kelvinpairenergy ] ) .
the appearance of a cusp and consequently of branch ii in the spectrum is also notable .
but the details of branch ii are different than those of fig .
[ fig : kelvin_energy ] and eq .
( [ eq : kelvinpairenergy2 ] ) .
there is now no upper limit in the velocity @xmath127 .
in fact , all @xmath267 and @xmath142 in eq .
( [ eq : a11 ] ) diverge in the limit of small @xmath104 and @xmath268 , which coincides with the dispersion @xmath269 of a free particle with mass equal to the total mass of the pair @xmath270 .
we have also carried out a stability analysis of the special kelvin - like solution ( [ eq : a6a ] ) to find that the motion is marginally stable along branch i but becomes unstable along branch ii .
this conclusion is in agreement with a similar result obtained in ref .
@xcite in the case of a vortex ring in a superfluid , as is further discussed in the concluding remarks of our section [ sec : kelvin ] . as a last example we consider the case of 2d motion of two like charges @xmath271
. then a special solution of eq .
( [ eq : a1 ] ) is given by @xmath272 which describes a pair rotating at constant radius @xmath273 and angular frequency @xmath274 where the velocity @xmath127 is determined from the algebraic equation @xmath275 that expresses the exact balance of the centrifugal , the magnetic , and the mutual force .
the conserved energy is still calculated from eq .
( [ eq : a3 ] ) with @xmath276 , but the conservation of the impulse @xmath243 of eq .
( [ eq : a4 ] ) is simply equivalent to the statement that rotation takes place around a fixed guiding center .
more relevant is now the conserved angular momentum ( impulse ) which is given by @xmath277 where the overdot denotes time derivative .
again , the angular momentum differs from its standard mechanical expression by important field - dependent terms .
now , for the specific choice of potential energy given by eq .
( [ eq : a10 ] ) , and constants @xmath278 and @xmath260 , the algebraic equation ( [ eq : a13 ] ) yields @xmath279 while the angular velocity @xmath167 , the energy @xmath29 , and the angular momentum @xmath179 , read @xmath280 in view of eq .
( [ eq : a15 ] ) all three quantities in ( [ eq : a16 ] ) are expressed in terms of a single parameter @xmath273 . as a check of consistency one
may verify the relation @xmath281 using eqs .
( [ eq : a15]-[eq : a16 ] ) .
the dependence of @xmath29 and @xmath179 on angular velocity @xmath167 as well as the @xmath29 vs @xmath179 dispersion are shown in fig .
[ fig : pp_energy ] .
again there exists a close analogy with the results of section [ sec : rotate ] on rotating va pairs .
in particular , for large diameter @xmath104 , eqs .
( [ eq : a16 ] ) yield @xmath282 , and @xmath283 with @xmath284 , which should be compared with the asymptotic results for va pairs given in eqs .
( [ eq : estimates2])([eq : rotatepairenergy ] ) . on the other hand ,
some quantitative differences arise at small @xmath104 , where the angular momentum vanishes as expected ( @xmath285 ) but the angular frequency diverges ( @xmath286 ) .
furthermore , the energy @xmath29 does not reach a finite value at @xmath287 ( as was the case for rotating va pairs ) but diverges logarithmically to minus infinity .
finally , an analysis of mechanical stability @xcite shows that circular motion of two like charges in a magnetic field is marginally stable for all values of @xmath104 , in contrast to the kelvin motion discussed earlier in this section which becomes unstable at small @xmath104 .
however , unlike kelvin motion which proceeds with no acceleration , a rotating pair is expected to radiate when full electrodynamics is turned on .
surely , this is also a source of instability and may indicate a similar instability for the rotating va pairs discussed in section [ sec : rotate ] .
b. v. waeyenberge , a. puzic , h. stoll , k. w. chou , t. tyliszczak , r. hertel , m. fhnle , h. brckl , k. rott , g. reiss , i. neudecker , d. weiss , c. h. back , and g. schtz , nature(london ) * 444 * , 461 ( 2006 ) . | we study the dynamics of vortex - antivortex ( va ) pairs in an infinitely thin ferromagnetic film with easy - plane anisotropy .
these are localized excitations with finite energy that are characterized by a topological ( skyrmion ) number @xmath0 .
topologically trivial @xmath1 va pairs undergo kelvin motion analogous to that encountered in fluid dynamics .
in contrast , topologically nontrivial @xmath2 va pairs perform rotational motion around a fixed guiding center .
we present the results of a detailed study in both cases and further demonstrate that in the presence of dissipation a rotating @xmath3 va pair shrinks to a point and is annihilated , due to the discreteness of the lattice , thus leading to a `` topologically forbidden '' @xmath4 process .
we argue that the latter process underlies the experimentally observed vortex core switching whereby the polarity of a single vortex is reversed after collision with an @xmath5 va pair created by a burst of an applied alternating magnetic field . |
options are financial contracts which give the holder the right to buy ( call options ) or sell ( put options ) commodities or securities for a predetermined exercise ( or strike ) price by a certain expiration date @xcite .
conventional european ( american ) options can be exercised only on ( at any time up to ) the expiration date .
since the option confers on its holder a right with no obligation , it should carry a price at the time of contract .
it is the classic work of black , scholes and merton @xcite which suggested a strategy for determining a fair price for the option in a risk - free environment .
closed - form valuation within the black - scholes - merton equilibrium pricing theory @xcite is only possible for a small subset of financial derivatives . in the majority of cases
one must appeal to numerical techniques such as monte carlo simulations , or finite difference methods and much of the effort in the field has been in developing efficient algorithms for numerically solving the black - scholes equation @xcite .
an alternative direction has been the evaluation of discrete - time , discrete - state stochastic models of the market on binomial and trinomial trees @xcite .
not only is this discrete approach intuitive and easily accessible to a less mathematically sophisticated audience ; but it also seems to us to be a more accurate description of market dynamics and better suited for evaluating more involved financial instruments .
moreover , the few exact black - scholes results available can be recovered in the appropriate continuous - time trading limit .
the main difficulty in pricing with binomial trees has been the non - monotonic numerical convergence and the dramatic increase in computational effort with increasing number of time steps @xcite .
for example , the state of the art calculations involve memory storage scaling linearly ( quadratically ) with the number of time steps , @xmath0 , for european ( american ) options , while the computation time increases like @xmath1 in both cases @xcite . in this paper
we reconsider valuation on binomial trees from what we call a forward looking " prospective : we imagine acting as well - educated consumers who attempt to eliminate risk and estimate the future expected value of an option according to some reasonable dynamical model .
we will regard the movement of the price on the tree as a random walk ( with statistical properties consistent with a risk - neutral world ) with walls " imposed by the nature of the option , such as the possibility of early exercise ( american options ) or the presence of barriers .
the resulting mathematical formulation then has two conceptually distinct components : the first ingredient is an explicit description of the possible walls " . for example , in the case of barrier american options both the barrier and the early exercise " surface need to be specified .
the second step will be to compute the probability that the price reaches particular values at every accessible point on the tree .
this involves counting the number of paths reaching that point in the presence of walls " , a somewhat involved but exactly solvable combinatorics problem .
once these two steps ( specifying the walls and computing the probabilities ) are accomplished the value of both european and american options , with and without barriers , can be written down explicitly . in an attempt to be pedagogical ,
we will limit ourselves to the simplest put options : european , simple american and european with a straight up - and - out " barrier . although the calculation can be simply extended to the barrier american option that discussion merits a separate publication .
as far as we know , in the case of trees explicit formulas like the ones we are proposing exist in the literature only in the simplest case of conventional european options @xcite . for the more complicated case of american options ,
the main issues are best summarized in the last chapter of neil chriss book @xcite : the true difficulty in pricing american options is determining exactly what the early exercise boundary looks like .
if we could know this _ a priori _ for any option ( e.g. , by some sort of formula ) , we could produce pricing formulas for american options . "
below we propose a solution to this problem in the context of binomial trees .
our formulation complements the earlier studies of american options in the limit of continuous - time trading @xcite which also focus on the presence of an early exercise boundary for the valuation of path - dependent instruments .
the study of the continuum limit of our formulas is instructive and will be left for a future publication .
to establish notation we begin by dividing the life of an option , @xmath2 , into @xmath0 time intervals of equal length , @xmath3 .
we assume that at each discrete time @xmath4 ( @xmath5 ) the stock price moves from its initial value , @xmath6 , to one of two new values : either up to @xmath7 ( @xmath8 ) or down to @xmath9 ( @xmath10 ) @xcite .
this process defines a tree with nodes labeled by a two dimensional vector , @xmath11 ( @xmath12 ) and characterized by a stock price @xmath13 , the price reached at time @xmath14 after @xmath15 up and @xmath16 down movements , starting from the original price @xmath17 .
the probability of an up ( down ) movement will be denoted by @xmath18 ( @xmath19 ) ; and thus each point on the tree is also characterized by the probability , @xmath20 , which represents the probability associated with a single path of @xmath21 time steps , @xmath15 ( @xmath16 ) of which involve an increase ( decrease ) in the stock price .
computing the probability of connecting the origin with point @xmath11 requires , in addition to the single path probability , a factor counting _ the number _ of such possible paths in the presence of a barrier and/or the possibility of early exercise .
the calculation of this degeneracy factor involves the details of each financial derivative and it will be discussed in turn for each of our examples .
the binomial tree model introduces three free parameters , @xmath22 and @xmath18 .
two of these are usually fixed by requiring that the important statistical properties of the random process defined above , such as the mean and variance , coincide with those of the continuum black - scholes - merton theory @xcite .
in particular , @xmath23 where @xmath24 is the risk - free interest rate , and the volatility , @xmath25 , is a measure of the variance of the stock price . we are left with one free parameter which can be chosen to simplify the theoretical analysis ; one might choose , for example , @xmath26 @xcite , which simplifies the tree geometry by arranging that an up motion followed by a down motion leads to no change in the stock price .
this condition together with ( [ statprop1 ] ) and ( [ statprop2 ] ) implies : @xmath27 we stress that equations ( 1 - 5 ) are to be regarded as short - time approximations where terms higher order in @xmath28 were ignored . with these definitions out of the way we can begin discussing the valuation of put options with strike price @xmath29 and expiration time @xmath2 .
the simple european put option is a good illustration of our forward looking " approach .
we are interested in all those paths on the tree which , at expiration time @xmath30 , reach a price , @xmath31 , for which the option should be exercised .
that implies that @xmath32 $ ] , where @xmath33 refers to the integer part of the quantity in square brackets .
the mean value of the option at expiration can then be written as a sum over all values of @xmath34 of the payoff at @xmath15 , @xmath35 , multiplied by the probability of realizing the price @xmath36 after @xmath0 time steps , @xmath37 $ ] .
as already mentioned above , @xmath37 = \aleph _ e [ n , j ] p_u ^j ( 1-p_u ) ^{n - j}$ ] , where @xmath38 $ ] counts the number of paths starting at the origin and reaching the price @xmath39 in @xmath0 time steps . for the case of conventional european options
this is just the number of paths of @xmath0 time steps , with @xmath15 up and @xmath40 down movements of the price , and is thus given by the binomial coefficient , @xmath41 ~ = ~\left ( \begin{array}{c } n\\j \end{array } \right ) ~=~\frac{n!}{j!(n - j)!}.\ ] ] the resulting expression for the mean value of the option at maturity is then discounted to the time of contract by the risk - free interest rate factor , @xmath42 , to determine the current expected value of the option : @xmath43 this expression is not new : it was first discussed by cox and rubinstein @xcite who also showed that in the appropriate continuous trading - time limit ( @xmath44 ) ( [ valueev ] ) reduces to the black - scholes result @xcite .
we are now ready to extend ( [ valueev ] ) into an exact formula for the mean value of an european put option with a barrier .
although our approach can be used for other barrier instruments , we consider the simplest case of an up - and - out " put option which ceases to exist when some barrier price , @xmath45 , higher than the current stock is reached . with the choice @xmath26 an explicit equation for the nodes of the tree which constitute the barrier can be written down : @xmath46 here , @xmath47 $ ] defines the first point just above the barrier , @xmath48 , and @xmath49 labels the last relevant point on the barrier corresponding to the time closest to the maturity of the option , i.e. , @xmath50 $ ] . since the probability that any allowed path starting with the present stock price , @xmath17 , reaches an exercise price at maturity , @xmath51 , is still @xmath52 ( with @xmath53 ) the average value of the european barrier option
can be written in a form similar to ( [ valueev ] ) : @xmath54~ p_u ^j ( 1-p_u)^{n - j } \left ( x - s_0 u^j d^{n - j}\right ) , \ ] ] where @xmath55 $ ] is the number of paths @xmath0 time - steps long involving @xmath15 up and @xmath40 down movements of the price excluding those paths reaching any of the points on or above the barrier ( [ barrier ] ) . as we will explain below , @xmath55 $ ] is given by @xmath56 ~=~ \left(\begin{array}{c } n\\j \end{array } \right ) ~-~\sum _ { h=0 } ^{h_m } \aleph _ { eb } ^{res}[j_b + 1 + 2h , j_b + 1+h ] \left ( \begin{array}{c }
n - j_b -1 -2h\\ j - j_b -1 -h \end{array}\right ) , \ ] ] where the second term on the right - hand side represents the contribution from the unwanted paths which hit the barrier ( [ barrier ] ) before reaching an exercise point @xmath57 . to understand the form of the excluded contribution in ( [ excluded ] ) we first note that reaching the excluded region requires that the path hits the barrier at least once .
one might think that the number of unwanted paths can then be calculated by ( i ) counting the number of paths connecting the origin to a given point on the barrier ; ( ii ) multiplying this by the number of paths connecting that point on the barrier with the exercise point @xmath57 @xcite this includes all paths which wander _ into _ the above - barrier region ; and finally ( iii ) summing over all points of the barrier ( [ barrier ] ) .
however , a particular path reaching a given point on the barrier might have already hit any of the previous barrier points , and thus it would also be counted in the contribution in ( ii ) from all paths starting at the first barrier point reached by the particular path under consideration .
thus , summing indiscriminately over barrier points would lead to overcounting unless , in ( i ) , we only include those paths which hit the barrier for the first time . in other words ,
( i ) must only include paths starting from the origin which reach the particular point on the barrier without having previously visited any other barrier point .
the number of such restricted paths ( reaching the point @xmath58 ) is what we denoted by @xmath59 $ ] in ( [ excluded ] ) .
also note that the final sum over the length of the barrier is restricted to @xmath60 with @xmath61 , corresponding to the fact that , in general , the exercise point @xmath57 can not be reached from all points on the barrier .
this completes our explanation of ( [ excluded ] ) . to solve equations
( [ recursione1 ] ) and ( [ recursione2 ] ) we first combine the sum on the right - hand side of ( [ recursione2 ] ) with the term on the left and rewrite the resulting equation in the form of a discrete convolution : @xmath68 \left ( \begin{array}{c } 2(h - l)\\ h - l \end{array } \right ) = \left ( \begin{array}{c } j_b + 1 + 2h\\ j_b + 1 + h \end{array } \right ) , \ ] ] where the boundary condition , @xmath69=1 $ ] , is already included as the @xmath70 contribution to ( [ convolutione ] ) .
note that ( [ convolutione ] ) can be solved by standard laplace transform ( or @xmath71-transform ) techniques @xcite .
since in applying these ideas to the more complicated american options we will lose the convolution form the kernel will depend on @xmath72 and @xmath73 separately and not only through the difference , @xmath74 we will proceed in a more general way and stay in configuration space " until the very end .
we prefer to regard ( [ convolutione ] ) as a matrix equation of the form : @xmath75 here @xmath76 and @xmath77 are @xmath78 dimensional vectors , with components @xmath79 $ ] and @xmath80 , @xmath81 , and the @xmath82 dimensional matrix , @xmath83 , can be written as , @xmath84_{h , l}~=~\left ( \begin{array}{c } 2(h - l)\\ h - l\end{array}\right ) \theta \left ( h - l\right ) .\ ] ] note that in ( [ matrixleb ] ) we have explicitly added a @xmath85 function ( @xmath86 for @xmath87 and vanishes otherwise ) to stress that @xmath88 is a lower triangular matrix with unity along and zeros above the diagonal .
this simple observation allows us to rewrite ( [ matrixeqeb ] ) in the convenient form , @xmath89 where @xmath90 is a nilpotent matrix of order @xmath91 , @xmath92 for @xmath93 ; and @xmath94_{h.l}=\left ( \begin{array}{c } 2(h - l)\\h - l\end{array}\right ) \theta ( h - l-1)$ ] .
the nilpotent property of @xmath95 allows us to write down the explicit solution for ( [ matrixeqeb ] ) , @xmath96^{-1}{\bf d}_{eb}~=~ \sum _ { r=0 } ^{h_m } ( -1)^r { \bf q}_{eb } ^r { \bf d}_{eb},\ ] ] which in turn leads to the following formula for the value of the option , @xmath97_{h , l}\left ( \begin{array}{c } j_b + 1 + 2l\\ j_b + 1 + l\end{array}\right ) \\ & \times & p_u ^j ( 1-p_u)^{n - j } \left ( x - s_0 u^j d^{n - j}\right ) \nonumber.\end{aligned}\ ] ] the lower limit , @xmath98 , on the external sum in ( [ priceres ] ) excludes all paths unaffected by the presence of the barrier ; also we have explicitly indicated the @xmath0 and/or @xmath15 dependence of the various quantities involved ; and have separated out the contribution to @xmath55 $ ] from unrestricted paths ( the first term on the right - hand side of ( [ excluded ] ) ) which simply leads to the value of the european put option given in ( [ valueev ] ) .
we expect that , since we have an analytical formula , we should be able to recover the exact solution of the continuum black - scholes theory for this simplest of barrier options @xcite as was already done for conventional european puts @xcite .
figure [ fig1 ] shows the numerical convergence of the binomial value of a representative up - and - out " european put option to its analytic value @xcite . the same general idea used in the case of european barrier options
will now be used to write down an exact formula for the price of a simple american option , regarding the latter as an option with an early - exercise barrier .
let us proceed in carrying out the program outlined above by assuming that the exb , i.e. , @xmath100 $ ] , is explicitly given . to begin our calculation we will need some very general properties of the barrier .
these follow from two simple characteristics of early exercise : ( i ) if the point @xmath11 is an early exercise point , then so are all points deeper in - the - money " , @xmath106 ; and ( ii ) if two adjacent points at the same time step , @xmath107 and @xmath108 , are both early exercise points so is the point @xmath11 .
( the latter property follows from a conventional
backwardation " argument @xcite which indicates that the average expected payoff at @xmath11 , discounted at the risk - free interest rate , is smaller than the actual payoff , thus making @xmath11 itself an early exercise point . )
it is not hard to see that ( i ) and ( ii ) guarantee that the inner part of the early exercise region can not be reached without crossing the exb .
thus , if we define @xmath109 to be the first time for which early exercise becomes possible and parametrize the points on the exb as @xmath110)$ ] with @xmath111 , it then follows that @xmath112 = 0 $ ] .
moreover , the structure of the tree ensures that @xmath100 $ ] is a non decreasing function of @xmath21 ; more precisely , for each time step , @xmath113 $ ] either increases by one or remains the same .
the formal expression for the price of an american option can be written down once one recognizes that once a path hits the exb the option expires and thus any point on the barrier can be reached at most once . as a result ,
the value of the option is a sum of ( appropriately discounted ) payoffs along the barrier , weighted by the probability of reaching each point on the barrier without having visited the barrier at previous times .
we can then write the expected value of an american option as : @xmath114 p_u ^{j_x [ i_a + h ] } ( 1-p_u ) ^{i_a + h - j_x [ i_a + h ] } \left ( x - s_0 u^{j_x [ i_a + h ] } d^{i_a + h -j_x [ i_a + h ] } \right ) , \ ] ] where @xmath115 denotes the number of paths reaching the exb in @xmath116 time steps without having previously visited any points on the barrier .
one last step is the determination of @xmath105 , the value of the american put at every point @xmath11 on the tree which , in turn , will allow us to derive the equation for the exb .
this is easily done by simply translating the origin in ( [ valueavfin ] ) : @xmath140_{h , l}\left ( \begin{array}{c } i_a + l - i\\ j_x [ i_a +
l ] -j \end{array } \right ) \nonumber \\ & \times & p_u ^{j_x [ i_a + h]-j } ( 1-p_u ) ^{i_a + h - j_x [ i_a + h ] -i+j}\left ( x - s_0 u^{j_x [ i_a + h ] } d^{i_a + h -j_x [ i_a + h ] } \right ) .\end{aligned}\ ] ] together with ( [ defsurface ] ) this then leads to the rather formidable - looking equation for the barrier height @xmath141 $ ] at the @xmath142-th time step ( @xmath143 ) , as a functional of the barrier position at all _ future _ time steps before expiration : @xmath144 } d^{i_a + k -\tilde{j } [ i_a + k ] } \right ) ~&=&~\sum _ { { h , l , m}=k } ^{n - i_a } e^{-r(h - k)\tau } ( -1)^m \left [ { \bf q}_a ^m\right ] _ { h , l}\left ( \begin{array}{c } l - k\\ j_x [ i_a + l ] - j_x [ i_a + k ] \end{array } \right ) \nonumber\\&\times & p_u ^{j_x [ i_a + h]-j_x [ i_a + k ] } ( 1-p_u ) ^ { h - k - j_x [ i_a + h ] + j_x [ i_a + k]}\\&\times&\left ( x - s_0 u^{j_x [ i_a + h ] } d^{i_a + h -j_x [ i_a + h ] } \right ) \nonumber\\ \label{exb2 } j_x [ i_a + k]~~&=&~ \mbox{int}\left \{\tilde{j } [ i_a + k ] \right \}.\end{aligned}\ ] ] [ it should now be clear that in ( [ valij ] ) @xmath15 must be restricted to points on the tree as the binomial coefficient @xmath145 -\tilde{j } [ x_a + l ] \end{array}\right ) $ ] would be ill - defined hence the choice ( [ defsurface ] ) . ] equations ( [ exb ] ) and ( [ exb2 ] ) for the boundary together with the formula for the value of the option , ( [ valueavfin ] ) , constitute an exact pricing strategy for a conventional american put .
a similar formula for an american put with an up - and - out " barrier will be discussed in a future publication .
it is instructive to consider equations ( [ valueavfin ] ) , ( [ exb ] ) and ( [ exb2 ] ) in the explicitly solvable case of a straight barrier .
we begin with the observation that , at expiration , @xmath146 , ( [ exb ] ) reduces to the equation for @xmath147 $ ] , already defined in the case of the european option , and thus , the barrier goes through the point @xmath148 .
moreover , starting from the exact point @xmath148 on the barrier and decreasing @xmath100 $ ] by one with each backward time step we reach @xmath149 along the straight line , @xmath100=i - n+j^*$ ] . recall that , since with each increasing time step , @xmath100 $ ] either increases by one or remains the same , this straight line represents a lower bound for the early exercise barrier . for this straight barrier ( [ inversion ] ) and ( [ valueavfin ] ) reduce to , @xmath150~p_u ^h \left ( 1-p_u \right ) ^{n - j^ { * } } \left ( x - s_0 u^h d^{n - j^{*}}\right ) \end{aligned}\ ] ] with @xmath151 we expect that the result for the true barrier should approach the straight line formula for coarse enough time steps , @xmath152 , ( where this @xmath109 is the first time of early exercise in the limit of continuous - time trading ) .
we have presented a scheme for pricing options with and without barriers on binomial trees . to the best of our knowledge
ours is the first explicit derivation of exact formulas treating barriers on binomial trees .
it is our expectation that in the limit of continuous - time trading we should be able to recover the few exact results available in the literature , especially for american options@xcite .
we also hope that our explicit formulas may provide a framework for improving the efficiency of numerical computations .
the authors dedicate this paper to professor ferdinando mancini , a remarkable teacher , colleague and friend , on the occasion of his 60th birthday .
we are grateful to stanko barle for reading the manuscript and bringing the work of references @xcite and @xcite to our attention .
finally , we acknowledge the hospitality of the nyu physics department where most of this work was conceived . | we reconsider the valuation of barrier options by means of binomial trees from a forward looking " prospective rather than the more conventional backward induction " one used by standard approaches .
this reformulation allows us to write closed - form expressions for the value of european and american put barrier - options on a non - dividend - paying stock . |
we would like to thank clemens ritter for his invaluable help with data collection and the epsrc ( grant number ep / c534654 ) for funding . | ni@xmath0sio@xmath1 , liebenbergite , is an example of a quasi - one - dimensional magnet made up of frustrated corner sharing triangles of ni@xmath2 ( @xmath3 ) ions that propagate parallel to the _ b _ axis .
ni@xmath0sio@xmath1 is isostructural with olivine , a common mineral of varying composition fe@xmath4mg@xmath5sio@xmath1 , and is described in the orthorhombic space group _
pnma_. a synthetic polycrystalline sample of ni@xmath0sio@xmath1 was studied using constant wavelength powder neutron diffraction .
diffraction spectra were collected above and below the antiferromagnetic ordering transition ( @xmath6 @xmath7 34k ) and were used to refine the atomic and magnetic structures of ni@xmath0sio@xmath1 .
corepresentational theory was used to determine the symmetry - allowed magnetic structures after the nel transition and the refined magnetic structure evidences both ferromagnetic and antiferromagnetic inter - chain interactions , and ferromagnetic intra - chain coupling . the competition between the magnetic interactions can be seen in the canting of the moments away from a collinear arrangement .
current theories of magnetism are challenged by highly correlated systems that show degeneracies and fluctuations .
the characterisation of ground states of new model systems is an important step for testing and confirming predictions , and geometrically frustrated magnets provide important opportunity for this .
a triangle of antiferromagnetically coupled spins is the archetypal example of a geometrically frustrated motif , due to the competing exchange interactions between the three moments .
there are a huge variety of ways to build these triangles into higher dimensional structures , with particular interest in structures with low connectivity , such as the vertex sharing triangle and tetrahedral geometries of the kagom @xcite and pyrochlore @xcite lattices respectively .
the olivine family of minerals ( @xmath8o@xmath1 where @xmath9 = an m@xmath2 transition metal ion , often with mixed occupancy and @xmath10 = a group iii or group iv atom , most commonly si ) consist of a one dimensional array of metal ions which form zig - zag chains of corner - sharing triangles . in ni@xmath0sio@xmath1
there are two symmetry distinct metal sites : ni(1 ) form linear chains of ions propagating parallel to the _ b_-axis , and ni(2 ) make up the apex of the triangles zig - zagging along the linear chains within the @xmath11-plane ( figure [ structure]b ) .
these metal sites are coordinated by an octahedra of oxygen , and the tetrahedral interstitial sites between the chains are occupied by group iii or group iv atoms ( figure [ structure]a ) .
preliminary inelastic powder neutron studies of this material show the presence of a spin gap at @xmath12@xmath13 that opens below t@xmath14k at the transition to the ordered state .
this gapped behaviour was first eluded to by inverse susceptibility measurements that were well fitted to a spin gap plus curie - weiss model@xcite that was predicted as a characteristic of a one dimensional chain of integer spin@xcite .
this is in contrast to the gapless behaviour predicted and observed for chains of non - integer spin . *
a * : a polyhedral representation of ni@xmath0sio@xmath1 with nio@xmath15 displayed as dark edge - sharing octahedra and sio@xmath1 displayed as light tetrahedra occupying the interstitial sites between the chains ; * b * : the magnetic structure of ni@xmath0sio@xmath1 refined from neutron diffraction data collected at 2k and @xmath16 showing the zig - zag chain arrangement of corner sharing isosceles triangles of ni@xmath2 ions propagating along the @xmath17-axis ; * c * : a view parallel to the @xmath17-axis giving the edge - lengths of the isosceles triangles formed from the ni@xmath2 ions ( generated using fpstudio @xcite ) . ] diffraction spectra from a 15 g sample of ni@xmath0sio@xmath1 @xcite , were recorded on the d1a powder neutron diffractometer at the ill , both above and below the nel transition using neutrons of wavelength @xmath16 .
the sample temperature was controlled using a standard ` orange ' cryostat and the sample was contained within a vanadium can .
the refinement of the collected spectra at 39k and 2k will be presented .
the refined nuclear structure of ni@xmath0sio@xmath1 at 39k was consistent with previously published powder diffraction results @xcite and the low temperature diffraction pattern there were new magnetic bragg peaks that could be indexed with a magnetic propogation vector * k@xmath18*=(@xmath19 ) in kovalev s notation @xcite , also consistent with previously published results @xcite , although the refined magnetic structure was found to differ from that reported . the different symmetry - types of magnetic structure were determined using corepresentation analysis and the program sara_h _ @xcite .
refinement of the magnetic structure was performed in terms of the basis vectors , @xmath20 , of a given irreducible corepresentation multiplied by a ( weighting ) coefficient , @xmath21 . in this formalism
the set of refined moments , @xmath22 , is generated from the linear combination of the basis vectors : @xmath23 . the observed and calculated neutron diffraction spectrum of ni@xmath0sio@xmath1 with nuclear and magnetic reflections , collected on d1a of the ill at 2k using neutrons of @xmath16@xcite .
the crosses correspond to experimental data ; the lines correspond to calculated values and the upper and lower tick - marks indicate the positions of the magnetic and crystallographic reflections respectively .
the goodness of fit parameters were @xmath24 and @xmath25 for 35 parameters . ]
the data could only be well fitted using a corepresentation formed by the combination of the representations @xmath26 and @xmath27 from kovalev s tables ( table [ bvtable ] ) @xcite .
it is notable that the 4 equivalent positions of each of the ni crystallographic sites form distinct groups , or _
orbits _ , formed from positions ( 1 + 3 ) and ( 2 + 4 ) and the combination of these representations defines moment components on all of the symmetry related positions .
the equivalence of these orbits prevents the sizes of the moments from being separately refined , and those moments of a given ni site were constrained to be equal in magnitude .
the refinement ( figure [ structure]c ) was not improved by the presence of components parallel to the @xmath17-axis , and the moments of both the ni(1 ) and ni(2 ) sites were subsequently restricted to the @xmath28-plane .
the inter - chain couplings are shown to be antiferromagnetic when along the @xmath29 direction and ferromagnetic along the @xmath30 direction , with ferromagnetic interactions dominating between the intra - chain spins .
the moments on the two sites form a non - collinear structure , with the canting of ni(1 ) and ni(2 ) moments away from the @xmath31-axis being @xmath32 and @xmath33 respectively .
the moments of the ni(1 ) and ni(2 ) sites refined to 2.03(3 ) and 2.08(4 ) @xmath34 , respectively , in good agreement with expectations for @xmath35 ni@xmath2 .
the refined structure differs from that previously reported in two respects.@xcite 1 . )
no mention of an attempt to refine the moments of the two crystallographically distinct ni@xmath2 sites to different values was made but we found that an improvement in the refinement could be achieved if these values were allowed to refine separately . 2 . )
symmetry analysis shows that @xmath36 and @xmath37 would need to be used to describe the previously reported structure , in contrast to @xmath26 and @xmath27 from our results .
the differences in the diffraction pattern between these two structures are subtle but our goodness of fit was noticeably reduced if the previously reported structure was used to model the data .
the subtle differences could easily have been masked by the resolution of the previously reported pattern along with the limited angular range reported ( 5 - 30 2@xmath38 ) in comparison to our high resolution data collected ( 5 - 160 2@xmath38 ) .
.basis vectors for representations @xmath26 and @xmath27 of the space group @xmath39 with * k*@xmath18=(@xmath19 ) following kovalev s notation@xcite .
the equivalent positions of ni(1 ) are defined according to 1 : @xmath40 , 2 : @xmath41 , 3 : @xmath42 , 4 : @xmath43 ; and for ni(2 ) according to 1 : @xmath44 , 2 : @xmath45 , 3 : @xmath46 , 4 : @xmath47 [ cols="^,^,^,^,^,^,^,^,^ " , ] in conclusion , we report a low temperature powder neutron diffraction study of ni@xmath0sio@xmath1 , a quasi-1-dimensional magnet made up of zig - zagging chains of ni@xmath2 triangles .
the magnetic structure is shown to reveal evidence of frustrated interactions with a significant canting away from collinearity .
we note also that corepresentational theory is required to fully describe the symmetry of the spin configuration which we found to have subtle differences to that previously published . |
insect development rates are applied not only in pest control management but also in forensic entomology @xcite .
several species of dipteran and coleopteran families infest decaying material in order to breed offspring , and this includes the colonization of animal carcasses as well as dead bodies @xcite . in homicide investigations , determination of the age of larvae feeding on a corpse can indicate a minimal post - mortem interval ( pmi@xmath0 ) @xcite .
this is often important in forensic case work @xcite .
the general life cycle of blow flies includes four stages : egg stage , larval stage , pupal stage and imago stage @xcite . during the larval stage , three instars
can be separated : 1st , 2nd and 3rd instar , where the latter is divided due to behavioral changes in feeding and post - feeding larvae .
blow flies deposit egg clutches directly on the food substrate , such as a dead body @xcite , in a position where the eggs are protected and in a moist environment . this ensures a food supply for the hatching 1st instar larvae .
the first three instars each undergo a moult to reach the next developmental stage ; the stages can be distinguished by the number of respiratory slits at the posterior end of the larvae .
the third instar stage lasts for longer than the first two and is divided in a feeding and a post - feeding phase .
the latter is a preparation for pupation .
therefore , the larvae leave the food source to find a suitable place for pupation , emptying their gut @xcite .
about one third of the pre - adult development time is spent in the post - feeding larval stage @xcite
. then pupation sets in and the imago develops within the pupal case till eclosion @xcite .
this last stage persists for about half of the time of the total development .
the larval growth rate depends on its body temperature , which is directly influenced by environmental conditions as ambient temperature and the heat generated by maggot aggregations @xcite . also , an important detail for pmi determination is that each species has its own temperature dependent growth rate . in forensic case work ,
two different methods are frequently used to calculate a pmi .
the first uses isomegalen or isomorphen diagrams , by which the lengths or the developmental stage of the larvae are combined as a function of time and mean ambient temperature in a single diagram @xcite . according to its originators ,
this method is optimal only if the body and therefore the larvae were not undergoing fluctuating temperatures , e.g. in an enclosed environment where the temperature was nearly constant .
the second method of calculating a pmi estimates the accumulated degree days or hours ( add or adh ) .
adh values represent a certain number of `` energy hours '' that are necessary for the development of insect larvae .
the degree day or hour concept assumes that the developmental rate is proportional to the temperature within a certain species - specific temperature range ( overview in @xcite ) .
however , the relationship of temperature and development rate ( reciprocal of development time ) is typically curvilinear at high and low temperatures and linear only in between .
the formula for calculating adh is given by @xmath2 where @xmath3 is the development time , @xmath4 is the ambient temperature , and the minimum developmental threshold temperature @xmath5 is a species - specific value , the so called development zero , which is the x - intercept , i.e. , an extrapolation of the linear approximation of the reciprocal of the developmental time .
this value has no biological meaning , it is the mathematical consequence of using a linear regression analysis @xcite .
one basic condition for using the adh method is that the adh value for completing a developmental stage stays constant within certain temperature thresholds .
for example a developmental duration for finishing a certain stage of 14 days at 25@xmath1c results in 238 add when a base temperature of 8@xmath1c is assumed .
a developmental duration of 19 days at 21@xmath1c results in 231 add , both add - values are in the same range .
we analyzed a published data - set for the development of _ lucilia sericata _
( meigen 1826 ) @xcite and calculated the corresponding adh values for these data .
[ f - temp_adh ] shows the calculated adh values for a base temperature of @xmath5=8@xmath1c ( as calculated by a linear regression analysis for the used data - set ) . in the figure
we see a new effect : for the younger and also shorter developmental phases the adh values are nearly constant over the complete range of temperatures , but for the post - feeding and the pupal stages the adh values are strongly temperature dependent .
in general , the adh method seems to give good results only when the larvae of interest have been exposed to temperatures similar to those used in generating the reference value applied in the pmi calculation @xcite .
moreover , the temperature range in which the development rate is actually linear is not wide enough to cover all temperatures during a typical summer in germany ( see also examples for may / june 2008 in fig .
[ f - temp_profile ] ) .
furthermore , neither developmental durations nor base temperatures for development have been calculated for species originating from germany .
the method must therefore be used carefully .
furthermore , it is highly problematic that uncertainties for temperature measurements from a crime scene can not be taken into account by either of the commonly used methods for pmi determination .
it is difficult to determine the actual temperature controlling the larvae at a real crime scene .
since temperature is the variable that most influences development , it is crucial to consider it as accurately as possible .
the standard procedure is to use temperatures of the nearest weather station for the desired time frame and correct them by applying a regression starting from temperatures measured at the crime scene , when taking the larvae as evidence @xcite . the corrected values still contain uncertainties that can not be accounted for by the methods currently used for pmi determination .
no information exists for either model about the quality of the method or the error intervals of the calculated pmis .
we analyzed developmental data for _ l. sericata _ at different temperatures and fitted an individual exponential function for each developmental stage .
data used as input to the model were published by grassberger and reiter ( 2001 ) and represent the minimal time in hours to complete each larval phase ( egg stage = stage 0 , 1st instar = stage 1 , 2nd instar = stage 2 , 3rd instar feeding = stage 3 , 3rd instar post - feeding = stage 4 and pupal stage = stage 5 ) until eclosion of the adult blow fly .
the used data - set is one of the rare sets which covers a lot of temperatures and the resulting growth curve seems to represent growth behavior well ( see original paper ) .
unfortunately , grassberger and reiter do not give any error values for their measurements , so we assumed an error for the developmental times of about 1 hour .
these authors used 250 g of raw beef liver in plastic jars , and placed 100eggs on the food substrate .
the jars were placed in a precision incubator . at each temperature regime
the procedure was repeated 10times .
every 4hours , four of the most developed maggots were removed from the plastic jars , killed in boiling water , and preserved in alcohol @xcite and then their stage of development was determined .
our new larval growth model is based on the data shown in fig .
[ f - temp_time ] , in which the duration of each developmental stage was measured as a function of temperature @xcite .
these data points were fitted with an exponential function of the form : @xmath6 where @xmath7 is the duration of one developmental stage @xmath8 as a function of temperature @xmath4 .
the parameters fitted for the different stages are shown in table [ t - fitpar ] .
the parameter @xmath9 defines how strongly the time interval depends on temperature ; the higher the parameter in table [ t - fitpar ] , the steeper is the gradient of the fitted curve .
@xmath10 represents the minimum time interval required for finishing a certain developmental stage and @xmath11 provides the absolute normalization .
the developmental stages of the maggots were determined every @xmath12h , such that time measurement errors are set to @xmath13 following an uniform distribution .
it is assumed that the maggot body temperature is known to an accuracy of 3% in order to take into account uncertainties about differences between ambient and maggot body temperature .
the parameters @xmath11 , @xmath9 and @xmath10 were determined by minimizing the sum of error squares . as seen in fig .
[ f - temp_time ] , the exponential function accurately models the behavior during all developmental stages and will be used below . in all stages
the developmental duration at temperatures below 24 @xmath1c starts to rise exponentially .
[ f - temp_adh ] shows the calculated adh values corresponding to eq .
( [ e - adh ] ) ( data points ) .
in addition , the figure shows the function @xmath14 ( lines ) .
@xmath15 is calculated by eq .
( [ e - fit ] ) with the previously fitted parameters ( table [ t - fitpar ] ) .
again , the functions give a reasonable description of the data .
nevertheless , the model is an empirical one , based on the observations of the data points generated by grassberger and reiter ( 2001 ) . for european and especially german temperatures , calculation of the total developmental duration must allow for non - linear temperature behavior in order to ensure accuracy .
the basic idea underlying a new approach in pmi determination is to follow an ambient time - temperature profile @xmath16 backwards in time starting from the time point @xmath17 at which the maggots of interest were collected .
the idea of backwards calculation is obviously similar to the adh method , but in the new model the important improvement is the way of calculating the larval age .
the latter is calculated successively during certain time steps using the fitted functions ( introduced in fig .
[ f - temp_time ] ) corresponding to the current developmental stage . in each stage
@xmath8 the relative developmental progress is @xmath18 ( values 0 - 1 ) where 0 is the beginning and 1 is the finishing point of each developmental stage ; e.g. a maggot in the middle of the post - feeding stage is @xmath19 , at the end of the post - feeding stage it is @xmath20 and so forth .
the developmental duration @xmath21 spent in each individual stage is calculated by solving the relation : @xmath22 where @xmath23 is the infinitesimal relative development .
the calculation starts with the developmental stage of the maggot at the time of collection , summing the developmental progress of each stage backwards until the beginning of the egg stage is reached .
the calculation for each collection stage uses @xmath24 .
the total development time @xmath25 or post - mortem interval ( pmi ) is then given by @xmath26 for the new model a program was written in c++ using root ( http://root.cern.ch/ ) .
this program includes all mentioned mathematical steps and produces the figures shown here as output . for each new pmi calculation ,
the corresponding temperature profile can be inserted and individually chosen uncertainties can be included . to explore the uncertainties in the total developmental duration , a monte - carlo simulation was applied , which is commonly used for simulations in life sciences @xcite .
it is a method for calculating one final uncertainty after considering all statistically independent uncertainties that influence e.g. the larval age .
the mean pmi with corresponding standard variation is calculated @xmath27 times taking into account and varying all uncertainties described in the following .
first , the developmental profiles @xmath15 have uncertainties due to the measurement procedure .
second , the time - temperature profile from the collection scene is not known precisely and must be approximated using temperature values from nearby weather stations .
the variations are introduced for each model as follows : development profile : : : the mean duration values of the temperature - time data are randomly smeared with a uniform distribution with corresponding error @xmath28 ; for the maggot body temperature @xmath29 a gaussian distribution is used .
new fits with the function in eq .
( [ e - fit ] ) are performed for each stage .
time - temperature profile : : : deviations between the temperature profile at the collection scene and the nearest weather station are accounted for by gaussian smearing of time @xmath30 and temperature @xmath4 , with the corresponding errors @xmath31 and @xmath32 as width for each data point .
@xmath32 can be inserted in the models calculation individually dependent on the differences between the temperatures at finding place and weather station .
we calculated the pmi for a mock crime scene with the following parameters : the error of the measurement of the original data @xmath33h , the errors of the data of the weather station @xmath34h and @xmath35c , the difference of the ambient temperature and larval body temperature @xmath36% for 10.000 models for a fixed collection stage progress of @xmath37 .
the results are shown in fig .
[ f - temp_profile ] _ ( upper part ) _ which is a direct output of the new program that calculates the pmi .
the lower time axis defines the progress of the temperature profile forward in time , representing the time frame of interest .
the temperature profile used here ( black line ) is taken from the minimum and maximum temperatures in may and june 2008 measured at cologne / bonn airport .
the right end of the diagram marks a fictional time point of maggot collection and therefore the starting point for pmi calculation .
the upper time axis depicts the pmi backwards in time starting from the moment of maggot collection . for each developmental stage
the pmi was calculated by following a linear interpolation between the maximum and minimum temperatures .
the histograms illustrate the pmi distribution for each stage and show a clear single peak structure .
the arrows on the top show the 1-standard deviation interval for each stage around the mean pmi value , and range between 0.1 and 1.2 days ( depending on the stage ) . since no data points below temperatures @xmath38c were measured , the functions @xmath15 were extrapolated to lower temperatures . as expected , the pmi and the corresponding standard deviation increase with higher developmental duration ( see arrows above histogram ) .
since the exact progress within the developmental stage at collection time is most of the time also unknown , a third uncertainty is introduced : stage progress : : : the developmental stage at collection time was determined only to integer precision , so that it is assumed the exact progress is an uniformly distributed value between 0 and 1 .
consequently , the starting value for the pmi calculation @xmath18 at time @xmath17 is randomly and uniformly chosen within the interval [ 0,1 ] for each model . fig .
[ f - temp_profile ] _ ( lower part ) _ shows the pmi calculation for the same parameters as before , but without setting the progress of the development for each stage to a fixed value .
the 1-standard deviation values increase by 0.3 to 3.3 days .
the resulting uncertainty in the progress of the stage contributes about 75% to the total pmi error interval .
in addition , the histograms show deviations from a clear single peak structure , e.g. for the pupal stage , implying that the pmi probabilities for 21days and 26days are nearly the same . to use the new model , the crucial parameter is therefore the correct determination of the progress of the developmental stage of maggots collected from a corpse .
the impact of correct temperature determination at the maggot collection scene is shown in fig .
[ f - stage_pmi_compare ] .
the data points represent the mean pmis with an error bar of 1 standard deviation as a function of collection stage for three different temperature profiles .
the triangles show the pmis for the original temperature profile as measured at cologne / bonn airport .
the bullets ( squares ) show the results for the same profile but subtracted ( and added ) by 2 @xmath1c .
as expected , the pmis and the corresponding standard deviations of the lower ( higher ) temperature profile increase ( decrease ) relative to the nominal profile .
these differences in temperature of 2 @xmath1c give rise to an effect of 15 - 30% .
that implies that a miscalculation of the temperature at the crime scene of 2 @xmath1c will result in a miscalculation of the pmi by 15 - 30% .
the later the stage , the greater the deviation from the actual pmi . in fig .
[ f - pmi_rel_compare ] , pmis calculated using the corresponding mean temperature values in the temperature interval @xmath39 $ ] are compared with pmis from our model using the three temperature profiles introduced previously .
the calculated mean temperatures were as follows ( calculated for the time frame till completion of each stage ) : stage 0 = 18@xmath1c , stage 1 = 19 @xmath1c , stage 2 = 20 @xmath1c , stage 3 = 19 @xmath1c , stage 4 = 16 @xmath1c , stage 5 = 17 @xmath1c .
the pmi values based on the temperature profile and those based on a mean temperature value agree to within about 5% for the high temperature value ( original profile + 2 @xmath1c ) in all stages .
the deviation between mean temperature and the original temperature profile exceeds the 10% level starting at the 3rd instar feeding stage , and increases to 25% in the pupal stage .
this effect becomes even larger for the low temperature profile ( original profile subtracted by 2 @xmath1c ) . starting from the 2nd instar stage , the deviation increases from about 10% up to about 65% for the pupal stage .
this means that use of mean temperature values overestimates the influence of low temperatures and underestimates periods of high temperatures .
the effect should be larger if the mean temperature during the development is lower still , e.g. in spring or fall . in general , more data points are needed for the developmental duration at low temperature ranges to provide more reliable statements .
we calculated the pmi in a real case where the actual pmi was known due to a confession of the offender . at the end of august 2007
the victim was killed in early morning and was found 4 days later also in the morning on a grassland .
this leads to a pmi of approximately 96 hours .
the victim was stabbed to death and had several wounds which would act as attractant to the blow flies .
it can be assumed that blow flies started ovipositing early after death occurred @xcite .
autopsy was performed directly after the corpse was recovered and several 2nd instar larvae of _
l. sericata _ were collected .
the largest larvae measured 6.1 mm .
hourly temperature values were taken from a weather station 10 km away .
the mean temperature was 16 @xmath1c .
using grassberger and reiters isomegalen diagram for a larvae measuring 6 mm and a mean temperature of 16 @xmath1c results in a time interval of 3.2 days plus 30 hours ( larval development time plus egg period ) . in total a pmi of 107 hours is indicated .
this would shift the time of oviposition to nighttime , which is a highly unlikely event @xcite .
the same data can be used to calculate the adh value for _ l. sericata _ for reaching 6 mm in order to calculate the pmi not based on the mean temperature but on hourly data .
as mentioned earlier , a regression analysis of the data set reveals a base temperature of 8 @xmath1c .
the corresponding adh value is therefore 856 , based on the equation : @xmath40 subtracting the hourly adh values , estimated by the temperature values from the weather station and the base temperature , from the starting value of 856 results in a pmi of 101 hours . to use the new model for calculating larval age in the real case ,
information about the progress of the 2nd instar larval stage was required . in the original work of grassberger and reiter ( 2001 , figure 1 )
a figure is included showing the growth of the larvae and also the time points for each moult . according to this figure , the 2nd instar stage sets in after the larvae have reached a size of approximately 4 mm and ends when the larvae have reached a size of approximately 8 mm . as the largest larvae we collected measured 6 mm , we chose p=0.5 as progress for the larval stage .
we included the hourly temperature profile and chose a temperature error of 1 @xmath1c .
the result of the calculation was a pmi of 99 hours ( sd = 3 hours ) .
these calculations of a pmi in a real case show that all three methods give reasonable results .
furthermore , it becomes obvious that the new model is a possible alternative for the existing methods with the benefit of directly providing a standard deviation for the calculation .
the new model improves the larval age calculation in specific ways .
it can be used in non - linear parts of the temperature dependent development , and includes individually defined uncertainties for a temperature profile determined retrospectively from the nearest weather station . in the new model the temperature profile plus the determination of the larval stage are translated into a mean pmi as well as a standard deviation .
pmi calculation using mean temperatures , however , can lead to severe deviations from the real pmi .
so far , the main uncertainty arises from the fact that the developmental stage is determined only on a 1 - 6 scale ( egg , 1st instar , 2nd instar , 3rd instar feeding , 3rd instar post - feeding and pupae ) .
as shown above , 75% of the uncertainties in the model depend on the exact determination of the developmental progress , and additional length values , as shown for the pmi calculation in the real case , will propably increase its accuracy leading to more accurate pmi calculations .
moreover , the next step is to produce own growth data with known error values to refine the inclusion of uncertainties that are only rough estimates at the present time and to improve the till now only empirical model . nevertheless , the new pmi calculation program is suitable for use in forensic case work as a general tool for pmi determination .
scientists from every country or climatic region can incorporate their own growth values for different species and ensure a high accuracy in pmi determination .
anderson g ( 2001 ) _ forensic entomology : the utility of arthropods in legal investigations _
, chapter insect succession on carrion and its relationship to determining time of death .
crc press , pp .
143175 archer m and elger m ( 2003 ) female breeding - site preferences and larval feeding strategies of carrion - breeding calliphoridae and sarcophagidae ( diptera ) : a quantitative analysis .
australian journal of zoology 51:165174 grassberger m and reiter c ( 2001 ) effect of temperature on lucilia sericata ( diptera : calliphoridae ) development with special reference to the isomegalen- and isomorphen - diagram .
forensic sci int 120(1 - 2):3236 | homicide investigations often depend on the determination of a minimum post - mortem interval ( pmi@xmath0 ) by forensic entomologists .
the age of the most developed insect larvae ( mostly blow fly larvae ) gives reasonably reliable information about the minimum time a person has been dead .
methods such as isomegalen diagrams or adh calculations can have problems in their reliability , so we established in this study a new growth model to calculate the larval age of _ lucilia sericata _ ( meigen 1826 ) .
this is based on the actual non - linear development of the blow fly and is designed to include uncertainties , e.g. for temperature values from the crime scene .
we used published data for the development of _
l. sericata _ to estimate non - linear functions describing the temperature dependent behavior of each developmental state . for the new model it is most important to determine the progress within one developmental state as correctly as possible since this affects the accuracy of the pmi estimation by up to 75% .
we found that pmi calculations based on one mean temperature value differ by up to 65% from pmis based on an 12-hourly time temperature profile .
differences of 2 @xmath1c in the estimation of the crime scene temperature result in a deviation in pmi calculation of 15 - 30% .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.